Kwint PDF

Photography:
Cover design and

lay-out inside work:
Printed by:
Patricia Nauta, Leiden, and Ruth Veneboer, s-Hertogenbosch

Francis te Nijenhuis, zonnezijn creaties, s-Hertogenbosch
Optima Grafische Communicatie, Rotterdam
The work presented in this thesis was performed at the Division of Pharmacoepidemiology & Clinical Pharmacology, Utrecht Institute for Pharmaceutical
Sciences (UIPS), Faculty of Science, Utrecht University, Utrecht; the Department of
Public Health and Primary Care, Leiden University Medical Center, Leiden and the
SIR Institute for Pharmacy Practice and Policy, Leiden.
CIP-gegevens Koninklijke Bibliotheek, Den Haag

Kwint, H.F.
Improving appropriate medication use for older people in primary care.
Thesis Utrecht University - with ref. - with summary in Dutch
ISBN/EAN: 978-94-6169-373-0
2013 Henk-Frans Kwint
For articles published or accepted for publication, the copyright has been transferred to the respective
publisher. No part of this thesis may be reproduced, stored in a retrieval system, or transmitted in any form
or by any means without the permission of the author or, when appropriate, the publisher of the manuscript.
I M P R OV I N G A P P R O P R IAT E M E D IC AT IO N U SE F O R
O L D E R P E O P L E I N P R I M A RY C A R E
O P T I M A L I SE R E N VA N J U I ST M E D IC I J N G E B RU I K
D O O R O U D E R E N I N D E E E R ST E L I J N S G E Z O N D H E I D S Z O R G
(met een samenvatting in het Nederlands)
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de
rector magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college
voor promoties in het openbaar te verdedigen op woensdag 29 mei des middags te
2.30 uur
door
HENDRIK FRANS KWINT
geboren op 17 januari 1970 te Meppel
P ROM OTOR E N :
Prof.dr. M.L. Bouvy

Prof.dr. J. Gussekloo
CO-PROMOTOR:
Dr. A. Faber
Financial support for conducting parts of the presented studies was provided by:
Apotheek Voorzorg, a provider of multidose drug dispensing systems, and the Royal Dutch Association
for the Advancement of Pharmacy (KNMP) (Chapter 3.2).
BENU Apotheken (formerly known as LLOYDS Apotheken); Royal Dutch Association for the
Advancement of Pharmacy (KNMP); Astra-Zeneca and the healthcare insurance companies Achmea
and Menzis (Chapter 4.1, 4.2 and 4.3).
Connecting Care cooperation (Chapter 4.4).
Financial support in general and/or for the publication of this thesis was provided by:
Apotheek Stevenshof, SIR Institute for Pharmacy Practice and Policy and Nederlands Bijwerkingen
Fonds.
Chapter 1 Introduction
Chapter 2 Collaboration in medication review
2
Chapter 3
The relationship between the extent of collaboration

of general practitioners and pharmacists and the
implementation of recommendations arising from
medication review. A systematic review.
9
21
23
Multidose drug dispensing
47
3.1 Medication adherence and knowledge of older

patients with and without multidose drug dispensing.
49
3.2 Effects of medication review on drug-related

problems in patients using multidose drug dispensing
systems. A pragmatic randomised controlled study.
65
Chapter 4 Home Medication Review

4.1 Home medication review in older patients.
A randomised controlled trial in primary care
93
95
4.2 The contribution of patient interviews to the

identification of drug-related problems in home
medication review.
115
4.3 Completeness of medication reviews provided by

community pharmacists
133
4.4 Evaluation of medication reviews in older patients

using STOPP-START criteria
145
Chapter 5 General discussion
159
Chapter 6 From Summary to About the author
185
6.1 Summary
187
6.2 Samenvatting
193
6.3 Dankwoord
199
6.4 List of co-authors
209
6.5 List of publications
211
6.6 About the author
213
I nt ro d uc t io n
| C hapte r 1
The proportion of people over 60 years of age is growing faster than any other age
group, as a result of longer life expectancy and declining fertility rates.1 While it
is important to remember that the increase in life expectancy is the consequence
of improvements in living conditions and medical advances, the ageing of the
population does create major challenges to health care systems.2 Patients aged
75 years and older often present a complex clinical picture involving coexisting
conditions and frailty.3 The implementation of guidelines for the management of
these conditions has resulted in an increase in the number of drugs prescribed to
older people.2,4 One out of three visitors of 75 years and older of Dutch community
pharmacies uses 5 chronic drugs or more, also known as polypharmacy.5
While the benefits of some individual therapeutic interventions are clear-cut, the
co-prescription of multiple drugs has considerable attendant risks.2 Adhering
to current clinical practice guidelines in caring for an older person may have
undesirable effects.6 Clinical practice guidelines especially fail to address the
needs of older patients with complex co-morbidity. Older people are at increased
risk of adverse drug reactions (ADR) secondary to age-related changes in
pharmacokinetics and pharmacodynamics, drug interactions, increased comorbidity and associated polypharmacy. At least 5% of hospital admissions are
directly related to ADRs 7-10 and higher rates have been reported for older people,
who are likely to be receiving multiple medications for long-term illnesses.8 The
Hospital Admission Related to Medication (HARM) study in The Netherlands
suggested that almost half of these medication related hospitalizations could be
avoided.7 One of the main recommendations of the HARM-study was that the
medication use of older patients with polypharmacy should be reviewed regularly
for potential drug-related problems (DRPs). This evaluation should also include
identification of barriers for medication regimen adherence and should provide
tools to facilitate the proper use of medication.7
This thesis focuses at both the medication management and the quality of
pharmacotherapy of older people with polypharmacy in primary care and ignores
older people in nursing homes and hospitals. Health policy encourages older people
to remain living in their own homes as long as possible. The proportion of older
people living in their own homes is increasing (e.g. from 80% in 2000 till 86% in
2010 for persons of 80 years and older).11 The ability to remain independent in ones
home partly depends on the ability to manage a complicated medication regimen 12
and the use of multiple medicines can give rise to various problems.
Clinical pharmacy interventions for community-dwelling older people with
polypharmacy can be divided in dispensing services (aimed at support of
11
medication management) and medication reviews (aimed at appropriateness of the

pharmacotherapy, see Figure 1).13
Fi g u re 1 Clinical pharmac y inter ventions for older people with
po l y pha r ma c y
Patients
Assessment
Intervention
65 years and 5 drugs
Medication
management
problems
Dispensing
service
Drugrelated
problems
Medication
review
DI SP E N SI N G SE RV IC E S
Medication management is a term that is generally used for both problems
and solutions related to administration of a medication regimen. In this thesis,
medication management is defined as the patients ability to self-administrate his or
her medication regimen.14 A dispensing service consists of the provision of support
(e.g. adherence aids) that enable patients to manage their own medication.13
A dispensing service is preceded by an assessment of patients medication
management problems (Figure 1). This consists of different elements. The first
and most important element is to assess patients actual medication use: which
prescribed drugs are actually used, which apparently discontinued prescription
drugs are still used and which non-prescription drugs (e.g. over-the-counter drugs,
complementary and alternative drugs) are used. Other elements may be discussion
of patients medication knowledge and assessment of risk factors for medication
12
I nt ro d uc t io n
| C hapte r 1
mismanagement (e.g. adherence and practical issues such as difficulties with

opening containers).13,15
Dispensing services consist of a variety of strategies to optimize medication
management including education, written medication schedules, medication
calendars, special packaging to organize medications, disposal of expired/unused
medications 13,14 and chronic medication services.16-18
The use of special packagings to organize medications is growing fast in The
Netherlands. The most frequently used special packaging is multidose drug
dispensing (MDD). MDD provides patients with robot-dispensed unit doses. All
drugs intended for one dosing moment are gathered in disposable bags and labelled
with patient data, drug contents, and the date and time for intake.19-21 In The
Netherlands, patients experiencing difficulties with managing their medications
are mostly identified as eligible for MDD by community pharmacists, general
practitioners (GPs) or relatives.19 Another development in The Netherlands is the
steep increase of chronic medication services (CMS) by community pharmacies.
This proactive service consists of synchronizing chronic medications and informing
patients when their refills are available.16,17
M E DIC AT ION R E V I E W
The term medication review is used to describe a plethora of interventions
aimed at the appropriateness of pharmacotherapy that might be carried out by
prescribers, pharmacists or nurses, both independently as well as in collaboration.
Medication review has been defined as a structured, critical examination of a
patients medicines with the objective of reaching an agreement with the patient
about treatment, optimising the impact of medicines, minimizing the number
of drug-related problems (DRPs) and reducing waste.22 The difference between
a medication review and a dispensing service is that medication review is an indepth analysis of patients drug-related problems including effectiveness and
safety issues and potential undertreatment whereas a dispensing service is aimed
at patients practical problems with medication use including adherence issues
(Figure 1). Ideally, medication reviews also include assessment of these medication
management problems.13
In 2002, the Medicines Partnership in the United Kingdom described different
levels of medication review based on the availability of patient data: Prescription
review (medication records), Treatment review (medication and medical records)
and Clinical medication review (medication records, medical records and patient
13
interview).22 The development of the Medicines Use Review (MUR) was an

important reason to introduce a new type of medication review in 2008. The MUR
was classified as a Concordance or compliance review with the scope on exploring
medicine taking including the patients self-reported use and beliefs of medicines.23
On the contrary Prescription review usually takes places without consulting the
patient.23 For a Clinical medication review, not only a patient interview is required,
but also access to patients medical notes and laboratory test results.23,24 All these
types and levels of medication review appear to be defined from the perspective of
the pharmacist (Figure 2).
Fi g u re 2 D ifferent t y pes of med icati on revi ew based on avai labi li t y of
d at a 22- 24
Patient
interview
Concordance
& compliance
review
Clinical
medication
review
Medication
records
Treatment
review
Medical
records
Prescription
review
Only few studies have evaluated medication reviews by physicians independently 25,26
while time limitations probably impede widespread implementation of these
reviews.25 Dependent on the availability of data more types of review appear to be
possible, but these are not mentioned in literature (e.g. a patient interview without
any pharmacy or prescriber data or a review solely based on prescriber data)
(Figure 2).
Clinical medication review has been adapted in different countries, like Home
Medicines Reviews in Australia,13,27 Medication Therapy Management services in
the United States 28 and Comprehensive Medication Reviews in Finland.29 Home
14
I nt ro d uc t io n
| C hapte r 1
Medicines Reviews (HMRs) in Australia are structured collaborative reviews

conducted by an accredited pharmacist in the patients home. The patient should
have at least one risk factor that necessitates a review. Examples of risk factors are:
currently taking five or more medications; significant changes to the medication
regimen in the preceding 3 months, recent hospital discharge; having difficulty
with organising medicines; symptoms suggestive of an adverse drug reaction;
taking medication with a narrow therapeutic index or requiring therapeutic
drug monitoring.13,27 The HMR process consists of: a formal GP referral to the
community pharmacy; HMR by an accredited pharmacist; formal report from the
pharmacist, which is discussed with the GP; patient and GP agreement with the
pharmaceutical care plan; implementation of the plan with monitoring and followup. At the home visit, the patient is asked to present all medicines (prescribed and
otherwise) to the pharmacist, who then reviews the patients knowledge, storage,
practical problems and perception of the effectiveness and adverse effects of each
medicine. The pharmacist also arranges for disposal of expired or unused medicines
as required.13
The Dutch clinical medication review process is described in the recent Dutch
multidisciplinary guideline Polypharmacy in the elderly.30 This process consists
of five steps, is continuous and occurs over multiple encounters of pharmacists,
patients and physicians (Figure 3).31
At the start of the process, data collection has to take place including medication
records from the community pharmacy (current medications, past medication use,
information on comorbidity and drug intolerances) and medical records from the
GP practice (current medical conditions, medical history and laboratory data). The
five steps are:
(1) Patient interview (pharmaceutical anamnesis)
During the patient interview the pharmacist verifies patients actual medication use
and potential use of earlier discontinued prescription drugs, over-the-counter drugs
and complementary and alternative drugs. In addition, practical problems and
perception of the effectiveness and adverse effects of each medicine are reviewed.
The patients beliefs about and experiences with pharmacotherapy are discussed in
a non-judgmental way.
(2) Assessment
The pharmacist identifies potential drug-related problems (DRPs) and associated
recommendations by assessing patient, medical and medication data using
both implicit and explicit criteria. A simple implicit approach is to match each
of the patients conditions with the medications that the patient is taking. Areas
of mismatch can highlight drugs that are used with no indication, underused or
15
Fi g u re 3 The medication review process according to Dutch multidisciplinary

guideline Polypharmacy in the elderly
1. Patient interview
5. Follow-up and
monitoring
2. Assessment
(identifying DRPs)
3. Pharmaceutical
care plan
4. Implementation
of actions
DRP = drug-related problem
misused.25 As explicit criteria for inappropriate prescribing, STOPP (Screening

Tool of Older Persons Prescriptions) is recommended which contains 65 explicit
rules for avoidance of certain drugs/drug classes. In addition, START (Screening
Tool to Alert doctors to Right Treatment) lists 22 common instances of potentially
prescribing omissions.30,32,33
(3) Pharmaceutical care plan
During a case-conference (preferably face-to-face between pharmacist and
physician) potential DRPs and associated recommendations are discussed. In a
pharmaceutical care plan the actions are formulated to resolve DRPs. Actions are
prioritized based on the patients perspective, focusing on DRPs that causes the
most concern. The pharmaceutical care plan should contain detailed information
on who is responsible for implementation and evaluation of actions (e.g. GP,
pharmacist or practice nurse).
(4) Implementation of action
The responsible care provider explains to the patient why an action is required,
for example a change in the medication regimen. In concordance, patient and care
provider decide if this adaptation of the regimen can be implemented or should
be adjusted. The number of changes during each patient encounter should be
16
I nt ro d uc t io n
| C hapte r 1
restricted. It is recommended that the agreed actions are written on an information

leaflet as well as the provision of an updated medication list.
(5) Follow-up and monitoring
The responsible care provider verifies if all agreed actions are implemented. Where
necessary the patient is consulted once more to examine whether DRPs have been
resolved. If needed, new actions will be planned.
OB J E C T I V E
Dispensing services and medication reviews are relatively new clinical pharmacy
interventions to increase the appropriate use of medicines especially for older
patients with polypharmacy. This offers an opportunity for pharmacists to fulfil
their professional roles as health care providers in collaboration with physicians and
patients. There is limited evidence that interventions aimed at appropriateness of
medication reduce drug-related problems in older people. However, it is not clear if
these result in clinical improvements.34,35 The heterogeneity in patient populations,
settings, interventions and outcomes in earlier studies make it difficult to draw
definitive conclusions. Therefore, more research data are needed in order to clarify
the effects of these new clinical pharmacy services.
In this thesis, studies will be presented with the objective to describe the effects
of clinical pharmacy interventions for older patients with polypharmacy in
primary care. Apart from the focus of medication review on appropriateness of the
pharmacotherapy, this thesis will provide data on the impact of a dispensing service
(i.e. multi dose drug dispensing systems) on the medication management of older
patients.
OU T L I N E OF T H E T H E SI S
This thesis consists of three parts. Chapter 2 presents a systematic literature review
that investigates how the degree of collaboration between the general practitioner
(GP) and the pharmacist impacts on the implementation of recommendations
arising from medication review.
Chapter 3 focuses on the impact of multi dose drug dispensing systems on the
medication management of older patients and the appropriateness of medication
in these systems.
17
In Chapter 3.1 we compared the self-reported medication adherence and knowledge

of older patients receiving their drugs via multidose drug dispensing (MDD-users)
with patients receiving manually-dispensed drugs (non-MDD-users).
In Chapter 3.2 the results are presented of a pragmatic randomized controlled
study on the effect of medication review for older patients using multidose drug
dispensing (MDD) on DRPs.
Chapter 4 focuses on the impact of medication review on different outcomes
and in-depth analyses of the sources for identification of DRPs, the pharmacists
conducting medication reviews and the presence of explicit criteria in DRPs.
Chapter 4.1 describes the results of a randomised controlled trial on the effect of
Home Medication Review on DRPs, disease-specific outcomes (blood pressure,
LDL-cholesterol and HbA1c) and health-related quality of life in older patients.
Chapter 4.2 focuses on the contribution of the patient interview to the identification
of DRPs within the Home Medication Review.
Chapter 4.3 describes the completeness of DRPs identified by community
pharmacists.
Chapter 4.4 aims to describe the number and types of explicit STOPP-START
criteria present in identified potential DRPs and their implementation rate.
Finally, the results of these studies are summarized and put into a broader
perspective in Chapter 5.
REFERENCES
1.
Cerreta F, Eichler HG, Rasi G. Drug policy for an aging population - the European Medicines
Agencys geriatric medicines strategy. N Engl J Med 2012;367:1972-4.
2.
Hubbard RE, OMahony MS, Woodhouse KW. Medication prescribing in frail older people.
Eur J Clin Pharmacol 2012;Sep [epub].
3.
Gnjidic D, Hilmer SN, Blyth FM, Naganathan V, Cumming RG, Handelsman DJ, et al.
High-risk prescribing and incidence of frailty among older community-dwelling men. Clin
Pharmacol Ther 2012;91:521-8.
4.
Hajjar ER, Cafiero AC, Hanlon JT. Polypharmacy in elderly patients. Am J Geriatr
Pharmacother 2007;5:345-51.
5.
Stichting Farmaceutische Kengetallen (SFK). Polyfarmacie bij een op de tien

apotheekbezoekers. Available from: http://www.sfk.nl/nieuws-publicaties/PW/2012/
polyfarmacie-voor-1-op-de-10-apotheekbezoekers (Accessed 3 Jan 2013).
6.
Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and
quality of care for older patients with multiple comorbid diseases: implications for pay for
performance. JAMA 2005;294:716-24.
18
I nt ro d uc t io n
| C hapte r 1
7.
Leendertse AJ, Egberts AC, Stoker LJ, van den Bemt PM. Frequency of and risk factors for
preventable medication-related hospital admissions in the Netherlands. Arch Intern Med
2008;168:1890-6.
8.
Kongkaew C, Noyce PR, Ashcroft DM. Hospital admissions associated with adverse drug
reactions: a systematic review of prospective observational studies. Ann Pharmacother
2008;42:1017-25.
9.
Beijer HM, Blaey CJd. Hospitalisations caused by adverse drug reactions (ADR): a metaanalysis of observational studies. Pharm World Sci 2002;25:46-54.
10. Howard RL, Avery AJ, Slavenburg S, Royal S, Pipe G, Lucassen P, et al. Which drugs
cause preventable admissions to hospital? A systematic review. Br J Clin Pharmacol
2007;63:136-47.
11. Centraal Bureau voor de Statistek (CBS). Demografie van de vergrijzing, Bevolkingstrends
2e kwartaal 2011. Available from: http://www.cbs.nl/nl-NL/menu/themas/bevolking/
publicaties/artikelen/archief/2011/2011-3434-wm.htm (Accessed 3 Jan 2013).
12. Marek KD, Antle L. Medication Management of the Community-Dwelling Older Adult. In:
Hughes RG, editor. Patient Safety and Quality: An Evidence-Based Handbook for Nurses.
Rockville (MD); 2008.
13. MacKeigan LD, Nissen LM. Clinical pharmacy services in the home. Disease Management
and Health Outcomes 2008;16:227-44.
14. Maddigan SL, Farris KB, Keating N, Wiens CA, Johnson JA. Predictors of older adults
capacity for medication management in a self-medication program: a retrospective chart
review. J Aging Health 2003;15:332-52.
15. Drenth-van Maanen AC, van Marum RJ, Knol W, van der Linden CM, Jansen PA.
Prescribing optimization method for improving prescribing in elderly patients receiving
polypharmacy: results of application to case histories by general practitioners. Drugs Aging
2009;26:687-701.
16. van Geffen K, Vos I, van Horssen N, Winters N, Bouvy ML. Apotheker proactiever. Enquete:
rol in herhaalmedicatie groeit. Pharm Weekbl 2011;146(20).
17. de Hertog NJ, Donker AA, Smith R. Patient therapietrouwer met herhaalservice. Significante
verbetering bij zeven geneesmiddelgroepen. Pharm Weekbl 2011;146(20).
18. The Scottish Government. Establishing Effective Therapeutic Partnerships - A
generic framework to underpin the Chronic Medication Service element of the
Community Pharmacy Contract. 2009. Available from: http://www.scotland.gov.uk/
Publications/2010/01/07144120/0 (Accessed 3 Jan 2013).
19. van Wijck F. Populariteit Baxterrol vraagt om goede leidraad. Pharm Weekbl 2011;146(21).
20. Wekre LJ, Spigset O, Sletvold O, Sund JK, Grimsmo A. Multidose drug dispensing and
discrepancies between medication records. Qual Saf Health Care 2010;19:e42.
21. Johnell K, Fastbom J. Multi-dose drug dispensing and inappropriate drug use: A nationwide
register-based study of over 700,000 elderly. Scand J Prim Health Care 2008;26:86-91.
22. Shaw J, Seal R, Pilling M. Task force on medicines partnership and the national collaborative
medicines management services programme. Room for review: a guide to medication review.
2002. Available from: http://www.npc.nhs.uk/review_medicines/intro/resources/room_for_
review.pdf (Accessed 1 Feb 2013).
19
23. Clyne W, Blenkinsopp A, Seal R. National prescribing centre. A guide to medication review.
2008. Available from: http://www.npc.nhs.uk/review_medicines/intro/resources/agtmr_
web1.pdf (Accessed 1 Feb 2013).
24. Blenkinsopp A, Bond C, Raynor DK. Medication reviews. Br J Clin Pharmacol
2012;74:573-80.
25. Steinman MA, Hanlon JT. Managing medications in clinically complex elders: Theres got to
be a happy medium. JAMA 2010;304:1592-601.
26. Fiss T, Dreier A, Meinke C, van den Berg N, Ritter CA, Hoffmann W. Frequency of
inappropriate drugs in primary care: analysis of a sample of immobile patients who received
periodic home visits. Age Ageing 2011;40:66-73.
27. Castelino RL, Bajorek BV, Chen TF. Retrospective evaluation of home medicines review by
pharmacists in older Australian patients using the medication appropriateness index. Ann
28. Bluml BM. Definition of medication therapy management: development of professionwide
consensus. J Am Pharm Assoc 2005;45:566-72.
29. Leikola SN, Virolainen J, Tuomainen L, Tuominen RK, Airaksinen MS. Comprehensive
medication reviews for elderly patients: findings and recommendations to physicians. J Am
Pharm Assoc 2012;52:630-3.
30. Vermeulen Windsant-van den Tweel AM, Verduijn MM, Derijks HJ, van Marum RJ.
Detection of inappropriate medication use in the elderly; will the STOPP and START criteria
become the new Dutch standards? Ned Tijdschr Geneeskd 2012;156:A5076.
31. Cipolle RJ. Pharmaceutical care practice: the clinicians guide. 2nd ed. New York: McGrawHill, Medical Pub. Division; 2004.
32. Gallagher P, Ryan C, Byrne S, Kennedy J, OMahony D. STOPP (Screening Tool of Older
Persons Prescriptions) and START (Screening Tool to Alert doctors to Right Treatment).
Consensus validation. Int J Clin Pharmacol Ther 2008;46:72-83.
33. Gallagher P, OMahony D. STOPP (Screening Tool of Older Persons potentially
inappropriate Prescriptions): application to acutely ill elderly patients and comparison with
Beers criteria. Age Ageing 2008;37:673-9.
34. Patterson SM, Bradley MC, Kerse N, Cardwell CR, Hughes CM. Interventions to improve the
appropriate use of polypharmacy for older people. Cochrane Database Syst Rev. 2012;5.
35. Holland R, Desborough J, Goodyer L, Hall S, Wright D, Loke YK. Does pharmacistled medication review help to reduce hospital admissions and deaths in older people?
A systematic review and meta-analysis. Br J Clin Pharmacol 2008;65:303-16.
20
A B ST R AC T
B ackg roun d
Many studies have investigated the effect of medication review on a variety of
outcomes, but the elements of the interventions have been quite diverse. Moreover,
implementation rates of recommendations also vary widely between studies.
Aim
The objective of this study was to investigate how the extent of collaboration between
the general practitioner (GP) and the pharmacist impacts on the implementation of
recommendations arising from medication review.
Me tho ds
MEDLINE, EMBASE and Web of Science were searched for studies published
between January 2000 and April 2012. Keywords included medication review,
medication therapy management, pharmaceutical services and drug utilization
review. Sixteen articles (describing 14 randomized controlled trials [RCTs]) out of
620 titles met the inclusion criteria. Inclusion criteria for the review were medication
review, RCT design, involvement of both pharmacist and GP, and communitydwelling patients (mean age > 70 years) who had not been recently discharged.
After quality assessment of the article, the presence of the following eight key
elements reflecting collaboration were scored for each intervention: pharmacist
with clinical experience, own pharmacist involved, sharing of medical records,
patient interview by pharmacist, invitation of patients by GP, case conference
between GP and pharmacist, action plan, follow-up. The primary outcome was
the implementation rate of recommendations. Meta-regression analysis was used
to assess the association between the implementation rate and the number of key
elements present.
Resu lts
Twelve RCTs were included after quality assessment. The mean number of key
elements within the intervention was 5.2 (range 18). The mean implementation
rate of recommendations was 50% (range 1786). The association between the
number of key elements present in the intervention and the implementation rate of
recommendations was significant: = 0.085 (95%CI 0.0520.128; P < 0.01).
24
G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew
| C hapte r 2
C on clu si on
This systematic review shows a significant association between the number of key
elements of the intervention reflecting collaborative aspects in medication review
and the implementation rate of recommendations.
I N T ROD U C T ION
Polypharmacy and drug-related morbidity is increasingly recognized as a major
public health problem among the elderly.1,2 Medication review has been proposed
as an important strategy to constrain the negative effects of polypharmacy, aiming
at safer and more effective use of medicines.3,4
Medication review has been defined as a structured, critical examination of a
patients medicines with the objective of reaching an agreement with the patient
about treatment, optimising the impact of medicines, minimizing the number
of medication-related problems and reducing waste.5 Three types of medication
review have been described, based on the purpose of the review: prescription
review technical issues related to prescription(s), concordance or compliance
review (issues relating to the patients medicine behaviour) and clinical medication
review (issues relating to the patients use of medicines in the context of their
condition).6 Concomitantly, efforts have been made to standardize medication
review.5-7
However, systematic reviews of pharmacist-led medication review have not shown
an effect on clinical outcomes such as hospital admissions or mortality.8-10 In
some studies, positive effects were reported on intermediate outcomes like drug
knowledge and adherence.8 The heterogeneity in patient populations, settings,
interventions and outcomes in these studies made it difficult to draw definitive
conclusions. There may be merit in combining the expertise of the pharmacist and
physician with shared decision-making involving the patient in order to improve
outcomes.11 Previous systematic reviews did not take into account the variability in
collaboration between pharmacists and general practitioners (GPs) in medication
reviews.
Studies on barriers and facilitators in medication review reveal collaborative
aspects that might be essential for conducting successful medication reviews.12-14
The most commonly cited facilitators were having an established pharmacist
physician relationship 13,14 and a face-to-face meeting (case conference) between
pharmacist and physician to discuss the pharmacists recommendations.13,15,16
Using a pharmacist other than the patients regular pharmacist was seen as a barrier,
25
as was inadequate clinical training of the pharmacist.13 Without access to medical

records, the pharmacist may make tentative or inappropriate recommendations
that are of little help.13 The GPPC (General PractitionerPharmacist Collaboration)
study 17 further suggested that a general practice-based service could be more
facilitating than a community pharmacy-based service.17 This could imply that
patients are approached for medication review by the GP practice, which is also
common in the Home Medicines Review (HMR) programme in Australia.12,13 They
further suggested that the pharmacist should meet the patient for interview about
their medicines in the physicians office,17-19 while a patient interview at home by
an accredited pharmacist is the predominant step of the HMR programme.12,13
Finally, it is important for a collaborative medication review that responsibilities
for implementation of the action plan and follow-up are clearly defined and divided
between physician and pharmacist.13
The aim of this systematic review was to investigate how the degree of collaboration
between the GP and the pharmacist impacts on the implementation of
recommendations arising from medication review.
M E T HOD S
S e arch
Our search strategy identified research on medication review interventions
involving pharmacists and GPs. MEDLINE, EMBASE and Web of Science were
searched for articles published between 1 January 2000 and 1 April 2012. These
dates were chosen because relatively few studies with an elaborate description of
the medication review process were published before 2000. Interventions were
identified using the following keywords and medical subject headings (MeSH):
medication review, medication therapy management, pharmaceutical services
and drug utilization review (see Appendix I for detailed search terms). Different
publications on the same group of patients were considered as one study.
Stu dy s el e c t i on
All titles were reviewed by two investigators (H.K. and L.B.). Studies were excluded
if both agreed that the title clearly indicated that the study did not concern
medication review and/or focussed on only one drug or drug class. H.K. and
L.B. assessed all remaining abstracts independently in this manner. Studies were
included if they fulfilled the following criteria: medication review, randomized
26
| C hapte r 2
clinical trial (RCT), pharmacist and GP involved, community-dwelling patients in

primary care, mean age 70 years, patients not recently discharged (< 1 month).
Only studies in English were included. Finally, full papers from potential studies
were assessed independently by the two investigators for their suitability for
inclusion. Differences were resolved by discussion, or a third investigator (either
A.F. or M.B.) was consulted.
Q u a l ity ass essment of th e stu di e s
Trial quality was assessed according to the Delphi list.20 This list consists of ten
criteria: randomization, treatment allocation, similar groups at baseline, eligibility
criteria, blinding of outcome assessor, blinding of care provider, blinding of
patient, point estimates and measures of variability, intention-to-treat analysis and
reporting of withdrawal/drop-out rate. In addition, we added power calculation to
this list. Studies with a low score on the quality assessment (5 or fewer items scored
yes) were excluded for analysis of outcomes.
Stu d y ch ar ac ter i st i cs
C ateg or i z ation
Studies were categorized by study author, year of publication, number of
pharmacists and GPs, country, number of patients, duration of the study, mean age
and sex of patients, mean number of drugs, description of the intervention, setting,
number of recommendations in the intervention group, the clinical, intermediate
and process outcomes assessed and the quality score.
O utcomes
The primary outcome was the implementation rate of recommendations
following drug-related problems (DRPs) identified during medication review.
The implementation rate was defined as the percentage of recommendations
fully or partly implemented and/or the percentage of DRPs resolved. Partial
implementation of recommendations means that an action other than that
originally proposed by the pharmacist was implemented. Fully and partly
implemented recommendations were counted equally. Data on clinical outcomes
(hospital admissions, quality of life), intermediate outcomes (adherence) and other
process outcomes (drug changes, number of drugs) were also extracted. The effect
on clinical, intermediate and process outcomes was described as a significant effect
in favour of the intervention group, a significant effect in favour of the control
group or no significant effect.
27
Ke y eleme nts of the inte r ve ntion

The intervention was characterized by the presence or absence of eight key elements
reflecting collaborative aspects between a GP and a pharmacist, based on the
aforementioned facilitators and barriers in medication review.12-14 The choice of the
key elements was supported by scientific discussion with experienced pharmacist
reviewers who regarded these elements as having face validity. The following
key elements were assessed: (1) pharmacist with clinical experience means that
the study pharmacist had adequate clinical training and expertise to perform
medication reviews; (2) own pharmacist involved means that the study pharmacist
is the patients regular pharmacist who has a longer lasting therapeutic relationship
with his or her patient; (3) sharing of medical records describes full access for
the care provider performing the medication review to GP data on diseases of the
patient and clinical values; (4) patient interview by pharmacist means a face-toface consultation between a pharmacist and a patient this pharmacist must have
a relationship with the GP; (5) invitation of the patients by GP means that the
patient is invited to the study or referred for medication review by the GP (practice);
(6) case conference GP and pharmacist indicates a face-to-face meeting between
at least the GP and the pharmacist to discuss the DRPs and recommendations for
specific patients; (7) action plan means that the study investigators reported that
the agreed recommendations were formulated as an action plan and that there were
designated persons responsible for implementation of this plan; and (8) follow-up
has taken place to assess whether the actions have been implemented, and to assess
the patients experience with these actions.
D ata sy nth esis an d an a lysis
For each trial, we extracted data on the primary outcome, implementation
rate. When the implementation rate was not present, we derived this rate from
the percentage of DRPs resolved or the decrease in the number of potentially
inappropriate prescriptions (PIPs). Trial quality and key elements of interventions
were assessed independently by the two investigators (H.K. and L.B.) for each
included study. Differences were resolved by discussion, or a third investigator
(either A.F. or M.B.) was consulted.
Meta-regression analysis was used to assess the association between the number of
key elements and the implementation rate, with the number of recommendations
in the different studies as possible effect moderator. This mixed-effects analysis was
conducted using the metafor statistical package in R (version 2.12.2, R Project for
Statistical Computing, Vienna, Austria, 2011, http://www.R-project.org).
28
| C hapte r 2
R E SU LT S
S e arch resu lts
A total of 620 titles were identified, 16 of which (describing 14 RCTs) met the
inclusion criteria and were included in this review (Figure 1).3,15-17,19,21-32
Fi g u re 1 Flow chart describing study selection and excluded studies
Potentially relevant publications

identified and titles screened
n = 620
Excluded: 467
because intervention is not
medication review, and/or
targeting specific drugs
Abstracts of potential
publications screened
n = 153
Potentially appropriate
publications
n = 22
describing 20 RCTs
Excluded: 131
no RCT (95)
secondary or tertiary care (20)
targeting specific disease (8)
discharge (7)
age too low (1)
Excluded: 6
age too low (3)
targeting specific drug classes (2)
medication review only part of intervention (1)
Publications included
in review
n = 16
describing 14 RCTs
RCTs = randomized controlled trials
Q u a l ity ass essment of stu d i es

The methodological quality of 12 of the 14 studies was assessed as adequate (i.e. 6
or more of 11 items scored yes)3,15-17,19,21,23-27,29-32 (see Table 1). The quality of the
trial in the two remaining studies was scored as low 22,28 and they were therefore
29
30
Method of
randomisation
performed
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Allard (2001)
Bernsten (2003)
Bryant (2011)
Denneboom (2007)
Grymonpre (2001)
Krska (2001)
Kwint (2011)
Lenaghan (2007)
Sellors (2003)
Sorensen (2004)
Sturgess (2003)
Volume (2001)
Williams (2004)
Zermansky (2001)
No
Yes
Yes
Yes
Not clear
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Not clear
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Not clear
Not clear
Not clear
No
Not clear
Not clear
Not clear
Not clear
Not clear
No
Not clear
Yes
No
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Treatment Similar
Eligibility Outcome Care
allocation groups at criteria
assessor provider
concealed baseline specified blinded
blinded
Quality assessment of studies
Study author
Table 1
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Yes
Not clear
Yes
No
No
Yes
No
No
No
Yes
Point of
Including
estimates intentionand mea- to-treat
sures of
variability
Not clear Yes
No
Yes
Yes
No
No
No
Patient
blinded
Yes (5%)
Yes (5%)
Yes (20%)
No (42%)
Yes (25%)
Yes (11%)
Yes (2%)
Yes (9%)
Yes (13%)
Yes (16%)
Yes (7%)
No (48%)
No (34%)
Yes (9%)
Withdrawal
rate (%)
unlikely to
cause bias?
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
Yes
Power
Quality
calculation Score
reported
| C hapte r 2
excluded from further analysis. These studies were related; one trial formed part of
a larger co-ordinated project with more countries, which was described in the other
paper.22,28
In all included studies, a method of randomization was performed and eligibility
criteria were specified. The majority of studies reported a method of treatment
allocation,3,15-17,22-26,29-31 while in three studies this was either not clearly described or
not conducted.19,21,27,32 All except two studies reported similar groups at baseline.28-30
An independent outcome assessor who was blinded to the intervention allocation
was clearly described in only two studies.17,21 Because of the nature of the studied
intervention, the care provider was never blinded to the intervention allocation.
The patient was blinded for the intervention allocation in three studies.3,15,16,23 In
two of these studies, patient interviews were conducted for both the intervention
and the control group, but a pharmaceutical care plan was implemented only for the
intervention group.3,23 In the third study, no patient interview was conducted and
there was no description of patient involvement with the study.15,16 Point estimates
and measures of variability were described in all studies. Intention-to-treat analysis
was conducted in five studies.19,21,23,25,27,32 The withdrawal rate was likely to have
caused bias in 3 of 14 studies.3,15,16,19,21,23-27,29-32
Stu d y ch ar ac ter i st i cs
In Table 2, the study characteristics and outcomes are presented for the 12 included
studies. The number of participants in these studies ranged from 118 to 1,188. The
mean age of the participants was 76.6 years (range 71.884.3) and 66% were females
(range 5690). The mean number of prescribed drugs was 7.2 (range 4.512). Three
of these studies were performed in the US,23,29-31 three in the UK,3,19,25,32 and two in
The Netherlands.15,16,24
O utcomes
Seven of 12 studies provided data on clinical outcomes.3,17,19,25-27,29,32 Six of these
studies reported on quality of life measured using the 36-item Short Form Health
Survey (SF-36)3,17,19,26,27,29,32 or the EuroQOL-5D/visual analogue scale (VAS).25 No
effects were found on total scores for quality of life, and one study reported only
negative effects on some domains in one study.17 Data on hospital admissions were
provided by four studies,19,25-27,32 and no significant effects were reported. Twelve
studies provided data on intermediate outcomes. Two studies reported on adherence,
either self-reported 29,30 or measured by refill rate,23 with no effect. Two studies
reported on DRPs in both the intervention and the control group, with positive
effects on DRPs resolved.3,24 Two studies 17,21 reported on potentially inappropriate
31
32
No. of
pharmacists
and GPs
(country)
Team of 2
physicians,
pharmacist
and nurse;
52 GPs
(Can)
26 CPs;
57 GPs
(NZ)
29 CPs;
84 GPs
(NL)
Pharmacists
NS;
35 GPs
(US)
Pharmacists
NS, 6 general
practices
(UK)
Allard et al.21
(2001)
Bryant et al.17
(2011)
Denneboom et
al.15
(2007)
Grymonpre et
al.23
(2001)
Krska et al.3
(2001)
332 (60.5%);
75.1
135 (79.3%);
77.0
738 (62.1%);
81.0
498 (59.0%);
75.4
266 (90%);
80.5
Patients
n (%F);
mean age
in years
3 mo
12 mo
9 mo
12 mo
12 mo
Duration
7.5
6.2
7.2
5.0
6.3
No. of
drugs
Home-based MR. Care

plan sent to GP, agreed
actions implemented by
pharmacist assisted by
practice staff (general
practices)
Home medication
history taken by lay
person and reviewed by
pharmacy consultant (one
interdisciplinary health
clinic)
MR without pt interview
(treatment review)
followed by CC with GP
(study arm A) or written
feedback (study arm B)
(community pharmacies)
Home MR followed by
CC with GP (community
pharmacies)
Nurse met pts at home

for inventory of drugs.
MR was performed by a
multidisciplinary team and
mailed to GPs (Faculty of
Medicine, University)
Intervention (setting)
1,206 (6.3)
794 (11.5)
(A) 141 (0.5)

(B) 128 (0.3)
462 (1.7)
147 (1.1)
No. of
recommendations (mean
per pt)
QoL (SF-36) (
QoL (SF-36):
emotional role ( );
social functioning
( ); other domains
( )
Clinical outcomes
(significant effect)
Description of included studies: effect on clinical, intermediate and process outcomes
Study
(year of
publication)
Table 2
Proportion of PCIs
resolved ( ); use of
health services ( )
Adherence (refill
rate) ( ); drug
knowledge ( );
no. of prescribed
drugs ( )
No. of drug
changes: 6 months
( ); 9 months ( )
MAI ( ); no. of PIMs

( ); no. of drug
changes ( )
No. of PIMs ( );
no. of pts with
1 PIM ( ); no.
of pts with drug
improvements ( );
no. of prescribed
drugs ( )
Intermediate and
process outcomes
Quality
score
No. of
pharmacists
and GPs
(country)
6 CPs;
GPs NS
(NL)
1 CP;
9 GPs
(UK)
24 CPs;
48 GPs
(US)
32 CPs;
84 GPs
(Aus)
Kwint et al.24
(2011)
Lenaghan et al.25
(2007)
Sellors et al.26
(2003)
Sorensen et al.27
(2004)
400 (63.8%);
71.8
889 (62.8%);
74.0
136 (65.7%);
84.3
118 (68.5%);
79.3
Patients
n (%F);
mean age
in years
6 mo
5 mo
6 mo
6 mo
Duration
8.0
7.9
7.4
10.1
No. of
drugs
Home visit and MR by

pharmacist and team
conferences with GP
and implementation at
next pt/GP visit (general
practices and community
pharmacies)
Clinic-based MR, then

telephone interviews
(family practices in 24 sites)
Home-based MR (one
dispensing general
practice)
MRs without pt interview

were performed by external
pharmacist reviewers at
distance, but the following
CC with GP was conducted
by pts own CPs (community
pharmacies)
564 (3.2)
1,093 (2.5)
71 (1.0)
249 (4.0)
No. of
per pt)
No. of hospital
admissions ( );
no. of hospital
services ( );
QoL (SF-36) ( );
severity of illness
(DUSOI-A) ( );
ADEs ( )
No. of hospital
admissions ( );
no. of drugrelated hospital
admissions ( );
QoL (SF-36) ( )
No. of hospital
admissions ( );
deaths ( ); QoL
(EQ-5D/VAS) ( )
Clinical outcomes
Study
(year of
publication)
Table 2
Table 2 continued
No. of daily units of

drugs taken ( );
no. of drugs taken
(
No. of prescribed
drugs ( )
No. of DRPs
leading to
recommendation
for drug change
( ); no. of drug
changes related to
a recommendation
( ); no. of drug
changes ( )
Intermediate and
process outcomes
Quality
score

| C hapte r 2
33
34
5 CPs;
GPs NS
(US)
1 CP,
1 GP
(US)
1 clinical
pharmacist;
4 GPs
(UK)
Volume et al.29
(2001)
Williams et al.31
(2004)
Zermansky et al.32
(2001)
1188 (56%);
73.5
140 (57.1%);
73.7
363 (66.9%);
74.0
Patients
n (%F);
mean age
in years
12 mo
6 wk
15 mo
Duration
4.7
12.0
4.5
No. of
drugs
One clinical pharmacist

held pt consultations in GP
office followed by clinical
MR (4 general practices)
Clinical MR (one health

centre ambulatory clinic)
Pharmaceutical care
services including
clinical MR (community
pharmacies)
502 (0.8)
257 (4.1)
559 (3.5)
No. of
per pt)
)
No. of hospital
admissions ( )
QoL (SF-36) (
Clinical outcomes
No. of drug
changes ( );
frequency of
dose ( ); no.
of prescribed
drugs ( ); no. of
hospital outpatient
attendances
( ); no. of GP
consultations ( )
No. of prescribed
drugs ( )
Adherence (selfreported) ( )
Intermediate and
process outcomes
Quality
score
GP = general practitioner; F = female; pt(s) = patient(s); Can = Canada; MR = medication review; PIMs = potentially inappropriate medications; CP = community pharmacist;
NZ = New Zealand; CC = case conference; QoL = quality of life; MAI = Medication Appropriateness Index; NL = Netherlands; NS = not specified; US = United States; UK =
United Kingdom; PCIs = pharmaceutical care issues; DRP = drug-related problem; VAS = visual analogue scale; Aus = Australia; SF-36 = 36-item Short Form Health Survey;
DUSOI-A = Dukes Severity of Illness Visual Analogue Scale; ADEs = adverse drug events
significant effect in favour of intervention group; significant effect in favour of control group; no significant difference
No. of
pharmacists
and GPs
(country)
Study
(year of
publication)
Table 2
Table 2 continued
| C hapte r 2
medications (PIMs), with positive effects for one study.17 Process outcomes
were reported in all studies. Two studies reported a reduction in the number of
(prescribed) drugs,19,31,32 while in four studies no effect was reported.21,23,25,26 Five
studies reported an increase in the number of drug changes.15-17,19,24,31,32
Implementation rates of recommendations in the intervention group are shown
in Table 2. The percentage of implemented recommendations was reported in
seven studies,15-17,19,23,25-27,32 while a percentage of resolved DRPs was mentioned in
three studies.3,24,29,30 In two studies, the implementation rate was derived from the
decrease in the number of PIPs compared with the total number of PIPs.21,31
Ke y eleme nts of the inte r ve ntion
Key elements of 13 interventions from the 12 studies are shown in Table 3. One
study compared outcomes between two intervention groups (case conference
and written feedback)15,16 and therefore both study arms (A and B) are shown.
Pharmacists had clinical experience in 10 of the 13 interventions.3,17,19,23,25-27,29-31
Pharmacists were accredited pharmacists,17,27 consultant pharmacists 23,31 or clinical
pharmacists.19,32 They followed a university accredited externship programme,26
were clinically trained,3 experienced in medication reviews 24,25 or had a postgraduate qualification in pharmacy practice.25 In 8 of 13 interventions, the
patients own GP was involved.15-17,25-27,29,30 In the other interventions, the study
pharmacist had no existing therapeutic relationship with the patient or this was not
described 3,19,21,23,31,32 (see also Table 1). Pharmacists had full access to GPs medical
records of the patient in 8 of the 13 interventions.3,17,19,25-27,29,32 Patient interviews
were conducted in 11 of 13 interventions, at home 3,17,21,23,25,27 or in the GPs office
or clinic.19,26,31,32 In 3 of 13 interventions, eligible patients were invited by the GP to
participate in the study.17,19,27,32 Case conferences between GPs and pharmacists were
conducted in 7 of 13 interventions.15-17,19,24-27,32 In three interventions, letters with
recommendations or care plans were sent to the GPs (written feedback).3,15,16,23 As
mentioned earlier, one study compared the process outcomes of case conferences
with written feedback.15,16 As part of two interventions, case conferences were
held by external multidisciplinary teams without the patients own GP 21,31 and
recommendations were mailed to the GP 21 or implemented with endorsement of
the GP.31 Action plans were used for implementation of agreed recommendations in
9 of 13 interventions.3,15-17,19,24-27,29,32 A follow-up of the implementation of actions
was described in 11 of 13 interventions,3,15-17,19,21,23,25-27,29,31,32 most often conducted
by a pharmacist.15,17,19,23,25,26,31,32
35
36
Implementation rate
h/CP
GPO/CP
NS/CP
h/CP
h/P
h/CP
GPO/P
Sorensen et al.27 (2004)
Sellors et al.26 (2003)
Volume et al.29 (2001)
Bryant et al.17 (2011)
Krska et al.3 (2001)
Lenaghan et al.25 (2007)
Zermansky et al.32 (2001)
GP+O
P+O
NS
GP
GP+CP
CP
CP
CP
GP
CP
86%
85%
83%
62%
58%
56 %
54%
33%
30%
30%
29%
GP = general practitioner; h = patients home; O = other care provider; P = pharmacist; NS = not specified; CP = community pharmacist; GPO = GPs office
NS/CP
NS
Kwint et al.24 (2011)
Williams et al.31 (2004)
NS
Denneboom et al.15 study arm A

(2007)
Follow
up (care
provider)
h/P+O
Action
plan (care
provider)
Grymonpre et al.23 (2001)
Case
conference
GP and
pharmacist
25%
Invitation
by GP
h/O
Patient
interview
(location/care
provider)
Allard et al.21 (2001)
Access to
medical
records
17%
Own
pharmacist
involved
Denneboom et al.15 study arm B

(2007)
Pharmacist
with
clinical
experience
Key elements of intervention versus implementation rate of recommendations
Study
(year of publication)
Table 3
12
NS
12
NS
NS
1.5
NS
12
Outcome
measurement
(months)
| C hapte r 2
Ass o ci ati on b et we en numb er of ke y el em ent s an d outc om e s

Key elements of the intervention and the implementation rate are shown in Table 2.
The mean number of key elements present in the interventions was 5.2 (range 18).
The mean implementation rate was 50% (range 1786).
The association between the number of key elements present in the intervention
and the implementation rate of recommendations was positive: an increase in
number of key elements was related to an increase in implementation rate metaregression, = 0.085 (95%CI 0.0520.128; P < 0.01) (Figure 2). Figure 3 is a
forest plot showing the observed and expected implementation rates, estimated
on the meta-analysis association between number of key elements and the
implementation rate. In all but three interventions, the observed implementation
rate was within the 95% confidence interval of the expected value.3,19,27,32
No meta-regression analyses were possible for the association between the number
of key elements and the number of hospital admissions (n = 4 studies), the number
of drug changes (n = 5 studies) and the number of prescribed drugs (n = 5 studies),
because of the low number of studies and participants with these outcomes.
DI S C U S SION
This systematic review shows a significant association between the number of key
elements of the intervention reflecting collaborative aspects in medication review
and the implementation rate. This suggests that more intensive collaboration
between GP and pharmacist in medication review leads to higher recommendation
implementation rates.
The expected implementation rate could be predicted from the number of key
elements estimated from this association (Figure 3). This model gives a good
prediction of the implementation rate for the majority of the studies. For three
studies, the expected implementation rate was different from the observed
value.3,19,27,32 A higher implementation rate than expected was observed in the
studies by Krska et al.3 and Zermansky et al.19,32 In the study by Zermansky et
al.,19,32 one pharmacist collaborated with only a few GPs, similar to the study by
Lenaghan et al.,25 resulting in comparably high implementation rates. The major
difference with the Lenaghan study was that the patients own pharmacist was not
involved. However, it is conceivable that this (clinical) pharmacist established a
good relationship with this small number of GPs and their patients whilst he was
consulting patients in the GPs office. Studies with pharmacists working at a GP
practice yielded high rates of acceptance of recommendations.18,19,26,32 Conversely,
37
Kwint et al.
(2011)
No. of elements
Denneboom et al.
study arm A (2007)
Williams et al.
(2004)
The size of the circles reflects the number of recommendations in the intervention group of the different studies.
Denneboom et al.
study arm B (2007)
Allard et al.
(2001)
Grymonpre
et al. (2001)
y = 0.0538 + 0.0850x
Volume et al.
(2001)
Krska et al.
(2001)
Sellors et al.
(2003)
Zermansky
et al. (2001)
Fig ure 2 Bubble plot of number of key elements of intervention vs. implementation rate of recommendations
Implementation rate (%)
80
60
40
20
38
8
Sorensen
et al. (2004)
Bryant et al.
(2011)
Lenaghan
et al. (2007)
30 (2237 )
33 (2839 )
54 (5059)
56 (5359 )
58 (5462)
62 (5766 )
83 (8185 )
85 (7693)
86 (8389)
Denneboom et al. study arm A (2007)
Williams et al. (2004)
Sorensen et al. (2004)
Sellors et al. (2003)
Volume et al. (2001)
Bryant et al. (2011)
Krska et al. (2001)
Lenaghan et al. (2007)
Zermansky et al. (2001)
Mean observed implementation rate with confidence intervals (95%CIs). The grey diamonds represent the 95%CIs of the expected implementation rate estimated on the
meta-regression analysis association between the number of key elements and the implementation rate.
30 (2435 )
Kwint et al. (2011)
100
29 (2632 )
Grymonpre et al. (2001)
25
50
75
Implementation rate (%)
25 (1832 )
Allard et al. (2001)
17 (1124 )
Denneboom et al. study arm B (2007)
Study (year of publication)
Fig ure 3 Forest plot of observed and expected implementation rate of recommendations

| C hapte r 2
39
Krska et al.3 was the only study with a high implementation rate without a case
conference. This face-to-face meeting between GPs and pharmacists to discuss
the pharmacists recommendations is often considered one of the most important
and key elements of the collaborative approach in medication review.13,15,16,33
In the study by Krska et al.,3 pharmacists were assisted by practice staff in the
implementation of accepted actions. Possibly, this partly explains the high
implementation rate. Furthermore, the nature and number of pharmacists and
their relationship with patients was not specified.3 On the other hand, a lower than
expected implementation rate was observed in the intervention by Sorensen et
al.27 In this study, a large number of pharmacists collaborated with an even larger
number of GPs, which could have made it difficult to achieve high implementation
rates. The implementation rate and numbers of GPs and pharmacists in Sorensen
et al.27 were similar to those in the study by Sellors et al.26 The major difference was
that, in Sorensen et al.27, patients were invited by the GP.
There have been no earlier systematic reviews investigating the implementation
rate of recommendations. We found a significant association between the number
of key elements reflecting collaborative aspects and the implementation rate. This
finding is in agreement with other medication review studies in secondary and
tertiary care where direct communication between healthcare providers revealed
higher acceptance rates of recommendations.34,35 For clinical and intermediate
outcomes, no association could be assessed, because the number of studies
reporting these outcomes was too low. Earlier systematic reviews reported no effect
on hospital admissions and quality of life.8,10 Compared with these reviews, our
scope was more focused, as we included only RCTs, community-dwelling patients
in primary care, a mean age of 70 years and no recent discharge, yielding only 12
trials after quality assessment. However, due to our inclusion criteria, patients in
our review were relatively healthy and not directly at risk for hospital admission.
In contrast, studies on medication review in hospitals were more successful in
preventing hospital (re)admissions because they generally reviewed patients who
were admitted to hospital and at high risk for readmission.36,37
There may be additional reasons for low implementation rates. Sellors et al.26
showed that these reasons might include patient reluctance, previous failed
attempts at the same strategy and a relatively short period for implementation
combined with the occurrence of more urgent issues. In particular, the periods over
which implementation rates were measured varied between the different studies
in this review. Nor do we know if the recommendations in the different studies
were clinically appropriate.17 GPs perceptions of pharmacists recommendations
in the GPPC study revealed that they generally found the recommendations
40
| C hapte r 2
useful although at times theoretical.38 Pharmacists recommendations may be less

appropriate if a high proportion of patients are already receiving the recommended
treatment, for example, in the MEDMAN study (ceiling effect).39
There were several strengths to this study. First of all, like in other systematic
reviews on medication review, trials reported very heterogeneous outcomes that
could not be pooled. In this systematic review, we could compare different trials
using the implementation rate as the common (process) outcome. Implementation
rates are also reported in many (observational) studies in community-dwelling
patients 40,41 as well as for patients in nursing homes.34,42 Implementation rate
has a greater significance than acceptance rate of pharmacists recommendations
by physicians because it includes enactment of the recommendation by a care
provider and the level of acceptance of recommendations by the patient. Secondly,
we described eight different elements of the intervention that reflect collaborative
aspects between GPs and pharmacists. These key elements were based on described
facilitators and barriers in medication review.12-14 Thirdly, the importance of patient
involvement in medication review was also reflected by the key elements own
pharmacist involved, patient interview and follow-up.
Our decision to consider all eight key elements of the intervention as equivalent
weighted determinants for the implementation rate could be seen as a limitation.
For example, the face-to-face discussion between pharmacists and GPs seemed a
key element that could have more weight. The small number of studies precluded
us from studying the association of the individual key elements in a multivariate
design. Also, there could have been other key elements reflecting collaborative
aspects that we may have missed. Furthermore, it was not possible to discriminate
between the clinical relevance of the implemented recommendations in the different
studies. This clinical relevance was only described in the study by Denneboom
et al.15,16 Finally, it cannot be ruled out that some relevant RCTs may have been
missed or excluded.
C ON C LU SION
This systematic review showed that the number of key elements reflecting
collaborative aspects of medication review was significantly associated with the
recommendation implementation rate. Further clinical trials could demonstrate
whether an increase in collaborative aspects leads to higher implementation rates.
Based on this model, future studies for elderly in primary care could consider these
key elements of intervention to design a standardized medication review process.
41
More research is needed to assess which key elements of this collaborative approach
are the most important and if there are additional elements that may influence
implementation rates. Next to the physician and the pharmacist, the patient is the
third main player in the medication review process. Future studies could focus on
the influence of the patient on the implementation rate. Large multicentre trials in
primary care are needed to draw definitive conclusions on whether a standardized
collaborative approach in medication review could affect clinical outcomes. Such
trials may be expensive, difficult to organize in practice settings and it may be
questioned how many and which elderly community-dwelling patients in primary
care are at greatest risk for negative clinical outcomes.
Ack n ow l e dgements - The authors thank J.C. Riemens-Louisse, student, for her
contribution to the data extraction forms and quality assessment forms and S.V. Belitser
for her contribution to the statistical analysis.
REFERENCES
1.
Rollason V, Vogt N. Reduction of polypharmacy in the elderly: a systematic review of the role
of the pharmacist. Drugs Aging 2003;20:817-32.
2.
Castelino RL, Bajorek BV, Chen TF. Targeting suboptimal prescribing in the elderly: a review
of the impact of pharmacy services. Ann Pharmacother 2009;43:1096-1106.
3.
Krska J, Cromarty JA, Arris F, Jamieson D, Hansford D, Duffus PRS, et al. Pharmacist-led
medication review in patients over 65: a randomized, controlled trial in primary care. Age
Ageing 2001;30:205-11.
4.
Hanlon JT, Schmader KE, Ruby CM, Weinberger M. Suboptimal prescribing in older
inpatients and outpatients. J Am Geriatr Soc 2001;49:200-9.
5.
Shaw J, Seal R, Pilling M. Task Force on Medicines Partnership and the National
Collaborative Medicines Management Services Programme. Room for review: a guide to
medication review. 2002. Available from: http://www.npc.nhs.uk/review_medicines/intro/
resources/room_for_review.pdf (Accessed 1 Feb 2013).
6.
Clyne W, Blenkinsopp A, Seal R. National Prescribing Centre. A guide to medication review.

7.
Lowe CJ, Petty DR, Zermansky AG, Raynor DK. Development of a method for clinical
medication review by a pharmacist in general practice. Pharm World Sci 2000;22:121-6.
8.
Holland R, Desborough J, Goodyer L, Hall S, Wright D, Loke YK. Does pharmacist-led

medication review help to reduce hospital admissions and deaths in older people? A
systematic review and meta-analysis. Br J Clin Pharmacol 2008; 65:303-16.
42
9.
| C hapte r 2
Royal S, Smeaton L, Avery AJ, Hurwitz B, Sheikh A. Interventions in primary care to reduce
medication related adverse events and hospital admissions: systematic review and metaanalysis. Qual Saf Health Care 2006;15:23-31.
10. Nkansah N, Mostovetsky O, Yu C, Chheng T, Beney J, Bond CM, et al. Effect of outpatient
pharmacists non-dispensing roles on patient outcomes and prescribing patterns. Cochrane
Database Syst Rev 2010;(7):CD000336.
11. Holland R, Smith R, Harvey I. Where now for pharmacist led medication review?
J Epidemiol Community Health 2006;60:92-93.
12. Urbis Keys Young. Evaluation of the home medicines review program (pharmacy
component): final report. 2005. Available from: http://beta.guild.org.au/uploadedfiles/
Medication_Management_Reviews/Overview/Urbis%20Keys%20Young%20evaluation.pdf
(Accessed 3 Feb 2013).
13. MacKeigan LD, Nissen LM. Clinical pharmacy services in the home. Dis Manage Health
Outcomes 2008;16:227-44.
14. Niquille A, Lattman C, Bugnon O. Medication reviews led by community pharmacists in
Switzerland: a qualitative survey to evaluate barriers and facilitators. Pharm Pract 2010;8:3542.
15. Denneboom W, Dautzenberg MG, Grol R, De Smet PA. Treatment reviews of older people
on polypharmacy in primary care: cluster controlled trial comparing two approaches. Br J
Gen Pract 2007;57:723-31.
16. Denneboom W, Dautzenberg MGH, Grol R, De Smet PAGM. Comparison of two methods
for performing treatment reviews by pharmacists and general practitioners for homedwelling elderly people. J Eval Clin Pract 2008;14:446-52.
17. Bryant LJ, Coster G, Gamble GD, McCormick RN. The General Practitioner-Pharmacist
Collaboration (GPPC) study: a randomised controlled trial of clinical medication reviews in
community pharmacy. Int J Pharm Pract 2011;19:94-105.
18. Isetts BJ, Brown LM, Schondelmeyer SW, Lenarz LA. Quality assessment of a collaborative
approach for decreasing drug-related morbidity and achieving therapeutic goals. Arch Intern
Med 2003;163:1813-20.
19. Zermansky AG, Petty DR, Raynor DK, Lowe CJ, Freemantle N, Vail A. Clinical medication
review by a pharmacist of patients on repeat prescriptions in general practice: a randomised
controlled trial. Health Technol Assess 2002;6:1-86.
20. Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, Bouter LM, et al. The Delphi list:
a criteria list for quality assessment of randomized clinical trials for conducting systematic
reviews developed by Delphi consensus. J Clin Epidemiol 1998; 51: 1235-41.
21. Allard J, Hebert R, Rioux M, Asselin J, Voyer L. Efficacy of a clinical medication review on
the number of potentially inappropriate prescriptions prescribed for community-dwelling
elderly people. CMAJ 2001;164: 1291-96.
22. Bernsten C, Bjorkman I, Caramona M, Crealey G, Frokjaer B, Grundberger E, et al.
Improving the well-being of elderly patients via community pharmacy-based provision
of pharmaceutical care: a multicentre study in seven European countries. Drugs Aging
2001;18:63-77.
23. Grymonpre RE, Williamson DA, Montgomery PR. Impact of a pharmaceutical care model
for non-institutionalised elderly: results of a randomised, controlled trial. Int J Pharm Pract
2001;9:235-41.
43
24. Kwint HF, Faber A, Gussekloo J, Bouvy ML. Effects of medication review on drug-related
problems in patients using automated drug-dispensing systems: a pragmatic randomized
controlled study. Drugs Aging 2011;28:305-14.
25. Lenaghan E, Holland R, Brooks A. Home-based medication review in a high risk elderly
population in primary care - The POLYMED randomised controlled trial. Age Ageing
2007;36:292-7.
26. Sellors J, Kaczorowski J, Sellors C, Dolovich L, Woodward C, Willan A, et al. A randomized
controlled trial of a pharmacist consultation program for family physicians and their elderly
patients. CMAJ 2003;169:17-22.
27. Sorensen L, Stokes JA, Purdie DM, Woodward M, Elliott R, Roberts MS. Medication reviews
in the community: results of a randomized, controlled effectiveness trial. Br J Clin Pharmacol
2004;58: 648-64.
28. Sturgess IK, McElnay JC, Hughes CM, Crealey G. Community pharmacy based provision of
pharmaceutical care to older patients. Pharm World Sci 2003;25:218-26.
29. Volume CI, Farris KB, Kassam R, Cox CE, Cave A. Pharmaceutical care research and
education project: patient outcomes. J Am Pharm Assoc 2001;41:411-20.
30. Kassam R, Farris KB, Burback L, Volume CI, Cox CE, Cave A. Pharmaceutical care research
and education project: pharmacists interventions. J Am Pharm Assoc 2001 ;41:401-10.
31. Williams ME, Pulliam CC, Hunter R, Johnson TM, Owens JE, Kincaid J, et al. The shortterm effect of interdisciplinary medication review on function and cost in ambulatory elderly
people. J Am Geriatr Soc 2004;52:93-8.
32. Zermansky AG, Petty DR, Raynor DK, Freemantle N, Vail A, Lowe CJ. Randomised
controlled trial of clinical medication review by a pharmacist of elderly patients receiving
repeat prescriptions in general practice. BMJ 2001;323:1340-3.
33. Almeida Neto AC, Chen TF. Exploring elements of interprofessional collaboration between
pharmacists and physicians in medication review. Pharm World Sci 2007;29:574-6.
34. Nishtala PS, McLachlan AJ, Bell JS, Chen TF. A retrospective study of drug-related problems
in Australian aged care homes: medication reviews involving pharmacists and general
practitioners. J Eval Clin Pract 2011;17:97-103.
35. Spinewine A, Swine C, Dhillon S, Lambert P, Nachega JB, Wilmotte L, et al. Effect of a
collaborative approach on the quality of prescribing for geriatric inpatients: a randomized,
controlled trial. J Am Geriatr Soc 2007;55:658-65.
36. Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund-Udenaes M, Toss H, et al. A
comprehensive pharmacist intervention to reduce morbidity in patients 80 years or older: a
randomized controlled trial. Arch Intern Med 2009;169:894-900.
37. Makowsky MJ, Koshman SL, Midodzi WK, Tsuyuki RT. Capturing outcomes of clinical
activities performed by a rounding pharmacist practicing in a team environment: the
COLLABORATE study. Med Care 2009;47:642-50.
38. Bryant L, Coster G, McCormick R. General practitioner perceptions of clinical medication
reviews undertaken by community pharmacists. J Prim Health Care 2010;2:225-33.
39. Bond C. The MEDMAN study: a randomized controlled trial of community pharmacy-led
medicines management for patients with coronary heart disease. Fam Pract 2007;24:189-200.
44
| C hapte r 2
40. Farris KB, Cote I, Feeny D, Johnson JA, Tsuyuki RT, Brilliant S, et al. Enhancing primary
care for complex patients: demonstration project using multidisciplinary teams. Can Fam
Physician 2004;50:998-1003.
41. Krass I, Smith C. Impact of medication regimen reviews performed by community
pharmacists for ambulatory patients through liaisons with general medical practitioners. Int
J Pharm Pract 2000;8:111-20.
42. Halvorsen KH, Ruths S, Granas AG, Viktil KK. Multidisciplinary intervention to identify
and resolve drug-related problems in Norwegian nursing homes. Scand J Prim Health Care
2010;28:82-8.
45
Appe n d ix I
Literature search terms in MEDLINE
(Referral and Consultation[MAJR] OR Pharmaceutical Services[MAJR] OR Drug utilization

review[MAJR] OR treatment review[TIAB] OR medication review[TIAB] OR drug review[TIAB]
OR medication therapy management[TIAB] OR medicine management[TIAB] OR monitoring
pharmacotherapy[TIAB] OR reviewing medication[TIAB] OR pharmaceutical care OR Health
Services for the Aged[MAJR]) AND (outcome[all fields] OR patient admission[mesh] OR
hospital admissions[all fields] OR emergency admissions[all fields] OR mortality[mesh] OR
mortality[all fields] OR adverse drug events[all fields] OR ADEs[all fields] OR adverse drug
reactions[all fields] OR drug related problems[all fields] OR DRPs[all fields] OR improving well
being[all fields] OR MAI[all fields] OR Beers criteria[all fields] OR STOPP criteria[all fields] OR
appropriateness[all fields] OR treatment outcome[mesh] OR quality of life[mesh] OR EQ 5D[all
fields] OR SF-36[all fields] OR potentially inappropriate prescriptions[all fields] OR PIPs[all
fields] OR potentially inappropriate medications[all fields] OR PIMs[all fields] OR Treatment
Refusal[Mesh] OR Patient Compliance[Mesh] OR Medication Adherence[all fields] OR drug
changes[all fields] OR medication changes[all fields] OR optimizing drug regimens[all fields] OR
optimising drug regimens[all fields] OR number of drugs[all fields] OR numbers of drugs[all
fields] OR drug knowledge[all fields] OR knowledge of medicines[all fields] OR Accidental
Falls[Mesh] OR falls[all fields] OR falling[all fields] OR clinical benefit[all fields] OR Costs and
Cost Analysis[Mesh] OR cost-benefit[all fields] OR cost benefit[all fields] OR cost effectiveness[all
fields] OR drug costs[all fields]) AND (aged[mesh] OR aged[all fields] OR older[all fields]
OR elderly[all fields] OR 60[all fields] OR 65[all fields] OR 70[all fields] OR 75[all fields] OR
80[all fields] OR polypharmacy[mesh] OR polypharmacy[all fields]) AND (Community[all
fields] OR community pharmacy services[mesh] OR General Practice[all fields] OR Family
practice[mesh] OR Family practice[all fields] OR primary health care[mesh] OR domiciliary[all
fields] OR ambulatory[all fields] OR home dwelling[all fields] OR Physicians[Mesh] OR General
Practitioners[all fields] OR General Practitioner[all fields] OR Family physician[all fields] OR Family
physicians[all fields] OR Interprofessional Relations[Mesh] OR Physicians Practice Patterns[Mesh]
OR pharmacists[mesh] OR pharmacists[all fields] OR pharmacist[all fields]) AND (randomized
controlled trial[pt] OR controlled clinical trial[pt] OR randomized controlled trials[mh] OR random
allocation[mh] OR double-blind method[mh] OR single-blind method[mh] OR clinical trial[pt] OR
clinical trials[mesh] OR clinical trial[tw] OR ((singl*[tw] OR doubl*[tw] OR trebl*[tw] OR tripl*[tw])
AND (mask*[tw] OR blind*[tw])) OR placebos[mesh] OR placebo*[tw] OR random*[tw] OR Research
Design[MeSH:noexp] OR comparative study[pt] OR follow-up studies[mesh] OR prospective
studies[mesh] OR prospectiv*[tw] OR volunteer*[tw] OR control*[tw] OR latin square [tw] OR crossover studies [mh] OR control[tw] OR Evaluation Studies [Publication Type] OR Evaluation Studies as
Topic[Mesh]) NOT (animal[mesh] NOT human[mesh]) AND (randomized controlled trial[Publication
Type] OR (randomized[Title/Abstract] AND controlled[Title/Abstract] AND trial[Title/Abstract]))
Limits: Publication Date from 2000/01/01
46
A B ST R AC T
Aim
We compared the self-reported medication adherence and knowledge of older
patients receiving their drugs via multidose drug dispensing (MDD-users) with
patients receiving manually-dispensed drugs (non-MDD-users).
Me tho ds
MDD-users ( 65 years, 5 oral chronic drugs) were randomly selected from
eight Dutch community pharmacies. Non-MDD-users ( 5 oral chronic drugs)
were matched on age and gender. Medication adherence was assessed by using the
Medication Adherence Reporting Scale (MARS) and medication knowledge by
asking the indication of drugs. Cognitive function was measured with Mini-Mental
State Examination (MMSE) for a sub selection of patients.
Resu lts
The percentage of patients being adherent to all drugs was higher for MDDusers (n = 119; 81%) compared to non-MDD-users (n = 96; 58%; P < 0.01).The
percentage of patients with adequate knowledge was lower for MDD-users (40%)
compared to non-MDD-users (79%; P < 0.01). The differences in adherence were
independent of knowledge and MMSE-scores.
C on clusi on
This study shows that older patients receiving their drugs via multidose drug
dispensing reported a higher medication adherence compared to patients receiving
manually-dispensed drugs, despite a lower knowledge and lower cognitive function
among patients receiving MDD.
50
Ad h e re nce and k nowledge of M DD - vs. no n-M D D -user s
| C hapter 3.1
I N T ROD U C T ION
Older people with polypharmacy may experience difficulties managing their
medications. These difficulties could be due to complicated therapeutic regimens
or practical problems (e.g. halving tablets and opening packaging).1,2 Dosing aids
may help patients with these practical problems to adhere to their therapeutic
regimens.3,4 The awareness of the availability of dosing aids by older patients may
vary considerably between, and even within countries.5,6
Multidose drug dispensing (MDD), also known as automated drug dispensing, is a
sophisticated dosing aid that provides patients with robot-dispensed unit doses. All
drugs intended for one dosing moment are gathered in disposable bags and labelled
with patient data, drug contents, and the date and time for intake.4,7,8 Most research
originates from the Scandinavian countries and the Netherlands where MDD is used
by community-dwelling patients and patients in nursing homes.4,8-10 The number of
community-dwelling MDD-users in the Netherlands increased strongly in recent
years till 360.000 in 2011.11 This increase is partly due to a change in legislation
which does not allow home health care employees to manage patients medications
anymore. Next to home health care, Dutch patients are mostly recruited for MDD
by community pharmacists, general practitioners (GPs) or family when older
patients experience difficulties managing their medications.4 GPs can also refer
patients for more specific reasons (e.g. decreased cognitive function, [suspected]
non-adherence or severe psychiatric problems).4 MDD is especially appropriate
for persons who chronically use several drugs without frequent medication
changes.9 Prescription lists provided by community pharmacies for MDD-users are
authorized manually by GPs and other prescribers. These prescriptions are ordered
through a community pharmacy, which electronically forwards the total orders to
an MDD supplier. Dispensed drugs are returned to the community pharmacy who
deliver MDD-systems to the patients.12,13 It has been suggested that MDD reduces
medication errors, increases medication adherence and decreases waste of unused
drugs.8,9 However, it may be questioned whether patients still know the indication
of the drugs in the MDD systems.
Previous studies showed that adequate medication knowledge differed from 60% to
72%.3,14,15 Most studies reported a positive association between patients knowledge
and adherence.3,15-18 It is thought that patient knowledge of manually-dispensed
drugs is essential for competently managing their medication regimen.19 However,
as adherence is frequently determined by multiple factors, knowledge plays only a
minor role.20 Studies into the relation between adherence and cognitive function
show conflicting results.3
51
The aim of our study was to assess the self-reported adherence and medication
knowledge of older patients receiving their drugs via multidose drug dispensing
(MDD-users) compared to patients receiving only manually-dispensed drugs (nonMDD-users).
M E T HOD S
Stu dy D esi g n
This was a cross sectional study. Patients were selected and interviewed between
October 2010 and January 2012. The majority of patients in the Netherlands are
registered at only one community pharmacy, independently of prescriber, and
patient medication records are virtually complete with regard to prescription
drugs.21
Pati ents
Patients were recruited from eight Dutch community pharmacies. MDD-users
were defined as patients with at least one drug dispensed by multidose drug
dispensing. MDD-users were eligible if they were aged 65 years, used at least
five different oral prescription chronic drugs, lived at home or in a residential care
home. MDD-users had to be able to take their own drugs to explore if they had
other medication management problems than patients that used only manuallydispensed drugs (non-MDD-users). Patients in nursing homes were therefore
excluded. Even if the MDD system is used, a large share of the MDD-users will
still also need manually-dispensed drugs e.g. insulin, eye drops and vitamin K
antagonists. MDD-users were selected randomly from participating community
pharmacies using computer-generated random numbers. MDD-users were
invited to participate in the study by one of the pharmacists or research assistants.
Non-MDD-users were selected using the same inclusion criteria as for the MDDusers, except no drugs were dispensed by an MDD system. For each MDD-user,
two non-MDD-users in the same community pharmacy were invited and matched
on age (plus or minus one year) and gender. The second patient was selected in
advance in case the first patient was not willing to participate in the study.
D at a c ol l e c ti on
The interviews in both groups were conducted by the patients own pharmacist or
one of the research assistants (either C.O. or G.S.). Prior to each patient interview,
drug dispensing records were collected from the community pharmacy. Besides the
52
| C hapter 3.1
pharmacy list with dispensed drugs, the interviewer evaluated other drugs that the
patient was using, including prescription drugs not on the list, over-the-counter
drugs and complementary and alternative drugs. The drugs actually taken by the
patient were used for analysis. 92% of patient interviews were conducted at the
patients home.
Main outcome me asures
The primary outcomes were self-reported medication adherence and knowledge.
Adherence was measured by the Medication Adherence Report Scale (MARS). The
MARS consists of five statements concerning self-reported adherence: forgetfulness,
altering the dosage, stopping taking medication, missing a dose, and taking less
than instructed. The statements have response categories on a five-point Likert scale
where 1 = always, 2 = often, 3 = sometimes, 4 = rarely and 5 = never.22,23 For each
drug, a sum for MARS was calculated ranging from 5 to 25. The MARS-score per
patient was the mean of the different MARS-scores for each of their drugs. As there
is no concordance in the cut-off point for adherent and non-adherent behaviour
and MARS scores generally show a very skewed distribution, a score 22 was
considered as non-adherent.22,23
Medication knowledge was measured by asking the patient for the indication of
their drugs. Medication knowledge was graded as knowing indication of drug if
the patient could tell the indication of the drug spontaneously or patient was able to
retrieve the indication from a patient information leaflet. Patients who use patient
information leaflets have access to relevant information in daily life and therefore
can be regarded as having knowledge of indication. When patients could not tell
the indication or mentioned a wrong indication, this was regarded as not knowing
indication of drug. Knowing the indication of minimal 75% of their drugs was
considered as adequate knowledge.
To explore the influence of cognitive function on both adherence and knowledge,
we added the assessment of Mini-Mental State Examination (MMSE) while the
study was on-going. Therefore MMSE-tests for MDD-users were conducted 6 to
12 months after the interview on adherence and knowledge. MMSE-tests for nonMDD-users were conducted simultaneously with the interview on adherence and
knowledge. Patients were divided in two subgroups: MMSE-score > 27 and MMSEscore 27.
C on f i d ent i a l it y
In order to protect the patients privacy, all medical data were anonymized by the
community pharmacist and research assistants using a randomly assigned unique
53
Fi g u re 1 Study flow chart
MDD-users a
selected
n = 186
Non-MDD-users a
matched at age and sex
n = 238
Not invited: 45
Not invited: 109
excluded by pharmacy (18) b

not reachable (5)
not approached (22)
second match (63)

not reachable (28)
started MDD (18)
Patients invited
n = 141
Patients invited
n = 129
Not participated: 14
Not participated: 33
care home/hospital (5)

not willing to participate (8)
other (1)
deceased (4)
illness/hospital (10)
not willing to participate (19)
Patients participated
n = 127
Patients participated
n = 96
Excluded for analysis: 8

< 65 years (2)
< 5 drugs (6)
Included for analysis
n = 119

n = 96
MDD = multidose drug dispensing

a) MDD-users are patients with at least one drug dispensed by multidose drug dispensing; Non-MDD-users are
patients who used only manually-dispensed drugs.
b) Admission to nursing home, hospital, deceased, not able to take own drugs.
number for each patient. Informed consent was obtained from all participants who
agreed in performing MMSE-tests.
Statisti c a l an a l ys es
Independent t-tests were used for continuous variables with a normal distribution
and non-parametric Mann Whitney U tests for other continuous variables with
54
| C hapter 3.1
a skewed distribution (age, MMSE, MARS-score per patient). The Pearson chisquared (2) tests were applied for each categorical variable. A P-value < 0.05 was
considered statistically significant.
Data were analyzed using database (Microsoft Access 2010; Microsoft Corporation,
Redmond, WA, USA) and statistical software (SPSS version 20.0; SPSS Inc.,
Chicago, IL, USA).
R E SU LT S
Pati ent f l ow an d ch ar ac teristi c s
A total of 177 MDD patients 65 years were invited to participate in the study
and119 patients (67%) accepted the invitation. Matching at age and gender resulted
in 238 eligible non-MDD-users of whom 96 participated (Figure 1).
Baseline characteristics of both MDD-users and non-MDD users are shown
in Table 1. No significant differences were seen in the total number of drugs per
patient. Furthermore, there were no significant differences in the percentages of
most used prescription drug classes, except for drugs used in diabetes.
Ad heren ce an d k n ow l e d ge of M DD - u s er s vs . n on - M DD - u s er s
Self-reported adherence and medication knowledge are shown in Table 2. Selfreported adherence was higher for MDD-users (median MARS score 25.0)
compared to non-MDD users (median 24.7; P < 0.01). The percentage of patients
being adherent to all drugs (MARS-score 23 for each of their drugs) was higher
for MDD-users compared to non-MDD-users (91% vs. 58%; P < 0.01). The mean
percentage of drugs for which patient knew the indication was lower for MDDusers compared to non-MDD-users (63% vs. 85%; P < 0.01). The percentage of
patients with adequate knowledge was 40% for MDD-users and 79% for nonMDD-users (P < 0.01).
After stratification for medication knowledge, adherence remained higher for
MDD-users in both subgroups (Table 3). No correlation was found between the
mean percentage of drugs for which patients knew the indication and the MARSscore per patient (Spearmans rho 0.054; P = 0.4).
MDD-users reported for 16 of 1128 drugs (1.4%) non-adherence (MARS-score
22) against 59 of 960 drugs (6.1%) for the non-MDD-users (P < 0.01). Nonadherence was most often reported for drugs for peptic ulcer and GORD (12%) and
high-ceiling diuretics (9%).
55
Ta b l e 1
Baseline characteristics of MDD-usersa and non-MDD-usersa

MDD-users
n = 119
Non-MDD-users
n = 96
P-Value
Female; n (%)
74 (62%)
63 (66%)
0.60b
Age in years; median (IQR)
80 (7683)
81 (7784)
0.23c
Number of drugs per patient; mean SD
9.7 3.0
10.0 3.2
0.61d
Number of manually-dispensed drugs per patient;

mean SD
2.8 2.1
10.0 3.2
< 0.01d
Number of drugs in MDD; mean SD
7.0 2.2
Most used prescription drug classes (ATC); n (%)

Antithrombotic agents (B01A)
114 (98%)
89 (92%)
0.63b
Agents acting on the Renin-Angiotensin System (C09)
95 (82%)
75 (78%)
0.72b
Lipid Modifying agents (C10A)
82 (71%)
55 (57%)
0.20b
Drugs for obstructive airway diseases (R03)
41 (35%)
41 (43%)
0.40b
Drugs used in diabetes (A10)
101 (87%)
38 (39%)
< 0.01b
Beta blocking agents (C07A)
69 (60%)
56 (58%)
0.88b
Drugs for peptic ulcer and GORD (A02B)
82 (71%)
60 (62%)
0.43b
Calcium channel blockers (C08C)
30 (26%)
33 (34%)
0.26b
Benzodiazepine derivatives (N05BA,N05CD)
26 (22%)
29 (30%)
0.27b
High-ceiling diuretics (C03C)
41 (35%)
23 (24%)
0.12b
Low-ceiling diuretics (C03A, C03B, C03E)
27 (23%)
23 (24%)
0.94b
MDD = multidose drug dispensing; IQR= interquartile range

b) Pearson chi-squared test.
c) Mann-Whitney U test.
d) Student t-test.
C omp ar i s on of M DD - d r u g s an d m anu a l l y d i sp ens e d d r u g s i n

M DD -us ers
86% of MDD-users (102 of 119) used also manually dispensed drugs. Medication
adherence and knowledge of MDD-drugs vs. manually dispensed drugs within the
group of MDD-users is shown in Table 4.
67% of MDD-users (80 of 119) used at least one manually dispensed drug
chronically. The distribution of MARS-scores was different (P < 0.01), while the
median MARS-scores were similar (25.0). The percentage of patients being adherent
56
| C hapter 3.1
Ta b l e 2
Self-reported medication adherence and knowledge at patient level of

MDD-usersa and non-MDD-users a
MDD-users
Non-MDD-users
P-Value
M e d i c at i o n a d h e re n ce
n = 119
n = 96
MARS score; median (IQR)
25.0 (24.825.0)
24.7 (24.025.0)
< 0.01b
Patients with MARS score 23 for each of their

drugs; n (%)
108 (91%)
56 (58%)
< 0.01c
M e d i c at i o n k n owle d g e
n = 119
n = 96
Percentage of drugs for which patient knew

indication; mean SD
63 31
85 23
< 0.01d
Patients with adequate knowledgee; n (%)
47 (40%)
76 (79%)
< 0.01c
MDD = multidose drug dispensing; IQR= interquartile range

b) Mann-Whitney U test.
c) Pearson chi-squared test.
d) Student t-test.
e) Adequate medication knowledge: knowing indication > 75% of drugs.
Ta b l e 3
Self-reported medication adherence of MDD-usersa and non-MDD-usersa

depending on medication knowledge
Non adequate knowledge
MDDusers
Non-MDDusers
Medication adherence
n = 72
n = 20
MARS score;
median (IQR)
25.0
(24.625.0)
24.6
(24.025.0)
Patients with MARS score

23 for each of their
drugs; n (%)
63 (88%)
13 (65%)
P-Value
Adequate knowledge
MDDusers
Non-MDD- P-Value
users
n = 47
n = 76
0.03b
25.0
(24.825.0)
24.8
(24.025.0)
< 0.01b
0.02c
45 (93%)
25 (58%)
< 0.01c
MDD = multidose drug dispensing; MARS = Medication Adherence Reporting Scale; IQR= interquartile range
57
Ta b l e 4
Medication adherence and knowledge at patient level of MDD-usersa for

MDD-drugs and manually-dispensed drugs
MDD-drugs
Manually-dispensed
drugs
M e d i c at i o n a d h e re n ce
n = 119
n = 80
MARS score; median (IQR)
25.0 (24.725.0)
25.0 (25.025.0)
< 0.01b
Patients with MARS score 23 for each of

their drugs; n (%)
111 (93%)
77 (97%)
< 0.01c
M e d i c at io n k n owle d g e
n = 119
n = 102
Percentage of drugs for which patient knew

indication; mean SD
50 39
95 18
< 0.01d
Patients with adequate knowledge ; n (%)
41 (35%)
94 (92%)
< 0.01c
P-Value
MDD = multidose drug dispensing; MARS = Medication Adherence Reporting Scale; IQR= interquartile range
a) MDD-users are patients with at least one drug dispensed by multidose drug dispensing.
d) Student t-test.
to all drugs (MARS-score 23 for each of their drugs) was lower for MDD-drugs
(93%) compared to manually dispensed drugs (97%; P < 0.01).
Within the group of MDD-users, the knowledge per patient for MDD-drugs was
lower (50%) compared to manually dispensed drugs (95%; P < 0.01). The percentage
of patients with adequate knowledge of MDD-drugs was lower (35%) compared to
the percentage of patients with adequate knowledge of manually dispensed drugs
(92%; P < 0.01).
C o g n it ive f u n c t i on
We were able to measure cognitive function by MMSE for a selection of the
included patients (58 MDD-users [49%], 73 non-MDD-users [76%]). Reasons
for not performing MMSE for MDD-users were deceased (n = 6), not willing to
participate (n = 11), unable to contact (n = 38) and other reasons (n = 4). Reasons
for not performing MMSE for non-MDD-users were admission to hospital (n = 2),
not willing to participate (n = 4) and unable to contact (n = 11). The median
MMSE score was different for MDD-users compared to non-MDD-users (27 vs 28;
P = 0.02).
58
Ta b l e 5
| C hapter 3.1
Medication adherence and knowledge of MDD-usersa and non-MDD-usersa

depending on cognitive function
MMSE 27
MMSE > 27
MDDusers
Non-MDD- P-Value
users
MDDusers
Non-MDD- P-Value
users
Medication adherence
n = 35
n = 30
n = 23
n = 43
MARS score;
median (IQR)
25.0
(24.825.0)
24.8
(24.225.0)
0.05b
25.0
(24.925.0)
24.8
(24.125.0)
Patients with MARS score

23 for each of their
drugs; n (%)
63 (88%)
13 (65%)
0.02c
45 (93%)
25 (58%)
Medication knowledge
n = 35
n = 30
n = 23
n = 43
Percentage of drugs for

which patient knew
indication; mean SD
65 31
80 25
0.04d
67 30
84 24
Patients with adequate

knowledgee; n (%)
16 (46%)
21 (70%)
0.04c
10 (44%)
34 (79%)
0.02b
< 0.01c
0.01d
< 0.01c
MDD = multidose drug dispensing; MMSE = Mini-Mental State Examination; MARS = Medication Adherence
Reporting Scale; IQR= interquartile range
d) Student t-test.
After stratification for cognitive function in subgroups of MMSE-scores < 27

and 27, self-reported adherence remained higher for MDD-users within both
subgroups. Likewise, medication knowledge remained lower for MDD-users in
both subgroups (Table 5).
DI S C U S SION
To our knowledge, this is the first study that showed that self-reported medication
adherence is higher for patients using MDD compared to patients using manuallydispensed drugs. This difference was independent of medication knowledge and
cognitive function. There is limited evidence that dosing aids in general can increase
medication adherence.24 However, there was no specific evidence for MDD which
59
is expected to increase medication adherence.8,9

Our study showed a low medication knowledge and a high adherence for MDDusers and no correlation between them. At first sight, this absence of an association
between knowledge and adherence for MDD-users seems contradictive to earlier
studies which report a positive association between knowledge and adherence.3,15-18
However, it could also indicate that the appropriate group of patients received
MDD. Without MDD, their medication adherence probably would have been
lower.15 Moreover, knowledge is only one of the many factors that could influence
adherence. Forgetfulness and practical difficulties with medication management
(e.g. removing medication from its primary packaging) might be more important
barriers to adequate adherence in this specific group of older patients.
Especially within the group of MDD-users, the knowledge of MDD-drugs was
low while the knowledge of non-MDD-drugs was very high. This may partly be
explained by patients not recognizing the different tablets within the disposable
plastic bags. The majority of tablets was white and can only be identified by form
and inscriptions. Furthermore, most MDD-users seemed not to bother about
the indications of their drugs. This patients remoteness of their drugs has been
reported earlier as a possible disadvantage of dosing aids.25 In contrast, MDD-users
had a high knowledge of drugs not dispensed by MDD. Chronic drugs not suitable
for MDD were especially non-oral dosage forms like insulin, inhalation drugs and
eye drops. Furthermore, laxatives (sachets or syrup) and vitamin K antagonists that
often require a special dosing scheme were not dispensed in MDD. These dosage
forms were generally well recognized by patients.
As expected, the median MMSE score was lower for MDD-users compared to the
non-MDD-users. This lower cognitive function may be expected to give difficulties
with medication management.26 However, the differences in adherence and
knowledge between the two groups were not influenced by the cognitive function
within the two groups. This suggests that these differences could be mainly
attributed to the differences in dispensing systems.
This study had several strengths. First of all, non-MDD users were matched at age
and sex and were selected within the same pharmacy. Secondly, we stratified for
cognition by MMSE scores.
The choice to measure adherence by self-reporting could be seen as a limitation
because it could overestimate adherence. There is no gold standard for the
assessment of adherence. Different measurement instruments may result in
different adherence estimates for the same patient and medication group.27
Measuring the refill rate is not possible for MDD-users, because they automatically
receive a new MDD-system. Although programmable MDD dispensers that can
60
| C hapter 3.1
monitor adherence have recently become available, these cannot be applied for
non-MDD-users.5 However, for comparing adherence in groups with different
dispensing systems, self-reporting seemed the most appropriate method. Recent
studies have shown that self-reported methods like MARS are concordant with
direct methods.18,23 Another limitation of our study was that MARS has not been
validated for use among people with cognitive impairment. Furthermore, especially
for the MDD-users, the response rate to perform a MMSE-test was low. This could
be due to the fact that this was not performed simultaneously with the interview on
knowledge and adherence for MDD-users. This might result in an underestimation
of MMSE for MDD-users, but does not influence adherence and knowledge
outcomes. Finally, the design of the study was cross-sectional. This means that we
could find a high adherence and low knowledge for MDD-users, but we cannot
assume a causal relationship.
Our findings may have implications for practice. The low knowledge of drugs in the
MDD-system may raise problems when medication has to be changed in the MDDsystem. For example, when a GP advices to (temporarily) stop taking a certain drug,
patients may not recognize which tablet to stop. Otherwise, more readable drug
information should be available to the patient while the information on the bags
of the MDD-systems is too limited. Furthermore, the patients remoteness of their
drugs may also lead to continuous use of drugs that otherwise would be stopped by
the patient after a certain time (for example analgesics, anthistamines). This leaves
a high responsibility for the GP and, in particular, the community pharmacist
to monitor the use of MDD-drugs of older patients in combination with regular
patient contact.
In conclusion, our study showed that older patients using multidose drug
dispensing (MDD-users) report higher medication adherence and have a lower
medication knowledge compared to patients using manually-dispensed drugs
(non-MDD-users). The differences in adherence were independent of medication
knowledge and cognitive function. These finding suggests that the higher adherence
of MDD-users could be attributed mainly to the MDD-system. Future intervention
studies are needed to determine whether older patients who are non-adherent on
manually-dispensed drugs become more adherent when they start with a MDDsystem.
Ack now l e dgements - The authors like to thank Laurens Biesma, Hinke Dorhout,
Rene Dull, Eva de Groot, Rik van der Meer and Ruud Timmerman for participation as
pharmacists in this study. The authors like to thank Cynthia Oosterlee for her work as
research assistant.
61
REFERENCES
1.
Tordoff JM, Bagge ML, Gray AR, Campbell AJ, Norris PT. Medicine-taking practices
in community-dwelling people aged > or = 75 years in new zealand. Age Ageing
2010;39:574-80.
2.
Granas AG, Bates I. Patients understanding and management of their illnesses and
prescribed medicines - a descriptive study. Pharm World Sci 2005;27:321-8.
3.
Barat I, Andreasen F, Damsgaard EM. Drug therapy in the elderly: What doctors believe and
patients actually do. Br J Clin Pharmacol 2001;51:615-22.
4.
Kwint HF, Faber A, Gussekloo J, Bouvy ML. Effects of medication review on drug-related
problems in patients using automated drug-dispensing systems: A pragmatic randomized
controlled study. Drugs Aging. 2011;28:305-14.
5.
Lakey SL, Gray SL, Borson S. Assessment of older adults knowledge of and preferences for
medication management tools and support systems. Ann Pharmacother 2009;43:1011-9.
6.
Nunney J, Raynor DK, Knapp P, Closs SJ. How do the attitudes and beliefs of older people
and healthcare professionals impact on the use of multi-compartment compliance aids?: a
qualitative study using grounded theory. Drugs Aging 2011;28:403-14.
7.
Wekre LJ, Spigset O, Sletvold O, Sund JK, Grimsmo A. Multidose drug dispensing and
discrepancies between medication records. Qual Saf Health Care 2010;19:e42.
8.
Johnell K, Fastbom J. Multi-dose drug dispensing and inappropriate drug use: A nationwide
9.
Reuther LO, Lysen C, Faxholm M, Salomon L, Hendriksen C. Multi-dose drug dispensing is

a challenge across the primary-secondary care interface. Dan Med Bull 2011;58:A4341.
10. Wekre LJ, Melby L, Grimsmo A. Early experiences with the multidose drug dispensing
system--a matter of trust? Scand J Prim Health Care 2011;29:45-50.
11. Stichting Farmaceutische Kengetallen (SFK). Data en feiten 2012: het jaar 2011 in cijfers.
Available from: http://www.sfk.nl/pdf-documenten/data-en-feiten/data-en-feiten-2012
(Accessed 3 March 2013).
12. Halvorsen KH, Granas AG, Engeland A, Ruths S. Prescribing quality for older people in
Norwegian nursing homes and home nursing services using multidose dispensed drugs.
Pharmacoepidemiol Drug Saf 2012;21:929-36.
13. Wekre LJ, Bakken K, Garasen H, Grimsmo A. GPs prescription routines and cooperation
with other healthcare personnel before and after implementation of multidose drug
dispensing. Scand J Public Health 2012;40:523-30.
14. Blenkiron P. The elderly and their medication: Understanding and compliance in a family
practice. Postgrad Med J 1996;72:671-6.
15. Modig S, Kristensson J, Ekwall AK, Hallberg IR, Midlov P. Frail elderly patients in primary
care - their medication knowledge and beliefs about prescribed medicines. Eur J Clin
Pharmacol 2009;65:151-5.
16. Okuno J, Yanagi H, Tomura S, Oka M, Hara S, Hirano C, et al. Compliance and medication
knowledge among elderly japanese home-care recipients. Eur J Clin Pharmacol 1999;55:1459.
62
| C hapter 3.1
17. Burge S, White D, Bajorek E, Bazaldua O, Trevino J, Albright T, et al. Correlates of

medication knowledge and adherence: Findings from the residency research network of
south texas. Fam Med 2005;37:712-8.
18. Tsai KT, Chen JH, Wen CJ, Kuo HK, Lu IS, Chiu LS, et al. Medication adherence among
geriatric outpatients prescribed multiple medications. Am J Geriatr Pharmacother
2012;10:61-8.
19. Lam P, Elliott RA, George J. Impact of a self-administration of medications programme on
elderly inpatients competence to manage medications: A pilot study. J Clin Pharm Ther
2011;36:80-6.
20. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353:487-9.
21. Buurma H, Bouvy ML, De Smet PA, Floor-Schreudering A, Leufkens HG, Egberts AC.
Prevalence and determinants of pharmacy shopping behaviour. J Clin Pharm Ther
2008;33:17-23.
22. Mardby AC, Akerlind I, Jorgensen T. Beliefs about medicines and self-reported adherence
among pharmacy clients. Patient Educ Couns 2007;69:158-64.
23. van den Bemt BJ, van den Hoogen FH, Benraad B, Hekster YA, van Riel PL, van Lankveld W.
Adherence rates and associations with nonadherence in patients with rheumatoid arthritis
using disease modifying antirheumatic drugs. J Rheumatol 2009;36:2164-70.
24. Mahtani KR, Heneghan CJ, Glasziou PP, Perera R. Reminder packaging for improving
adherence to self-administered long-term medications. Cochrane Database Syst Rev
2011;9:CD005025.
25. Raynor DK, Nunney JM. Medicine compliance aids are partial solution, not panacea. BMJ
2002;324:1338.
26. Maddigan SL, Farris KB, Keating N, Wiens CA, Johnson JA. Predictors of older adults
capacity for medication management in a self-medication program: A retrospective chart
review. J Aging Health 2003;15:332-5.
27. Ammassari A, Trotta MP, Murri R, Castelli F, Narciso P, Noto P, et al. Correlates and
predictors of adherence to highly active antiretroviral therapy: overview of published
literature. J Acquir Immune Defic Syndr 2002;31:S123-7.
63
A B ST R AC T
B ack g roun d
There are concerns that multidose drug dispensing (MDD) may increase
inappropriate drug use. MDD could lead to perpetual repeating of drug therapies
without the necessary re-evaluation.
Aim
The aim of this study was to examine the effect of a pharmacist-led medication
review on drug-related problems (DRPs) in older patients receiving their drugs via
MDD.
Me tho ds
This was a pragmatic randomized controlled study conducted in primary care.
Patients were recruited from six Dutch community pharmacies. They were eligible
if they lived at home, were aged 65 years, and used five or more different drugs, of
which at least one had to be dispensed via an MDD system. Patients were randomly
allocated to receive a medication review at the start of the study (intervention
group) or after 6 months (waiting-list group). Each patient was independently
reviewed by two pharmacist reviewers. The results of these medication reviews
were sent to the community pharmacist to be discussed with the patients general
practitioner (GP). The primary outcome measure was the number of DRPs leading
to a recommendation for drug change. Secondary outcomes were the total number
of drug changes and the number of drug changes related to a recommendation. In
order to analyse drug changes, medication records were collected 6 months after
the medication review or index date in the waiting-list group. Potential DRPs were
classified using the DOCUMENT classification.
Re su lts
There were no baseline differences between the 63 patients in the intervention
group and the 55 patients in the waiting-list group with respect to age, sex, number
of drugs per patient and type of drug prescribed. The mean number of DRPs per
patient at baseline in the intervention group and waiting list combined was 8.5, with
no difference between the groups. At baseline, the mean number of DRPs leading to
a recommendation for drug change was 4.5 per patient and did not differ between
the two groups. After 6 months, the number of DRPs leading to a recommendation
for drug change decreased by 30% in the intervention group versus 5% in the
waiting-list group (P < 0.01).
66
E f fe c t s o f m e d icat io n re view o n M DD
| C hapter 3.2
C onclu si ons
This study shows that patients using MDD have a high number of DRPs. Medication
review decreases the number of DRPs among these patients. We recommend that
all patients with multidose drug dispensing should have a thorough medication
review by pharmacists and prescribers.
I N T ROD U C T ION
At least 5% of hospital admissions are directly related to adverse drug reactions
(ADRs).1-4 Higher rates have been reported among elderly patients, who are likely
to be receiving multiple medications for long-term illnesses.1 In most studies,
these ADRs were not only side effects but also drug-related problems (DRPs)
such as prescribing errors, poor adherence and insufficient monitoring.4 A recent
study in the Netherlands suggested that almost half of these medication-related
hospitalizations could be avoided.2,5 That study also identified a relationship
between drug-related hospital admissions and decreased cognition and poor
medication adherence.2,5
Dosing aids may help patients adhere to their therapeutic regimens. Multidose drug
dispensing (MDD) is a sophisticated dosing aid that provides patients with robotdispensed unit doses. All drugs intended for one dosing moment are gathered in
disposable bags and labelled with patient data, drug contents, and the date and time
for intake.6
MDD is more likely to be offered to patients with a high probability of inappropriate
drug use but cannot be considered a panacea for all such patients.7,8 Firstly, for
practical reasons not all dosage forms (e.g. powders, inhalers, ointments) can
be dispensed using the distribution robot.6,9 Secondly, MDD may not solve
inappropriate drug use, and may even aggravate it. MDD could lead to perpetual
repeating of prescribed therapies without the necessary re-evaluation. It is therefore
suggested that MDD should be combined with regular patient counselling and
medication review.8 In Finland, a medication review performed by the pharmacist
is required before a patient can be enrolled in an MDD programme.10
Community pharmacists in the Netherlands have limited experience with
medication review. Therefore, we decided to support community pharmacists with
an expert panel of pharmacist reviewers with experience in identifying DRPs in a
pharmaceutical care plan.
The aim of the present study was to examine the effect of a community pharmacistled medication review of DRPs in older patients receiving their drugs via MDD.
67
M E T HOD S
Stu d y D esig n
This was a pragmatic randomized controlled study. Patients were enrolled between
October 2007 and February 2008.
Pati ents
Patients were recruited from six Dutch community pharmacies. Pharmacists were
a convenience sample. They were invited to participate based on information from
the provider of multidose drug dispensing systems that these pharmacists had a
sufficient number of home-dwelling patients using multidose drug dispensing
systems. Each community pharmacist invited two general practitioners (GPs) to
participate in the study.
Patients were eligible if they used five or more medicines, were aged 65 years and
lived at home. At least one of their medicines had to be dispensed via an MDD
system. In the Netherlands, patients are recruited for MDD mostly by referral from
a GP when he/she suspects inappropriate drug use (e.g. complicated medication
regimens, decreased cognition, [suspected] non-adherence or severe psychiatric
problems).
Patients with MDD systems were informed about the medication review service by
their pharmacist. Patients could decline participation. Because the pharmacists and
GPs were introducing a new service, it was not feasible to review all the patients at
the same time. Therefore, selected patients were randomized into two groups per
pharmacy using computer-generated random numbers. Patients were allocated to
an intervention group, which received a medication review at the start of the study,
or to a 6-month waiting-list group.
Ethi cs an d Pat i ent C on f i d enti a lit y
As the study did not involve a major invasion of the participants autonomy and
no standard care was withheld, no formal ethical approval was needed. Patients
received written information about the study and were able to decline participation.
community pharmacist using a randomly assigned unique number for each patient.
Inter vent i on
For each patient, data from both the community pharmacy and the GP were
collected by the community pharmacist and included drug dispensing records,
information on co-morbidity and/or drug intolerance, relevant patient notes, and
laboratory data (e.g. blood pressure, glycosylated haemoglobin and creatinine
68
| C hapter 3.2
clearance). No face-to-face interviews with patients were conducted by the

community pharmacists.
The data for each patient were independently reviewed by two independent
pharmacists from a pool of five pharmacist reviewers (A.F., H.B., H.K., J.K.D. and
M.B.). The pharmacist reviewers used in this study had several years of experience
of medication review as well as in-depth knowledge of national clinical guidelines.
The reviewers used both implicit and explicit criteria to identify potential DRPs.
Explicit criteria consisted of a list of clinical rules based on Dutch treatment
and prescription guidelines. Examples of clinical rules were Lack of appropriate
treatment for secondary prevention for CHD (coronary heart disease) (antiplatelet,
lipid-lowering, -blocker [-adrenoceptor antagonist], ACE-inhibitor), Lack of
appropriate treatment for patients with diabetes mellitus and LDL (low-density
lipoprotein) cholesterol > 2.5 mmol/L (lipid lowering therapy) or No available
monitoring data for blood pressure, lipids, glucose, BMI (body mass index) or data
> 1 year old in patients for whom these measurements are indicated in treatment
guidelines.11 Implicit criteria for identifying DRPs were based on a structural
assessment by Cipolle 12 according to a rational order of indication, effectiveness,
safety and compliance. The two reviewers reached consensus in a case conference.
If no consensus was achieved, a third reviewer was consulted until consensus was
reached. The results of these medication reviews were sent to the community
pharmacist to be discussed in a case conference with the patients GP within 4
weeks.
The waiting-list group patients underwent their medication review at time t = 6
months, i.e. 6 months after the intervention group. As a consequence, the waitinglist group was reviewed retrospectively by pharmacist reviewers using the data
available at time t = 0. Before discussing the pharmaceutical care plan for the
waiting-list group with the GP, the community pharmacist checked whether any
medication changes had taken place between t = 0 and t = 6. We assumed that any
medication change at t = 6 in the waiting-list group was a result of usual care. The
follow-up of recommendations for the waiting-list group was not included in this
study.
D at a C l assif i c at i on
Drugs were classified using the Anatomical Therapeutic Chemical (ATC)
classification. Potential DRPs were classified using the DOCUMENT classification
system 33 with modifications as described. Information concerning the DRPs
consisted of the type and subtype of problem (see Appendix I, which shows the
DOCUMENT classification system).13,14 The original DOCUMENT system was
69
Fi g u re 1
Study flow chart
Randomization
n = 125
Excluded before t = 0
7 patients
Intervention group at t = 0
n = 63
Lost to follow-up:
deceased (2)
moved (4)
admission to nursing home (2)
Intervention group at t = 6
n = 55
Waiting-list group at t = 0
n = 55
Lost to follow-up:
deceased (2)
Waiting-list group at t = 6
n = 53
t = 0 signifies baseline (0 months); t = 6 signifies 6 months.
developed in accordance with the major requirements for a DRP classification

system.14,15 In addition to the original DOCUMENT classification system, we
added five DRP subtypes. Within the DRP type Drug selection, DRP subtypes Lack
of indication or unclear indication, Lack of effectiveness and Contra-indication/
intolerance were added. Within the DRP type Toxicity or adverse reaction, DRP
subtypes Risk of adverse effects and Possible drug treatment in response to
adverse effect were added. See Appendix II, for a summary of these modifications.
All coding and classification was undertaken by one investigator (H.K.) and a
student investigator (D.T.). Each investigator coded all data independently. When
there were differences in coding, the investigators reached consensus in a case
conference with a third investigator (either A.F. or M.B.).
O utc ome Me asu res
After 6 months (t = 6), medication records were collected in order to analyse the
drug changes. The outcomes could only be assessed for patients with complete
medication records from t = 0 until t = 6. The primary outcome measure was the
70
| C hapter 3.2
reduction in the number of DRPs leading to a recommendation for drug change.

Secondary outcome measures were the total number of drug changes and the
number of drug changes related to a recommendation.
Drug changes refer to prescription drugs and were defined as a cessation of a
drug, a dose change, an addition of a drug, a replacement of a drug, a change in
dose frequency/schedule or a drug formulation change. A drug change could be
related to a recommendation from the treatment review, but could also be related
to the process of usual care (not related to a recommendation). The total number
of drug changes is the sum of drug changes related to a recommendation and drug
changes not related to a recommendation. A qualitative process evaluation was
performed by means of a questionnaire completed by the participating community
pharmacists. The aim of this questionnaire was to reveal the opinion of the
community pharmacists about the medication review process.
S ampl e Siz e C a l c u l ati on
Based on a literature review we assumed that our intervention would result in an
additional mean reduction of 1.5 DRPs (related to a drug-change modification)
per patient in the intervention group. With a 95% confidence interval, power of 0.8
and a standard deviation of 2.5 for the number of medication-related problems per
patient, we needed 45 patients in both study groups. Expecting a drop-out rate of
25%, we aimed at including 60 patients in both groups.
Statist i c a l An a l ysis
All data were analysed using database (Microsoft Access , 2003; Microsoft
Corporation, Redmond, WA, USA) and statistical software (SPSS version 17.0; SPSS
Inc., Chicago, IL, USA). Descriptive statistics were used for basic characteristics.
Independent sample t-tests were applied for sex, age, number of prescription drugs
and number of DRPs to test the difference between groups. Pearson chi-squared
(2) tests were used for each categorical variable. A paired t-test was used to test
for differences in DRPs per patient between t = 0 and t = 6. A P-value < 0.05 was
considered statistically significant.
R E SU LT S
Six pharmacies recruited 125 patients aged 65 years. Seven patients were excluded
after randomization but before the actual start of the medication review (t = 0)
(Figure 1). Reasons for exclusion were death (2/125), admission to hospital (1/125),
71
Ta b l e 1
Baseline sociodemographic and drug-related characteristics of patients in

the intervention and waiting-list group
Characteristic
Intervention
group (n = 63)
Waiting-list
group (n = 55)
P-value
S o c i o d e m o gra p h i c
Female; n (%)
48 (76%)
33 (60%)
0.06
Age in years; mean SD
78.7 6.8
80.0 7.2
0.34
Number of prescribed drugs; mean per patient SD
10.3 3.1
9.8 3.6
0.41
6.7 2.2
7.2 2.6
0.27
46 (73%)
42 (76%)
0.68
35 (56%)
38 (69%)
0.13
36 (56%)
31 (56%)
0.93
29 (46%)
30 (55%)
0.36
Benzodiazepine derivatives (N05BA, N05CD)
33 (52%)
23 (42%)
0.26
28 (44%)
28 (51%)
0.49
29 (46%)
27 (49%)
0.75
30 (48%)
24 (44%)
0.67
20 (32%)
14 (26%)
0.46
17 (27%)
16 (29%)
0.80
Number of MDD drugs; mean per patient SD

D r u g c l a sse s (ATC ) at t = 0 ; n (%)
N u m b e r o f D R Ps ; total (mean per patient SD)
539 (8.6 2.8)
455 (8.4 2.8) 0.76
D (rug Selection)
O (ver or underdose)
C (ompliance)
U (ntreated conditions)
M (onitoring required)
T (oxicity)
141
129
0.41
78
48
0.06
0.64
86
72
0.94
181
145
0.55
49
59
0.06
t = 0 signifies baseline (0 months).

SD = standard deviation; MDD = multidose drug dispensing; ATC = Anatomical Therapeutic Chemical
classification; GORD = gastro-oesophageal reflux disease; DRPs = drug-related problems
admission to nursing home (1/125), start of multidose dispensing after t = 0 (2/125)

and no medication data (1/125). At baseline (t = 0), 63 patients remained in the
intervention group and 55 patients remained in the waiting-list group.
72
| C hapter 3.2
Ta b l e 2
Drug-related problems (DRPs) with a recommendation for drug change for

patients in the intervention and waiting-list groups at t = 0 and t = 6
DRP type and subtype
To ta l (m e a n p e r p at i e nt )
D (r ug
s e le c t i o n )
Duplication
Intervention group
(n = 55)
Waiting-list group
(n = 53)
t=0
t=6
t=0
t=6
249 (4.5)
175 (3.2)
231 (4.4)
221 (4.2)
92
64
87
84
Drug interaction
Wrong dosage form
Lack of indication or unclear indication a
39
28
39
38
Lack of effectiveness a
31
21
26
24
57
39
40
37
30
24
18
16
Contra-indication/ intolerance a
O (ve r
o r u n d e rd o s e )
Dosage too high

Dosage too low
12
14
14
Dose frequency/schedule
15
10
C (o m p l i a n ce )
Taking too little
Taking too much
60
42
61
59
Condition not adequately treated
47
37
45
43
Preventive therapy required
13
16
16
U (ntre ate d
co n d i t i o n s )
M (o n i to r i n g
re q u i re d )
Laboratory monitoring
Non laboratory monitoring
36
27
41
40
T (ox i c it y)
Toxicity evident
Risk on adverse effects a
Possible drug treatment in response to
adverse effect a
30
23
32
31

a) Modifications made to the original DOCUMENT system for the current study.
73
At baseline, there were no differences in age, sex and numbers of prescribed drugs
in both groups (Table 1). Furthermore, there were no significant differences in
the types of drugs prescribed. Antithrombotic agents were the most commonly
prescribed drugs in both groups. In the intervention group, a mean of 8.6 potential
DRPs per patient were found compared with 8.4 in the waiting-list group (P = 0.8).
In both groups, DRPs were most often classified as either Monitoring or Drug
selection (Table 1).
Between t = 0 and t = 6, eight patients in the intervention group were lost to followup: two had died, four had moved away and two patients were admitted to a nursing
home. Two patients in the waiting-list group died between t = 0 and t = 6. At t = 6,
55 patients in the intervention group and 53 in the waiting-list group had complete
medication records (Figure 1).
Patients in the intervention group (n = 55) had a mean of 4.5 potential DRPs with
a recommendation for drug change versus 4.4 in the waiting-list group (n = 53).
At t = 6, mean DRPs per patient had decreased by 30% in the intervention group
versus 5% in the waiting-list group (P < 0.01) (Table 2).
Between t = 0 and t = 6, the mean number of drug changes per patient was
significantly higher among patients in the intervention group compared with
patients in the waiting-list group (120/51 vs. 51/53, P = 0.02). Of these drug
changes, 62% (74/120) in the intervention group were related to a recommendation
versus 20% (10/51) in the waiting-list group (p < 0.01) (Table 3).
Ta b l e 3
Proportion of drug changes between between t = 0 en t = 6 related to a

recommendation
Intervention group
(n = 55)a
Waiting-list group
(n = 53)a
P-value
Ty p e o f d r u g c h a n g e
Overall
74/120 ( 62%)
10/51 (20%)
< 0.01
Cessation of drug
32/39 ( 82%)
5/9 (44%)
0.09
Dose change
16/30 ( 53%)
2/13 (15%)
0.02
Addition of drug
15/34 ( 44%)
2/23 ( 9%)
< 0.01
Replacement of drug
9/15 ( 60%)
1/6 (17%)
0.07
Dose frequency or schedule change
1/1 (100%)
0/0 ( 0%)
Drug formulation change
0/1 (
0/0 ( 0%)
0%)

a) Number of drug changes related to a recommendation/total drug changes.
74
| C hapter 3.2
Ta b l e 4
Opinions of participating community pharmacists about the medication

review process
Disagree
Not disagree nor agree
Agree
In general I agreed with the DRPs which were

formulated by the pharmacist reviewers
In general I agreed with the recommendations

which were formulated by the pharmacist
reviewers
The pharmacist reviewers had a lack of relevant

information to give specific recommendations
There was good cooperation with the GPs in

medication review
The GP was not willing to share relevant clinical

information which is important for medication
review
The GP showed reluctance concerning the

recommendations
Discussing DRPs and recommendations with the

GP was difficult
It was difficult to make appointments with the

GP about the pharmaceutical care plan of the
patient
The patient was resistant to medication review and

medication changes due to medication review
It is efficient to get a care plan from pharmacist

reviewers
I prefer to conduct medication reviews myself

instead of receiving care plans
I want to follow courses on medication review
Opinion
DRPs = drug-related problems; GP = general practitioner
Interviews with the pharmacists revealed problems in the implementation of

discussion points raised at case conferences between pharmacists and GPs.
This included difficulties in making appointments as well as differences in the
receptiveness of GPs to adapt the medication regimen in order to solve DRPs.
Table 4 shows the participating community pharmacists opinions about the
medication review process.
75
DI S C U S SION
This study shows the relevance of medication review in elderly patients receiving
drugs via MDD. These patients had a high mean number of DRPs (8.5 per patient).
We identified only one study with a comparable high mean number of DRPs (7.8
per patient);16 most studies identified a mean of less than five DRPs per patient in
home-dwelling ambulant populations.14,17-19
Medication review increased the number of drug changes and decreased the mean
number of potential DRPs per patient by 30%. This raises the question whether
drug therapy is appropriately monitored in older patients receiving MDD. In
the Netherlands, prescriptions for such patients are often repeated based on
medication lists, without critical re-evaluation.6 This is confirmed by the finding
that medication review by study pharmacists revealed a mean of 4.5 potential DRPs
with a recommendation for a drug change per patient.
The most commonly suggested recommendation was cessation of a particular drug.
Similar findings on the type of recommendations have been reported previously.18,20
In the present study, 30% of the recommendations led to an actual change in
drug therapy. Although this seems a rather low percentage, it is comparable
with the results of earlier studies in the Netherlands in which 2830% of the
recommendations suggested by community pharmacists were implemented after
case conferences.17,18 Studies in other countries have revealed a somewhat higher
implementation rate (5558%).19,21,22 Under more controlled conditions (in
hospital or using experienced clinical pharmacists), 7578% of suggested actions
were implemented.16,20 The low proportion of recommendations that resulted in
drug changes is presumably a reflection of the acceptance rate by GPs. This could
partly be explained by the process of medication review in the present study. The
reviews were conducted by pharmacist reviewers at a distance and the GP was not
involved in an early stage of the medication review process. Furthermore, there
is a structural lack of shared expectations of collaboration between community
pharmacists and GPs, and a lack of routine face-to-face interactions.23 Stricter
structuring of the cooperative relationship between GPs and pharmacists will result
in more opportunities for pharmacists in terms of their advisory and interventional
roles.24-26 GPs also seem to be more reluctant to change a medication regimen
in patients with a more complex medical profile, as was the case with patients in
the present study.18 Moreover, recommendations are based on general treatment
guidelines that GPs might not always find appropriate for the elderly population (e.g.
initiating HMG-CoA reductase inhibitor [statin] therapy might not be considered
relevant in a patient with a relatively short life expectancy). There are no evidencebased standards for the treatment of patients with multiple pathologies, leaving the
76
| C hapter 3.2
GP to treat the individual patient to the best of his/her knowledge and judgement.27
Research into the underlying reasons for low acceptance of recommendations is
therefore necessary.
In the present study, we found no indication that patients showed any resistance
toward changes in medication. Future studies should investigate whether this
assumption is valid and whether face-to-face interviews with patients using
multidose dispensing might identify additional DRPs.
The number of drop-outs was higher in the intervention group (n = 8) than in the
control group (n = 2). Although medication review could lead to identification of
clinical problems that result in admission to hospital or a nursing home, in this
study, we have no reason to assume that medication review actually resulted in such
admissions in the intervention group. It is more likely that the differences were due
to chance.
The medication review process as presented in this study is probably not a suitable
method for daily clinical practice. Half of the community pharmacists stated that
the pharmacist reviewers had a lack of information. Almost all of them said they
preferred to conduct medication reviews themselves in the future. However, the
question arises as to whether community pharmacists have sufficient expertise to
perform such reviews. A practical solution could be for community pharmacists,
who start the medication review, to participate in a course in which feedback on
reviews from expert reviewers (and portfolio building) plays an essential role.
There are some potential limitations associated with this study. First, we used
intermediate primary outcomes (i.e. the change in the number of potential DRPs and
drug changes). There is no guarantee that reducing DRPs will have a positive impact
on all clinical outcomes (e.g. hospitalizations and quality of life).28-30 However, a
correlation between the presence of DRPs and the control of cardiovascular risk
factors, quality of life and healthcare costs has recently been established.31 In our
study, the reduction in the number of DRP subtype Risk of adverse effects was due
to the cessation of (mostly) CNS drugs. This might be correlated with a reduction in
falls, because medication review in care homes has been associated with a reduction
in the number of falls.22 In the latter study, almost one-third of the medicines that
were ceased were CNS drugs, which are a well-established cause of falls.22 It is
also important to note that only drug changes could be measured; the follow-up
of recommendations with solutions other than drug changes (e.g. instructions to
patients or additional monitoring) was not documented.
Secondly, the medical information for each patient (e.g. indications and laboratory
data) was limited because we used only readily available data. A full clinical
77
medication review, including complete medical records and a patient interview,

might have revealed additional DRPs.32
Thirdly, each medication review was conducted by two reviewers from a pool of
five well trained pharmacist reviewers. This raises the question of whether less
experienced reviewers would have identified identical DRPs. On the other hand,
the use of experienced reviewers probably led to more complete and standardized
medication reviews.
Finally, patients in both groups were treated by the same pharmacists and GPs.
Recommendations in the intervention group could have led to some contamination
in the waiting-list group. However, because there were very few changes in the
waiting-list group, we believe that any such contamination would have been
relatively small.
C ON C LU SION S
This study indicates that the quality of pharmacotherapy for patients with multidose
drug dispensing can be improved. We recommend that all patients with multidose
drug dispensing should undergo a thorough medication review by pharmacists and
prescribers.
Future research should focus on the impact of a clinical medication review in
patients with automatic drug dispensing. Patient interviews can reveal user-related
problems with automatic drug dispensing, and may also serve to check the use of
drugs for self-administration.
The optimal frequency for performing medication reviews and follow-up will
probably differ between individual patients. Future studies may help provide
recommendations on the timing and frequency of medication reviews for patients
with automatic drug dispensing.
Ack n ow l e dgements - The authors thank all the participating community pharmacies,
the pharmacist reviewers J.M. Krijger-Dijkema and H. Buurma, and D. Titre for her
contribution to the data classification. The authors also wish to thank Dr Peterson and
colleagues, University of Tasmania, Australia, for their permission to include the adapted
version of their DOCUMENT classification system in the Appendix that accompanies this
chapter.
78
| C hapter 3.2
REFERENCES
1.
2008;42:1017-25.
2.
2008;168:1890-6.
3.
Beijer HM, Blaey CJD. Hospitalisations caused by adverse drug reactions (ADR): a metaanalysis of observational studies. Pharm World Sci 2002;25:46-54.
4.
Howard RL, Avery AJ, Slavenburg S, Royal S, Pipe G, Lucassen P, et al. Which drugs
cause preventable admissions to hospital? A systematic review. Br J Clin Pharmacol
2007;63:136-47.
5.
De Smet PA. Hospital admissions related to medications and implementing guidelines. Arch
Intern Med 2009;169:810-1.
6.
Johnell K, Fastbom J. Multi-dose drug dispensing and inappropriate drug use: a nationwide
7.
Gould ON, Todd L, Irvine-Meek J. Adherence devices in a community sample: how are
pillboxes used? Can Pharm J 2009;142:28-35.
8.
Larsen AB, Haugbolle LS. The impact of an automated dose-dispensing scheme on user
compliance, medication understanding, and medication stockpiles. Res Social Adm Pharm
2007;3:265-84.
9.
van den Bemt PM, Idzinga JC, Robertz H, Kormelink DG, Pels N. Medication administration
errors in nursing homes using an automated medication dispensing system. J Am Med
Inform Assoc 2009;16:486-92.
10. Bell JS, Vaananen M, Ovaskainen H, H, Narhi U, Airaksinen MS. Providing patient
care in community pharmacies: practice and research in Finland. Ann Pharmacother
2007;41:1039-46.
11. Krska J, Avery AJ, Community Pharmacy Medicines Management Project Evaluation Team.
Evaluation of medication reviews conducted by community pharmacists: a quantitative
analysis of documented issues and recommendations. Br J Clin Pharmacol 2008;65:386-96.
12. Cipolle RJ. Pharmaceutical care practice: the clinicians guide. 2nd ed. New York: McGrawHill, Medical Pub. Division, 2004.
13. Tenni P, Stafford A, Peterson G, Jackson S. A comparison of drug-related problems identified
in RMMRs and HMRs. Aust Pharm 2007;26:414-20.
14. Rasmussen M, Stafford AC, Tenni PC, Peterson GM, Jackson SL, Hejlesen A, et al. Drugrelated problems identified in medication reviews by Australian pharmacists. Pharm World
Sci 2009;31:216-23.
15. van Mil JW, Westerlund LO, Hersberger KE, Schaefer MA. Drug-related problem
classification systems. Ann Pharmacother 2004;38:859-67.
16. Krska J, Cromarty JA, Arris F, Jamieson D, Hansford D, Duffus PRS et al. Pharmacist-led
Ageing 2001;30:205-11.
79
on polypharmacy in primary care: cluster controlled trial comparing two approaches.
Br J Gen Pract 2007;57:723-31.
18. Vinks TH, Egberts TC, de Lange TM, de Koning FH. Pharmacist-based medication
review reduces potential drug-related problems in the elderly: the SMOG controlled trial.
Drugs Aging 2009;26:123-33.
2004;58:648-64.
20. Mrlin C, Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund-Udenaes M, et al. A
comprehensive pharmacist intervention to reduce morbidity in patients 80 years or older: a
randomized controlled trial. Arch Intern Med 2009;169:894-900.
21. Sellors J, Kaczorowski J, Sellors C, C, Dolovich L, Woodward C, Willan A, et al. A
randomized controlled trial of a pharmacist consultation program for family physicians and
their elderly patients. CMAJ 2003;169:17-22.
22. Zermansky AG, Alldred DP, Petty DR, Raynor DK, Freemantle N, Eastaugh J, et al. Clinical
medication review by a pharmacist of elderly people living in care homes: randomised
controlled trial. Age Ageing 2006;35:586-91.
23. Almeida Neto AC, Chen TF. Exploring elements of interprofessional collaboration between
pharmacists and physicians in medication review. Pharm World Sci 2007;29:574-6.
24. Knottnerus JA, Muijrers PEM, Grol RPTM, Sijbrandij J, Janknegt R. Pharmaceutical care
and its relationship to prescribing behaviour of general practitioners. Pharm World Sci
2006;28:302-8.
25. Muijrers PE, Grol RP, Sijbrandij J, Janknegt R, Knottnerus JA. Differences in prescribing
between GPs: impact of the cooperation with pharmacists and impact of visits from
pharmaceutical industry representatives. Fam Pract 2005;22:624-30.
26. Niquille A, Lattman C, Bugnon O. Medication reviews led by community pharmacists
in Switzerland: a qualitative survey to evaluate barriers and facilitators. Pharm Pract
2010;8:35-42.
27. Klarin I, Wimo A, Fastbom J. The association of inappropriate drug use with hospitalisation
and mortality: a population-based study of the very old. Drugs Aging 2005;22:69-82.
28. Holland R, Lenaghan E, Harvey I, Smith R, Shepstone L, Lipp A. Does home based
medication review keep older people out of hospital? The HOMER randomised controlled
trial. BMJ 2005;330:293.
30. RESPECT Trial Team. Effectiveness of shared pharmaceutical care for older patients:
RESPECT trial findings. Br J Gen Pract 2010;60:e10-9.
31. Niquille A, Bugnon O. Relationship between drug-related problems and health outcomes: a
cross-sectional study among cardiovascular patients. Pharm World Sci 2010;32:512-9.
80
| C hapter 3.2
33. Peterson G, Tenni P. Identifying, prioritizing and documenting drug-related problems.

Aust Pharm 2004;23:706-9.
81
Appe n d ix I
DOCUMENT CLASSIFICATION SYSTEM
The following classification system for drug-related problems and their resolution has been modified
from the original DOCUMENT system with permission from G. Peterson and colleagues, University of
Tasmania, Australia.
# Indicates a new or modified subcategory, see Appendix II.
D rug
selec tion
These are problems related to the choice of drug prescribed or taken

Inappropriate drug selection can occur in a number of situations, including the prescribers selection
of the drug and the pharmacists selection of a drug in the dispensing process. There are many factors
which may have led to inappropriate selection, for example a patient with multiple prescribers or
communication post-discharge or a dispensing error.
Duplication (D1)
When to Use:
Duplication refers to where a patient is prescribed and/ or is using two drugs from the same therapeutic
class.
When there are no obvious adverse clinical effects of the two drugs together, but it is either inappropriate or
very unusual to see them prescribed or used together as they are from the same therapeutic class.
This also covers the specific compliance situation where a person may be inappropriately taking two
brands of the same drug.
Examples of when to use:
Patient prescribed ranitidine plus pantoprazole
Patient prescribed both terbutaline and salbutamol inhalers
Patient taking generic amiodarone and brand name Cordarone at the same time
Patient taking a sample provided by a doctor as well as the same drug dispensed at the pharmacy
When Not to Use:
If the drugs involved are not of the same therapeutic class, then use Drug selection - Drug interaction
(D2).
If there is evidence of an adverse clinical effect use Toxicity - related to drug interaction (T2)
Drug interaction (D2)
When to Use:
When there are no obvious adverse clinical effects of the drug interaction, but the interaction is serious
enough to check if the doctor knows of it.
When there is a likely serious interaction between the patients existing therapy and a newly prescribed or
used drug, but the patient has NOT yet commenced taking the new drug.
Patient commenced on tramadol who is already taking fluoxetine
Patient ceases amiodarone while continuing on warfarin
Patient requests to purchase an over the counter antacid when taking tetracycline
When Not to Use:
If the interacting drug is of the same therapeutic class as a drug currently taken by the patient, then use
Drug selection - Duplication (D1).
If the interaction is causing, or is suspected of causing a noticeable effect of any sort, then use Toxicity Caused by drug interaction (T2).
Wrong drug (D3)
When to Use:
Where it was clear the intention of the prescriber was for a different drug from that prescribed for the
patient.
82
| C hapter 3.2
When the prescription was intended to mean a different drug and there was an error or,
When the drug being taken was prescribed correctly but was dispensed as the wrong drug.
Patient supplied with and taking Lasix (furosemide) 20 mg, while doctor prescribed Losec
(omeprazole) 20 mg
Doctor prescribes chlorpromazine 200 mg twice daily but intended carbamazepine 200 mg twice
daily
When Not to Use:
If the drug is felt to be inappropriate because of specific patient parameters such as poor renal function,
then use Contra-indication/intolerance (D5c).
If the drug prescribed is unavailable for dispensing (either because your pharmacy has no stock or the
manufacturer/distributor has no stock) then use Non-clinical (N0).
If the patient is experiencing symptoms of toxicity use Toxicity evident (T3).
Wrong dosage form (D4)
When to Use:
When the formulation of the product is inappropriate or incorrect when considering its intended use.
Also covers the specific situation where an error by the prescriber results in an absurd set of instructions (eg,
salbutamol inhaler, apply three times a day).
Vancomycin oral capsules prescribed to treat systemic infection
Ear drop product ordered or supplied for an eye problem
When Not to Use:
If the patient has a physical problem with the administration of the dosage form as it is intended to be
used (e.g. swallowing a particular form of the medication whole, cannot appropriately insert suppositories,
arthritis limiting the use of an inhaler) then use Compliance - Difficulty using dosage form (C4).
If the difficulty is related to a technical problem with the use of an administration device such as an
aerohaler, then use Education - Demonstration of device (E3).
If the difficulty is not a technical one, and related to lack of understanding of the use of the dose form, then
use Education - Confusion about therapy (E2).
Lack of indication or unclear indication (D5a)#
When to Use:
When a drug is felt to be unnecessary based on the conditions the patient has.
Patient commenced omeprazole when they were taking celecoxib for a sore knee. Celecoxib has
been ceased, but they are still taking omeprazole.
Lack of effectiveness (D5b)#
When to Use:
When you believe a more effective drug is available and you suggest it instead of the current therapy.
When current guidelines recommend another therapy.
Use of drugs such as promethazine, betahistine, mebeverine
Doxazosin as monotherapy for hypertension in patients with diabetes
When Not to Use:
Dosage is too low for drug to be effective, choose O2: Dosage too low.
Contraindication/intolerance (D5c)#
When to Use:
When there is a contraindication to the use of a drug because of an underlying condition in the patient.
When a drug or drug group is prescribed for the patient to which there has previously been a major adverse
reaction.
Propranolol in patients with diabetes
Patient has an history of allergic reactions for a certain antibiotic and receives a repeat prescription
for this antibiotic
83
When Not to Use:

If the patient is currently experiencing symptoms of toxicity, then use Toxicity/Adverse reaction - Toxicity
evident (T3).
Other drug selection problem (D0)#
When to Use:
When a less expensive or alternative brand is substituted purely for cost reasons.
When the drug being used is out of date or deteriorated in some other way.
Patient has nitroglycerine tablets for use that are over 2 years old and have been stored incorrectly
When Not to Use:
If a less expensive brand is substituted because the ordered brand is unavailable, then use Non-clinical
(N0).
O ver
or underdose prescrib ed
Problems related to the prescribed dose or schedule of the drug

These problems relates to an inappropriate dose, either the individual dose or the cumulative dose of
the drug. These are situations where the dose prescribed is not appropriate either based on previous
dosage or reference dose ranges.
Dose too high (O1)
When to Use:
When the total daily dose of a medication prescribed is too high for the patient, either based on previous
dosage or reference dose ranges. Includes situations where the dose is too high for the patient in relation to
their renal function or weight, age etc.
This also includes situations where the dose that is prescribed is too high by unintentional error.
Patient is prescribed gliclazide MR 180 mg in the morning
Patient is prescribed dexamethasone 50 mg daily (doctor was thinking of prednisolone dose)
Patient prescribed spironolactone 100 mg bd for heart failure
When Not to Use:
If the patient is taking too high a dose as a result of not following the appropriate directions, then the
situation relates to compliance use Compliance - Taking too much (C2).
Dose too low (O2)
When to Use:
When the dose prescribed is either too low based on reference dose ranges or too low based on previous
therapy.
This includes situations where the dose that is prescribed is too low by unintentional error.
Locum doctor prescribes irbesartan 150 mg daily, when previous therapy was 300mg daily
Prescription for prazosin 0.5 mg twice daily for hypertension
When Not to Use:
If the actual dose per day is suitable, but the duration is too short, then use Over/under dose - Other (O0)
If the patient is taking too low a dose of a drug as a result of not following the appropriate directions, then
the situation relates to compliance use Compliance - Taking too little (C1).
Other Dose Problem (O3)
When to Use:
When the duration of use of the product is inappropriate - either it is too short or too long.
When the total dose of a medication is suitable, but the frequency or the dosage schedule is inappropriate.
Patient prescribed cephalexin 500 mg three times a day for 3 days for cystitis
Simvastatin ordered as 40 mg in the morning
Gliclazide MR prescribed as three times daily
84
| C hapter 3.2
When Not to Use:

If the patient is not taking the appropriate dose of a product as a result of a lack of understanding of the
dosage regimen,then a compliance related code would be more appropriate.
C ompliance
Problems related to the way the patient takes the medication
These are situations where a problem arises because of the way in which the patient takes the
medication. The influences behind compliance issues are numerous. If the patient decides to take the
medication in a way other than prescribed by the doctor then the issue is one of compliance.
Taking too little (C1)
When to Use:
When the patient uses too little of a medication as a result of forgetfulness or lack of understanding of the
dosage regimen prescribed. (refill rate < 0.8)
Patient taking metformin only when required rather than regularly
Patient using Transiderm-Nitro patches only every few days, not regularly
Patient not taking medication because they believe it will stop working later on
When Not to Use:
If the underuse is appropriate because of the resolution of symptoms or a condition, then use Other drug
selection problem (D0) and specify that the drug may no longer be required.
If the patient has a physical problem with the administration of the dosage form as it is intended to be
used (e.g. swallowing a particular form of the medication whole, cannot appropriately insert suppositories,
arthritis limiting the use of an inhaler) then use Compliance -Difficulty using dosage form (C4).
Taking too much (C2)
When to Use:
When the patient uses too much of a medication as a result of forgetfulness or lack of understanding of the
dosageregimen prescribed. (refill rate > 1.2)
Patient presents requesting a second salbutamol inhaler 11 days after the previous one was
provided
Patient continuing to take 50 mg daily of prednisolone, had forgotten to commence a dose
reduction schedule as instructed by the doctor
Patient believes they have forgotten a medication and takes a second dose on the same day
When Not to Use:
If the overuse is due to an appropriate increase in use because of increased symptoms, then use Untreated
indication - Condition not adequately treated (U1).
If the overuse consists of inappropriately taking two different brands or forms of the same ingredient
unknowingly, then use Drug selection - Duplication (D1).
Intentional drug misuse (C3)
When to Use:
When there is an intentional overuse of a particular, potentially abusable, product. Includes the situation
where the prescription appears to be a forgery. Situations where misuse is suspected can also be recorded.
Patient presents a third prescription for paracetamol/codeine within 2 weeks, each of the
prescriptions was written by a different doctor
Patient presents a hand written prescription for morphine slow release and the quantity of tablets
appears to have been altered
When Not to Use:
If the overuse is due to an appropriate increase in use because of increased symptoms, then use Untreated
indication - Condition not adequately treated (U1).
85
Difficulty using dosage form (C4)

When to Use:
When the patient has a physical problem with the administration of the dosage form or device as it is
intended to be used.
Patient cannot swallow her slow release diltiazem capsules
Patient with scoliosis cannot insert suppositories
Controlled release tablet ordered for a patient who must crush all oral medications
Arthritis is limiting the patients use of their salbutamol inhaler.
When Not to Use:
If the difficulty is related to a technical problem with the use of an administration or monitoring device such
as an aerohaler, then use Education - Demonstration of device (E3).
If the difficulty is not a technical one, and related to lack of understanding of the use of the dose form, then
use Education - Confusion about therapy (E2).
Other Compliance Problem (C5)
When to Use:
When the patient is aware of the way to take the drug, is physically able to take the drug, and understands
its purpose,but does not wish to take it.
Patient unwilling to use mirtazapine after reading the package insert
When Not to Use:
If the compliance issue results in two drugs of the same therapeutic class being taken inadvertently, then
use Drug selection - Duplication (D1).
If the patient does not wish to take the medication because it is causing an adverse event of some sort, then
a Toxicity or adverse event category would be appropriate.
U ntreated
indic ations
Problems relating to actual or potential conditions that require management

This category encompasses the situations where the current disease management is suboptimal. It
is also relevant where there are potential conditions which could be avoided through preventative
therapy.
Condition not adequately treated (U1)
When to Use:
When the patient has a symptom or disease condition that is either not being treated, or is not being
treated adequately.
Patient taking triamterene/hydrochlorthiazide and perindopril for high blood pressure, but blood
pressure continues to be high
Patient develops nausea as part of a viral illness and requires addition of antinauseant medication
When Not to Use:
If the patient requires additional therapy as a preventative strategy (e.g. potassium when on a loop
diuretic), then use Untreated indication - Preventive therapy required (U2).
Preventive therapy required (U2)
When to Use:
When the patient requires additional therapy to prevent a likely adverse event as a result of the patients
therapy, coexisting diseases or risk factors.
Patient commences on morphine slow release and you suggest the addition of a stool softener
You suggest addition of omeprazole to antiplatelet therapy in a person older than 80
86
| C hapter 3.2
When Not to Use:

If the patient already has treatment for a particular problem, but it is not effective enough, then use
Untreated indication - Condition not adequately treated (U1).
Other Untreated indication Problem (U3)
When to Use:
When you think the patient has any other problem relating to actual or potential conditions that requires
management.
M onitoring
Problems related to monitoring the efficacy or adverse effects of a drug
Where the situation primarily involves monitoring the effect of the drug, this can include laboratory
monitoring (International Normalized Ratio [INR], drug level) or non-laboratory monitoring (blood
pressure).
Laboratory Monitoring (M1)
When to Use:
When, in the absence of any adverse effects, you believe that a laboratory test is required (e.g. potassium,
creatinine, white cell count, INR).
Also covers the situation where you undertake the monitoring in question and provide a recommendation
following the result. (e.g. INR monitoring and suggesting a change of warfarin dose).
When, in the absence of any adverse effects, you believe that drug level monitoring is required.
Patient recently increased furosemide dose from 40 mg daily to 120 mg daily without a change in
potassium replacement
Patient commenced on amiodarone and you recommend a thyroid function test
Elderly woman on digoxin, who has not had a blood test for two years
When Not to Use:
If there are adverse effects associated with the parameter, then use Other Toxicity problem (T0), and specify
the parameter to be tested and the symptom or sign (e.g. patient with leg cramps, suggest magnesium
level).
If the need for laboratory level monitoring comes about as a result of a newly commenced drug, then use
Drug selection - Drug interaction (D2). The monitoring then becomes a recommendation, not the primary
problem.
If the patient is experiencing an adverse effect of some sort, which you believe is due to elevated drug levels,
then use Toxicity - Caused by dose (T1).
Non-Laboratory Monitoring (M2)
When to Use:
When, in the absence of any adverse effects, you believe that non-laboratory monitoring is required. (e.g.
blood pressure, blood sugar level, temperature, weight).
Also covers the situation where the test is undertaken as a screening process.
A patient with heart failure has an appropriate increase in his dose of furosemide and you advise
him to weigh himself each day for the next week
A patient has recently changed blood pressure medication and you encourage her to keep a record
of blood pressure
When Not to Use:
If you recommend monitoring of a parameter (e.g. weight, BSL) as a result of another drug problem, then
that recommendation should be recorded in the Recommendation code section. The type of problem that
leads to this recommendation may vary.
Other Monitoring Problem (M3)
When to Use:
When the patient has another problem related to the monitoring of his drugs for either efficacy or adverse
87
effects.
When the patient should be having monitoring done, but has problems with attending the laboratory, or
paying for the test or equipment needed.
E duc ation
or inform ation
Problems related to knowledge of the disease or its management

This category includes situations where the patient is seeking information from the pharmacist.
This information relates to their current therapy either in relation to the management or the disease
process. When pharmacists are approached for information from other health professionals these
should be recorded under this category.
Patient drug information request (E1)
When to Use:
When the patient has a reasonable understanding of their condition, but requests further information
about their medication.
Patient requests information about alendronate and you provide a CMI
Patient is travelling OS and asks you what immunisations he would require
When Not to Use:
If the patient requests information primarily about the disease state, rather than a drug, then use
Education - Disease management or advice (E4).
If the patient does not request the information, but you discover that they need the information in the
course of your routine dispensing, then use Education - Confusion about therapy (E2).
If the request is not about a specific drug, but a therapeutic device, then use Education - Demonstration of
device (E3).
Confusion about therapy (E2)
When to Use:
When the patient does not understand the reasons for the use of a medication, but they still take the
medication as directed (i.e. correct dose and time).
When providing a new prescription for metoprolol for a patient with newly diagnosed
hypertension, you find that she believes that the drug may cure the condition and she can stop the
drug in a few months
Patient has been prescribed two antibiotics and they ask you why they need to take both
When Not to Use:
If the patient requests further information, then use either Education - Drug information request (E1) or
Education - Disease management or advice (E4) as appropriate.
If the confusion would have (or did) resulted in a change in compliance (either taking too much or too little
of the medication), then an appropriate compliance code should be selected.
If the request is not about a specific drug, but a therapeutic device, then use Education - Demonstration of
device (E3).
Demonstration of device (E3)
When to Use:
When the patient has a technical problem with the use of an administration or monitoring device.(eg
inhaler, BSL Monitor, Turbuhaler).
A patient is changed from a metered dose inhaler to an aerohaler and requests a demonstration of
how to use the device
Patient has decided to start using a dosage administration aid and they ask the pharmacist how to
use it correctly
88
| C hapter 3.2
When Not to Use:

If the patient understands how to use the device, but has a physical reason for not being able to use it, then
use Compliance - Difficulty using dosage form (C4).
Disease management or advice (E4)
When to Use:
When the primary purpose of the interaction with the patient was to inform them of critical aspects of the
management or prevention of a disease or condition.
Also covers the situation where the patient requests the information or where there is confusion about a
fundamental aspect of a condition they have
You counsel a patient with heart failure about fluid restriction
You provide information about weight loss or smoking cessation for a person who has
cardiovascular disease
When Not to Use:
If the patient request information primarily regarding a drug, then use Education - Drug information
request (E1).
Other Education or Information Problem (E0)
When to Use:
When another health care worker (e.g. a doctor or another pharmacist) requests information. Also covers
any other education or information related problem.
N on- clinic al
Problems related to administrative aspects of the prescription
These problems are essentially administrative, in particular where details of the prescription need
some clarification.
Not sub-classified (N0)
When to Use:
When an illegible prescription requires clarification.
When the prescription does not meet requirements.
When the drug is unavailable from the manufacturer or is out of stock temporarily.
When the dose of the medication is not specified on the prescription.
When the prescriber is not authorised to prescribe that particular medication.
When the patient has problems getting to the pharmacy or collecting prescriptions.
Physeptone 5 mg tablets not available, substitute 10 mg tablets with dose adjustment
When Not to Use:
If a less expensive or alternative brand is substituted purely for cost reasons, then use Other drug selection
problem (D0) and specify brand substitution for cost reasons.
T oxicit y
Problems related to the presence of signs or symptoms which are suspected to be related to an
adverse effect of the drug. In the presence of a sign or symptom of toxicity the situation must be
recorded under this category. It also includes situations where it is suspected that the issue is an
adverse drug reaction.
Dose related (T1)
When to Use:
When the patient has signs or symptoms that suggest an adverse reaction that is likely to be dose related.
Also where compliance issues have lead to symptoms of toxicity.
89

Patient has increased their dose of tramadol and develops headache, sweating and agitation
Promethazine and amitriptyline together causing worsening of dry mouth
Patient prescribed Diamicron MR three times daily and has significant hypoglycaemic symptoms
Patient intentionally misusing medication presents with signs or symptoms of toxicity
When Not to Use:
If the patient does not have any signs or symptoms of adverse effects and you believe the dose is too high,
then use Over or underdose - Dose too high (O1).
Caused by drug interaction (T2)
When to Use:
When the patient has signs or symptoms that suggest an adverse reaction that relates to the presence of an
interacting drug.
Patient taking warfarin develops an elevated INR after commencing metronidazole
Patient taking perindopril and frusemide, who commences diclofenac and develops renal
dysfunction
When Not to Use:
If the patient has an interacting drug present, but there are NO signs or symptoms of the interaction
causing an adverse effect, then use Drug selection - Drug interaction (D2). Also where the patient has been
prescribed interacting drugs but has not taken the medications.
Toxicity/ Adverse reaction evident (T3)
When to Use:
When there are symptoms of toxicity but the cause is not due to interaction or dose but there is a suspected
medication cause.
Patient on captopril develops a dry cough
Patient develops hypotension after commencing prazosin, even though the dose is controlling the
prostatic hypertrophy
When not to Use:
If there are no signs or symptoms of toxicity or adverse reaction.
Risk of adverse effects(T4)#
When to Use:
When you believe (prolonged) use of a drug gives a high risk of adverse effects (or has a low risk of severe
adverse effects.)
Patient is taking diazepam for years and has a increased risk of falling
Rosiglitazone has been associated with an increased risk of heart attacks
Possible drug treatment in response to adverse effect (T5)#
When to Use:
When the patient is using a drug which seems indicated for an adverse effect of another drug.
Patient on captopril is also using codeine
Patient on diltiazem has recently started a laxative
Other Toxicity/Adverse Effect problem (T0)
When to Use:
When the patient has signs or symptoms but the situation doesnt fit under the other toxicity subcategories.
Expand on the situation in the notes section.
90
| C hapter 3.2
Ap pe n d ix II Modifications to the original DOCUMENT classification system
D rug
selec tion
We have divided the original subcategory Other drug selection problem (D0) into four new
subcategories.
Th e o r i gi n a l s u b c ate g o r y wa s :
Other drug selection problem (D0)
When to Use:
reaction.
Maxolon prescribed and doctor contacted to change to Pramin
Patient has Anginine tablets for use that are over 2 years old and have been stored incorrectly
Patient commenced omeprazole when they were taking Celebrex for a sore knee. Celebrex has
been ceased, but they are still taking omeprazole
When Not to Use:
If a less expensive brand is substituted because the ordered brand is unavailable, then use Non-clinical
(N0).
If the patient is currently experiencing symptoms of toxicity, then use Toxicity / Adverse reaction - toxicity
evident(T3).
Th i s s u b c ate g o r y wa s re p l a ce d by :
Lack of indication or unclear indication (D5a)
When to Use:
Patient commenced omeprazole when they were taking Celebrex for a sore knee. Celebrex has
been ceased, but they are still taking omeprazole
Lack of effectiveness (D5b)
When to Use:
When current guidelines recommend another therapy.
Use of drugs such as promethazine, betahistine, mebeverine
Doxazosin as monotherapy for hypertension in patients with diabetes
When Not to Use:
Dosage is too low for drug to be effective, choose O2: Dosage too low.
Contra-indication/intolerance (D5c)
When to Use:
reaction.
Propranolol in patients with diabetes
Patient has a history of allergic reactions for a certain antibiotic and receives a repeat prescription
for this antibiotic
91
When Not to Use:

If the patient is currently experiencing symptoms of toxicity, then use Toxicity / Adverse reaction - toxicity
evident(T3).
Other drug selection problem (D0)
When to Use:
Patient has Anginine tablets for use that are over 2 years old and have been stored incorrectly
If a less expensive brand is substituted because the ordered brand is unavailable, then use Nonclinical (N0)
T oxicit y
We have added two new subcategories.
Th e n e w s u b c ate g o r i e s a re :
Risk of adverse effects (T4)
When to Use:
When you believe (prolonged) use of a drug gives a high risk of adverse effects (or has a low risk of severe
adverse effects).
Patient is taking diazepam for years and has a increased risk of falling
Rosiglitazone has been associated with an increased risk of heart attacks
Possible drug treatment in response to adverse effect (T5)
When to Use:
When the patient is using a drug which seems indicated for an adverse effect of another drug.
Patient on captopril is also using codeine
Patient on diltiazem has recently started a laxative
92
A B ST R AC T
Aim
To examine the effect of Home Medication Review on drug-related problems
(DRPs), disease-specific outcomes (low-density lipoprotein [LDL]-cholesterol,
systolic blood pressure [SBP] and glycosylated hemoglobin [HbA1c]) and healthrelated quality of life (EQ-5D/VAS).
Me tho ds
Community-dwelling patients ( 65 years, 5 oral chronic drugs including at least
one cardiovascular or anti-diabetic drug) were randomized at general practitioner
(GP)-level in ten Dutch community pharmacies. The community pharmacist
interviewed patients at home about their medication. A pharmaceutical care plan
was implemented after face-to-face discussion and agreement with the patients GP.
Control patients received no medication review.
Resu lts
Intervention patients had a mean of 5.8 potential DRPs leading to a recommendation
for drug change of which 1.6 (28%) were resolved after 12 months. Compared
to the control group, patients in the intervention group more often achieved
LDL-cholesterol treatment goals (odds ratio [OR] 2.0; 95% confidence interval
[95%CI] 1.13.6; P = 0.020) and had lower LDL-cholesterol levels (mean 0.18
mmol/l; 95%CI 0.33 to 0.04; P = 0.015). Effects on LDL-cholesterol were more
pronounced for the secondary prevention subgroup. No differences were found for
SBP- and HbA1c treatment goals, the combined endpoint of all treatment goals and
EQ-5D utility and VAS scores.
C on clusi on
Home medication review resolved a substantial number of drug-related problems
and improved LDL-cholesterol, but did not lead to improvements in other diseasespecific outcomes and health-related quality of life. Future studies in medication
review could focus on strategies that detect those patients for whom DRPs have the
most negative impact (Trialregister.nl number, NTR1036).
96
H o m e M ed icat io n R e view : R C T
| C hapter 4.1
I N T ROD U C T ION
Older patients increasingly receive multiple drugs to manage common chronic
diseases,1 resulting in complex medication regimens.2 Adverse drug reactions
(ADR) are a leading cause of morbidity; at least 5% of hospital admissions are
directly related to adverse drug reactions.1,3-5 Almost half of these medicationrelated hospitalizations have been suggested to be avoidable.5,6
Medication review has been recommended for older patients to prevent hospital
admissions.6,7 Studies on the effects of medication reviews in community
pharmacies have repeatedly shown positive effects on intermediate outcomes such
as resolution of drug-related problems and medication adherence. 8-10 Moreover,
studies in specific disease areas have shown improved control of disease-specific
outcomes such as for hypertension,11-14 dyslipidemia 12,15-17 and diabetes.18,19
Regarding dyslipidemia, pharmacist interventions with a collaborative approach
were effective in optimizing cholesterol management especially for patients with
an indication for secondary prevention of cardiovascular events, i.e. patients with
coronary artery disease or diabetes mellitus.15-17 Until now, medication review
studies in primary care did not show an effect on more general clinical outcomes
such as hospital admissions and health-related quality of life.20-27
Interventions in previous studies of medication review in primary care were very
heterogeneous, especially in the degree of collaboration between pharmacists
and general practitioners (GPs). For example, pharmacists did not always have
access to medical data, a patient interview was not always conducted and faceto-face case-conferences between GP and pharmacists were often lacking.9
The aim of the study was to examine the effect of Home Medication Review (HMR),
a clinical medication review 28 that incorporates all collaborative aspects including a
patient interview at home,23,29,30 on drug-related problems (DRPs), disease-specific
outcomes (low-density lipoprotein [LDL]-cholesterol, systolic blood pressure
[SBP], glycosylated hemoglobin [HbA1c]) and health-related quality of life (EQ-5D/
VAS in older community-dwelling patients.
M E T HOD S
Stu dy p ar t i cip ant s
Participants were recruited from between February 2008 and August 2010.
Patients, GPs and community pharmacists of 10 pharmacies in The Netherlands
participated. Participating community pharmacists approached at least two GPs
of different GP practices to join the study. Using pharmacy medication records,
97
patients were identified who were 65 years of age, and were chronically using
at least five oral prescription drugs, including at least one cardiovascular or one
anti-diabetic drug. Patients were excluded if 80% of their drugs were prescribed
by specialists or patients were living in nursing homes. Using computer-generated
random numbers, a selection of eligible patients were invited by letter to participate.
Community pharmacists were asked to telephone non-responders after two weeks.
Formal ethical approval was obtained from the Medical Ethical Review Board of
the Utrecht University Medical Centre. Patients gave written informed consent.
Stu dy pro ce du res
Community pharmacists who were aware of the intervention status implemented
study procedures. At baseline, 6 months and 12 months, community pharmacists
visited patients in both study groups at home for measurement of blood pressure
and assessment of health-related quality of life. Blood pressure was measured at
home by the community pharmacist according to a pretested protocol 31 by means
of the automated and validated Omron M6 Comfort monitor.32 Health-related
quality of life was measured with the EQ-5D utility score based on the British EQ5D tariff 33 and the Visual Analogue Scale (VAS). Patients were requested to visit
clinical laboratories at baseline, 6 months and 12 months to give blood samples
for laboratory measurements of HbA1c (reported in the International Federation of
Clinical Chemistry [IFCC] units) and a fasting lipid panel (total cholesterol, lowdensity lipoprotein [LDL], high-density lipoprotein [HDL] and triglycerides).
R and om is ati on an d blin d ing
A cluster randomised design with the general practitioner as the unit of
randomisation was chosen to limit contamination between patients as much as
possible. Pharmacists were blinded for the status of the GP during recruitment
of patients. After recruitment, block randomisation was performed using opaque
envelopes.
Inter venti on
The intervention was based on the Australian HMR model and has been
described elsewhere.34 Briefly, for patients in the intervention group, the patients
community pharmacist visited the patient at home for an interview about the
patients drugs. Potential DRPs were identified by the community pharmacist using
both the patients pharmacy medication records, GP medical records, additional
98
| C hapter 4.1
measurements (laboratory values, blood pressure, health-quality of life) and the

data from the patient interview. These pharmaceutical care plans were evaluated and
completed by an independent expert reviewer panel. Care plans were implemented
after face-to-face discussion and agreement between the community pharmacist
and patients GP.
E n hance d u su a l c are
Patients in the control group were also visited at home for outcome measurements
by the community pharmacist but no medication review was performed. If
necessary, the pharmacist answered questions of the patient. GPs were informed
with the patients permission when a systolic blood pressure above 180 mmHg was
measured.35,36
O utc om e me asures
The primary outcomes were the percentage of resolved drug-related problems
(DRPs) leading to recommendations for drug change in the invention group,
measured by analysing drug changes between 0 and 12 months, and the percentage
of patients achieving treatment goals on all disease-specific outcomes (LDLcholesterol 2,5 mmol/l, SBP 140 mmHg) and, for patients with diabetes, HbA1c
53 mmol/mol.35,37 Secondary outcomes were changes in the mean values of
LDL-cholesterol, SBP, HbA1c and health-related quality of life (EQ-5D/VAS). Due
to time constraints and the burden for the participating community pharmacists,
medication reviews after 12 months to identify DRPs in the control group were not
performed
D ata cl assif i c ati on
classification. Potential DRPs were classified according to the D.O.C.U.M.E.N.T.
system.34,38 All coding and classification were undertaken by two investigators
(H.K. and Y.A). When there were differences in coding, the investigators reached
consensus during a case conference with a third investigator (either A.F. or M.B.).34
S ampl e siz e c a l c u l ati on
It was expected that 50% of patients would achieve all treatment goals for LDLcholesterol, SBP and HbA1c at baseline.11-13,15-17,39-45 The intervention was expected
to result in an absolute increase of achievement of goals of 12.5% to 67.5%. With
a power of 80% and alpha of 0.05, 123 patients were needed in both study groups.
Because randomisation took place at the level of GP, we corrected the sample size
99
Fi g u re 1 Flow Consort diagram
Eligible patients
n = 801
Responders who did not gave
informed consent: 163
deceased (19)
medical conditions (29)
moved (19)
no interest (96)
Patients who withdrew: 66

deceased (6)
moved (6)
no follow-up pharmacist (9)
loss of interest/other (36)
Non-responders: 287
Responders who gave

informed consent
n = 351
Randomisation per
general practitioner
n = 285
Baseline visit + measurement

Intervention group
n = 155
Lost to follow-up: 29
deceased (2)
moved (4)
Follow-up at 6 months
n = 122
(no measurement: n = 4)
deceased (2)
moved (1)
n = 115
100
Baseline visit + measurement

Control group
n = 130
deceased (1)
moved (1)
n = 110
(no measurement: n = 6)
deceased (3)
moved (2)
n = 93
| C hapter 4.1
for the variation between GPs. We assumed a total of 10 patients for each GP, a
standard deviation of 0.05 between GPs and a standard deviation of 0.49 between
intervention and control group. This resulted in a correction factor of 1.21. With
this factor the number of patients in both study groups increased to 149 (1.21123).
Expecting a drop-out rate of 25%, we aimed at including 200 patients in both
groups.
Statist i c a l an a l ysi s
All changes of disease-specific outcomes and health-related quality of life after 6 and
12 months were analysed on an intention-to-treat basis. A sensitivity analysis was
performed for the subgroup of patients receiving secondary prevention (defined
as patients using drugs for diabetes [A10] and/or platelet aggregation inhibitors
[B01AC]).
First, outcomes were analysed at the patient level. Independent t-tests were used
for continuous variables with a normal distribution and non-parametric Mann
Whitney U tests for continuous variables with a skewed distribution. The Pearson
chi-squared (2) tests were applied for categorical variables.
Second, to account for clustering at the level of the general practitioner and to adjust
for baseline scores, linear mixed models (LMM) were used to analyse changes in
continuous parameters such as disease-specific outcomes and health-related quality
of life. A Generalized Estimating Equation (GEE) approach for binary data was used
to analyse differences in achieving treatment goals for disease-specific outcomes
with adjustment for clustering per general practitioner and baseline scores.
A P-value < 0.05 was considered statistically significant. All data were analysed
using Microsoft Access 2010 (Microsoft Corporation, Redmond, WA, USA) and
SPSS version 20.0 (SPSS Inc., Chicago, IL, USA).
R E SU LT S
Pati ents
Of 801 patients approached, 514 (64%) responded and 351 (43%) consented to
participate (Figure 1). The main reason not to give informed consent among
responding patients (n = 163) was no interest (n = 96). Sixty-six patients withdrew
before the actual start of the study and a total of 285 participants were randomized.
As a consequence of the randomisation on GP level, an unequal number of patients
were randomised to either the intervention group (155 patients from 21 GPs) or
the control group (130 patients from 20 GPs). Patients in both groups had similar
101
Ta b l e 1
Baseline characteristics of patients
Baseline characteristic
Intervention group
n = 155
Control group
n = 130
Female; n (%)
84 (54%)
57 (44%)
Age in years; median (interquartile range)
76 (72-81)
77 (72-81)
9.1 3.6
8.9 2.9
109 (72%)
80 (62%)
104 (68%)
88 (68%)
97 (64%)
83 (64%)
78 (52%)
75 (58%)
63 (41%)
73 (57%) a
49 (32%)
35 (27%)
42 (28%)
30 (23%)
47 (31%)
44 (34%)
35 (22%)
25 (17%)
31 (20%)
30 (23%)
30 (20%)
22 (17%)
LDL-cholesterol (mmol/l)
2.64 0.98
2.57 0.91
Secondary prevention
2.45 0.85
2.40 0.82
Other
3.10 1.13
3.05 0.96
M o s t p re sc r i b e d d r ug gro up s (ATC ) at t=0; n ( % )
D i s e a s e - s p e c i f i c o u tco m e s at t = 0 ; m e a n S D
SBP (mmHg)
148 22.6
149 25.6
HbA1c (mmol/mol) b
54.7 13.3
53.9 11.9
EQ-5D utility
0.73 0.26
0.74 0.23
VAS
72.2 14.4
70.1 13.2
H e a l t h - re l ate d q u a l i t y o f l i fe ; m e a n S D
SD = standard deviation; ATC = Anatomical Therapeutic Chemical classification; GORD= gastro-oesophageal

reflux disease; LDL = low-density lipoprotein; SBP = systolic blood pressure; VAS = Visual Analogue Scale
a) P < 0.05.
b) Only for patients with drugs used in diabetes (A10) (Intervention group, n = 48 and Control group, n = 45).
102
| C hapter 4.1
Ta b l e 2
Resolution rate of potential DRPs with a recommendation for a drug change

To ta l
n resolved (%)
905
250 (28%)
221
71 (32%)
Duplication
12
5 (42%)
Drug interaction
14
8 (57%)
D (r ug
S e le c t i o n )
Contra-indications apparent
23
8 (35%)
161
46 (29%)
11
4 (36%)
164
54 (33%)
Prescribed dosage too high
33
11 (33%)
Prescribed dosage too low
71
7 (10%)
Incorrect or unclear dosing instructions
60
36 (60%)
61
33 (54%)
Taking too little
25
16 (64%)
Taking too much
4 (57%)
29
13 (45%)
384
81 (21%)
281
61 (22%)
Condition untreated
67
8 (12%)
36
12 (33%)
75
11 (15%)
75
11 (15%)
No indication apparent
Other drug selection problem
O (ve r
C (o m p li a n ce )
Difficulty using dosage form
U (n d e r t re ate d )
Condition undertreated
T (ox i c i t y)
Toxicity, allergic reaction or adverse effect present
DRPs = drug-related problems
demographic characteristics and baseline drug use was similar except for drugs
for peptic ulcer and gastro-oesophageal reflux disease (Table 1). Baseline diseasespecific outcomes (LDL-cholesterol, SBP and HbA1c) and health-related quality
of life were similar in both groups. At 6 months, the number of drop-outs was 29
(19%) in the intervention group and 14 (11%) in the control group (Figure 1). The
main reason for this difference was lack of follow-up by pharmacists who did not
provide data for 12 patients. At 12 months an additional 11 patients (7%) were lost
103
Ta b l e 3
Impact of medication review on achieving LDL cholesterol, systolic blood

pressure and HbA1c goals
Goal
OR a (95%CI)
LDL-cholesterol 2,5 mmol/l
P-value
2.0 (1.13.6)
0.02
2.4 (1.24.7)
0.02
Other
1.3 (0.44.1)
0.65
1.2 (0.731.9)
0.48
SBP 140 mmHg

HbA1c 53 mmol/mol
All goals c
0.95 (0.412.2)
0.90
1.5 (0.773.0)
0.23
LDL = low-density lipoprotein; OR = odds ratio; 95%CI = 95% confidence interval; SBP = systolic blood pressure
a) ORs were obtained by a Generalized Estimating Equation (GEE) approach.
c) For patients with drugs used in diabetes (A10) SBP, LDL and HbA1c goals, for all other patients only SBP and
LDL-cholesterol goals.
to follow up in the intervention group and 23 (16%) in the control group leaving
115 patients in the intervention group and 93 patients in the control group. The
main reason for this difference was drop-out of 6 patients in the control group due
to medical conditions (e.g. hospital admission). The total drop-out rate was 26% in
the intervention group and 28% in the control group (P = 0.62).
D r ug-rel ate d probl ems
A total of 1565 potential DRPs were identified for 155 intervention patients. Nine
hundred five of these DRPs (58%) were followed by a recommendation for a drug
change (mean 5.8 per patient) (Table 2). 28% of these DRPs were resolved (mean
1.6 per patient). The most frequent identified DRP subtypes were Condition
undertreated (n = 281; 22% resolved) and No indication apparent (n = 161; 29%
resolved).
Re comm en d at i ons on d i s e a s e - sp e c i f i c outc om e s i n i nter vent i on
g roup
Eighty-nine recommendations were aimed at decreasing LDL-cholesterol of which
30% were implemented. The most common recommendations were Addition of a
drug (n = 37; 22% implemented) and Dose increase (n = 21; 24% implemented).
Nine recommendations regarding lipid lowering drugs were aimed at management
of adverse drug reactions which comprised mainly Cessation of a drug (n = 8; 25%
implemented).
104
| C hapter 4.1
Out of 94 recommendations related to blood pressure control 32 (34%) were

implemented. The most common recommendations were Replacement
of drug(n = 40; 30% implemented) and Addition of a drug (n = 31; 23%
implemented). Forty-five recommendations regarding antihypertensive drugs
were aimed at management of adverse drug reactions, mainly Cessation of a drug
(n = 35; 37% implemented).
In 45 patients with diabetes, 11 out of 43 recommendations (28%) to improve
diabetes control were implemented. The most common recommendation was
Addition of a drug (n = 22; 14% implemented). Nine recommendations were
aimed at management of adverse drug reactions; these were mainly Cessation of
drug (n = 6; 33% implemented).
D is e as e- sp e cif i c outcomes
The percentage of patients that achieved all treatment goals was 20.7% at baseline,
19.0% after 6 months and 23.8% after 12 months in the intervention group and
21.8% at baseline, 15.5% after 6 months and 19.2% after 12 months in the control
group (GEE; OR 1.5; 95%CI 0.773.0; P = 0.23).
The percentage of patients that achieved LDL-cholesterol goals in the intervention
group increased from 51.7% (baseline) to 56.6% (6 months) and 62.1% (12 months)
compared to a change from 54.1% (baseline) to 44.9% (6 months) and 55.1%
(12 months) in the control group. The results of the GEE model indicated that
participants in the intervention group had a two-fold increased chance of achieving
the LDL-cholesterol goal at each time point during follow-up compared to the
control group (Table 3, OR 2.0; 95%CI 1.13.6; P = 0.020).
Mean LDL-cholesterol levels in the intervention group were lower at each time
point during follow-up compared to the control group (Table 4, LMM; difference
0.18 mmol/l; 95%CI 0.33 to 0.04; P = 0.015). In the intervention group, also
lower mean total cholesterol levels were found at each time point (0.20 mmol/l;
95%CI 0.36 to 0.04; P = 0.013). No differences were observed for HDLcholesterol (0.13; 95%CI 0.31 to 0.05; P = 0.16) and triglycerides (+0.02; 95%CI
0.13 to 0.17; P = 0.81).
A higher implementation rate of recommendations aimed at lowering LDLcholesterol for intervention patients was observed in the secondary prevention
subgroup compared to the other subgroup (13/42 [31%] vs. 0/30 [0%]; P < 0.01).
The percentage of intervention patients achieving LDL-cholesterol goals was higher
(Table 3) and mean LDL-cholesterol levels were lowered in the intervention group
(Table 4) compared to controls in the subgroup of secondary prevention.
105
Ta b l e 4
Impact of medication review on disease-specific outcomes and healthrelated quality of life
Dependent variable
Effect size (95%CI)
P-value
LDL-cholesterol (mmol/l)
0.18 (0.33 to 0.04)
0.02
0.29 (0.46 to 0.11)
< 0.01
0.08 (0.18 to 0.33)
0.55
Total cholesterol (mmol/l)
Other
0.20 (0.36 to 0.04)
0.01
HDL-cholesterol (mmol/l)
0.13 (0.31 to 0.05)
0.16
0.02 (0.13 to 0.17)
0.76
SBP (mmHg)
0.80 (2.9 to 4.5)
0.67
DBP (mmHg)
0.17 (2.5 to 2.1)
0.89
0.02 (3.7 to 3.6)
0.99
Triglycerides (mmol/l)
HbA1c (mmol/mol)
EQ-5D Utility
VAS
0.03 (0.01 to 0.07)
0.16
0.8 (1.3 to 2.9)
0.44
95%CI = 95% confidence interval; LDL = low-density lipoprotein; HDL = high-density lipoprotein; SBP = systolic
blood pressure; DBP = diastolic blood pressure; VAS = Visual Analogue Scale
a) Results were obtained by linear mixed-effects models with control group as reference.
The percentage of intervention patients that achieved SBP-goals was 41.7% at

baseline, 45.8% after 6 months and 45.1% after 12 months compared to 45.7%,
44.0% and 44.0% for control patients (Table 3, GEE; OR 1.3; 95%CI 0.731.9;
P = 0.48). Mean SBP levels also did not differ between the intervention and the
control group (Table 4, LMM; difference +0.80 mmHg; 95%CI 2.9 to 4.5; P = 0.67).
The percentage of intervention patients that achieved HbA1c-goals was 55.3% at
baseline, 43.6% after 6 months and 52.5% after 12 months compared to 51.2%,
47.2% and 56.7% for control patients (Table 3, GEE; OR 0.95; 95%CI 0.412.2;
P = 0.90). Also, mean HbA1c-levels were not different between both groups (Table 4,
LMM, difference 0.02 mmol/mol; 95%CI 3.7 to 3.6; P = 0.99).
He a lth-rel ate d qu a lity of life
The mean EQ-5D utility score was 0.73 (standard deviation [SD] 0.26) at baseline
and 0.75 (SD 0.24) after 12 months in the intervention group compared to 0.74
(SD 0.23) at baseline and 0.71 (SD 0.29) after 12 months in the control group.
The absolute EQ-5D utility scores after 12 months were similar in both groups
(P = 0.38), as were the changes in EQ-5D utility scores (Table 4, LMM; difference
+0.03; 95%CI 0.01 to 0.07; P = 0.16).
106
| C hapter 4.1
The mean VAS-score was 72.2 (SD 14.4) at baseline and 71.0 (SD 11.8) after 12
months in the intervention group compared to 70.1 (SD 13.2) at baseline and 69.9
(SD 13.9) after 12 months in the control group. The absolute VAS-scores in both
groups were similar after 12 months (P = 0.55), as were the changes in VAS scores
(Table 4, LMM; difference 0.9; 95%CI 1.2 to 3.0; P = 0.41).
DI S C U S SION
This study shows that home medication review can resolve a substantial number
of drug-related problems for older community-dwelling patients. However, home
medication review did not have a beneficial effect on all treatment goals. For LDLcholesterol, the proportion of patients that achieved treatment goals increased and
mean levels were lowered, which can be fully attributed to the subgroup of patients
receiving secondary prevention. No significant changes were observed for systolic
blood pressure, HbA1c and health-related quality of life.
Only 28% of all DRPs with a recommendation for drug change were resolved. This
relatively low implementation rate is comparable with earlier Dutch studies on
medication review.46-48 Nevertheless a higher implementation rate would have been
expected because the study-intervention comprised many collaborative aspects of
medication review.9 Interviews with GPs in our study revealed that they generally
considered the recommendations of the community pharmacists clinically relevant.
However, a GP may decide not to implement a recommendation after weighing
the potential benefit of a drug change for an individual patient in the light of the
severity of the disease, the necessity of treatment and several patient characteristics
including prognosis. Furthermore, the amount of almost 6 potential DRPs and
recommendations per patient may be too high to be implemented at once, however
an implementation process lasting more than 12 months may be too long. The
implementation of recommendations may not have been monitored sufficiently.
Other reasons might include patient reluctance, previous failed attempts of the
same strategy and the occurrence of more urgent issues.22
The improvement of LDL-cholesterol outcomes was fully attributable to the
subgroup of patients receiving secondary prevention. This was in agreement with a
higher implementation rate of recommendations for patients in this subgroup. Our
results are in line with earlier studies on disease-specific reviews in dyslipidaemia
that showed that collaborative interventions of pharmacists with GPs could improve
cholesterol outcomes for secondary prevention patients with increased lipid
values.15-17 Our results with respect to LDL-cholesterol outcomes are remarkable,
107
because in our study, the interventions that were aimed to improve lipid values were
part of general medication review recommendations, and our patient population
consisted not solely of secondary prevention patients with increased lipid values.
Interestingly, the implementation rate of recommendations to change
hyperlipidemia treatment was relatively low and could therefore not account fully
for the achievement of LDL-cholesterol goals in our study. The effect of achievement
of LDL-cholesterol goals in our intervention group may perhaps be explained by an
increased overall medication adherence due to the pharmacist home visits.
No significant changes were found for SBP and HbA1c goals in our study. In earlier
studies of disease-specific medication reviews, pharmacists managed to improve
blood pressure control for patients with hypertension 42,43,45 and to improve HbA1c in
patients with diabetes.39-41 The interventions for patients with hypertension in these
earlier studies were more intensive compared to our study with monthly meetings
with pharmacists who were allowed to make appropriate changes in prescribed
drugs, adjusted dosages and provided drug counselling.43 Pharmaceutical care for
patients with diabetes in those studies was also characterized by more frequent
patient contacts and monitoring.41 In all these studies pharmacists appeared to
implement medication changes without face-to-face meetings with prescribers.39-43,45
Moreover, baseline values of SBP and HbA1c in these studies were mostly much
higher than in our study, giving more opportunity for improvement. Patients in
the control group with high blood pressure measurements were referred to their
GP. Finally, an equal number of recommendations concerning antihypertensive
drugs were aimed at intensifying treatment to increase effectiveness as were aimed
at discontinuing treatment because of side effects. This is not expected to improve
blood pressure management, but could decrease patients experienced side effects.
In our study, we did not find a difference in health-related quality of life between
the intervention group and the control group. The EQ-5D utility was measured
according to the British tariff, but measurements with the Dutch tariff 49 showed
comparable results. The EQ-5D utility scores in our study were relatively high and
comparable with another study with community-dwelling patients not recently
discharged from hospital.50 These high EQ-5D utility scores in our study make
it more difficult to achieve an improvement in health-related quality of life. For
hospitalised patients with lower baseline scores improvement has been shown after
medication review.51 However, our results are in line with earlier studies in primary
care with comparable patient populations which also showed no effect on total
scores of quality of life measured by SF-36 20-26 or EQ-5D/VAS.27,50 Still, the results
of our study also indicate that the intervention group experienced no worsening of
quality of life compared to the control group.
108
| C hapter 4.1
Our study has several strengths. First of all, the study was conducted in a practice
setting with patients own pharmacists and GPs. Second, the intervention
comprised many key elements reflecting collaborative aspects (pharmacist with
clinical experience, own pharmacist involved, sharing of medical records, patient
interview by pharmacist, case-conference between GP and pharmacist, action plan,
follow-up).34 Third, next to the availability of all clinical records of the patients,
additional laboratory values and blood pressure values were obtained as part of the
RCT. Additional DRPs could be identified from these measurements.
There are some limitations. First, a considerable part of the eligible patients were
classified as non-responder. It was unclear if all community pharmacists actually
performed the telephone reminder after two weeks. It is also possible that nonresponders were contacted by telephone, but the community pharmacist forgot to
ask or register the reason for not giving informed consent. Second, disease-specific
goals for cholesterol, blood pressure and HbA1c were based on general prescription
guidelines.35-37 It may be questioned if these goals can always be applied for older
people, as some classical determinants of disease and mortality in middle age, such
as hypertension and high LDL-cholesterol levels, have been shown to reverse or
disappear in old age.52-54 Third, we obtained detailed and complete information on
intermediate outcomes, which were the primary outcome of this study, but we did
not measure clinical outcomes, such as medication-related hospital admissions. Not
only would it be difficult to obtain detailed data on these hospital admissions and
to establish a causal relation between medication use and these hospital admissions,
our study would also be underpowered to detect any differences in hospital
admissions between the intervention group and control group because of its small
sample size. Finally, although cross contamination of GPs was prevented as much as
possible, cross contamination at pharmacist level could have occurred.
The scope of the medication review in this study was much broader than
the three disease-specific outcomes measured. For cholesterol, blood
pressure and glycosylated hemoglobin, disease-specific reviews may be more
effective.11,12,15-17,41,42,44 It is therefore also questionable if health-related quality
of life is the appropriate outcome for general medication review in communitydwelling patients. Many DRPs related to other diseases were identified in this study
that could lead to mixture of improved short and long term health outcomes in
patients. Whereas effects on short term health outcomes could lead to an improved
health-related quality of life, this might be offset by preventive efforts to improve
long term health outcomes. The latter may lead to an increase in drug use and
related minor side-effects that have a negative effect on quality of life. Medication
review with a more focussed approach on patients needs, concerns and complaints
109
could possibly have more effect on health-related quality of life. For example, pain
management may improve the pain/complaints dimension of the EQ-5D healthrelated quality of life. We therefore suggest that future medication review studies
find strategies to detect those patients for whom DRPs have the most negative
impact.
C ON C LU SION
In conclusion, home medication review in older patients in primary care resolves
a substantial number of drug-related problems and improved LDL-cholesterol,
but does not lead to improvement of other disease-specific outcomes, like systolic
blood pressure and glycosylated hemoglobin nor health-related quality of life.
Ack n ow l e dgements This study was conducted in pharmacies of BENU Apotheken
(formerly known as LLOYDS Apotheken). The authors like to thank all ten participating
community pharmacies, L.A. Schul of BENU Apotheken for her coordinating activities, J.M.
Krijger-Dijkema as a pharmacist reviewer and Y. Amarouchi, student, for his contribution
to the data classification.
REFERENCES
1.
2008;42:1017-25.
2.
Moen J, Antonov K, Larsson CA, Lindblad U, Nilsson JL, Rastam L, et al. Factors associated
with multiple medication use in different age groups. Ann Pharmacother 2009;43:1978-85.
3.
Howard RL, Avery AJ, Slavenburg S, Royal S, Pipe G, Lucassen P, et al. Which drugs cause
preventable admissions to hospital? A systematic review. Br J Clin Pharmacol 2007;63:13647.
4.
Beijer HM, Blaey CJ de. Hospitalisations caused by adverse drug reactions (ADR): a metaanalysis of observational studies. Pharm World Sci 2002;25:46-54.
5.
2008;168:1890-6.
6.
De Smet PA. Hospital admissions related to medications and implementing guidelines. Arch
Intern Med 2009;169:810-1.
110
| C hapter 4.1
7.
Leendertse AJ, de Koning FH, Goudswaard AN, Jonkhoff AR, van den Bogert SC, de Gier
HJ, et al. Preventing hospital admissions by reviewing medication (PHARM) in primary
care: design of the cluster randomised, controlled, multi-centre PHARM-study. BMC Health
Serv Res 2011;11:4.
8.
Holland R, Desborough J, Goodyer L, Hall S, Wright D, Loke YK. Does pharmacist-led

systematic review and meta-analysis. Br J Clin Pharmacol 2008;65:303-16.
9.
Kwint HF, Bermingham L, Faber A, Gussekloo J, Bouvy ML. The Relationship between the
Extent of Collaboration of General Practitioners and Pharmacists and the Implementation
of Recommendations Arising from Medication Review: A Systematic Review. Drugs Aging
2013;30:91-102.
10. Geurts MM, Talsma J, Brouwers JR, de Gier JJ. Medication review and reconciliation with
cooperation between pharmacist and general practitioner and the benefit for the patient: a
systematic review. Br J Clin Pharmacol 2012;74:16-33.
11. Bosworth HB, Powers BJ, Olsen MK, McCant F, Grubber J, Smith V, et al. Home blood
pressure management and improved blood pressure control: results from a randomized
controlled trial. Arch Intern Med 2011;171:1173-80.
12. Bunting BA, Smith BH, Sutherland SE. The Asheville Project: clinical and economic
outcomes of a community-based long-term medication therapy management program for
hypertension and dyslipidemia. J Am Pharm Assoc 2008;48:23-31.
13. Tjia J, Fischer SH, Raebel MA, Peterson D, Zhao Y, Gagne SJ, et al. Baseline and Follow-up
Laboratory Monitoring of Cardiovascular Medications. Ann Pharmacother 2011;45:1077-84.
14. Machado M, Bajcar J, Guzzo GC, Einarson TR. Sensitivity of patient outcomes to pharmacist
interventions. Part II: Systematic review and meta-analysis in hypertension management.
Ann Pharmacother 2007;41:1770-81.
15. Straka RJ, Taheri R, Cooper SL, Smith JC. Achieving cholesterol target in a managed care
organization (ACTION) trial. Pharmacotherapy 2005;25:360-71.
16. Tsuyuki RT, Johnson JA, Teo KK, Simpson SH, Ackman ML, Biggs RS, et al. A randomized
trial of the effect of community pharmacist intervention on cholesterol risk management:
the Study of Cardiovascular Risk Intervention by Pharmacists (SCRIP). Arch Intern Med
2002;162:1149-55.
17. Tsuyuki RT, Olson KL, Dubyk AM, Schindel TJ, Johnson JA. Effect of community pharmacist
intervention on cholesterol levels in patients at high risk of cardiovascular events: the
Second Study of Cardiovascular Risk Intervention by Pharmacists (SCRIP-plus). Am J Med
2004;116:130-3.
18. Machado M, Bajcar J, Guzzo GC, Einarson TR. Sensitivity of patient outcomes to pharmacist
interventions. Part I: systematic review and meta-analysis in diabetes management. Ann
19. Santschi V, Chiolero A, Paradis G, Colosimo AL, Burnand B. Pharmacist Interventions to
Improve Cardiovascular Disease Risk Factors in Diabetes: A systematic review and metaanalysis of randomized controlled trials. Diabetes Care 2012;35:2706-17.
20. Krska J, Cromarty JA, Arris F, Jamieson D, Hansford D, Duffus PRS, et al. Pharmacist-led
Ageing 2001;30:205-11.
111
21. Bryant LJ, Coster G, Gamble GD, McCormick RN. The General Practitioner-Pharmacist
2004;58:648-64.
2007;36:292-7.
31. van de Steeg-van Gompel CH, Wensing M, De Smet PA. Implementation of adherence
support for patients with hypertension despite antihypertensive therapy in general practice: a
cluster randomized trial. Am J Hypertens 2010;23:1038-45.
32. Topouchian JA, El Assaad MA, Orobinskaia LV, El Feghali RN, Asmar RG. Validation of two
automatic devices for self-measurement of blood pressure according to the International
Protocol of the European Society of Hypertension: the Omron M6 (HEM-7001-E) and the
Omron R7 (HEM 637-IT). Blood Press Monit 2006;11:165-71.
33. Dolan P. Modeling valuations for EuroQol health states. Med Care 1997;35:1095-1108.
34. Kwint HF, Faber A, Gussekloo J, Bouvy ML. The contribution of patient interviews to the
identification of drug-related problems in home medication review. J Clin Pharm Ther
2012;37:674-80.
35. Smulders YM, Burgers JS, Scheltens T, van Hout BA, Wiersma T, Simoons ML, et al. Clinical
practice guideline for cardiovascular risk management in the Netherlands. Neth J Med
2008;66:169-74.
36. Burgers JS, Simoons ML, Hoes AW, Stehouwer CD, Stalman WA. Guideline Cardiovascular
Risk Management. Ned Tijdschr Geneeskd 2007;151:1068-74.
112
| C hapter 4.1
37. Bouma M, Rutten GE, de Grauw WJ, Wiersma T, Goudswaard AN, Nederlands Huisartsen
Genootschap. Summary of the practice guideline Diabetes mellitus type 2 (second revision)
from the Dutch College of General Practitioners. Ned Tijdschr Geneeskd 2006;150:2251-6.
38. Williams M, Peterson GM, Tenni PC, Bindoff IK, Stafford AC. DOCUMENT: a system for
classifying drug-related problems in community pharmacy. Int J Clin Pharm 2011;34:4352.
39. Berringer R, Shibley MC, Cary CC, Pugh CB, Powers PA, Rafi JA. Outcomes of a community
pharmacy-based diabetes monitoring program. J Am Pharm Assoc 1999;39:791-7.
40. Cioffi ST, Caron MF, Kalus JS, Hill P, Buckley TE. Glycosylated hemoglobin, cardiovascular,
and renal outcomes in a pharmacist-managed clinic. Ann Pharmacother 2004;38:771-5.
41. Clifford RM, Davis WA, Batty KT, Davis TM, Fremantle Diabetes Study. Effect of a
pharmaceutical care program on vascular risk factors in type 2 diabetes: the Fremantle
Diabetes Study. Diabetes Care 2005;28:771-6.
42. Garcao JA, Cabrita J. Evaluation of a pharmaceutical care program for hypertensive patients
in rural Portugal. J Am Pharm Assoc 2002;42:858-64.
43. Vivian EM. Improving blood pressure control in a pharmacist-managed hypertension clinic.
Pharmacotherapy 2002 ;22:1533-40.
44. Chabot I, Moisan J, Gregoire JP, Milot A. Pharmacist intervention program for control of
hypertension. Ann Pharmacother 2003;37:1186-93.
45. Zillich AJ, Sutherland JM, Kumbera PA, Carter BL. Hypertension outcomes through blood
pressure monitoring and evaluation by pharmacists (HOME study). J Gen Intern Med
2005;20:1091-6.
Gen Pract 2007;57:723-31.
49. Lamers LM, McDonnell J, Stalmeier PF, Krabbe PF, Busschbach JJ. The Dutch tariff: results
and arguments for an effective design for national EQ-5D valuation studies. Health Econ
2006;15:1121-32.
50. Community Pharmacy Medicines Management Project Evaluation Team. The MEDMAN
study: a randomized controlled trial of community pharmacy-led medicines management
for patients with coronary heart disease. Fam Pract 2007;24:189-200.
51. Bladh L, Ottosson E, Karlsson J, Klintberg L, Wallerstedt SM. Effects of a clinical pharmacist
service on health-related quality of life and prescribing of drugs: a randomised controlled
trial. BMJ Qual Saf 2011;20:738-46.
52. Weverling-Rijnsburger AW, Jonkers IJ, van Exel E, Gussekloo J, Westendorp RG. Highdensity vs low-density lipoprotein cholesterol as the risk factor for coronary artery disease
and stroke in old age. Arch Intern Med 2003;163:1549-54.
113
53. van Bemmel T, Gussekloo J, Westendorp RG, Blauw GJ. In a population-based prospective
study, no association between high blood pressure and mortality after age 85 years.
J Hypertens 2006;24:287-92.
54. de Ruijter W, Westendorp RG, Assendelft WJ, den Elzen WP, de Craen AJ, le Cessie S, et al.
Use of Framingham risk score and new biomarkers to predict cardiovascular mortality in
older people: population based observational cohort study. BMJ 2009;338:a3083.
114
A B ST R AC T
Aim
To what extent patient interviews contribute to the identification of drug-related
problems (DRPs) in home medication review in terms of number, type and clinical
relevance.
Me tho ds
We performed a cross sectional study within the intervention arm of a randomised
controlled trial (RCT). Patients were recruited from ten Dutch community
pharmacies. Patients were eligible if they were home-dwelling, aged 65 years and
over and used five or more different drugs, including at least one cardiovascular or
anti-diabetic drug.
The community pharmacist interviewed the patient at home about the medicines
and identified potential DRPs in combination with medication and clinical records.
This medication review was completed by an independent pharmacist reviewers
panel.
Outcomes were the number and type of DRPs and recommendations and
percentage of clinical relevant DRPs. Clinical relevance of DRPs was assessed by
DRPs assigned a high priority, DRPs followed by recommendations for drug change
and DRPs followed by implemented recommendations for drug change.
Re su lts
A total of 1565 potential DRPs and recommendations (10 per patient) were
identified for 155 patients (median age 76 years, 54% female). 58% of all recommendations involved a drug change.
27% of all DRPs were identified during patient interviews and 74% from medication
and clinical records. Compared to DRPs identified from patient medication and
clinical records, DRPs identified during patient interviews were more frequently
assigned a high priority (odds ratio [OR] 1.8; 95% confidence interval [95%CI]
1.42.2), were more frequently associated with recommendations for drug change
(OR 2.4; 95%CI 1.93.1) and implemented recommendations for drug change
(OR 2.8; 95%CI 2.13.7).
C onclusi on
This study shows that more than a quarter of all DRPs were identified during
patient interviews. DRPs identified during patient interviews were more frequently
assigned a higher clinical relevance.
116
Co nt r i b u t io n o f p at ie nt inte r v ie ws
| C hapter 4.2
I N T ROD U C T ION
Studies have identified potential drug-related problems (DRPs) during different
types of medication review.1-8 Medication reviews can be solely based on patient
medication and clinical records,1,2 but can also be combined with patient
interviews.3-8 This more extensive review is known as a clinical medication review.9
Patient interviews have been performed in several settings such as hospitals (3),
pharmacies,8 GPs offices 4,7 or at the patients home.5,6 In Australia, a patient
interview at home by an accredited pharmacist is the predominant method of
clinical medication review.5,10-12 In Europe, the patient is often invited to the
community pharmacy for an interview, like the Medicine Use Review (MUR) in
England.13 Different European studies in primary care included patient interviews
at home 6,14-16 but these were not always conducted by a clinically well-trained
pharmacist and in close cooperation with a general practitioner like in Australia.14-16
Although DRPs identified during patient interviews have been shown to be
clinically relevant, it is unclear to what extent additional DRPs are identified
when complete clinical and medication records are available. Moreover, limited
knowledge is available on the clinical relevance of DRPs identified through patient
interviews compared to DRPs identified from clinical and medication records.3
This study examines the contribution of a patient interview to the identification
of DRPs in home medication review with the availability of complete dispensing
and clinical records. Moreover, we aimed to compare the clinical relevance of DRPs
identified during patient interviews to DRPs identified by the combination of
patient medication and clinical records.
M E T HOD S
Stu d y d esig n
This was a cross sectional study within the intervention arm of a randomised
controlled trial (RCT) in a primary care setting. This RCT aimed to assess whether
home medication reviews could reduce the number of DRPs and increase the
proportion of patients with adequate control of blood pressure, cholesterol and
glycosylated hemoglobin. Patients were enrolled between February 2008 and
August 2010.
Pati ent s
Patients were recruited from ten Dutch community pharmacies. Patients were
eligible if they were home-dwelling, aged 65 years and over and used at least five oral
117
prescription drugs, including at least one cardiovascular or one anti-diabetic drug.

Consenting patients were visited and interviewed face-to-face by a pharmacist at
home. Patients were excluded if the majority of drugs were prescribed by specialists.
Formal ethical approval was obtained from the medical ethical review board of
the Utrecht University Medical Centre. Patients gave written informed consent.
Inter venti on
For all patients complete patient medication records from the community pharmacy
including drug dispensing records, information on co-morbidity, drug intolerances
and other relevant patient notes were available. Since the majority of patients in
The Netherlands are registered at only one community pharmacy, independently
of prescriber, patient medication records are virtually complete with regard to
prescription drugs.17 The community pharmacists collected data from the clinical
records of the patient with the help from GP practice, including medical history
and laboratory data. As part of the study protocol, patients were offered additional
laboratory measurements of HbA1c, cholesterol, sodium, potassium and creatinine
and blood pressure measurement. The patients community pharmacist interviewed
the patient at home aiming to identify possible DRPs.
During the home visit the community pharmacist evaluated all medications
patients kept at home including discontinued prescription drugs, over-thecounter drugs and complementary and alternative medicines (patients indicated
whether each medication was currently taken). Community pharmacists had
limited experience with medication review. Therefore, they received a two day
training course in medication review as a part of this study. Within this course,
pharmacists were taught how to perform an structured medication review and
how to communicate with patients about adherence and understanding of the
drug therapy regimen and about patients experiences and concerns regarding
drug therapy (in particular possible adverse effects). GPs did not receive additional
training in medication review.
A pharmaceutical care plan was proposed by the community pharmacist using
both the patient medication and clinical records (including additional laboratory
data and blood pressure collected as part of the study protocol) and the data from
the patient interview. These pharmaceutical care plans were evaluated, if necessary
adjusted, and completed by two independent pharmacists from a pool of three
pharmacist reviewers (A.F; J.K.D; and H.K.). Pharmacist reviewers had several
118
| C hapter 4.2
years of experience with medication review as well as in-depth knowledge of

national clinical guidelines.
Reviewers used both implicit and explicit criteria to identify potential DRPs.
Explicit criteria consisted of a list of clinical rules based on Dutch treatment
and prescription guidelines. Examples of clinical rules were Lack of appropriate
treatment for secondary prevention for CHD (antiplatelet, lipid-lowering,
-blocker, ACE-inhibitor), Lack of appropriate treatment for patients with diabetes
and LDL > 2.5 mmol (lipid lowering) or No available monitoring data for blood
pressure, lipids, glucose, BMI or data > 1 year old in patients for whom these
measurements are indicated in treatment guidelines.18
Implicit criteria for identifying DRPs were based on a structural assessment as
proposed by Cipolle et al. in the rational order of indication, effectiveness, safety
and compliance.19,20 The two reviewers reached consensus in a case conference.
If no consensus was met, the third reviewer was consulted until consensus was
reached. For example, when the two reviewers could not agree (e.g. on the necessity
of gastric protection with a proton pump inhibitor in a geriatric patient using
acetysalicylic acid, or on the need to change the dosing moment for simvastatin
from the afternoon to the evening).
DRPs were prioritized by the pharmacist reviewers as high, medium or low from
patients perspective, with the highest priority on those that cause the most concern
(20) and needs action. For example, high priority was assigned to recommendations
that could directly relief patient complaints, or to recommendations following on
measurement of a deviating laboratory value or blood pressure.
The pharmaceutical care plans were sent to the community pharmacist to be
discussed in a case conference with the patients GP within four weeks. DRPs
with high priority were asked to be discussed first with the GP, followed by DRPs
with medium priority while DRPs with low priority were considerations with low
urgency.
D at a cl assi f i c at i on
classification. Potential DRPs were classified using the D.O.C.U.M.E.N.T.
classification system 1,21,22 using the recently updated version.23,24
All coding and classification was independently undertaken by one investigator
(H.K.) and a student investigator (Y.A.). When there were differences in coding,
the investigators reached consensus in a case conference with a third investigator
(either A.F. or M.B.).
119
O utc om e m e asu res .

After 12 months (t = 12), medication records were collected in order to analyse
the drug changes. The outcomes could only be assessed for patients with complete
medication records available for at least six months. The total number of DRPs and
recommendations were assessed by the main investigator (H.K.). Clinical relevance
was assessed by the percentage of DRPs assigned a high priority, the percentage
of recommendations for drug change and the percentage of implemented
recommendations for drug change.
Statisti c a l an a l ysis
All data were analysed using databases (Microsoft Access 2007; Microsoft
Corporation, Redmond, WA, USA) and statistical software (SPSS version 17.0; SPSS
Inc., Chicago, IL, USA). Descriptive statistics were used for basic characteristics.
Pearson Chi-square tests were used for each categorical variable. An independent
t-test was used for comparison of the mean number of DRPs per patient. A P-value
< 0.05 was considered statistically significant. Differences between the percentages
of clinical relevant DRPs and recommendations identified during patient interviews
and those identified by medication and clinical records were assessed by odds ratios
(ORs) and corresponding 95% confidence intervals (95%CIs).
R E SU LT S
Pati ent f l ow
Patients were recruited for the intervention group in ten community pharmacies.
A total of 481 patients were eligible for participation in the intervention group of
the study (Figure 1). 396 patients were sent an invitation to participate. Of patients
invited to participate 188 patients (47%) gave Informed Consent. 33 patients gave
informed consent, but did not actually participate in the study. The reasons for not
participating were loss of interest (n = 22), patient died (n = 4), health deterioration
(n = 4) and hospital admission (n = 3). Finally 155 patients were included in this
analysis (response 39%).
Pati ent ch ar ac teristi cs
The median age of the patients was 76 years and 54% was female (Table 1). A mean
of 4.2 diagnoses were registered per patient. The most common registered diagnoses
were hypertension (52%), diabetes mellitus (37%) and hyperlipidaemia (23%).
The mean number of prescribed drugs per patient was 9.0 [range 5-33]. The most
120
| C hapter 4.2
Fi g u re 1 Patient flow chart
Eligible patients
n = 481
No invitation sent: 85
death (14)
moved away (12)
GP (7)
unknown (52)
Invited
n = 396
No informed consent: 208
no interest (26 )
too burdensome (14)
unknown (168 )
Informed consent
n = 188
Did not participate: 33
death (4)
hospital admission (3)
worsened health (4)
loss of interest (22)
Participating
n = 155
GP = general practitioner
commonly prescribed drug groups were agents acting on the renin-angiotensin

system (ACE-inhibitors and AII antagonists) (72%), antithrombotic agents (70%)
and lipid modifying agents (68%).
DRPs an d re com m en d at i ons i n gen er a l
For 155 patients, a total of 1565 potential DRPs were identified, a mean of 10 DRPs
per patient (range 4-21) (Table 2). The most frequently observed types of DRPs
were Drug Selection (28%), Undertreated (26%), and Monitoring required
(23%). DRP subtypes were most often classified as No indication apparent (21%),
Condition undertreated (18%) and Laboratory monitoring (17%).
121
Ta b l e 1
Baseline sociodemographic, medical and drug-related characteristics of 155

patients
Female; n (%)
84 (54%)
76 (7281)
Number of prescription drugs; mean per patient SD
9.0 3.6
M e d i c a l h i sto r y ; n ( % )
Hypertension
75 (52%)
Diabetes mellitus
53 (37%)
Hyperlipidaemia
33 (23%)
Coronary artery disease
27 (19%)
Pulmonary disease
23 (16%)
Arrhythmia
24 (17%)
Cerebral vascular lesion, past or TIA
25 (17%)
Cataract
20 (14%)
Osteoporosis
16 (11%)
Artrosis
16 (11%)
Heart failure
12 ( 8%)
M o s t p re sc r i b e d d r ug gro up s (ATC ) ; n ( % )
112 (72%)
109 (70%)
106 (68%)
80 (52%)
68 (44%)
50 (32%)
51 (33%)
48 (31%)
35 (23%)
33 (21%)
31 (20%)
SD = standard deviation; TIA = transcient ischaemic attack; ATC = Anatomical Therapeutic Chemical
Classification
Out of 1565 recommendations 905 (58%) comprised a recommendation for drug

change (mean 5.8 per patient) (Table 3). The most common recommendations
for drug change were Addition of a drug (18%) and Cessation of a drug
122
| C hapter 4.2
Ta b l e 2
Comparison of number and type of DRPs identified from medication and

clinical records compared to patient interviews for 155 patients
O ve ra l l
Medication and
clinical records
n (%)
1150 (100%)
M e a n p e r p at i e nt SD
Patient
interviews
n (%)
P-value
415 (100%)
7.4 3.0
2.7 2.0
317 (28%)
118 (28%)
0.74
Duplication
9 (0.8%)
3 (0.7%)
0.91
Drug interaction
9 (0.8%)
6 (1.5%)
0.23
D (r ug
s e le c t i o n )
7 (1.7%)
< 0.01
102 (25%)
0.02
11 (1.0%)
0 (0.0%)
0.05
111 (10%)
58 (14%)
0.02
24 (2.1%)
9 (2.2%)
0.92
60 (5.2%)
12 (2.9%)
0.05
27 (2.3%)
37 (8.9%)
< 0.01
18 ( 1.6%)
78 (19%)
< 0.01
Taking too little
4 (0.3%)
39 (9.4%)
< 0.01
Taking too much
2 (0.2%)
6 (1.4%)
0.02
12 (1.0%)
33 (8.0%)
< 0.01
324 (28%)
78 (19%)
< 0.01
0.04
67 (5.8%)
221 (19.2%)
O (ve r
U (n d e r t re ate d )
223 (19.4%)
56 (14%)
Condition untreated
57 (5.0%)
20 (4.8%)
0.94
34 (3.0%)
2 (0.5%)
< 0.01
350 (30%)
9 ( 2.2%)
< 0.01
260 (22.6%)
1 (0.2%)
< 0.01
90 (7.8%)
8 (1.9%)
< 0.01
o r I n fo r m at i o n
2 (0.2%)
7 (1.7%)
< 0.01
Disease management or advice
2 (0.2%)
7 (1.7%)
< 0.01
29 (2.5%)
66 (16%)
< 0.01
29 (2.5%)
66 (16%)
< 0.01
M (o n i to r i n g )
Nonlaboratory monitoring
E (d uc at i o n )
T (ox i c i t y)
Toxicity, allergic reaction or adverse

effect present
DRPs = drug-related problems; SD = standard deviation
123
Ta b l e 3
Comparisons of number and type of recommendations identified from

medication and clinical records compared to patient interviews for 155
patients
Type of recommendation
Overall
Medication and
clinical records
n (%)
Patient
interviews
n (%)
P-value
1150 (100%)
415 (100%)
Mean per patient SD
7.4 (3.0%)
2.7 (2.0%)
Recommendations for drug change
603 (52%)
302 (73%)
< 0.01
Cessation of drug
135 (12%)
70 (17%)
0.01
Dose increase
75 (6.5%)
23 (5.5%)
0.48
Dose decrease
38 (3.3%)
25 (6.0%)
0.02
Addition of drug
219 (19%)
56 (14%)
0.01
Replacement of drug
103 (9.0%)
61 (15%)
< 0.01
28 (2.4%)
46 (11%)
< 0.01
5 (0.4%)
19 (4.6%)
< 0.01
Recommend dose administration aid
0 (0.0%)
3 (0.7%)
< 0.01
547 (48%)
113 (27%)
< 0.01
Education/counselling session
16 (1.4%)
33 (8.0%)
< 0.01
Monitoring: Non laboratory
86 (7.5%)
13 (3.1%)
< 0.01
Monitoring: Laboratory
280 (24%)
11 (2.7%)
< 0.01
Adjustment of patient records
156 (14%)
40 (9.6%)
0.04
9 (0.8%)
14 (3.4%)
< 0.01
Dose frequency/schedule change
Other recommendations
Other
SD = standard deviation
(13%). Frequently recorded recommendations not involving a drug change were

Monitoring: Laboratory (19%) and Adjustment of patient records (13%).
Out of 905 recommendations for a drug change, 264 (29%) recommendations were
implemented as a drug change (Table 4). The most implemented drug changes were
Cessation of a drug (58, 23%) and Addition of a drug (53, 21%).
DR Ps an d re com m en d at i ons i n p at i ent i nter v i e ws
More than a quarter of DRPs and following recommendations (415; 27%) were
identified during patient interviews (Table 2 and Table 3).
The DRP types Compliance (19% vs. 1.6%, P < 0.01), Toxicity (16% vs. 2.5%,
P < 0.01),Over or underdose (14% vs. 9.7%, P = 0.02) and Education (1.7% vs.
124
| C hapter 4.2
Ta b l e 4
Comparison of the clinical relevance of drug-related problems identified

from medication and clinical records and during patient interviews
Clinical relevance of DRPs
Medication and
clinical records
Patient
interviews
n (%)
n (%)
1150 (100%)
415 (100%)
reference
With high priority
445 (39%)
219 (53%)
1.8 (1.42.2)
< 0.01
With recommendations for drug

change
603 (52%)
302 (73%)
2.4 (1.93.1)
< 0.01
With implemented recommendations

for drug change
145 (13%)
120 (29%)
2.8 (2.13.7)
< 0.01
Overall (reference)
OR (95%CI)
P-value
DRPs = drug-related problems; OR = odds ratio; 95%CI = 95% confidence interval
0.2%, P < 0.01) were relatively more identified during patient interviews, whereas
Monitoring (30.4% vs. 2.2%, P < 0.01) and Undertreated (28% vs. 19%, P < 0.01)
were more frequently identified from medication and clinical records (Table 2).
Toxicity, allergic reaction or adverse effect present (16% vs. 3%, P < 0.01), Taking
too little (9% vs. 0.3%, P < 0.01) and Incorrect or unclear dosing instructions (9%
vs. 2%, P < 0.01) were the main DRP subtypes that were more frequently identified
during patient interviews (Table 2). Examples of specific DRP subtypes in patient
interviews are shown in Box 1.
Dose frequency/schedule change (11% vs. 2.4%, P < 0.01) and Drug formulation
change (5% vs. 0.4%, P < 0.01) were more often recommended based on patient
interviews (Table 3). Both recommendations were also more often implemented:
Dose frequency/schedule change (7.4% vs. 1.5%, P < 0.01) and Drug formulation
change (1.4% vs. 0.1%, P = 0.03).
C lini c a l rel e van ce of DR Ps
DRPs assigned a high priority were more likely to be identified during patient
interviews than from medication and clinical records (OR 1.8; 95%CI 1.4-2.2;
P < 0.01) (Table 4). Examples of DRP subtypes identified during patient interviews
with high, medium and low priority are shown in Box 1.
Furthermore, DRPs followed by recommendations for a drug change were more
likely to be identified during patient interviews (OR 2.4; 95%CI 1.9-3.1; P < 0.01)
(Table 3 and 4).
125
126
Despite use of two antihypertensives (nifedipine 30 mg retard once daily, candesartan 16

mg once daily), measurement at home shows a very high systolic blood pressure of 198
mmHg.
Uses nitroglycerine almost daily at 5:00 PM. because of dyspnoe. She experiences flushes.
Also uses isosorbidemononitrate 60 mg once daily and diltiazem 60 mg three times a day.
Next to cardiovascular medication uses a combined budesonide and salmeterol inhaler
two times a day.
Uses betamethason ointment every day without using an emollient. Complains about
delayed healing of wounds.
Has restarted alendronic acid without consulting physician. Thought this was a preventive
measure during use of iron tablets. Stopped taking prednisolone a half year before the
interview and does not suffer from osteoporosis according to the clinical data.
Takes levothyroxin at 9:00 AM. Did not know that this has to be taken half an hour before
breakfast.
Intermittently uses furosemide 60 mg for facial edema.
Uses a coumarin and experiences severe bleeding during blood sampling. Suffers from
itching in the evening, especially when the heater goes on, and suspects this is an adverse
effect.
Experiences hoarseness during use of beclomethason 100 g three doses at once after
breakfast.
, 79
, 83
, 78
, 70
, 65
, 75
, 76
Example of DRP
Examples of DRPs identified during patient interviews
, 79
Gender,
age (years)
B ox 1
Toxicity, allergic reaction or

adverse effect present

Incorrect or unclear dosing

instructions
DRP subtype
Medium
Medium
Medium
High
High
High
High
High
Assigned
priority
Uses six different medicines (fosinopril pantoprazol, metoprolol, acetylsalicylic acid,

atorvastatin and alfuzosin) at different dosing moments whereas these could be taken at
the same time.
Uses half a tablet of 80 mg sotalol twice daily whereas 40 mg tablets are available and
would be more convenient.
Experiences coughing by captopril. However, this is not disturbing.
, 76
, 75
, 79
DRPs = drug-related problems
Uses metformine 850 mg once daily instead of the prescribed 3 times a day because of
gastrointestinal problems. Is of the opinion that he gets prescribed too many medicines.
Except metformin uses a long acting insulin before the night and a short acting insulin
before meals.
Example of DRP
Examples of DRPs identified during patient interviews
, 73
Gender,
age (years)
B ox 1
Box 1 continued


instructions

instructions
Taking too little
DRP subtype
Low
Low
Low
Medium
Assigned
priority
| C hapter 4.2
127
Finally, DRPs followed by implemented recommendations for drug change were

more likely to be identified during patient interviews (28% vs. 12%, OR 2.8; 95%CI
2.1-3.7; P < 0.01 ) (Table 4).
DI S C U S SION
This study shows that patient interviews at home contribute significantly to the
identification of clinical relevant DRPs. Not only were more than a quarter of all
DRPs identified during patient interviews, the DRPs identified during patient
interviews were also assigned higher priorities and more frequently led to
recommendations involving a drug change and were more enacted compared to
DRPs identified from clinical and medication records.
The relative contribution of patient interviews to identification of all DRPs in our
study was comparable to findings of Krska (29%) in primary care.6 Studies with
patient interviews in other settings reported higher percentages (GPs office 73%,
hospital 40%), but the intervention and population in these studies were also quite
different.3,7 In addition, in our study 53% of the DRPs in patient interviews were
assigned a high priority and 73% were followed by a recommendation involving
a drug change. The only study that previously looked into the clinical relevance of
DRPs identified by patient interviews was performed in hospitalised patients and
assessed that 65% of these DRPs were of high relevance.3
Toxicity, allergic reaction or adverse effect present was the most frequent DRP
subtype that was identified in patient interviews. This was also the most common
pharmaceutical care issue in the study of Krska.6 This finding gives support to
the assumption that the concerns of the patient were sufficiently addressed by the
community pharmacists in the home visits. This is not always obvious, because
analysis of taped consultations from the HOMER-study suggested that pharmacist
reviewers were primarily concerned with compliance and knowledge of drugs.16,25
Taking too little was the second DRP subtype that was significantly more
identified during patient interviews. Compliance issues are mentioned as DRPs in
many studies with patient interviews.6,26,27 Sturgess showed that clinical medication
review including patient interviews even improved compliance measured by selfreporting and refill rate.27 Repeated domiciliary visits after initial medication review
may also enhance compliance as was measured by pill counts.28 Their explanation
was that compliance issues might be discussed easier in a patients home than in a
busy community pharmacy.26
128
| C hapter 4.2
Incorrect or unclear dosing instructions was the third DRP subtype that
was identified significantly more in patient interviews. The corresponding
recommendation Dose frequency/schedule change was often implemented. For
many patients, minor changes in their dose schedule can diminish the frequency
of dosing. As shown in an example in Box 1, some patients are taking medication
throughout the whole day whereas these could be taken at the same time. In an
earlier domiciliary interviewing study half of the patients judged that medication
management was a major daily preoccupation and spouses were often required to
assist.28
It could be questioned whether patient interviews at home reveal additional or
other type of DRPs than patient interviews in a consulting area in the community
pharmacy or GP practice, however this was not the aim of this study. On the one
hand, patient interviews at home may elicit more and other DRPs, because patients
might feel more comfortable at home and therefore are more likely to share their
experiences and concerns about their medicines. This is illustrated by the finding
that compliance issues and adverse effects were frequently identified by patient
interviews in our study and Krskas study.6 Furthermore, all medicines were available
at home, whereas patients invited to the pharmacy or GP practice might forget to
bring part of their medicines, especially those that are used intermittently.8 Finally,
certain medication-risk factors for example lack of medication administration
routine, multiple storage locations, hoarding and medication storage conditions
seem only to be identified by home visits.26,28-30 On the other hand, costs of home
visits are also higher. Future studies should focus on cost effectiveness of patient
interviews at home compared to interviews conducted at GP practice or pharmacy.
This study had several strengths. Firstly, the intervention in our study comprised
a clinical medication review meaning that all data were available to conduct a
medication review. Next to the availability of all clinical records of the patients,
additional laboratory values and blood pressure values were obtained as part of
the RCT. Despite the availability of all these data, more than a quarter of all DRPs
were identified from patient interviews. Secondly a very detailed and accurate
description of DRPs was possible, because all pharmaceutical care plans and reports
of patient interviews were electronically sent by the community pharmacists to the
pharmacist reviewers. This made it also possible to distinct very clearly between
DRPs identified from medication and clinical records and from patient interviews
without any overlapping. Thirdly, DRPs were assigned a priority by the pharmacist
reviewers panel and this could be seen as an indicator of clinical relevance. Fourthly,
129
patient interviews were conducted by community pharmacists. More than half

of the DRPs identified from these patient interviews had a high priority. In an
earlier study, in which patient interviews were conducted by pharmacy or practice
assistants, only a quarter of identified DRPs had potential clinical relevance.31
Finally, the use of experienced pharmacist reviewers probably led to more complete
and standardized pharmaceutical care plans.
There are some limitations associated with this study. Firstly, a considerable part
of the invited patients (55%) did not give informed consent. However, our target
group is an older population who have a considerable disease burden. In particular,
the extra lab monitoring which was part of the study was not appreciated by
many potential participants. Secondly, community pharmacists had a short
course (2-days) in medication review and had limited experience in conducting
medication reviews. Finally, each initial pharmaceutical care plan by a community
pharmacist was adjusted and completed by well-trained pharmacist reviewers
which had no relationship with patients. Although standardisation is a strength of
the study, it may limit generalizability to daily clinical practice where pharmacists
might identify less DRPs, but have a relationship with patients. Future research
should further elaborate on the role of both the community pharmacists and the
pharmacist reviewers regarding the identification of DRPs.
Patient interviews by community pharmacists have a major contribution in the
identification of DRPs. This implies that, in general, medication review without a
patient interview may lead to fewer clinical relevant recommendations.
Funding - This study was conducted in pharmacies of BENU Apotheken. The
study received unrestricted research funding by the Royal Dutch Association of
Pharmacists (KNMP), Astra-Zeneca and the healthcare insurance companies
Achmea and Menzis. This work was performed totally independently from these
funders.
Ack n ow l e dgements - The authors like to thank all ten participating community
pharmacies, L.A. Schul of LLOYDS Apotheken for her coordinating activities, J.M. KrijgerDijkema as a pharmacist reviewer and Y. Amarouchi, student, for his contribution to the
data classification.
130
| C hapter 4.2
REFERENCES
1.
Kwint HF, Faber A, Gussekloo J, Bouvy ML. Effects of medication review on drug-related
2.
Denneboom W, Dautzenberg MG, Grol R, De Smet PA. Treatment reviews of older people
Gen Pract 2007;57:723-31.
3.
Viktil KK, Blix HS, Moger TA, Reikvam A. Interview of patients by pharmacists contributes
significantly to the identification of drug-related problems (DRPs). Pharmacoepidemiol.
Drug Saf 2006;15667-74.
4.
Zermansky AG, Petty DR, Raynor DK, Freemantle N, Vail A, Lowe CJ. Randomised
5.
Sorensen L, Stokes JA, Purdie DM, Woodward M, Elliott R, Roberts MS. Medication reviews
2004;58:648-64.
6.
Ageing 2001;30:205-11.
7.
Jameson JP, VanNoord GR. Pharmacotherapy consultation on polypharmacy patients in

ambulatory care. Ann Pharmacother 2001;35:835-40.
8.
Nathan A, Goodyer L, Lovejoy A, Rashid A. Brown bag medication reviews as a means of

optimizing patients use of medication and of identifying potential clinical problems. Fam
Pract 1999;16:278-82.
9.
Clyne W, Blenkinsopp A, Seal R. A guide to medication review 2008. Available from: http://
www.npci.org.uk/medicines_management/review/medireview/library/library_good_
practice_guide1.php (Accessed 2010 Apr 19).
10. Castelino RL, Bajorek BV, Chen TF. Retrospective evaluation of home medicines review by
pharmacists in older Australian patients using the medication appropriateness index. Ann
11. Castelino RL, Bajorek BV, Chen TF. Are interventions recommended by pharmacists during
Home Medicines Review evidence-based? J Eval Clin Pract 2011;17:104110.
12. Castelino RL, Hilmer SN, Bajorek BV, Nishtala P, Chen TF. Drug Burden Index and
potentially inappropriate medications in community-dwelling older people: the impact of
Home Medicines Review. Drugs Aging 2010;27:135-48.
13. Latif A, Pollock K, Boardman HF. The contribution of the Medicines Use Review (MUR)
consultation to counseling practice in community pharmacies. Patient Educ Couns
2011;83:336-44.
14. Grymonpre RE, Williamson DA, Montgomery PR. Impact of a pharmaceutical care model
for non-institutionalised elderly: Results of a randomised, controlled trial. Int J Pharm Pract
2001;9:235-41.
131

2001;18:63-77.
16. Holland R, Lenaghan E, Harvey I, Smith R, Shepstone L, Lipp A, et al. Does home based
medication review keep older people out of hospital? The HOMER randomised controlled
trial. BMJ 2005;330:293-5.
17. Buurma H, Bouvy ML, De Smet PA, Floor-Schreudering A, Leufkens HG, Egberts AC.
Prevalence and determinants of pharmacy shopping behaviour. J Clin Pharm Ther
2008;33:17-23.
19. Cipolle RJ. Pharmaceutical care practice: the clinicians guide. 2nd ed. New York: McGrawHill, Medical Pub. Division; 2004.
20. Leendertse AJ, de Koning FH, Goudswaard AN, Jonkhoff AR, van den Bogert SC, de Gier
Serv Res 2011;11:4.
22. Rasmussen M, Stafford AC, Tenni PC, Peterson GM, Jackson SL, Hejlesen A, et al. Drugrelated problems identified in medication reviews by Australian pharmacists. Pharm World
Sci 2009;31:216-23.
23. Williams M, Peterson GM, Tenni PC, Bindoff IK, Curtain C, Hughes J, et al. Drug-related
problems detected in Australian Community Pharmacies: The PROMISe Trial. Ann
classifying drug-related problems in community pharmacy. Int J Clin Pharm 2011;34:43-52.
25. Salter C, Holland R, Harvey I, Henwood K. I havent even phoned my doctor yet. The
advice giving role of the pharmacist during consultations for medication review with patients
aged 80 or more: qualitative discourse analysis. BMJ 2007;334:1101.
26. Begley S, Livingstone C, Hodges N, Willamson V. Impact of domiciliary phamacy visits on
medication management in an elderly population. Int J Pharm Pract 1997;5:111-21.
28. Coleman DJ, Portlock J, Brown D. Delivering domiciliary pharmaceutical care from a health
centre pharmacy. Int J Pharm Pract 2001;9:127-37.
29. Sorensen L, Stokes JA, Purdie DM, Woodward M, Roberts MS. Medication management at
home: medication-related risk factors associated with poor health outcomes. Age Ageing
2005;34:626-32.
30. Sorensen L, Stokes JA, Purdie DM, Woodward M, Roberts MS. Medication management
at home: medication risk factor prevalence and inter-relationships. J Clin Pharm Ther
2006;31:485-91.
31. Denneboom W, Dautzenberg MG, Grol R, De Smet PA. User-related pharmaceutical care
problems and factors affecting them: the importance of clinical relevance. J Clin Pharm Ther
2005;30:215-23.
132
A B ST R AC T
Aim
To investigate the completeness of drug-related problems (DRPs) in terms of
number, type and clinical relevance identified by community pharmacists in Home
Medication Review (HMR).
Me tho ds
A cross sectional study within the intervention arm of a RCT among communitydwelling patients ( 65 years, 5 drugs) in ten Dutch community pharmacies.
Community pharmacists, who were inexperienced in medication review, received a
two days training in medication review. These pharmacists interviewed patients at
home about their medicines and identified potential DRPs and recommendations
in combination with medication and clinical records. Expert reviewers completed
the number of potential DRPs and recommendations by reviewing all available
information, including patient interview reports. Clinical relevance of DRPs was
assessed by DRPs assigned a high priority, DRPs followed by recommendations
for drug change and DRPs followed by implemented recommendations for drug
change.
Resu lts
In 155 patients, community pharmacists identified a mean of 3.6 (standard
deviation [SD] 2.8) potential DRPs and expert reviewers added 6.5 (SD 3.2) DRPs.
Community pharmacists formulated 2.6 (SD 2.3) recommendations per patient
and reviewers added 7.5 (SD 3.3) recommendations. Community pharmacists
identified a higher proportion of clinical relevant DRPs compared to pharmacists
reviewers, as assessed by DRPs with high priority (odds ratio [OR] 1.8; 95%
confidence interval [95%CI] 1.42.2) DRPs associated with recommendations for
drug change (OR 1.9; 95%CI 1.52.3) and implemented recommendations for drug
change (OR 2.1; 95%CI 1.62.7).
C on clusi on
This study shows that inexperienced community pharmacists with limited training
identified a lower number of potential DRPs and a higher proportion of clinical
relevant DRPs compared to expert reviewers. The results suggest that community
pharmacists need intensive post-graduate training to improve the identification of
DRPs.
134
Co mp le te n e s s o f m e d ic at io n re v ie ws
| C hapter 4.3
I N T ROD U C T ION
Several randomized controlled trials (RCTs) have demonstrated that clinical
medication review can resolve drug-related problems (DRPs). However, these
RCTs differed on the expertise of participating pharmacists who were either highly
trained 1-10 or inexperienced in medication review.11-13 Medication review is a
complex intervention, which requires both knowledge and skills.14,15 Medication
review skills have to be developed to explore patients experiences and beliefs
about medicines, to identify potential DRPs, formulate recommendations for a
pharmaceutical care plan and to discuss this care plan with physicians.15
No extensive post-graduate courses in medication review were available at the
start of this RCT on home medication review in The Netherlands.16 Therefore,
community pharmacists, who were inexperienced in medication review, received
a two day training course as a part of this study where pharmacists were taught
on pharmacotherapy and medication review skills. As part of the intervention,
pharmaceutical care plans by the community pharmacists were evaluated by
independent expert reviewers which were pharmacists with several years of
experience in medication review as well as applying clinical knowledge of national
prescription guidelines in practice.16
Relatively little is known to what extent community pharmacists are able to identify
DRPs in medication review.17,18 In particular, no research has been performed into
the clinical relevance of these DRPs. The aim of this study was to investigate the
completeness of DRPs in terms of number, type and clinical relevance identified by
community pharmacists in Home Medication Review (HMR).
M E T HOD S
Stu dy d esi g n
A cross sectional study within the intervention arm of an RCT on home medication
review in a primary care setting. The study involved the collection of completed
pharmaceutical care plans pertaining to 155 community-dwelling patients ( 65
years and 5 drugs including at least one cardiovascular or one anti-diabetic drug)
in ten community pharmacies.
Inter venti on
Inexperienced community pharmacists in medication review received a two
days training. Pharmacists had access to complete medication records from the
pharmacy and collected medical history and laboratory data of the patient with
135
the help from the general practitioner (GP) practice. As part of the study protocol,
patients were offered additional laboratory measurements of HbA1c, cholesterol,
sodium, potassium and creatinine and blood pressure measurement. Pharmacists
interviewed patients at home about their medicines, identified potential DRPs and
made recommendations. Subsequently expert reviewers completed the number
of potential DRPs and recommendations by reviewing all available information,
including patient interview reports. More details about the intervention have been
described elsewhere.16
D at a c ol l e c ti on an d cl assif i c ati on
Data extracted from the pharmaceutical care plans included medication and
medical information, patient interview reports, identified potential DRPs and
associated recommendations as well as the prioritization for implementation
(high, medium or low). Complete patient medication records from the community
pharmacy including drug dispensing records until at least 6 months after the patient
interview were collected separately.
Each potential DRP and recommendation was classified as being identified by
community pharmacist or added during completion by pharmacist reviewer.
When the description of a DRP or recommendation was incomplete and refined
by a pharmacist reviewer, this DRP was assigned to the community pharmacist.
All potential DRPs and associated recommendations were classified using the
D.O.C.U.M.E.N.T. classification system using the most recent version.16,19,20
Clinical relevance was assessed by the percentage of DRPs assigned a high priority,
the percentage of recommendations for drug change and the percentage of
implemented recommendations for drug change. Implemented recommendations
for drug change were assessed by analysing drug dispensing records for medication
changes.
All coding and classification was independently undertaken by one investigator
(H.K.) and a student investigator (Y.A.). When there were differences in coding, the
investigators reached consensus in a case conference with a third investigator (A.F.
or M.B.).
Statisti c a l an a l ysi s
Pearson Chi-square tests were used to compare categorical variables. Independent
t-tests were used to compare the mean number of DRPs per patient. Differences
between the percentages of clinical relevant DRPs and recommendations identified
by community pharmacists and those added by expert reviewers were compared by
odds ratios (ORs) and corresponding 95% confidence intervals (95%CIs).
136
| C hapter 4.3
Ta b l e 1
Comparison of number and type of DRPs identified by community

pharmacists and added by expert reviewers for 155 patients
Community Pharmacists
n (%) a
O ve ra l l
Expert Reviewers
n (%) a
P-value
553 (36%)
1012 (64%)
3.6 2.8
6.5 3.2
D (r ug
156 (36%)
279 (64%)
0.79
7 (58%)
5 (42%)
0.09
S e le c t i o n )
Duplication
Drug interaction
5 (33%)
10 (67%)
0.87
28 (38%)
46 (62%)
0.65
106 (33%)
217 (67%)
0.29
10 (91%)
1 ( 9%)
< 0.01
64 (38%)
105 (62%)
0.47
14 (42%)
19 (58%)
0.39
31 (43%)
41 (57%)
0.16
19 (30%)
45 (70%)
0.33
52 (54%)
44 (46%)
< 0.01
Taking too little
25 (58%)
18 (42%)
< 0.01
Taking too much
5 (63%)
3 (38%)
0.11
22 (49%)
23 (51%)
0.05
159 (40%)
243 (60%)
0.04
O (ve r
C (o m p l i a n ce )
U (n de r t re ate d )
117 (40%)
172 (60%)
0.04
Condition untreated
28 (36%)
49 (64%)
0.85
14 (39%)
22 (61%)
0.65
75 (21%)
284 (79%)
< 0.01
58 (22%)
203 (78%)
< 0.01
Non laboratory monitoring
17 (17%)
81 (83%)
< 0.01
o r I n fo r m at i o n
1 (11%)
8 (89%)
0.13
1 (11%)
8 (89%)
0.13
46 (48%)
49 (52%)
< 0.01
46 (48%)
49 (52%)
< 0.01
E (d uc at i o n )
T (ox i c it y)
Toxicity, allergic reaction or adverse

effect present
DRPs = drug-related problems; SD = standard deviation
a) % is the percentage within DRP type or subtype.
137
A P-value < 0.05 was considered statistically significant. All data were analysed
using Microsoft Access 2007 (Microsoft Corporation, Redmond, WA, USA) and
SPSS version 17.0 (SPSS Inc., Chicago, IL, USA).
R E SU LT S
Pharma cists
Ten community pharmacies were initially recruited. During the study period
16 different community pharmacists were involved in different phases of the
medication review process.
Id entif i e d DR Ps an d re commen d at i ons
In 155 patients, community pharmacists identified 553 potential DRPs (mean
per patient 3.6; SD 2.8) and expert reviewers added 1012 potential DRPs (mean
per patient 6.5; SD 3.2, Table 1). Community pharmacists formulated 398
recommendations (mean 2.6, SD 2.3 per patient) and expert reviewers added 1167
recommendations (mean per patient 7.5; SD 3.3; P < 0.01, Table 2). Considerable
variations between the community pharmacists were observed in the mean
proportion of identified DRPs per patient (range 13%57%) and formulated
recommendations (range 10%46%).
All DRP types, except Compliance, were in absolute numbers more frequently added
by expert reviewers. DRP types Toxicity (P < 0.01) and Undertreated(P = 0.04)
were relatively more frequently identified by community pharmacists (Table 1).
All recommendations were in absolute numbers more frequently added by expert
reviewers, except Drug formulation change. Addition of drug was relatively more
frequently recommended by community pharmacists (P = 0.02) (Table 2).
DRPs i d entif i e d in p ati ent inter v i e ws
Of 415 potential DRPs originating from patient interviews, community pharmacists
identified 171 DRPs (mean per patient 1.1; SD 1.1) and expert reviewers added 244
DRPs (mean per patient 1.6; SD 1.7) (P < 0.01). No indication apparent (n = 64;
26%), Condition undertreated (n = 33; 14%), Toxicity, allergic reaction or adverse
effect present (n = 30; 12%) and Incorrect or unclear dosing instructions(n=29;
12%) were the main DRP subtypes added by expert reviewers from patient
interviews.
138
| C hapter 4.3
Ta b l e 2
Comparisons of number and type of recommendations identified by

community pharmacists and added by expert reviewers for 155 patients
Community Pharmacists
n (%) a
Expert Reviewers
n (%) a
P-value
O ve ra l l
398 (25%)
1167 (75%)
2.5 2.3
7.5 3.3
R e co m me n d at i o n s fo r d rug c h an g e
272 (30%)
633 (70%)
< 0.01
Cessation of drug
59 (29%)
146 (71%)
0.24
Dose increase
31 (32%)
67 (68%)
0.15
Dose decrease
16 (25%)
47 (75%)
0.99
Addition of drug
85 (31%)
190 (69%)
0.02
Replacement of drug
49 (30%)
115 (70%)
0.17
Dose frequency/schedule change
20 (27%)
54 (73%)
0.75
12 (50%)
12 (50%)
< 0.01
Recommend dose administration aid

O t h e r re co m m e n d at i o n s
0 ( 0%)
3 (100%)
0.31
126 (19%)
534 (81%)
Education/counselling session
10 (20%)
39 (80%)
0.41
Monitoring: Non laboratory
14 (14%)
85 (86%)
< 0.01
Monitoring: Laboratory
59 (20%)
232 (80%)
0.03
Adjustment of patient records
38 (19%)
158 (81%)
0.04
4 (17%)
19 (83%)
0.37
Other
< 0.01
SD = standard deviation
a) % is the percentage within type or subtype of recommendation.
C lini c a l rel e van c e

Community pharmacist identified a higher proportion of DRPs with a high priority
compared to expert reviewers (OR 1.8; 95%CI 1.42.2; P < 0.01). Furthermore, a
higher proportion of DRPs followed by recommendations for a drug change was
identified by community pharmacists compared to expert reviewers (OR 1.9; 95%CI
1.52.3; P < 0.01). Finally, a higher proportion of DRPs followed by implemented
recommendations for drug change was identified by community pharmacists
(OR 2.1; 95%CI 1.62.7; P < 0.01) (Table 3).
139
Ta b l e 3
Comparison of clinical relevance of drug-related problems (DRPs) identified

by community pharmacists and added by expert reviewers
Clinical relevance of DRPs
Community
Pharmacists
Expert
Reviewers
n (%)
n (%)
O ve ra l l (re fe re n ce )
553 (100%)
1012 (100%)
With high priority
285 (52%)
379 (37%)
1.8 (1.4 2.2)
< 0.01
With recommendations for drug

change
375 (68%)
530 (52%)
1.9 (1.52.3)
< 0.01
With implemented recommendations

for drug change
132 (24%)
133 (13%)
2.1 (1.62.7)
< 0.01
OR (95%CI)
reference
P-value
OR = odds ratio; 95%CI = 95% confidence interval
DI S C U S SION
This study shows that expert reviewers added almost twice the amount of potential
DRPs already identified by community pharmacists and threefold the amount
of recommendations. Interestingly, expert reviewers identified more potential
DRPs from patient interviews than did community pharmacists. However, DRPs
identified by community pharmacists were more often clinical relevant compared
to DRPs added by pharmacists reviewers.
Expert reviewers almost doubled the amount of DRPs identified by community
pharmacists. The highest difference was seen for the DRP type Monitoring.
Reviewers especially added DRPs related to appropriate monitoring of hypertension,
dyslipidemia, diabetes 21,22 and other diseases. Community pharmacists rarely
identified these. Monitoring related problems accounted for more than two
DRPs per patient, but were mostly assigned a low priority and not followed by a
recommendation for drug change. Monitoring problems and some other DRPs
identified by expert reviewers might be perceived by community pharmacists
and GPs as too theoretical or textbook advices.23 This may partly explain that
community pharmacists identified a lower number but relatively more clinically
relevant DRPs.
There was a considerable variation in the completeness of the reviews. Although
community pharmacists were stimulated to conduct full medication reviews, some
pharmacists may have relied on the expert reviewers to complete the medication
reviews. This may partly be explained by the fact that the registration of all research
data was experienced as time-consuming by the participating pharmacists.
140
| C hapter 4.3
Furthermore, changes of pharmacists during the study hampered continuity of care

around the medication review process.
Surprisingly, a considerable part of issues that were discussed with the patients
were not formally identified as potential DRP by community pharmacists. The
most frequently DRP added by expert reviewers was Indication not apparent.
This refers to patients using drugs without knowing the indication which was
also lacking in the GP record or refers to patients using drugs not intended for
prolonged use. In general, these type of DRPs are followed by a recommendation
to discontinue the drug. Condition undertreated was also frequently added by
reviewers (e.g. patients indicated that their pain treatment was suboptimal, whereas
community pharmacists had not suggested a change of drug or dosing regimen).
Furthermore, adverse effects were frequently described in patient interview reports,
but surprisingly not always identified as potential DRPs. Possibly community
pharmacists did not recognize complaints as caused by adverse effects or did
recognize these, but considered these as inevitable. Finally, Incorrect or unclear
dosing instructions were also frequently missed as potential DRPs by community
pharmacists. Community pharmacists may consider these medication management
problems (e.g. time of intake) as typical pharmacist issues and of minor interest
for discussion with GP. Often these issues were directly solved during the patient
interview by an advice (e.g. change time of intake).
The percentage of identified DRPs by community pharmacists in our study (36%)
was comparable to findings of Krska et al. (34%) with a clinical pharmacist and
an experienced GP as expert reviewers.17 In the study of Krska, training of the
community pharmacists was limited to 2 days, specifically designed for the study
and no formal assessment of their competency to conduct medication reviews was
made.17 Laaksonen et al. investigated the performance of community pharmacists
who completed a more intensive postgraduate course (i.e. five 60-hour distance
learning modules in clinical therapeutics). These trained community pharmacists
managed to identify 75% of the DRPs found by a clinical pharmacist.18
This study had several strengths. First, a very detailed and accurate description
of DRPs was available, because community pharmacists sent all pharmaceutical
care plans to the expert reviewers. This enabled us to distinct very clearly between
DRPs identified by community pharmacists and expert reviewers and also provided
insight into the clinical relevance of DRPs. Second, the availability of detailed
patient interview reports enabled us to assign DRPs originating from patient
interviews.
There are some limitations associated with this study. First, expert reviewers in
this study could only refine or add to the DRPs already identified by community
141
pharmacists. Blinding of the reviewers for DRPs identified by community

pharmacists, would have enabled a more optimal comparison between community
pharmacists and expert reviewers. However, assuming that expert reviewers
would have identified the majority of DRPs detected by community pharmacists,
this would not have a major impact on the findings of the study. Second, expert
reviewers were dependent on the provided documentation by community
pharmacists for identifying DRPs and did not conduct patient interviews. On the
one hand, expert reviewers at distance might have identified some potential DRPs
in this study that they possibly might have neglected if they knew the patient. On
the other hand, expert reviewers might have identified more or other DRPs from
patient interviews. Third, the number of DRPs identified by community pharmacists
may have been underestimated, as a limited number of issues were directly solved
during the patient interview, but not registered in the pharmaceutical care plan by
the pharmacists.
C ON C LU SION
This study shows that inexperienced community pharmacists with limited training
identified a lower number of potential DRPs and a higher proportion of clinical
relevant DRPs compared to expert reviewers. This suggests that a two days training
in medication review may be too limited. A practical solution could be to enrol
pharmacists inexperienced in medication review in an intensive post-graduate
course in which feedback on reviews from expert reviewers (and portfolio building)
plays an essential role.
Ack n ow l e dgements - This study was conducted in pharmacies of BENU Apotheken
(formerly known as LLOYDS Apotheken). The authors like to thank all participating
community pharmacists, L.A. Schul of BENU Apotheken for her coordinating activities,
J.M. Krijger-Dijkema as a pharmacist reviewer and Y. Amarouchi, student, for his
contribution to the data classification.
REFERENCES
1.
142
Bryant LJ, Coster G, Gamble GD, McCormick RN. The General Practitioner-Pharmacist
| C hapter 4.3
2.
Grymonpre RE, Williamson DA, Montgomery PR. Impact of a pharmaceutical care model
2001;9:235-41.
3.
Ageing 2001;30:205-11.
4.
Lenaghan E, Holland R, Brooks A. Home-based medication review in a high risk elderly

2007;36:292-7.
5.
Sellors J, Kaczorowski J, Sellors C, Dolovich L, Woodward C, Willan A, et al. A randomized

6.
Sorensen L, Stokes JA, Purdie DM, Woodward M, Elliott R, Roberts MS. Medication reviews
2004;58:648-64.
7.
Volume CI, Farris KB, Kassam R, Cox CE, Cave A. Pharmaceutical care research and
8.
Kassam R, Farris KB, Burback L, Volume CI, Cox CE, Cave A. Pharmaceutical care research
and education project: pharmacists interventions. J Am Pharm Assoc 2001;41:401-10.
9.
Williams ME, Pulliam CC, Hunter R, Johnson TM, Owens JE, Kincaid J, et al. The ShortTerm Effect of Interdisciplinary Medication Review on Function and Cost in Ambulatory
Elderly People. J Am Geriatr Soc 2004;52:93-8.
2001;18:63-77.
on polypharmacy in primary care: cluster controlled trial comparing two approaches.
Br J Gen Pract 2007;57:723-31.
13. Bond C. The MEDMAN study: A randomized controlled trial of community pharmacy-led
14. Laaksonen R, Mills ER, Duggan C, Davies JG, Bates M, C.A. The effect of training and
service provision on the self-assessed competence of community pharmacists. Int J Pharm
Pract 2007;15:141-7.
2010;8:35-42.
2012;37:674-80.
143
18. Laaksonen R, Duggan C, Bates I. Performance of Community Pharmacists in Providing
Clinical Medication Reviews. Ann Pharmacother 2010;44:1181-90.
20. Williams M, Peterson GM, Tenni PC, Bindoff IK, Curtain C, Hughes J, et al. Drugrelated problems detected in Australian Community Pharmacies: The PROMISe Trial.
Ann Pharmacother 2011;45:1067-76.
21. Smulders YM, Burgers JS, Scheltens T, van Hout BA, Wiersma T, Simoons ML, et al. Clinical
practice guideline for cardiovascular risk management in the Netherlands. Neth J Med
2008;66:169-74.
22. Bouma M, Rutten GE, de Grauw WJ, Wiersma T, Goudswaard AN, Nederlands Huisartsen
Genootschap. Summary of the practice guideline Diabetes mellitus type 2 (second revision)
from the Dutch College of General Practitioners. Ned Tijdschr Geneeskd 2006;150:2251-6.
23. Bryant L, Coster G, McCormick R. General practitioner perceptions of clinical medication
reviews undertaken by community pharmacists. J Prim Health Care 2010;2:225-33.
144
A B ST R AC T
B ackg roun d
Explicit criteria have been formulated to identify potentially inappropriate
medicines and potential prescribing omissions respectively called Screening Tool
of Older Persons Prescriptions (STOPP) and Screening Tool to Alert doctors to
Right Treatment (START). However, there are no studies that have evaluated to
what extent these inappropriate medicines and prescribing omissions are already
identified by implicit criteria during medication review.
Aim
To determine the number and types of START-STOPP criteria present in identified
drug-related problems (DRPs) and recommendations of medication reviews.
Me tho ds
In 13 Dutch community pharmacies, the patients community pharmacist
conducted medication reviews in community-dwelling patients ( 65 years and 5
drugs). Community pharmacists identified potential DRPs and recommendations
by implicit criteria. Recommendations were implemented after agreement with
patients general practitioner. After completion, all DRPs and recommendations
identified were compared with START- and STOPP-criteria.
Resu lts
The total number of potential DRPs identified by community pharmacists was 1656
in 457 patients (mean 3.6 per patient). The percentage of START-criteria present in
identified DRPs was higher than the percentage of STOPP-criteria (13% vs. 5.7%;
P < 0.01). 82% of DRPs was not associated with STOPP-START-criteria.
The implementation rate for recommendations associated with STOPP-criteria
was higher compared to recommendations associated with START-criteria (56%
vs. 39%; P < 0.01). The three most common inappropriate medications according
to STOPP-criteria were duplicate drug classes, benzodiazepines and vasodilator
drugs. The three most common prescribing omissions according to START-criteria
were calcium and vitamin D, statins and -blockers.
C on clusi on
This study shows a higher prevalence of START-criteria compared to STOPPcriteria in identified DRPs of community-dwelling older patients while STOPPcriteria were implemented more frequently. The majority of DRPs was not
146
S TOPP- STA R T cr ite r ia
| C hapter 4.4
associated with STOPP-START criteria which underlines that implicit criteria

remain the most important instrument to identify resolvable DRPs.
I N T ROD U C T ION
Polypharmacy and inappropriate medication use by older people increase the risk
of adverse drug reactions.1 Inappropriate medications are defined as medications
for which the potential risk outweighs the potential benefit and for which a good
alternative drug is available.2 Next to inappropriate medications, older patients with
polypharmacy may also be susceptible to under-prescribing. Under-prescribing of
medications refers to the omission of a drug when there is a clear indication and no
contra-indication.3
Several tools are available to evaluate inappropriate prescribing in older patients,
including implicit and explicit criteria. Implicit criteria like the Medication
Appropriateness Index (MAI) may rely on expert professional judgment for their
application. Although the MAI has demonstrated good inter-rater reliability and
validity,4,5 the use of MAI may be rather time-consuming in practice and does not
assess under-prescribing.5 A structural assessment according to a rational order of
indication, effectiveness, safety and compliance based on Cipolle 6-8 was used in
different versions to identify drug-related problems (DRPs) in many RCTs.9-12
The first explicit criteria published for potentially inappropriate medications were
the Beers-criteria.2,13-15 These are the most widely cited criteria for inappropriate
medications, but the applicability outside the U.S. is limited due to differences in
in types of drugs and guidelines.16-18 In 2008, two sets of European-based criteria
(STOPP and START) were formulated to address the perceived deficiencies of
Beers criteria and prescribing omissions as well.19 STOPP (Screening Tool of Older
Persons Prescriptions) contains a list of 65 potentially inappropriate medications
or medication classes. START (Screening Tool to Alert doctors to Right Treatment)
lists 22 potential prescribing omissions (PPOs) in patients with particular medical
conditions.19
Studies in older patients showed that STOPP-criteria were more sensitive than
Beers-criteria in identifying potentially inappropriate medications.20-22 However,
there are no studies that have evaluated to what extent these inappropriate medicines
and also prescribing omissions are already identified by implicit criteria. Therefore,
the aim of our study was to determine the number and types of START-STOPP
criteria present in identified DRPs and recommendations found by medication
review with implicit criteria.
147
M E T HOD S
Stu dy pro ce du res an d p opu l at i on
Thirteen Dutch community pharmacies recruited community-dwelling patients
between January and December 2011. Patients were invited for medication review
if they were aged 65 years and older and used at least five oral prescription drugs.
Complete medication records including drug dispensing records for each patient
were collected from their pharmacies at the end of the study period. In order to
protect the patients privacy, all medical data were anonymized by the community
pharmacists using a randomly assigned unique number.
Me di c at i on re v i e w
The patients community pharmacist interviewed the patient at home or in the
pharmacy about the patients drugs. Patients concerns and experiences regarding
drug therapy (in particular perception of the effectiveness and potential adverse
effects), adherence issues, practical problems and understanding of their medication
regimen were addressed during this interview. A pharmaceutical care plan was
proposed by the community pharmacist using both the patients medication
records from the pharmacy, general practitioner (GP)s medical records and the
data from the patient interview. A structural assessment according to a rational
order of indication, effectiveness, safety and compliance based on Cipolle, Hepler
and Strand 6,7 was used as implicit criteria to identify potential DRPs and associated
recommendations in the care plan. Recommendations were implemented after
agreement between both the community pharmacist, the patients GP and the
patient. Follow-up of implemented recommendations was monitored by the
community pharmacists.
Participating community pharmacists were relatively inexperienced in performing
medication reviews. Therefore, they received a two day training course in
medication review and communication skills. In addition, pharmacists participated
in monthly web conference sessions moderated by a teacher in medication review.
During these sessions, case-studies and a treatment guideline were discussed.
D at a cl assif i c ati on
Classification System (11th edition, 2008) formulated by the World Health
Organization Collaborating Centre for Drug Statistics Methodology. Potential
DRPs and recommendations were classified according to the D.O.C.U.M.E.N.T.
system by the community pharmacists.12,23-25
148
| C hapter 4.4
Appl i c at i on of STA RT- STOPP c r iter i a

The primary outcomes were number, types and implementation rate of STOPPand START-criteria applicable to identified DRPs and associated recommendations
during medication review. STOPP- and START-criteria were retrospectively
applied independently by two investigators (H.K. and S.V.). Differences were
discussed until consensus was reached. The implementation rate was defined as the
percentage of recommendations that was fully or partly implemented according to
the community pharmacist.
Statist i c a l an a l ysi s
All implementation rates of recommendations were analyzed on an intention-totreat basis. Descriptive statistics were used for basic characteristics. Pearson chisquare tests were used for each categorical variable. A P-value < 0.05 was considered
Fi g u re 1 Flow chart of the medication review process
Patient interview
n = 533
Drop-outs: 72
Poor or no registration
Pharmaceutical care plan
(DRPs identified)
n = 461
Drop-outs: 4
Death (3)
Hospital admission (1)
Face-to-face discussion
with GP and follow-up
n = 457

n = 457
DRP = drug-related problem; GP = general practitioner
149
statistically significant. All data were analysed using Microsoft Access and Excel
2010 (Microsoft Corporation, Redmond, WA, USA) and SPSS version 20.0 (SPSS
Inc., Chicago, IL, USA).
Ta b l e 1
Baseline characteristics of participants ( 65 years and 5 drugs)
Characteristic
Female; n (%)
n = 457 (100%)
274 (60%)
77 (7381)
8.7 3.2
M o s t p re sc r i b e d d r ug c la sse s (ATC ) at t =0 ; n ( % )
n = 446 (100%)
308 (69%)
302 (68%)
264 (59%)
258 (57%)
256 (57%)
149 (33%)
137 (31%)
122 (27%)
107 (24%)
99 (22%)
100 (22%)
ATC = Anatomical Therapeutic Chemical Classification
R E SU LT S
D emo g r aph i cs
Twenty-one community pharmacists in 13 pharmacies collaborated with 65 GPs in
this study. The pharmacists conducted patient interviews with 533 of their patients.
A complete medication review was performed for 461 patients and 457 patients
were included for analysis (Figure 1). The median age was 77 years (interquartile
range: 7381) and 60% was women (Table 1). The most commonly prescribed
drug classes were Antithrombotic agents (69%) and Agents acting on the ReninAngiotensin System (68%).
150
| C hapter 4.4
Ta b l e 2
Classification of identified DRPs by STOPP- and START-criteria
STOPP
START
n (%) a
n (%) a
1656
94 (5.7%)
214 (13%)
303
59 (20%)
5 (1.7%)
Duplication
20
17 (85%)
Drug interaction
11
1 (9.0%)
Contra-indication apparent
16
7 (44%)
2 (13%)
256
34 (13%)
3 (1.2%)
n
O ve ra l l
D (r ug
s e le c t i o n )
O (ve r
200
1 (0.5%)
64
1 (1.6%)
52
84
142
1 (0.7%)
2 (1.4%)
72
1 (1.4%)
2 (2.8%)
507
13 (2.6%)
205 (40%)
380
12 (3.2%)
153 (40%)
Condition untreated
127
1 (0.8%)
52 (41%)
Taking too little

Taking too much
U(ndertreated)
Non-laboratory monitoring
E (d uc at i o n
8
62
222
152
70
o r i n fo r m at i o n )
62
62
N (o n - c li n i c a l)
1 (0.5%)
1 (1.4%)
55
1 (1.8%)
Other
55
1 (1.8%)
T (ox i c i t y)
165
19 (12%)
1 (0.6%)
165
19 (12%)
1 (0.6%)
Toxicity, allergic reaction or adverse effect present
STOPP = Screening Tool of Older Persons Prescriptions; START = Screening Tool to Alert doctors to Right
Treatment
a) % is the percentage within DRP type or subtype.
151
STOPP- an d START- criteri a among i d ent i f i e d DR Ps

A total of 1656 potential DRPs were identified for 457 patients (mean 3.6 per
patient) (Table 2). The percentage of START-criteria present in identified DRPs was
higher than the percentage of STOPP-criteria (13% vs. 5.7%; P < 0.01). The majority
of STOPP-criteria were associated with DRP type Drug selection (59/94; 63%),
while the majority of START-criteria were associated with DRP type Undertreated
(205/214; 96%).
82% of DRPs was not associated with STOPP-START-criteria. The most common
types of DRPs not associated with STOPP-START-criteria were: Undertreated
(289/1350; 21%), Drug selection (239/1350; 18%) and Monitoring (222/1350;
16% of DRPs).
Impl ement at i on r ate of re c om m en d at i ons a ss o c i ate d w it h
STOPP-START
The implementation rate for recommendations associated with STOPP-criteria
was higher compared to recommendations associated with START-criteria (56%
vs. 39%; P < 0.01). Both implementation rates of STOPP and START were lower
compared to recommendations not associated with STOPP-or START-criteria
(66%; P < 0.05 and P < 0.01 respectively) (Table 3).
STOPP-criteria were applicable to 79 of 338 recommendations to cease a drug
(23%). The implementation rate for STOPP-criteria was not different compared
Ta b l e 3
Implementation rate of recommendations associated with STOPP-START

criteria
O ve ra l l
STOPP
START
n (%) a
n (%) a
No STOPP-START
n (%) a
P-value b
94 (56%)
1350 (66%)
< 0.05
214 (39%)
1350 (66%)
< 0.01
Cessation of drug
79 (58%)
259 (51%)
0.23
Addition of a drug
197 (38%)
78 (54%)
0.02
15 (47%)
138 (51%)
0.75
17 (53%)
138 (51%)
0.83
Replacement of drug
STOPP = Screening Tool of Older Persons Prescriptions; START = Screening Tool to Alert doctors to Right
Treatment
a) % is the implementation rate.
b) Comparison of STOPP- or START-criteria vs. no STOPP-START-criteria.
152
| C hapter 4.4
Ta b l e 4
Implementation rates of most frequent potentially inappropriate

medications according to STOPP-criteria
Inappropriate medication
Any duplicate drug class prescription
19
Implementation
rate (%)
(47%)
Drugs that adversely affect fallers: benzodiazepines
12
(67%)
Drugs that adversely affect fallers: vasodilator drugs
12
(50%)
Long-term (i.e. > 1 month), long-acting benzodiazepines
(71%)
Aspirin not indicated
(33%)
Oestrogens without progestogen in patients with intact uterus
(100%)
NSAID with heart failure
(80%)
Long-term NSAID or colchicine for chronic treatment of gout no

contraindication to allopurinol
(80%)
-Blockers and frequent hypoglycaemic episodes
(67%)
Long-term opiates in those with recurrent falls
(50%)
STOPP = Screening Tool of Older Persons Prescriptions; NSAID = non-steroidal anti-inflammatory drug
to other recommendations to cease a drug (58% vs. 51%; P = 0.23). STARTcriteria were applicable to 197 of 275 recommendations to add a drug (72%). The
implementation rate for the subgroup START-criteria recommendations was lower
compared to other recommendations to add a drug (38 vs. 54%; P = 0.02) (Table 3).
One-hundred-eighty-six of the remaining 259 other recommendations not
associated with STOPP-START-criteria to cease a drug (72%) were associated with
DRP subtype No indication apparent Seventy-four of the remaining 78 other
recommendations not associated with STOPP-START-criteria to add a drug (95%)
were associated with DRP type Undertreated.
Pre v a l en ce an d ty p es of STOPP- STA RT c r iter i a
STOPP-criteria were present in 80 patients (18%). Sixty-nine patients (15%) had
one potentially inappropriate medicine and 11 (2%) had more than one. STARTcriteria were present in163 patients (36%). One-hundred-twenty-two patients
(27%) had one potential prescribing omission and 41 (9.0%) had more than one.
Nine types of STOPP-criteria accounted for 82% of the total and 25 of the 65
available types of STOPP-criteria (38%) were applied. The most common potentially
inappropriate medicines according to STOPP-criteria were duplicate drug classes
(n = 19; 20%), benzodiazepines (n = 12; 13%) and vasodilator drugs (n = 12; 13%)
(Table 4).
153
Ta b l e 5
Implementation rates of most frequent potential prescribing omissions

according to START criteria
Omitted medication medical condition
Calcium and Vitamin D supplement osteoporosis
58
(57%)
Statin therapy history of coronary, cerebral or peripheral vascular disease
31
(26%)
-Blocker angina, acute MI or heart failure
20
(15%)
Proton pump inhibitor ASA ( 100 mg) and > 80 years, NSAID and > 70
years or reflux
19
(79%)
Bisphosphonates corticosteroids or osteoporosis
16
(13%)
Statin therapy diabetes mellitus
13
(23%)
ACE inhibitor heart failure
(44%)
Metformin type 2 diabetes or metabolic syndrome
(22%)
ACE inhibitor or angiotensin receptor blocker diabetes with

nephropathy
(50%)
Antihypertensive therapy systolic blood pressure > 160 mmHg
(25%)
Implementation
rate (%)
START = Screening Tool to Alert doctors to Right Treatment; MI = myocardial infaction; ASA = acetylsalicylic acid;
NSAID = non-steroidal anti-inflammatory drug; ACE = angiotensin-converting enzyme
Ten types of START-criteria accounted for to 89% of the total and 18 of the
22 available START-criteria (82%) were applied. The most common potential
prescribing omissions according to START-criteria were calcium and vitamin D
in osteoporosis (n = 58; 27%), statins in coronary, cerebral or peripheral vascular
disease (n = 31; 14%) and -blockers in angina, acute myocardial infarction (MI) or
heart failure (n = 20; 9%) (Table 5).
DI S C U S SION
This study shows that START-criteria were two times more often applicable to
identified DRPs compared to STOPP-criteria among community-dwelling older
people. The implementation rate of recommendations concerning STOPP-criteria
was higher compared to recommendations concerning START-criteria. The
majority of identified DRPs was not associated with STOPP-START criteria.
STOPP-criteria in our study were applicable to 18% of the patients which is slightly
lower than the findings of Ryan et al. in a comparable primary care population.26
In our study 25 of 65 STOPP-criteria were used which is comparable to the study
154
| C hapter 4.4
of Ryan using 28 STOPP-criteria. Ryan et al. had access to more complete medical
information and could therefore have identified more STOPP-criteria.26,27 Our
study may underestimate the prevalence of STOPP-criteria while we could not
systematically apply the STOPP-criteria on medication records combined with a list
of diagnoses.
START-criteria were applicable to 36% of patients which is was considerably higher
than Ryan et al.(23%).26 Eighteen of 22 criteria accounted for the prescribing
omissions in our study while this was 15 in the study of Ryan.26 The high
prevalence of START-criteria in our study together with the high proportion of
recommendations to add a drug associated with START-criteria suggests a good
practical applicability of this tool for older patients with polypharmacy in primary
care.
However, studies in secondary care showed higher prevalence rates of STOPPSTART criteria. A study in six European hospitals revealed prevalence of STOPPand START-criteria above 50%.20 A recent study in nursing homes also identified a
prevalence of respectively 60% (STOPP) and 42% (START).28
The implementation rate of recommendations associated with STOPP was
comparable to other recommendations to cease a drug. On the contrary,
recommendations to add a new drug based on START were less frequently
implemented compared to other recommendations to add a drug. Especially,
recommendations to add cardiovascular drugs (e.g. statins, angiotensin-converting
enzyme [ACE] inhibitors and beta-blockers) were poorly implemented. It is likely
that GPs are cautious to change cardiovascular treatment of patients who are
concurrently seeing a specialist. Furthermore, non-acceptance may be caused by the
fact that patients previously experienced adverse effects on these drugs. Finally, GPs
may be reluctant to add preventive drugs in the oldest old, because high cholesterol
levels and hypertension for those patients may not be related to mortality.29,30 On
the contrary, addition of proton pump inhibitors and, to a lesser extent, calcium
and vitamin D had high implementation rates. These drugs are characterized by a
direct effect or by the absence of serious adverse effects.
Only 23% of all recommendations to cease a drug comprised a STOPP-criterion.
The most important reason to cease a drug was that there was no indication
apparent for a drug. Furthermore all DRPs with STOPP- and START-criteria were
already identified by community pharmacists by an implicit tool without using
explicit criteria. These findings underline the importance of using implicit criteria
for medication review and education of the community pharmacist to develop the
required medication review skills to use them.31
155
Although STOPP-START criteria were present in a minority of all DRPs identified,

the criteria do seem applicable as a screening tool for medication review. It has also
been suggested to incorporate STOPP and START in existing prescribing systems
in primary care.8,26 Our study suggests that the list of 65 STOPP-criteria may be too
extensive for the Dutch primary care setting while the 22 START-criteria seem to
fit. Future research should establish if a shorter list of STOPP-criteria may be more
appropriate for use in primary care.
Our study had several strengths. First of all, the elaborate description of DRPs
and recommendations in the pharmaceutical care plans gave us the opportunity
to retrospectively identify STOPP-START-criteria. Second, a high number of
community pharmacists and patients were enrolled in this study. Third, it was a
pragmatic study with no extra documentation or patient follow-up beyond what
would normally be expected.
There were some limitations to the study. First, the lack of medical histories,
especially current lists of diagnoses, prevented us from applying retrospectively
the STOPP and START-criteria on medication records. Although for a selected few
numbers of STOPP-criteria clinical information is not required, access to the full
clinical record is recommended for the majority of STOPP- and START-criteria.27
Second, the implementation rate of recommendations was based on self-report by
the community pharmacists and not on measurement of medication changes in
dispensing records.
C ON C LU SION
This study shows a higher prevalence of START-criteria compared to STOPPcriteria in identified DRPs of community-dwelling older patients while STOPPcriteria are implemented more frequently. The majority of identified DRPs was not
associated with STOPP-START criteria. These findings suggest that implicit criteria
remain the most important instrument to identify resolvable DRPs in primary care.
Future research should further establish the usefulness of STOPP-START criteria in
medication review by incorporation of the tool into the intervention.
Ack n ow l e dgements - This study was conducted in pharmacies of the cooperation
Connecting Care. The authors like to thank all 13 participating community pharmacies.
156
| C hapter 4.4
REFERENCES
1.
Willcox SM, Himmelstein DU, Woolhandler S. Inappropriate drug prescribing for the
community-dwelling elderly. JAMA 1994;272:292-6.
2.
Beers MH. Explicit criteria for determining potentially inappropriate medication use by the
elderly. An update. Arch Intern Med 1997;157:1531-6.
3.
Hanlon JT, Schmader KE, Ruby CM, Weinberger M. Suboptimal prescribing in older
inpatients and outpatients. J Am Geriatr Soc 2001;49:200-9.
4.
Castelino RL, Bajorek BV, Chen TF. Targeting Suboptimal Prescribing in the Elderly: A
Review of the Impact of Pharmacy Services. Ann Pharmacother 2009;43:1096-106.
5.
Spinewine A, Schmader KE, Barber N, Hughes C, Lapane KL, Swine C, et al. Appropriate
prescribing in elderly people: how well can it be measured and optimised? Lancet
2007;370:173-84.
6.
Serv Res 2011;11:4.
7.
Cipolle RJ. Pharmaceutical care practice: the clinicians guide. 2nd ed. New York: McGrawHill, Medical Pub. Division; 2004.
8.
Vermeulen Windsant-van den Tweel AM, Verduijn MM, Derijks HJ, van Marum RJ.
9.

2004;58:648-64.
13. Beers MH, Ouslander JG, Rollingher I, Reuben DB, Brooks J, Beck JC. Explicit criteria for
determining inappropriate medication use in nursing home residents. UCLA Division of
Geriatric Medicine. Arch Intern Med 1991;151:1825-32.
14. Fick DM, Cooper JW, Wade WE, Waller JL, Maclean JR, Beers MH. Updating the Beers
criteria for potentially inappropriate medication use in older adults: results of a US consensus
panel of experts. Arch Intern Med 2003;163:2716-24.
15. American Geriatrics Society 2012 Beers Criteria Update Expert Panel. American Geriatrics
Society updated Beers Criteria for potentially inappropriate medication use in older adults. J
Am Geriatr Soc 2012;60:616-31.
16. Levy HB, Marcus EL, Christen C. Beyond the beers criteria: A comparative overview of
explicit criteria. Ann Pharmacother 2010;44:1968-75.
157
17. OMahony D, Gallagher PF. Inappropriate prescribing in the older population: need for new
criteria. Age Ageing 2008;37:138-41.
18. Laroche ML, Charmes JP, Merle L. Potentially inappropriate medications in the elderly: a
French consensus panel list. Eur J Clin Pharmacol 2007;63:725-31.
20. Gallagher P, Lang PO, Cherubini A, Topinkova E, Cruz-Jentoft A, Montero Errasquin B, et
al. Prevalence of potentially inappropriate prescribing in an acutely ill population of older
patients admitted to six European hospitals. Eur J Clin Pharmacol 2011;67:1175-88.
22. Ryan C, OMahony D, Byrne S. Application of STOPP and START criteria: interrater
reliability among pharmacists. Ann Pharmacother 2009;43:1239-44.
24. Williams M, Peterson GM, Tenni PC, Bindoff IK, Curtain C, Hughes J, et al. Drug-Related
Problems Detected in Australian Community Pharmacies: The PROMISe Trial. Ann
2012;37:674-80.
26. Ryan C, OMahony D, Kennedy J, Weedle P, Byrne S. Potentially inappropriate prescribing in
an Irish elderly population in primary care. Br J Clin Pharmacol 2009;68:936-47.
27. Ryan C, OMahony D, ODonovan DO, OGrady E, Weedle P, Kennedy J, et al. A comparison
of the application of STOPP/START to patients drug lists with and without clinical
information. Int J Clin Pharm 2012 Dec [epub].
28. Ryan C, OMahony D, Kennedy J, Weedle P, Cottrell E, Heffernan M, et al. Potentially
inappropriate prescribing in older residents in Irish nursing homes. Age Ageing
2013;42:116-20.
29. Weverling-Rijnsburger AW, Jonkers IJ, van Exel E, Gussekloo J, Westendorp RG. Highdensity vs low-density lipoprotein cholesterol as the risk factor for coronary artery disease
and stroke in old age. Arch Intern Med 2003;163:1549-54.
30. van Bemmel T, Gussekloo J, Westendorp RG, Blauw GJ. In a population-based prospective
study, no association between high blood pressure and mortality after age 85 years.
J Hypertens 2006;24:287-92.
31. Niquille A, Lattman C, Bugnon O. Medication reviews led by community pharmacists in
Switzerland: a qualitative survey to evaluate barriers and facilitators. Pharm Pract (Internet)
2010;8:35-42.
158
G e n e ra l d iscu s s io n
| C hapte r 5
Health policy encourages older people to remain living in their own homes as long
as possible. However, the ability to remain independent in ones home sometimes
partly depends on the ability to manage a complicated medication regimen1 and the
use of multiple medicines can give rise to various drug-related problems (DRPs).
There is clearly a need for clinical pharmacy services that improve the appropriate
medication use of older people with polypharmacy in primary care.
In this thesis, a variety of studies have been presented each highlighting different
aspects of clinical pharmacy services, i.e. dispensing services and medication
reviews. We studied the effect of a specific dispensing service (i.e. multi dose drug
dispensing systems [MDD]) on the medication management of older patients. In
particular, the impact of these systems on medication adherence, knowledge and
the appropriateness of drugs in MDD were evaluated. Regarding medication review,
a literature review showed the relationship between collaborative aspects of this
intervention and the implementation rate of recommendations. Two randomized
controlled trials (RCT) investigated the added value of different types of medication
review on different outcomes. Finally, specific aspects of the medication review
process including the patient interview, pharmacists ability to identify DRPs and
the presence of explicit criteria in DRPs were highlighted in cross-sectional studies,
partly linked to these RCTs.
M A I N F I N DI N G S OF T H E T H E SI S
In general, the presented studies provided insight into the effects of the different
clinical pharmacy services in daily practice. The most important findings in the
studies on dispensing services were:
Older patients receiving their drugs via MDD reported a higher medication
adherence and lower knowledge compared to patients using manually-dispensed
drugs. The higher medication adherence was independent of medication
knowledge and cognitive function (Chapter 3.1).
Within the group of MDD-users, patients reported a higher knowledge of
their manually-dispensed drugs compared to their multidose dispensed drugs
(Chapter 3.1).
The appropriateness of prescribing within MDD systems can be improved
suggesting that all patients with MDD could undergo a thorough medication
review (Chapter 3.2).
161
The main findings in the studies on medication reviews were:

A systematic review suggested that a more intensive collaboration of general
practitioner and pharmacist results in higher implementation rates of
recommendations in medication review (Chapter 2).
Different types of medication review (prescription review and clinical medication
review) in different patient groups (MDD-users vs. patients using manuallydispensed drugs) identified high numbers of DRPs (Chapter 3.2 and 4.1).
Identification of a higher number of DRPs seemed associated with a lower
proportion of implemented recommendations (Chapter 3.2, 4.1 and 4.4).
Clinical medication review improved LDL-cholesterol, especially for patients
receiving secondary prevention. No significant changes were observed for systolic
blood pressure, HbA1c and health-related quality of life (Chapter 4.1).
More than a quarter of all DRPs were identified during patient interviews despite
an abundance of medical and clinical data and were more frequently assigned a
higher clinical relevance (Chapter 4.2).
The completeness of DRPs identified by community pharmacists can be
improved. Experts reviewers added almost twice the amount of already identified
DRPs by community pharmacists. However, community pharmacists did identify
relatively more DRPs with a higher clinical relevance (Chapter 4.3).
The prevalence of START (Screening Tool to Alert doctors to Right Treatment)criteria among identified potential DRPs is higher compared to STOPP
(Screening Tool of Older Persons Prescriptions)-criteria. On the contrary, the
implementation rate of STOPP-criteria was higher compared to START-criteria
(Chapter 4.4).
In this final chapter these findings will be put into a broader perspective by discussing
the eligible patient groups, interventions and different outcomes (Figure 1). At the
end of these sections the implications of our research on dispensing services and
medication review for practice and future research are described.
E L IG I B L E PAT I E N T S
In all conducted studies, patients were eligible when they were communitydwelling, older than 65 years and used 5 chronic drugs or more (= polypharmacy).
However, it could be questioned if this is the most appropriate patient group.
Therefore we attempted to study the effects of our clinical pharmacy interventions
in sub selections of these patients.
162
| C hapte r 5
Fi g u re 1 Eligible patients, intervention and outcomes
65 years and 5 drugs

and risk factors
Eligible patients
Intervention
Outcomes
Clinical pharmacy services
Process
Intermediate
Clinical
One sub selection was represented by patients receiving their drugs via multidose
drug dispensing (MDD-users, Chapter 3). The majority of MDD-users were
invited to receive their medication by MDD, because they experienced difficulties
in managing their manually-dispensed drugs in the perception of the community
pharmacist, general practitioner (GP) or family. GPs could also refer patients
for more specific reasons (e.g. decreased cognitive function, (suspected) nonadherence or severe psychiatric problems).2 The finding that MDD-users had a
lower mean MMSE-score compared to patients receiving manually-dispensed
drugs (Chapter 3) suggests that MDD-users might indeed have decreased cognitive
function.
Another sub selection were patients in the Home Medication Review study
(Chapter 4.1, 4.2, 4.3) who had to use at least one cardiovascular or anti-diabetic
drug, because of the chosen disease-specific outcomes (LDL-cholesterol, systolic
blood pressure, HbA1c). The majority of these patients received manually-dispensed
drugs. No additional eligibility risk-factors were specified in the study on STOPPSTART criteria (Chapter 4.4).
Age ( 65 years) and polypharmacy together with non-adherence, decreased
cognitive function, renal impairment, 4 co-morbidities and living alone were
identified as risk-factors for medication-related hospital admissions in the Dutch
HARM-study.3 Use of certain type of drug classes (ATC A and ATC B) was most
often associated with these admissions. When these drug classes combined with
163
three risk-factors (age, polypharmacy, non-adherence) are applied to an average

Dutch community pharmacy, 450 patients (6.4% of the pharmacy population)
are eligible for medication review.4 The Dutch multidisciplinary polypharmacy
guideline recommends the following additional risk-factors: renal impairment,
decreased cognitive function, increased risk of falling and signs of non-adherence.3,5
Still, it may be questioned if all these patients are at high-risk for drug-related
problems.
The majority of RCTs in medication review for home-dwelling older patients had
no additional eligibility criteria 6-9 or even applied lower polypharmacy criteria
than 5 drugs or more.7,8,10,11 Some RCTs applied additional criteria aiming to
select patients at increased risk for adverse drugs events and/or health decline. In
the RCT of Sorensen, one out of ten additional eligibility criteria was sufficient to
receive a medication review, e.g. taking 12 or more doses of medication per day;
suffering from three or more medical conditions; suspected by GPs to be nonadherent with medication or at risk in mismanaging medication due to language
difficulties, dexterity problems or impaired sight.12 Lenaghan included patients
over 80 years old with polypharmacy and at least one of the following additional
risk factors: living alone; confused mental state, vision or hearing impairment;
prescribed medicines associated with medication-related morbidity; or prescribed
> 7 regular oral medicines.13 Allard selected patients with a 6-item questionnaire
aiming to detect older people who are likely to experience functional decline over
the next year.14,15
The latter two studies include risk factors that are commonly found in frail
patients. Frailty is a syndrome associated with functional impairment and
increased susceptibility to disease, disability, and mortality.16 Frailty is related to
increasing age and co-morbidities, yet frailty is not synonymous with either age or
disease.17,18 There is no consensus on the exact definition of frailty. A widely used
phenotype according to Fried defines frailty as the presence of 3 of five criteria:
weight loss, exhaustion, weak grip strength, slow walking speed, low physical
activity.17,18 Gnjidic et al. showed that frail men were more likely to be exposed to
high-risk prescribing.16 There was a relationship between exposure to high-risk
prescribing (in this study defined by polypharmacy and exposure to medications
with anticholinergic and sedative effects) and the development of frailty over 2
years of follow-up. The authors suggested that reducing high-risk prescribing is an
important management strategy in frail elderly.16 It is therefore surprising that no
studies on medication review have been conducted in frail home-dwelling older
individuals with polypharmacy.
164
| C hapte r 5
Impli c ati ons for pr ac t i ce

The studies in this thesis did not provide sufficient data to answer the question which
patients are at high risk of DRPs. The number of 5 drugs as a strict cut-off to identify
polypharmacy may be of limited value in predicting a potential risk for occurrence
of DRPs.19 This may also account for the cut off of 65 years for age. Therefore, new
approaches to select patients eligible for medication review should be the subject of
future research. Selection criteria could be derived from the different data sources
for clinical medication review (patient interview, medication records and medical
records) and possibly by the combining these sources (Figure 2).
Some risk-factors mentioned in the Dutch polypharmacy guideline can be derived
from patients medication records (e.g. refill adherence and the use of fall-risk
increasing drugs).20 Moreover pharmacists increasingly have access to laboratory
data such as renal function. Additional risk-factors that could be identified with
medication records are other potential inappropriate medicines (for example
Fi g u re 2 Potential data sources for medication-related risk-factors of older patients
tBHF
FHZFBST
tQPMZQIBSNBDZ FHESVHT
tMJWJOHTJUVBUJPO FHMJWJOHBMPOF
Patients
Patient interview
Data sources
for potential
risk-factors
FHNFEJDBUJPONBOBHFNFOU
"%&TDPODFSOT FFDUJWFOFTT
Medication records
FHBEIFSFODF
JOBQQSPQSJBUFESVHT
OPPGJOUBLFTEPTFTQFSEBZ
Intervention
%JTQFOTJOH
TFSWJDF
FH.%%
.FEJDBUJPO
SFWJFX
Medical records
FHDPNPSCJEJUJFT
SFOBMGVODUJPO
DPHOJUJWFGVODUJPO
No
JOUFSWFOUJPO
165
STOPP-START criteria and drugs with a narrow therapeutic range).21-23 Finally the
complexity of the medication regimen or intake burden could be a criterion for
medication review.24
Medical records of the GP could provide additional information about the riskfactors mentioned in the Dutch polypharmacy guideline: renal impairment,
decreased cognitive function and history of falling. In addition, medical records
of the GP could also identify patients with multiple chronic co-morbidities. Frail
patients who experience complex problems may be identified among these patients,
but not necessarily.18 Additional information may be needed to identify these frail
patients who may be more susceptible to adverse drug events (ADE).16
Medication-related risk-factors may also be identified by a short patient interview
by pharmacy technicians or practice nurses. This interview may contain short
questions to check whether patients have symptoms suggestive of ADEs; medication
management problems; concerns about their medicines and the effectiveness
of medicines. Home health care employees could also detect a number of these
medication-related risk-factors by a recently developed red flag instrument. They
could report their findings to GP, pharmacist or nurse practitioner.25,26
A combined approach of the different sources of data in selecting patients will help
to identify patients at risk. Future research is needed to determine which combined
approach is satisfactory for identifying patients at highest risk of clinical relevant
DRPs.
Eligibility criteria for MDD and medication review may overlap and certain
patients are eligible for MDD as well as medication review. Furthermore medication
review will sometimes identify medication management problems that make
patients eligible for MDD. The steep increase in the number of MDD-users in
The Netherlands 27 suggests that initiating MDD is more related to polypharmacy
in general than to specific medication management problems. Other dispensing
services (e.g. chronic medication services)28-30 may be more appropriate for patients
with polypharmacy without medication management problems. This indicates a
need for a clear assessment of medication management problems. Potential criteria
for this assessment could be decreased cognitive function, low adherence and the
complexity of the medication regimen. Future research is needed to assess which
criteria predict medication management problems that can be resolved by the start
of MDD.
166
| C hapte r 5
I N T E RV E N T ION S
This thesis focuses on two clinical pharmacy interventions for older people with
polypharmacy in primary care: multi-dose drug dispensing (Chapter 3) and
medication review (Chapters 2, 3.2 and 4).
Mu lti d os e d r ug d isp ensing
We did not investigate MDD as an intervention per se, but studied different
aspects of the MDD system and its users. As described in Chapter 3.2, the
appropriateness of prescribing for patients receiving MDD (MDD-users)
could be improved (8.5 potential DRPs per patient). This finding suggests
that MDD could lead to perpetual repeat prescriptions without necessary reevaluation. The study presented in Chapter 3.1 shows that the self-reported
adherence is higher for MDD-users compared to patients receiving manuallydispensed drugs, but conversely MDD-users had a lower medication knowledge.
The differences in adherence were independent of knowledge and cognitive
function. These findings suggest that MDD-dispensing can increase adherence
in older patients with decreased cognitive function. As this was an observational
study, a formal evaluation study of the benefits of MDD is recommended.
The studies described in Chapter 3 suggest that MDD cannot be considered a
panacea for all older patients. The findings described in Chapter 3.2 suggest that all
MDD-users should regularly receive a medication review. The optimal frequency
for medication review will probably differ between individual patients dependent
on their morbidity and/or the frequency of medication changes.
Me di c ati on re v i e w
Several aspects of medication review were studied (Chapter 3.2 and 4). The
general discussion will discuss the different types of medication review used,21 the
instruments used for identifying DRPS (implicit and explicit criteria), the barriers
and facilitators for medication review, the involved health professionals and the role
of the patient.
Ty pes of medication re v ie w
The type of medication review in the study among MDD-users (Chapter 3.2) could
be characterised as a prescription review while no patient interview was conducted
and no medical data were available (see Introduction). This type of review, with
only medication records available, surprisingly yielded already 8.5 potential DRPs
per patient. The Home Medication Review (HMR) described in Chapter 4.1 was a
clinical medication review conducted in patients using mostly manually-dispensed
167
drugs and yielded 10.1 potential DRPs per patient. The addition of patient interview
at home and availability of medical data revealed more DRPs, but the additional
number compared to the MDD study was less than expected. However, MDDusers were generally older and used more drugs compared to the patients in the
HMR study. Moreover it is likely that patients using MDD at the time of the study
were selected for specific reasons (e.g. cognitive or severe psychiatric problems).
They may represent a group with a higher disease burden and a relatively higher
proportion of frail patients and may therefore be at higher risk for DRPs.
The patient interview in the last study (Chapter 4.4) could either take place in the
pharmacy or at home. This medication review yielded only 3.6 potential DRPs
per patient. This number was similar to the number independently identified by
community pharmacists in the HMR-trial (Chapter 4.3). The difference was that
all DRPs were identified by the community pharmacists themselves while in the
former two studies all DRPs were either identified by expert reviewers (MDD-users,
Chapter 3.2) or were identified by community pharmacists and completed by expert
reviewers (HMR-trial, Chapter 4.3). Community pharmacists may have ignored
some potential DRPs expecting that associated recommendations were unlikely to
be implemented based on their earlier experiences with GPs and patients.
Implicit and e x plic it c r ite r ia
In the presented studies DRPs were identified by both implicit criteria (judgmentbased) and explicit criteria (criterion-based).31 A structural assessment according
to a rational order of indication, effectiveness, safety and compliance based on
Cipolle, Hepler and Strand 4,5,32 was used as implicit criteria in all our medication
review studies (Chapter 3.2 and 4). Next to these implicit criteria, a set of explicit
criteria was available as back-ground information in the HMR-trial (Chapter 4)
which consisted of a list of clinical rules based on Dutch treatment and prescription
guidelines. This list had many similarities with the Australian Prescribing Indicator
Tool which consist mainly of optimum as well as inappropriate medication choices
for a large number of common medical conditions in the elderly.33,34 No explicit
criteria were used for identification of DRPs in the most recent study (Chapter 4.4),
but DRPs and recommendations were retrospectively analysed on presence of the
explicit STOPP- and START-criteria.35,36
The most frequently identified DRPs in our studies were No indication apparent
and Condition undertreated (Chapter 3.2 and 4). These DRPs are mainly identified
by the implicit approach that matches each of the patients conditions with the
current medications. Retrospective analysis of DRPs and recommendations
revealed that START-criteria were present in the majority of DRPs classified as
168
| C hapte r 5
Condition undertreated (Chapter 4.4). The prevalence of STOPP-criteria was

relatively low and not related to a specific DRP subtype.
START-criteria could serve as a tool to identify specific undertreatment next
to the described implicit approach. The extensive list of STOPP-criteria for
identification potential inappropriate medication appears to be of lower practical
value for patients in primary care. This list is of higher significance for patients
in nursing homes and hospitals where patients have a higher level of chronic comorbidities.37,38 Other tools for inappropriate prescribing such as the widely cited
U.S.-based Beers-criteria 39-42 or the derived German Priscus-list 43 are less sensitive
than STOPP-criteria in European studies.43,44 Next to the STOPP-START-list
for inappropriate prescribing, there is no validated explicit tool available in The
Netherlands. This could be developed on the basis of tools in other countries e.g.
the Australian Prescribing Indicator Tool,33,34 U.K. indicators for preventable drugrelated morbidity 45 and the recently developed medication assessment criteria in
the U.K.46 Explicit criteria could also serve as eligibility criteria, but should always
be judged on their clinical relevance before adding these as potential DRPs. The
studies in this thesis suggest that implicit criteria remain the most important
instrument to identify potential DRPs.
Health care professional s
The majority of studies in medication review involved a pharmacist. At the start
of our HMR-trial in 2007 (Chapter 4.1), community pharmacists were relatively
inexperienced in performing medication reviews. Therefore, they received a two
day training course in medication review and communication skills. They identified
a mean of 3.6 potential DRPs per patient while expert reviewers added 6.5 potential
DRPs per patient. DRPs identified by community pharmacist, however, were more
often clinically relevant (Chapter 4.3). In the most recent study without expert
reviewers, community pharmacists also identified 3.6 potential DRPs per patient
(Chapter 4.4). These pharmacists participated in monthly web-conference sessions
moderated by a teacher in medication review next to a two day introductory
training course. On the one hand, these findings indicate that the completeness
of medication reviews by community pharmacists can be improved and that two
days training on medication review may be too limited. On the other hand, it
could be questioned if identifying DRPS by pharmacist reviewers at distance is
a suitable method for daily clinical practice of medication review. Half of the
community pharmacists in Chapter 3.2 stated that the expert reviewers had a lack of
information. Furthermore, a part of potential DRPs identified by expert reviewers
were mainly based on general treatment and prescription guidelines which are not
169
always appropriate for older patients. Therefore not all DRPs identified by expert
reviewers might be perceived as clinically relevant by community pharmacists and
GPs.
These findings indicate that the education for community pharmacists should be
more extensive with a focus on the pharmacotherapy and medication review skills.47
A practical solution could be for community pharmacists, who start performing
medication review, to participate in a course in which feedback on reviews from
expert reviewers (and portfolio building) plays an essential role. This concept has
now evolved in an extensive post-graduate 9 day medication review course for
community pharmacists in The Netherlands.48
The systematic review suggested that collaboration of GPs and pharmacist is
important for a successful medication review (Chapter 2). The involvement of GPs in
our studies (Chapter 3.2, 4.1 and 4.4) was different. The reviews in Chapter 3.2 were
conducted by expert reviewers at distance and the GP was not involved in an early
stage of the medication review process. Interviews with the pharmacists revealed
difficulties in making appointments for case conferences as well as discussing DRPs
with the GPs. GPs also seemed to be reluctant to change a medication regimen in
patients with a complex medical profile, as was the case with the MDD-users in
Chapter 3.2. In the RCT presented in Chapter 4.1, GPs were asked in advance to
share patients medical records and to have case-conferences with the pharmacists.
Although the GPs in this latter study were more involved in the medication review
process, the implementation rate of recommendations was still relatively low. This
finding seemed not in agreement with GPs perceptions to medication review
in this study (Chapter 4.3). Almost all GPs considered pharmacists DRPs and
recommendations as clinically relevant in general. This suggests that the amount
of recommendations may be too high to be implemented simultaneously and also
may not be monitored sufficiently.
Finally, Chapter 4.4 describes a clinical medication review process in daily practice
where both GPs and community pharmacists were reimbursed. The implementation
rate in this study was relatively high (62%). The higher implementation rate is
probably partly related to the fact that pharmacists did less recommendations
and might already have anticipated that some recommendations will not be
implemented beforehand. Furthermore, GPs are becoming more conscious of
the need for medication review and are motivated by the national organisation of
GPs. However, this implementation rate must be interpreted with some caution
as medication changes were not measured in dispensing records as in the other
studies.
170
| C hapte r 5
The results of the last study confirm the suggestion of the systematic review
(Chapter 2) that a higher extent of collaboration between GPs and pharmacist leads
to higher implementation rates. The described key elements reflecting collaborative
aspects in Chapter 2 were not present to the same degree in the different studies
performed. For example, case-conferences between GP and pharmacists were
always conducted in the clinical medication review studies (Chapter 4) and
irregularly in the prescription review study among MDD-users (Chapter 3.2). In
all studies, patients were invited by community pharmacists and not by GPs. We
suggest that for future studies patients should receive a joint invitation from GP
and pharmacist.6 Another option might be a referral for medication review with the
pharmacist by the GP. Collaboration of pharmacist and GP could also contribute
to a selection of patients who could benefit most from a collaborative medication
review. For example, a pharmacist could provide an initial patient selection list and
together with GP could make an assessment if a medication review is needed (see
Eligible patients).
Involve me nt of patie nt s
In our systematic review (Chapter 2), the importance of patient involvement in
medication review was reflected by the key elements patient interview, own
pharmacist involved, and follow-up. The patient can express his or her beliefs,
experiences, attitudes and concerns of the pharmacotherapy in a patient interview.49
This face-to-face consultation should preferably be conducted by the patients
own pharmacist who generally has a longer lasting therapeutic relationship with
the patient. If this interview is conducted by a pharmacist other than the patients
regular pharmacist, this pharmacist will often not be involved in implementation
and follow-up.22,23 Adequate follow-up is recommended to evaluate patients
experiences with implemented actions.
Although in all our practice studies patients received information on the medication
review and were able to decline participation, further patient involvement was
variable. MDD-users in the first study (Chapter 3.2) were not actively involved in
the medication review process due to the absence of a patient interview. In contrast,
patients in the HMR-trial (Chapter 4.1) were visited at home for a comprehensive
patient interview. As part of the study protocol, follow-up visits were conducted
by pharmacists at 6 and 12 months both for measurements and to follow-up on
implementation of actions. Patients in the most recent practice study (Chapter 4.4)
were invited for a patient interview, either at the pharmacy or in the patients home.
No structural follow-up meetings with the patient were scheduled in this study,
although follow-up did take place in some cases.
171
Patients were satisfied with the medication review service in our HMR-trial
(Chapter 4.1), but a considerable proportion did not fully understood the meaning
of a HMR. This reflects the importance of improving patients involvement by
providing sufficient knowledge and awareness of the medication review process.
Much attention has been paid to increasing pharmacists consultations skills for
conducting these interviews in all medication review courses of our studies. The
same cannot be said for patients. For patients in The Netherlands there is limited
information how to communicate with health care providers.50 Interestingly,
an education programme in the U.K was developed for patients to improve
communication skills with health care professionals and to develop effective
partnerships with health care providers.49
B ar r iers and fac ilitators
Medication review services in daily practice need excellent workflow organization,
for example delegation of tasks within pharmacy teams and planning of medication
reviews.47 In our HMR-trial, this organization of medication review was often a
barrier (Chapter 4.1). Pharmacists were often distracted by the day-to-day processes
in the pharmacy which disturbed the planning of medication reviews. Moreover,
during the study period of the HMR-trial insurance companies enacted restrictions
on the remuneration of medicines. This did not only impose an administrative
burden on the participating pharmacists, but also led to budgetary restrictions
making delegation of tasks within pharmacy teams more difficult.
Another facilitator for medication review services by community pharmacies is
the co-location of a GP practice and an established relationship of the pharmacist
with the GPs. Almost all community pharmacies in our HMR-trial were co-located
to the participating GP practice permitting prompt access to patient information.
However, the existing working relationship between pharmacists and GPs before
the start of the study was quite different. The high rate of withdrawal and changing
of the pharmacists during the HMR-trial did not help to reinforce the relationship
between pharmacists and GPs.
Thirdly, supporting software for shared data gathering and efficient registration of
pharmaceutical care plans preferably integrated with GP and pharmacy computer
systems is a facilitator for medication review services that still has to be developed.
In the HMR-trial, the registration of all research data on forms was experienced as
time-consuming by the participating pharmacists.
Finally, remuneration of community pharmacists and GPs for conducting
medications reviews could be a facilitator. Both pharmacists and GPs received a
remuneration in the most recent practice study (Chapter 4.4). However, most
172
| C hapte r 5
community pharmacists in The Netherlands perceived the remuneration of

medication reviews as too low regarding the time consuming nature of the
medication reviews.51 This is partly related to the earlier mentioned inefficient and
time-consuming registration of medication reviews.
Studies on barriers and facilitators in medication review reveal collaborative
aspects that might be essential for conducting successful medication reviews.22,23,47
In our systematic review of the literature (Chapter 2) we characterized the
interventions in recent RCTs on medication review by the presence or absence
of eight key elements. These key elements were reflecting collaborative aspects
between a GP and a pharmacist based on the described facilitators and barriers
in medication review.22,23,47 Briefly, these key elements were (1) pharmacist with
clinical experience; (2) own pharmacist involved; (3) sharing of medical records;
(4) patient interview by pharmacist; (5) invitation of the patients by GP; (6) case
conference GP and pharmacist; (7) action plan; and (8) follow-up. This systematic
review showed an association between the number of key elements and the
implementation rate of recommendations. Future studies could consider these key
elements to design a medication review process.
Impl i c at i ons for cl i n i c a l ph arm ac y s er v i c e s in pr ac ti c e
The number of MDD-users is growing fast in The Netherlands 27 while the
widespread implementation of clinical medication reviews by community
pharmacists is hampered, according to community pharmacists mainly by
reimbursement problems and time constraints.51 However, not all pharmacists
seem to be prepared to change their profession to pharmaceutical care providers
requiring collaboration with GPs. Combined with our findings, these developments
indicate that next to the earlier mentioned eligibility criteria, both interventions
should be clearly described in guidelines and implemented. Moreover, alternative
approaches to the comprehensive clinical medication review should be considered.
As outlined earlier, eligibility criteria for MDD should be more restricted and other
dispensing services may be considered as an alternative. When a patient is being
invited for MDD, a medication management assessment could be conducted by
another pharmacy employee (e.g. a specialized technician) with the aim to verify
patients actual medication use (see Introduction). This assessment could also
determine which medication is appropriate for the multidose system and which
medication still has to be dispensed manually. This verification of patients actual
drug use has many similarities with the medication reconciliation which is defined
as the process of obtaining and maintaining a complete and accurate list of the
current medication in use by the patient.52 After initiating MDD, the use of MDD173
drugs should be regularly evaluated with the patient, especially when medication
changes occur.
There is a running debate on which patients are most likely to benefit from
medication review. Our medication review studies showed a great variation in
the numbers of DRPs per patient indicating that a clinical medication review as
described in the multidisciplinary guideline 5 (see Introduction) is not necessary
for each older patient. Dependent on the number and type of eligibility criteria,
a limited patient interview could be conducted in the community pharmacy or
a comprehensive patient interview at home. Discriminating medication review
interventions in a short and extensive type could help the implementation in daily
clinical practice. A patient interview should always be part of the medication review
process. Supportive software for efficient registration of DRPs and data gathering
preferably combined with integrated explicit criteria could further accelerate
the medication review process. Future research should investigate whether
discriminating patients who are eligible for different types of medication review is
feasible.
OU TC OM E S
Pro c e ss outcomes
The majority of RCTs in our systematic review (Chapter 2) reported improvement
of process outcomes such as an increase in the number of drug changes 6,11,53-57 and
reduction in the number of (prescribed) drugs 11,56,57 while in 4 studies no effect
on number of drugs was reported.7,9,13,14 In our practice study on MDD-users
(Chapter 3.2), the mean number of drug changes per patient was significantly
higher among patients in the intervention group compared with patients in the
waiting-list group.
Inter me d i ate outcom es
Studies of medication reviews in community pharmacies have shown positive
effects on intermediate outcomes such as resolution of drug-related problems
(DRPs), adherence 8,55,58,59 and disease-specific outcomes such as blood pressure,
lipid levels and HbA1c.60 The effects of medication reviews on costs are mixed,
while 4 RCTs 7-9,53,54 in our systematic review reported no effect and two reported a
positive effect.12,56
DRPs were frequently measured in all studies (Chapter 3.2 and 4) with an average
per patient between 3.6 and 10. Not all DRPs led to a recommendation for a drug
174
| C hapte r 5
change. In the studies on MDD-users (Chapter 3.2) and HMR (Chapter 4.1) DRPs
with such a recommendation were separately determined yielding 4.5 and 5.8 DRPs
per patient respectively. The resolution rates of these DRPs were determined by
implemented drug changes and were almost similar in the latter two studies (30%
and 28%) yielding 1.3 and 1.7 resolved DRPs per patient. Despite the lower number
of DRPs in the most recent medication review study (Chapter 4) the resolution rate
in this study was much higher (62%) yielding 2.2 resolved DRPs per patient.
Potential DRPs and recommendations were classified according to the DOCUMENT
system 61-63 by the investigators in the studies on MDD-users (Chapter 3.2) and
HMR (Chapter 4.1) and by the community pharmacists in the most recent study
(Chapter 4.4). The most identified subtypes of DRPs were No indication apparent,
Condition undertreated and Laboratory monitoring. The latter is not associated
with recommendations for a drug change, but is still important to check whether
each older patient receives the appropriate care according to disease-specific
guidelines. Moreover, the outcomes of these measurements (e.g. impaired renal
function) could give rise to subsequent changes in the medication regimens.
Furthermore, these measurements add to the completeness of the patient records
and thereby help to anticipate on future medication changes.
The reduction of DRPs is not the ultimate goal. Therefore, the effect of HMR in
our RCT (Chapter 3.1) was also examined on three disease-specific intermediate
outcomes (blood pressure, LDL-cholesterol and HbA1c). This did not result in an
effect on all treatment goals. The number of patients that achieved target levels for
LDL-cholesterol increased and mean levels were lowered, especially for patients
receiving secondary prevention, while no significant changes were observed for
systolic blood pressure and HbA1c. The scope of the medication review in our study
was much broader than the three disease-specific outcomes measured while the
interventions in disease-specific reviews for hypertension and diabetes were more
intensive compared to our study.64,65
We did not examine the effect of medication review on costs in our studies. The first
studies on medication review were aimed at reducing the number of drugs 66 which
could theoretically lower the costs of drugs. However, nowadays all medication
review studies also focus on undertreatment including recommendations to add
a drug. The frequency of recommendations for cessation of a drug were more
frequently adopted than recommendations to add a new drug in all our studies
(Chapter 3.2, 4.2 and 4.4). Theoretically by discontinuing drugs that are not
indicated or cause ADRs and simultaneously initiating drugs that are indicated, the
overall cost effectiveness of treatment is likely to increase.
175
C lini c a l outc om e s
There is an increasing call for evidence of the beneficial effects of medication review
on clinical outcomes such as hospital admissions, mortality and health-related
quality of life.67 Only few studies are reporting these outcomes and the majority did
not have sufficient statistical power to show an effect on hospital admissions.9,11-13
Systematic reviews did not find evidence for medication reviews reducing hospital
admissions and mortality in RCTs.68,69 However, overall hospital admissions may
not be an appropriate outcome measure, as analysis of the causes of admission
showed that only one in five are related to drugs and only one in 10 were judged
possibly preventable.67 Originally, we did aim to measure medication-related
hospital admissions in our HMR-trial (Chapter 4.1) to exclude a negative effect.
Given the fact, however, that it was difficult to collect sufficiently detailed data to
detect a causal relation between hospital admission and medication, this outcome
was dropped. Moreover, the Dutch HARM-study found that 2.5% of hospital
admissions was medication-related and avoidable.3 If we assume a prevalence of
10% hospital admissions at baseline in our study population which means that only
0.25% is avoidable, we would need more than 10,000 patients to achieve sufficient
power.
We measured the effect of HMR on health-related quality of life by the EQ-5D/VAS
in our RCT (Chapter 4.1). As with hospitalizations we did not expect a positive
effect on quality of life, but we wanted to exclude a negative effect. EQ-5D utility
scores improved slightly in the intervention group and decreased slightly in the
control group, but, like the VAS, these effects were not statistically significant and
not clinically relevant. This means at least that the intervention group did not
experience decreased quality of life. Until now there are no RCTs that report a
significant positive effect on quality of life of community-dwelling older patients.68
Medication review with a more focused approach on patients needs, concerns and
complaints could possibly have more effect on health-related quality of life. For
example, pain management may improve the pain/complaints dimension of the
EQ-5D health-related quality of life.
Until now, only generic and disease-specific instruments are available for the
measurement of health-related quality of life. There were some attempts to
develop an instrument for the measurement of pharmaceutical care, but this
instrument is not available yet.69 Generic instruments (e.g. EQ-5D) are not
sensitive enough to detect effects due to pharmaceutical care.69 Medication
review directed at patients with specific complaints such as pain, anxiety
and impaired activities of daily living, may result in changes of these generic
instruments. Although disease-specific instruments would probably be
176
| C hapte r 5
even more sensitive to measure positive effects in these types of complaints.

Other possible clinical outcomes are adverse drug events and falls. There is
inconclusive evidence that medication reviews in primary care could reduce
adverse effects.8,12,70 Medication reviews as part of multi-facetted interventions
could reduce falls in institutionalised patients,71 but there is no evidence in primary
care.72
Whi ch outcomes are appropri ate for m e d i c at i on re v i e w ?
The majority of process and intermediate outcomes remain useful for future studies
on medication review. Future studies still can consider DRPs, medication changes
and implementation rate as important outcomes of medication review. Diseasespecific outcomes are more appropriate for disease-specific medication reviews
such as interventions in patients with diabetes,60 hyperlipidemia,73 hypertension,74
pulmonary disease 75 or heart failure.76
The effects of medication review on clinical outcomes (mortality, hospital
admissions and quality of life) are variable. This is probably caused by the fact
that many studies lacked adequate statistical power to find effects on these type
of outcomes. Moreover mortality, hospital admissions and quality of life are all
multifactorial.77 Medication-related hospital admissions could be an appropriate
clinical outcome, but as these are relatively rare, very high numbers of patients are
needed and sufficiently detailed data on hospital admissions should be gathered.
This would make such studies when initiated in primary care extremely expensive.
Another approach could be to initiate these studies in high risk populations (e.g.
patients with recent drug-related hospitalizations). As the baseline risk on events
will be much higher in such a population, less patients will have to be included,
making the studies more feasible. Medication review with a more focused approach
on patients needs, concerns and complaints could possibly have more effect on
health-related quality of life. For example, pain management may improve the
pain/complaints dimension of the EQ-5D health-related quality of life. Other
more specific instruments to measure health-related quality of life outcomes
of medication review may be developed. Adverse drug events is another clinical
outcome that could be more responsive to medication review, but is rarely used
until now. Alternatively, it may be interesting to think of a composite endpoint (e.g.
any medication-related event, such as hospital admission, visit to the emergency
department or visit to GP). One thing is definitely clear: medication review is still
looking for the most appropriate outcome.
177
C ON C LU SION
This thesis presented a series of studies on medication review which showed that
medication review in Dutch community-dwelling older patients with polypharmacy
identifies a large number of drug-related problems. Although not all drug-related
problems can be resolved, an average of 12 problems per patient are resolved and
could potentially lead to improved health outcomes. Although we were able to
show some positive effects on intermediate disease-specific outcomes, it remains
difficult to translate the resolution of drug-related problems to health benefits.
The studies presented on multidose drug dispensing systems showed high
adherence, low medication knowledge and considerable inappropriateness of drugs
in these systems. These findings in combination with the steep growth of MDDusers emphasizes the need for sufficient pharmaceutical care around multidose
drug dispensing systems, including medication review. A comprehensive clinical
medication review may not be necessary for each older patient with polypharmacy.
Future studies should elaborate which patients benefit the most from a clinical
medication review using the appropriate clinical outcomes.
Despite the fact that there is still need for additional research into the benefits of
medication review for older patients using multiple medications, it is clear that there
is a potential public health problem. More than one million patients over 65 years
are using more than 5 chronic medications. Due to the ageing of the population this
number will further increase in the next decades. Moreover among those over 65
years the proportion of very old patients will increase even further. Extrapolating
on the findings of this thesis this would mean millions of potential drug-related
problems still have to be evaluated.
Nevertheless the widespread implementation of clinical medication reviews
in primary care is hampered. A joint effort of health care providers, insurance
companies and policy makers is needed to address this inappropriate use of
medicines. Although many community pharmacists have followed post-graduate
medication review courses in the past five years, there is room for improvement in
their clinical experience. Involvement of general practitioners is still very variable.
Health policy makers and insurance companies should encourage a joint approach
of medication review by pharmacists and physicians.
178
| C hapte r 5
REFERENCES
1.
Marek KD, Antle L. Medication Management of the Community-Dwelling Older Adult. In:
Hughes RG, editor. Patient Safety and Quality: An Evidence-Based Handbook for Nurses.
Rockville (MD): Agency for Healthcare Research and Quality (US); 2008.
2.
van Wijck F. Populariteit Baxterrol vraagt om goede leidraad. Pharm Weekbl 2011;146(21).
3.
2008;168:1890-6.
4.
Serv Res 2011;11:4.
5.
Vermeulen Windsant-van den Tweel AM, Verduijn MM, Derijks HJ, van Marum RJ.
6.
7.
Grymonpre RE, Williamson DA, Montgomery PR. Impact of a pharmaceutical care model
2001;9:235-41.
8.
Ageing 2001;30:205-11.
9.
Sellors J, Kaczorowski J, Sellors C, Dolovich L, Woodward C, Willan A, et al. A randomized

2004;58:648-64.
2007;36:292-7.
14. Allard J, Hebert R, Rioux M, Asselin J, Voyer L. Efficacy of a clinical medication review on
the number of potentially inappropriate prescriptions prescribed for community-dwelling
elderly people. CMAJ 2001;164:1291-6.
15. Hebert R, Robichaud L, Roy PM, Bravo G, Voyer L. Efficacy of a nurse-led multidimensional
preventive programme for older people at risk of functional decline. A randomized
controlled trial. Age Ageing 2001;30:147-53.
179
16. Gnjidic D, Hilmer SN, Blyth FM, Naganathan V, Cumming RG, Handelsman DJ, et al.
High-risk prescribing and incidence of frailty among older community-dwelling men. Clin
Pharmacol Ther 2012;91:521-8.
17. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older
adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56:M146-56.
18. Hubbard RE, OMahony MS, Woodhouse KW. Medication prescribing in frail older people.
Eur J Clin Pharmacol 2012 Sep [epub].
19. Viktil KK, Blix HS, Moger TA, Reikvam A. Polypharmacy as commonly defined is an
indicator of limited value in the assessment of drug-related problems. Br J Clin Pharmacol
2007;63:187-95.
20. Bauer TK, Lindenbaum K, Stroka MA, Engel S, Linder R, Verheyen F. Fall risk
increasing drugs and injuries of the frail elderly - evidence from administrative data.
Pharmacoepidemiol Drug Saf 2012;21:1321-7.
24. Mansur N, Weiss A, Beloosesky Y. Looking beyond polypharmacy: quantification of
medication regimen complexity in the elderly. Am J Geriatr Pharmacother 2012;10:223-9.
25. Leeuw de, M. Thuiszorg als ogen en oren van apothekers. Pharm Weekbl 2012;146:(41).
26. Sino GMS, Dooren AA, Haverkamp AM, Schuurmans MJ. Recognition of drug related
problems by home healthcare employees: A Dutch observational study with self reports.
J Nurs Educ Prac 2013;8:41-9.
27. College voor zorgverzekeringen (CVZ). GIPeilingen 2011. Ontwikkelingen genees- en
hulpmiddelengebruik. Available from: http://www.cvz.nl/binaries/live/cvzinternet/hst_
content/nl/documenten/gipeilingen/gip2011.pdf (Accessed 18 Jan 2013).
28. The Scottish Government. Establishing Effective Therapeutic Partnerships - A
generic framework to underpin the Chronic Medication Service element of the
Community Pharmacy Contract. 2009. Available from: http://www.scotland.gov.uk/
Publications/2010/01/07144120/0 (Accessed 3 Jan 2013).
29. van Geffen K, Vos I, van Horssen N, Winters N, Bouvy ML. Apotheker proactiever. Enquete:
rol in herhaalmedicatie groeit. Pharm Weekbl 2011;146:(20).
30. de Hertog NJ, Donker AA, Smith R. Patient therapietrouwer met herhaalservice. Significante
verbetering bij zeven geneesmiddelgroepen. Pharm Weekbl 2011;146:(20).
31. Spinewine A, Schmader KE, Barber N, Hughes C, Lapane KL, Swine C, et al. Appropriate
prescribing in elderly people: how well can it be measured and optimised? Lancet
2007;370:173-84.
32. Cipolle RJ. Pharmaceutical care practice : the clinicians guide. 2nd ed. New York: McGrawHill, Medical Pub. Division; 2004.
180
| C hapte r 5
33. Basger BJ, Chen TF, Moles RJ. Inappropriate medication use and prescribing indicators
in elderly Australians: development of a prescribing indicators tool. Drugs Aging
2008;25:777-93.
34. Basger BJ, Chen TF, Moles RJ. Validation of prescribing appropriateness criteria for older
Australians using the RAND/UCLA appropriateness method. BMJ Open 2012;2:5.
35. Ryan C, OMahony D, Kennedy J, Weedle P, Byrne S. Potentially inappropriate prescribing in
an Irish elderly population in primary care. Br J Clin Pharmacol 2009;68:936-47.
37. Ryan C, OMahony D, Kennedy J, Weedle P, Cottrell E, Heffernan M, et al. Potentially
inappropriate prescribing in older residents in Irish nursing homes. Age Ageing
2013;42:116-20.
38. Gallagher P, Lang PO, Cherubini A, Topinkova E, Cruz-Jentoft A, Montero Errasquin B, et
al. Prevalence of potentially inappropriate prescribing in an acutely ill population of older
patients admitted to six European hospitals. Eur J Clin Pharmacol 2011;67:1175-88.
39. Beers MH, Ouslander JG, Rollingher I, Reuben DB, Brooks J, Beck JC. Explicit criteria for
determining inappropriate medication use in nursing home residents. UCLA Division of
Geriatric Medicine. Arch Intern Med 1991;151:1825-32.
40. Beers MH. Explicit criteria for determining potentially inappropriate medication use by the
elderly. An update. Arch Intern Med 1997;157:1531-6.
41. Fick DM, Cooper JW, Wade WE, Waller JL, Maclean JR, Beers MH. Updating the Beers
criteria for potentially inappropriate medication use in older adults: results of a US consensus
panel of experts. Arch Intern Med 2003;163:2716-24.
42. American Geriatrics Society 2012 Beers Criteria Update Expert Panel. American Geriatrics
Society updated Beers Criteria for potentially inappropriate medication use in older adults.
J Am Geriatr Soc 2012;60:616-31.
43. Siebert S, Elkeles B, Hempel G, Kruse J, Smollich M. The PRISCUS list in clinical routine :
Practicability and comparison to international PIM lists. Z Gerontol Geriatr 2013;46:35-47.
45. Morris CJ, Rodgers S, Hammersley VS, Avery AJ, Cantrill JA. Indicators for preventable drug
related morbidity: application in primary care. Qual Saf Health Care 2004;13:181-5.
46. Dreischulte T, Grant AM, McCowan C, McAnaw JJ, Guthrie B. Quality and safety of
medication use in primary care: consensus validation of a new set of explicit medication
assessment criteria and prioritisation of topics for improvement. BMC Clin Pharmacol
2012;12:5.
2010;8:35-42.
48. de Leeuw M. Medicatiebeoordeling in opmars. Pharm Weekbl 2009;144:16-17.
49. Bond C, Blenkinsopp A, Raynor DK. Prescribing and partnership with patients. Br J Clin
Pharmacol 2012;74:581-8.
181
50. Nederlandse Consumenten- en PatientenFederatie (NPCF). Patientveiligheidskaarten.

Available from: http://www.mijnzorgveilig.nl/kaart/patienten/index.php (Accessed 5 Feb
2013).
51. Pijpker MJ, de Gier JJ. Prestatiebekosting verdient nieuwe impuls. Tarief
medicatiebeoordeling doet geen recht aan tijdsbesteding. Pharm Weekbl 2012;147:(36).
52. Karapinar-Carkit F, Borgsteede SD, Zoer J, Smit HJ, Egberts AC, van den Bemt PM.
Effect of medication reconciliation with and without patient counseling on the number
of pharmaceutical interventions among patients discharged from the hospital. Ann
Gen Pract 2007;57:723-31.
56. Williams ME, Pulliam CC, Hunter R, Johnson TM, Owens JE, Kincaid J, et al. The ShortTerm Effect of Interdisciplinary Medication Review on Function and Cost in Ambulatory
Elderly People. J Am Geriatr Soc 2004;52:93-8.
59. Vinks THAM, Egberts TCG, de Lange TM, de Koning FHP. Pharmacist-Based Medication
Review Reduces Potential Drug-Related Problems in the Elderly The SMOG Controlled
Trial. Drugs Aging 2009;26:123-33.
60. Santschi V, Chiolero A, Paradis G, Colosimo AL, Burnand B. Pharmacist Interventions to
Improve Cardiovascular Disease Risk Factors in Diabetes: A systematic review and metaanalysis of randomized controlled trials. Diabetes Care 2012;35:2706-17.
62. Williams M, Peterson GM, Tenni PC, Bindoff IK, Curtain C, Hughes J, et al. Drug-Related
Problems Detected in Australian Community Pharmacies: The PROMISe Trial. Ann
64. Vivian EM. Improving blood pressure control in a pharmacist-managed hypertension clinic.
Pharmacotherapy 2002;22:1533-40.
65. Clifford RM, Davis WA, Batty KT, Davis TM, Fremantle Diabetes Study. Effect of a
pharmaceutical care program on vascular risk factors in type 2 diabetes: the Fremantle
Diabetes Study. Diabetes Care 2005;28:771-6.
182
| C hapte r 5
66. Rollason V, Vogt N. Reduction of polypharmacy in the elderly: a systematic review of the role
of the pharmacist. Drugs Aging 2003;20:817-32.
67. Blenkinsopp A, Bond C, Raynor DK. Medication reviews. Br J Clin Pharmacol
2012;74:573-80.
68. Holland R, Desborough J, Goodyer L, Hall S, Wright D, Loke YK. Does pharmacist-led
systematic review and meta-analysis. Br J Clin Pharmacol 2008;65:303-16.
69. Royal S, Smeaton L, Avery AJ, Hurwitz B, Sheikh A. Interventions in primary care to reduce
medication related adverse events and hospital admissions: systematic review and metaanalysis. Qual Saf Health Care 2006;15:23-31.
70. Kheir NM, van Mil JW, Shaw JP, Sheridan JL. Health-related quality of life measurement in
pharmaceutical care. Targeting an outcome that matters. Pharm World Sci 2004;26:125-8.
71. Bond C. The MEDMAN study: A randomized controlled trial of community pharmacy-led
72. Oliver D, Connelly JB, Victor CR, Shaw FE, Whitehead A, Genc Y, et al. Strategies to prevent
falls and fractures in hospitals and care homes and effect of cognitive impairment: systematic
review and meta-analyses. BMJ 2007;334:82.
73. Tsuyuki RT, Johnson JA, Teo KK, Simpson SH, Ackman ML, Biggs RS, et al. A randomized
trial of the effect of community pharmacist intervention on cholesterol risk management:
the Study of Cardiovascular Risk Intervention by Pharmacists (SCRIP). Arch Intern Med
2002;162:1149-55.
74. Bosworth HB, Powers BJ, Olsen MK, McCant F, Grubber J, Smith V, et al. Home blood
pressure management and improved blood pressure control: results from a randomized
controlled trial. Arch Intern Med 2011;171:1173-80.
75. Weinberger M, Murray MD, Marrero DG, Brewer N, Lykens M, Harris LE, et al.
Effectiveness of pharmacist care for patients with reactive airways disease: A randomized
controlled trial. J Am Med Assoc 2002;288:1594-1602.
76. Koshman SL, Charrois TL, Simpson SH, McAlister FA, Tsuyuki RT. Pharmacist care of
patients with heart failure: a systematic review of randomized trials. Arch Intern Med
2008;168:687-94.
77. Spinewine A, Fialova D, Byrne S. The role of the pharmacist in optimizing pharmacotherapy
in older people. Drugs Aging 2012;29:495-510.
183
Older people often present a complex clinical picture with multi-morbidity

accompanied by frailty. The implementation of guidelines for the management of
the individual conditions of these patients has resulted in an increase in the number
of drugs prescribed. More than one million patients over 65 years are using more
than 5 chronic medications. Due to the ageing of the population this number
will further increase in the next decades. The HARM-study showed that 5.6% of
hospitalizations in The Netherlands are drug-related and recommended a regular
medication review in older patients with polypharmacy. This evaluation should
also include identification of barriers for medication regimen adherence and should
provide tools to facilitate the proper use of medication.
This thesis focuses at both the medication management and the quality of
pharmacotherapy of older people with polypharmacy in primary care and ignores
older people in nursing homes and hospitals. Clinical pharmacy interventions for
older people with polypharmacy can be divided in dispensing services (aimed
at support of medication management), e.g. multidose dispensing systems,
and medication reviews (aimed at appropriateness of the pharmacotherapy).
The objective of this thesis is to describe the effects of these clinical pharmacy
interventions on the medication management and quality of pharmacotherapy of
older patients with polypharmacy in primary care.
This thesis consists of three parts. Chapter 2 presents a systematic literature review
that investigates how the extent of collaboration between the general practitioner
(GP) and the pharmacist impacts on the implementation of recommendations
arising from medication review. Chapter 3 focuses on the impact of multi dose
drug dispensing systems on the medication management of older patients and the
appropriateness of medication in these systems. Chapter 4 focuses on the impact
of medication review on different outcomes and in-depth analyses of the sources
for identification of drug-related problems (DRPs), the pharmacists conducting
medication reviews and the presence of explicit criteria in DRPs.
187
Chapter 2 describes a systematic review where 12 RCTs on medication review

for community-dwelling older patients have been included. The presence of
the following eight key elements reflecting collaboration was scored for each
intervention: pharmacist with clinical experience, own pharmacist involved, sharing
of medical records, patient interview by pharmacist, invitation of patients by GP,
case conference between GP and pharmacist, action plan, follow-up. The primary
outcome was the percentage of recommendations arising from medication review
that was implemented (implementation rate). This review showed an association
between the number of these key elements reflecting collaborative aspects in
medication review and the implementation rate of recommendations: = 0.085
(95% confidence interval [95%CI] 0.0520.128; P < 0.01). This suggests that a more
intensive collaboration between GP and pharmacist in medication review leads to
higher recommendation implementation rates.
An increasing proportion of older patients in Dutch community pharmacies is
using multidose drug dispensing (MDD) systems. MDD is a sophisticated dosing
aid with all drugs intended for one dosing moment gathered in disposable bags.
It has been suggested that MDD increases medication adherence, but specific
evidence is lacking. In Chapter 3.1 we compared the self-reported medication
adherence and knowledge of older patients receiving their drugs via multidose drug
dispensing (MDD-users) with patients receiving manually-dispensed drugs (nonMDD-users). We showed that the percentage of patients being adherent to all their
drugs was higher for MDD-users (81%) compared to patients receiving manuallydispensed drugs (58%). These differences were independent of knowledge and
cognitive function. The percentage of patients with adequate medication knowledge
was lower for MDD-users (40%) compared to patients receiving manuallydispensed drugs (79%).
There are concerns that MDD could lead to perpetual repeating of drug therapies
without the necessary re-evaluation which may increase inappropriate drug use. In
Chapter 3.2 the results are presented of a pragmatic randomized controlled study
on the effect of medication review for older patients using MDD on DRPs. This
study with only medication records available (prescription review) identified
a high number of potential DRPs among MDD-users (mean 8.5 per patient).
Medication review decreased the number of DRPs leading to a recommendation for
drug change among MDD-users 30% after 6 months compared to a spontaneous
decrease of 5% among MDD-users who initially did not receive a medication
review. We recommend that all MDD-users should have a thorough medication
review by pharmacists and prescribers.
188
S um m a r y
| C hapter 6.1
Chapter 4.1 describes the results of a randomised controlled trial on the effect
of Home Medication Review, a clinical medication review that incorporates all
collaborative aspects including a patient interview at home, on DRPs, diseasespecific outcomes (systolic blood pressure [SBP], LDL-cholesterol and glycosylated
hemoglobin [HbA1c]) and health-related quality of life in older patients. Intervention
patients had a mean of 5.8 potential DRPs leading to a recommendation for drug
change We observed a similar decrease in the number of these DRPs (28%) as in
the study described in Chapter 3.2. Compared to the control group, patients in the
intervention group more often achieved LDL treatment goals (odds ratio [OR] 2.0
[1.13.6]; P = 0.020) and had lower LDL levels (mean 0.18 mmol/l [0.33 to 0.04];
P = 0.015). Effects on LDL-cholesterol were more pronounced for the secondary
prevention subgroup. No differences were found for SBP- and HbA1c treatment
goals, the combined endpoint of all treatment goals and EQ-5D utility and VAS
scores. We may conclude that this medication review with all patient data available
resolved a substantial number of drug-related problems and improved LDLcholesterol, but did not lead to improvements in other disease-specific outcomes
and health-related quality of life.
Chapter 4.2 focuses on the contribution of the patient interview to the identification
of DRPs within the Home Medication Review. Medication review for 155 patients
resulted in a mean of 10 DRPs per patient. This revealed that 27% of these DRPs
were identified by patients interviews and 74% by medication and clinical records.
Compared to DRPs identified from medication and clinical records, DRPs identified
during patient interviews were more frequently assigned a high priority (OR 1.8
[1.42.2]), were more frequently associated with recommendations for drug change
(OR 2.4 [1.93.1]) and implemented recommendations for drug change (OR 2.8
[2.13.7]). We conclude that more than a quarter of all DRPs were identified during
patient interviews despite the availability of complete medication and clinical
records and were more frequently assigned a higher clinical relevance.
Chapter 4.3 provides insight in the completeness of DRPs identified by community
pharmacists within the Home Medication Review study. Community pharmacists
were inexperienced in medication review and received a two-day training course at
the start of the study. Community pharmacists identified a mean of 3.6 (standard
deviation [SD] 2.8) potential DRPs per patient while pharmacists with clinical
experience (expert reviewers) added 6.5 (SD 3.2) DRPs per patient. Community
pharmacists formulated 2.6 (SD 2.3) recommendations per patient and reviewers
added 7.5 (SD 3.3) recommendations. Community pharmacists identified
relatively more clinical relevant DRPs (for example DRPs assigned a high priority).
Community pharmacists identified a lower number of potential DRPs from patient
189
interviews (1.1 per patient) compared to pharmacist reviewers (1.6 per patient). The
results suggest that community pharmacists need intensive post-graduate training
to improve the identification of DRPs.
DRPs may be identified by implicit criteria (judgement-based: general questions
like Is there an indication for the drug?) and explicit criteria (criterion-based:
lists of potential inappropriate drugs). There are no studies that have evaluated to
what extent explicit criteria are identified by implicit criteria during medication
review. Chapter 4.4 aims to describe the number and types of explicit STOPP
(Screening Tool of Older Persons Prescriptions)-START (Screening Tool to Alert
doctors to Right Treatment) criteria present in identified potential DRPs and their
implementation rate. The percentage of START-criteria present in identified DRPs
was higher than the percentage of STOPP-criteria (13% vs. 5.7%; P < 0.01). 82%
of DRPs was not associated with STOPP-START-criteria. The implementation
rate for recommendations associated with STOPP-criteria was higher compared
to recommendations associated with START-criteria (56% vs. 39%; P < 0.01).
This study shows a higher prevalence of START-criteria compared to STOPPcriteria among identified DRPs in community-dwelling older patients. In contrast,
STOPP-criteria were implemented more frequently. The majority of DRPs was
not associated with STOPP-START criteria which underlines that implicit criteria
remain the most important instrument to identify resolvable DRPs.
In Chapter 5 (General discussion) the findings of our studies are considered into
a broader perspective by discussing the eligible patient groups, interventions and
different outcomes. The implications of our research on dispensing services and
medication review for both daily clinical practice and future research are also
discussed.
Eligibility criteria for MDD and medication review may overlap and certain patients
are eligible for MDD as well as medication review. Selection criteria for medication
review could be derived from the different data sources for clinical medication
review (patient interview, medication records and medical records) and possibly
by combining these sources. Potential criteria for an assessment of medication
management problems could be decreased cognitive function, low adherence and
the complexity of the medication regimen. The studies in this thesis did not provide
sufficient data to identify these and other potential criteria and this should be the
subject of future research.
Our medication review studies showed a great variation in the numbers of DRPs
per patient indicating that a comprehensive clinical medication review is not
necessary for each older patient. Discriminating medication review interventions in
190
S um m a r y
| C hapter 6.1
a short and extensive type could help the implementation in daily clinical practice.
Dependent on the number and type of eligibility criteria, a limited patient interview
could be conducted in the community pharmacy or a comprehensive patient
interview at home. Future research should investigate whether discriminating
patients who are eligible for different types of medication review is feasible.
Different outcome measures were used in our studies. The majority of our process
outcomes (e.g. medication changes and implementation rate) and intermediate
outcomes (e.g. DRPs) remain useful for future studies on medication review.
Disease-specific outcomes (e.g. cholesterol, blood pressure, HbA1c) are more
appropriate for disease-specific medication reviews such as interventions in
patients with diabetes, hyperlipidemia and hypertension. Health-related quality
of life remains an important clinical outcome as well as mortality and hospital
admissions, but these are all multifactorial. Medication review with a more focused
approach on patients needs, concerns and complaints could possibly have more
effect on health-related quality of life. Other more specific instruments to measure
health-related quality of life outcomes of medication review may be developed.
Medication review is still looking for the most appropriate outcome.
In conclusion, this thesis presented a series of studies on medication review which
showed that medication review in Dutch community-dwelling older patients with
polypharmacy identifies a large number of drug-related problems and some positive
effects on intermediate disease-specific outcomes. An average of 12 problems per
patient are resolved and could potentially lead to improved health outcomes. This
translation of drug-related problems into health benefits remains difficult.
The studies presented on multidose drug dispensing systems showed high
adherence, low medication knowledge and considerable inappropriateness of drugs
in these systems. These findings in combination with the steep growth of MDDusers emphasize the need for sufficient pharmaceutical care around multidose drug
dispensing systems, including medication review.
191
Het klinisch beeld bij ouderen is vaak complex door een combinatie van
meerdere gezondheidsproblemen, die vaak samen gaan met een sterk afgenomen
algemeen functioneren (kwetsbaarheid). De implementatie van uiteenlopende
behandelrichtlijnen bij deze patinten heeft geleid tot een toename van het gebruik
van geneesmiddelen. Inmiddels gebruikt meer dan 1 miljoen ouderen 5 of meer
geneesmiddelen, ook wel polyfarmacie genoemd. Door de vergrijzing zal dit aantal
in de komende decennia verder toenemen. De HARM-studie liet zien dat 5.6% van
de ziekenhuisopnames in Nederland geneesmiddel-gerelateerd zijn. Mede op grond
daarvan wordt een jaarlijkse medicatiebeoordeling aanbevolen voor ouderen met
polyfarmacie. Daarbij wordt ook aanbevolen aandacht te besteden aan mogelijke
oorzaken voor therapieontrouw en zo nodig hulpmiddelen aan te reiken om het
juist gebruik van geneesmiddelen te bevorderen.
Het doel van dit proefschrift is om de effecten te beschrijven van interventies
op het medicatiebeheer en de kwaliteit van de farmacotherapie van ouderen
met polyfarmacie in de eerstelijnsgezondheidszorg. Ouderen in verpleeghuizen
en ziekenhuizen worden buiten beschouwing gelaten. Farmacotherapeutische
interventies voor ouderen kunnen worden onderverdeeld in ondersteuning bij
het medicatiebeheer, bijvoorbeeld met gendividualiseerde distributievormen, en
de medicatiebeoordeling, die voornamelijk gericht is op het optimaliseren van de
farmacotherapie.
Dit proefschrift bestaat uit drie delen. Hoofdstuk 2 is een systematische
literatuurbeoordeling naar de invloed van de intensiteit van de samenwerking
tussen huisarts en apotheker op de implementatie van aanbevelingen die
voortkomen uit een medicatiebeoordeling. Hoofdstuk 3 beschrijft de impact van
gendividualiseerde distributievormen (GDV) op het medicatiebeheer van ouderen
en de juistheid van de geneesmiddelen in deze systemen. Hoofdstuk 4 beschrijft
de effecten van medicatiebeoordeling in zowel een gerandomiseerd gecontroleerd
onderzoek als een grootschalige praktijkevaluatie. In dit hoofdstuk is tevens
gekeken naar de wijze waarop farmacotherapie-gerelateerde problemen (FTPs)
193
worden gedentificeerd, de mate waarin relatief ongetrainde apothekers in staat zijn

FTPs te identificeren en de waarde van expliciete criteria bij het beoordelen van
FTPs.
Hoofdstuk 2 beschrijft een systematische literatuurbeoordeling van 12 gerandomiseerde gecontroleerde onderzoeken op het gebied van medicatiebeoordeling
voor thuiswonende ouderen. De aanwezigheid van de volgende acht elementen,
die de samenwerking tussen apotheker en huisarts weerspiegelen, werd gescoord:
apotheker met klinische ervaring, betrokkenheid van de eigen apotheker,
de beschikbaarheid van medische gegevens, gesprek met de patint door de
apotheker, uitnodigen van patinten door de huisarts, face-to-face overleg
tussen huisarts en apotheker, behandelplan en follow-up. De primaire uitkomst
was de mate waarin de aanbevelingen van de medicatiebeoordeling, werden
doorgevoerd (implementatiegraad). Uit deze literatuurbeoordeling bleek dat meer
samenwerking geassocieerd was met een hogere implementatiegraad ( = 0.085
(95% betrouwbaarheidsinterval [95%BI] 0.0520.128; P < 0.01).
Een steeds groter gedeelte van de oudere patinten in Nederlandse apotheken
gebruikt een GDV, zoals een medicatierol (vaak Baxterrol genoemd). Bij
een medicatierol zijn de geneesmiddelen voor een individuele patint per
toedieningstijdstip verpakt in n zakje. Deze GDVs zouden de therapietrouw
verhogen, maar tot voor kort ontbrak daarvoor specifiek bewijs. In hoofdstuk 3.1
vergeleken we de zelf-gerapporteerde therapietrouw en medicatiekennis van 119
oudere gebruikers van medicatierollen met die van 96 oudere gebruikers van
reguliere verpakkingen van geneesmiddelen in acht apotheken. Het percentage
patinten dat therapietrouw was voor alle geneesmiddelen was hoger voor
de gebruikers van medicatierollen (81%) dan voor gebruikers van reguliere
verpakkingen (58%). Dit verschil was onafhankelijk van medicatiekennis en
cognitie. Het percentage patinten met adequate medicatiekennis was lager
voor gebruikers van medicatierollen (40%) dan voor gebruikers van reguliere
verpakkingen (79%).
Er zijn zorgen dat geneesmiddelen in GDVs worden herhaald zonder dat
herbeoordeling van de noodzaak en effecten van het geneesmiddel plaats vindt.
In hoofdstuk 3.2 laten we de resultaten zien van een pragmatisch gerandomiseerd
onderzoek in 6 apotheken naar medicatiebeoordeling bij 118 oudere gebruikers van
medicatierollen. Patinten werden at random toegewezen aan een interventiegroep,
waarbij direct een medicatiebeoordeling werd uitgevoerd, of aan de controlegroep,
waarbij dit pas na afloop van het onderzoek plaatsvond. In dit onderzoek werd
194
S a me nvatting
| C hapter 6.2
in beide groepen een hoog aantal potentile FTPs gedentificeerd (gemiddeld

8.5 per patint), waarvan ongeveer de helft werd gevolgd door een aanbeveling
voor een medicatiewijziging (gemiddeld 4.5 per patint). Medicatiebeoordeling
verlaagde het aantal FTPs met een aanbeveling voor een medicatiewijzing met
30% in de interventiegroep ten opzichte van een spontane daling van 5% in de
controlegroep. We bevelen daarom voor iedere oudere medicatierolgebruiker een
medicatiebeoordeling aan.
Het farmacotherapeutische thuisconsult is een bijzondere vorm van medicatiebeoordeling, waarbij het gesprek van de apotheker met de patint over de
ervaringen en zorgen rondom geneesmiddelen thuis plaats vindt. Deze vorm
van medicatiebeoordeling is gebaseerd op het Australische Home Medicines
Review. Hoofdstuk 4.1 beschrijft een gerandomiseerd gecontroleerd onderzoek
in 10 apotheken naar het effect van het farmacotherapeutisch thuisconsult op het
aantal FTPs, ziekte-gerelateerde uitkomsten (systolische bloeddruk [SBD], LDLcholesterol en geglycosyleerd hemoglobine [HbA1c]) en de kwaliteit van leven
van 285 ouderen. Ook hier werden patinten at random toegewezen aan een
interventiegroep, waarbij een medicatiebeoordeling inclusief een medicatiegesprek
thuis plaats vond, of aan een controlegroep met alleen reguliere zorg. Gemiddeld
hadden patinten in de interventiegroep 5.8 FTPs per patint met een aanbeveling
voor een medicatiewijziging. We vonden een vergelijkbare daling van deze
FTPs (28%) als in het onderzoek naar medicatiebeoordeling bij de medicatierol
(hoofdstuk 3.2). In vergelijking met de controlegroep haalden patinten in de
interventiegroep vaker behandeldoelen voor LDL-cholesterol (odds ratio [OR]
2.0 [1.13.6]; P = 0.020) en bereikten lagere LDL-cholesterolspiegels (gemiddeld
0.18 mmol/l [0.33 tot 0.04]; p = 0.015). Deze effecten op het LDL-cholesterol
waren meer uitgesproken bij patinten die al eerder een cardiovasculair event
hadden gehad (hartinfarct, CVA of TIA). Er werden geen verschillen gevonden
bij de behandeling van bloeddruk (SBD), de instelling van diabetes (HbA1c) en de
scores op gezondheid-gerelateerde kwaliteit van leven (EQ-5D/VAS). We mogen
concluderen dat deze intensieve vorm van medicatiebeoordeling een aanzienlijk
deel van de FTPs oplost en LDL-cholesterol verbetert, maar geen verbetering geeft
van andere ziekte-gerelateerde uitkomsten en van kwaliteit van leven.
Het gesprek met de patint over diens geneesmiddelengebruik, ook wel
farmacotherapeutische anamnese genoemd, vormt een essentieel onderdeel van
de medicatiebeoordeling. In hoofdstuk 4.2 beschrijven we een dwarsdoorsnede
onderzoek binnen de interventiegroep van het gerandomiseerde onderzoek naar
het farmacotherapeutisch thuisconsult uit hoofdstuk 4.1. Medicatiebeoordeling bij
195
155 patinten leverde gemiddeld 10 FTPs per patint op. Zeventwintig procent van
deze FTPs bleek afkomstig uit het patintgesprek en de overige kwamen voort uit
de analyse van medicatie- en medische gegevens. Vergeleken met FTPs afkomstig
uit medicatie- en medische gegevens werd aan FTPs die voortvloeiden uit
patintgesprekken vaker een hogere klinische relevantie toegekend, zoals FTPs met
een hoge prioriteit (OR 1.8 [1.42.2] en FTPs met gemplementeerde aanbevelingen
voor een medicatiewijziging (OR 2.8 [2.13.7]). We mogen concluderen dat bij
een medicatiebeoordeling zonder een patintgesprek een aanzienlijk deel van de
klinisch relevante problemen zal worden gemist.
Hoofdstuk 4.3 betrof eveneens een dwarsdoorsnede onderzoek binnen het
onderzoek naar het farmacotherapeutische thuisconsult (zie hoofdstuk 4.1).
De openbaar apothekers in dit onderzoek hadden beperkte ervaring met
medicatiebeoordelingen en volgden een 2-daagse cursus bij het begin van het
onderzoek. Openbaar apothekers signaleerden gemiddeld 3.6 (standaard deviatie
[SD] 2.8) potentile FTPs per patint, terwijl gespecialiseerde apothekers (expert
reviewers) daar 6.5 (SD 3.2) FTPs per patint aan toevoegden. Openbaar
apothekers formuleerden gemiddeld 2.6 (SD 2.3) aanbevelingen per patint en
expert reviewers voegden daar 7.5 (SD 3.3) aanbevelingen aan toe. De meer klinisch
relevante FTPs (bijvoorbeeld FTPs met hoge prioriteit) werden relatief vaker
opgespoord door de openbaar apothekers. Openbaar apothekers identificeerden
minder FTPs uit de patintgesprekken (gemiddeld 1.1 per patint) ten opzichte van
expert reviewers (1.6 per patint). Op grond van dit onderzoek kunnen we voor
elke openbare apotheker een intensieve postacademische nascholing aanbevelen
voor het uitvoeren van medicatiebeoordelingen.
Het identificeren van FTPs kan plaatsvinden met behulp van impliciete
criteria (algemene vragenlijsten, bijvoorbeeld Is er nog een indicatie voor
het geneesmiddel?) en expliciete criteria (lijsten van potentieel ongewenste
geneesmiddelen). Het was niet eerder onderzocht in welke mate FTPs die door
het gebruik van impliciete criteria zijn opgespoord ook gevonden hadden kunnen
worden met expliciete criteria. Het doel van het onderzoek in hoofdstuk 4.4 was om
het aantal en type expliciete STOPP (Screening Tool of Older Persons Prescriptions)
en START (Screening Tool to Alert doctors to Right Treatment) criteria te
beschrijven die van toepassing zijn op reeds gedentificeerde FTPs. Het percentage
START-criteria aanwezig in FTPs was hoger dan het percentage STOPP-criteria
(13% vs. 5.7%; P < 0.01). STOPP-criteria werden vaker gemplementeerd dan
START-criteria (56% vs. 39%; P < 0.01). 82% van de FTPs was niet geassocieerd
met STOPP-START-criteria. Hieruit concludeerden we dat impliciete criteria het
belangrijkste instrument zijn voor het opsporen van FTPs.
196
S a me nvatting
| C hapter 6.2
In de algemene beschouwing (hoofdstuk 5) worden de bevindingen van onze

studies in een breder perspectief geplaatst van selectiecriteria voor patinten,
interventies en verschillende uitkomstmaten. Ook worden de implicaties van het
onderzoek voor zowel de dagelijkse praktijk als voor vervolgonderzoek besproken.
Selectiecriteria voor de medicatierol en medicatiebeoordeling hebben een zekere
mate van overlap waardoor sommige patinten zowel voor een medicatierol als een
medicatiebeoordeling in aanmerking komen. Zowel een gesprek met de patint,
kenmerken van de medicatie als klinische gegevens kunnen aanleiding geven tot
een medicatiebeoordeling. Potentiele criteria voor een beoordeling van problemen
met medicatiebeheer zouden kunnen zijn: verminderde cognitieve functie, lage
therapietrouw en de complexiteit van het medicatieregime. De onderzoeken in dit
proefschrift hebben onvoldoende gegevens opgeleverd om deze criteria verder uit
te werken. Dit is dan ook belangrijk voor vervolgonderzoek.
Het uitgevoerde onderzoek naar medicatiebeoordeling liet een grote variatie
zien in de aantallen FTPs per patint, wat suggereert dat een uitgebreide
medicatiebeoordeling niet per se nodig is voor elke oudere patint. Een
onderscheid in een korte en een uitgebreide vorm van medicatiebeoordeling zou
de implementatie in de dagelijkse praktijk kunnen bevorderen. Afhankelijk van
het aantal en type selectiecriteria kan een kort patintgesprek worden gevoerd in
de apotheek of een uitgebreid gesprek bij de patint thuis. Vervolgonderzoek zou
moeten uitwijzen of een onderverdeling van medicatiebeoordelingen uitvoerbaar
is.
In onze studies hebben we verschillende uitkomstmaten gebruikt. De
meerderheid van onze procesuitkomsten (bijvoorbeeld medicatiewijzigingen en
implementatiegraad) en intermediaire uitkomsten (bijvoorbeeld FTPs) blijven
bruikbaar voor vervolgonderzoek naar medicatiebeoordeling. Ziekte-gerelateerde
uitkomstmaten (bijvoorbeeld cholesterol, bloeddruk en HbA1c) zijn beter
toepasbaar voor ziekte-gerelateerde medicatiebeoordelingen, zoals interventies
in patinten met diabetes, hyperlipidemie en hoge bloeddruk. Gezondheidgerelateerde kwaliteit van leven blijft een belangrijke klinische uitkomstmaat net
als mortaliteit en ziekenhuisopnames, maar allen zijn multifactorieel bepaald.
Een medicatiebeoordeling, die nog meer focust op de specifieke behoeften,
zorgen en klachten van de patint, heeft mogelijk meer effect op de kwaliteit
van leven. Ook zouden meer instrumenten ontwikkeld kunnen worden, die de
kwaliteit van leven specifiek als uitkomst van medicatiebeoordeling meten. Voor
de medicatiebeoordeling zijn we dus nog steeds op zoek naar de meest geschikte
uitkomstmaat.
197
Concluderend worden in dit proefschrift meerdere onderzoeken naar

medicatiebeoordeling beschreven. Ze tonen aan dat medicatiebeoordeling voor
Nederlandse thuiswonende ouderen met polyfarmacie een aanzienlijk aantal
farmacotherapie-gerelateerde problemen identificeert en positieve effecten op
ziekte-gerelateerde uitkomsten heeft. Het blijft echter moeilijk om het oplossen van
farmacotherapie-gerelateerde problemen te vertalen in gezondheidswinst.
De onderzoeken naar de gebruikers van medicatierollen laten een hoge
therapietrouw, een lage medicatiekennis en een hoog aandeel potentieel ongeschikte
geneesmiddelen in deze systemen zien. Deze bevindingen in combinatie met de
toename van het aantal gebruikers van medicatierollen onderstrepen de behoefte
aan gestructureerde farmaceutische patintenzorg rondom deze systemen, inclusief
een medicatiebeoordeling.
198
Een van de aanleidingen voor dit promotieonderzoek was het overlijden

van mijn vader. Een week voor zijn overlijden voerde ik bij hem mijn eerste
medicatiebeoordeling uit en verwoordde mijn vragen in een brief aan zijn huisarts
(zie achteraan dit dankwoord). Dit resulteerde in een ziekenhuisopname, maar
heeft voor mijn vader uiteindelijk niet mogen baten.
In het jaar daarna bleef het prettig om te werken als apotheker bij apotheek
Stevenshof, maar begon ik na te denken over de toekomst. Bij de zusterorganisatie
SIR Institute for Pharmacy Practice and Policy kreeg ondertussen het project
farmacotherapeutisch thuisconsult gestalte. Daarmee ontstond de mogelijkheid
voor een promotieonderzoek naar medicatiebeoordeling en zon mooie uitdaging
kon ik niet laten lopen.
Daarmee begon de marathon naar een finish die jarenlang niet zichtbaar was.
Hardlopen is al jarenlang mijn ding en in 2009 liep ik daadwerkelijk de marathon.
De marathon zelf is het zo lang mogelijk je eigen tempo aanhouden. Het laatste
stuk is overleven, wetend dat de finish nog steeds ver is, maar toch dichterbij komt.
Met name dan kan er veel gebeuren waardoor het alsnog mis kan gaan: hongerklop,
uitdroging, misselijkheid, diarree. Ook al gebeurt dat, opgeven is geen optie.
Gelukkig staat er op het laatste stuk ook meer publiek en wordt er harder voor je
gejuicht. Het is een fantastisch gevoel als je uiteindelijk de finish overgaat.
In tegenstelling tot een marathon lukt een promotietraject je niet alleen. Je hebt
veel mensen nodig: mensen die je oorspronkelijke werk overnemen, mensen die
je de kneepjes van het onderzoek leren, mensen die onderzoek voor je uitvoeren,
mensen die je motiveren en mensen waarmee je kunt ontspannen. Al deze mensen
waren er tijdens mijn promotietraject en vaak waren ze een mix.
Allereerst wil ik daarom de twee mensen bedanken, die mij in 1996 hebben
aangenomen als apotheker bij apotheek Stevenshof. Als eerste Henk Buurma,
199
directeur van SIR en apotheek Stevenshof, maar bovenal founding father

van de eerste academische openbare apotheek in Nederland. Jouw hart ligt bij
het farmaceutische praktijkonderzoek en je weet steeds anderen hiervoor te
enthousiasmeren. Wat een geweldige club mensen heb jij om je heen verzameld.
Jij bent altijd zeer enthousiast om een promotietraject op te starten. In mijn
geval wist je dat ik een nieuwe uitdaging zocht en toverde je een prachtig traject
rondom medicatiebeoordeling uit de hoed. Toen het wat moeizamer ging tijdens
het onderzoek wist je me ook te vertellen dat er geen weg terug was. Dus ben ik
doorgegaan en ik ben je er heel dankbaar voor.
Marcel Bouvy is natuurlijk die andere die ik op de voorgrond wil zetten. Ook jij
stond aan de basis van SIR en de academische apotheek Stevenshof. Jij bent mijn
eerste promotor, de apotheker-professor van de afdeling Farmaco-epidemiologie
en Klinische Farmacologie van de Universiteit Utrecht en altijd positief en
optimistisch. Je hebt me bij het onderzoek op alle mogelijke manieren geholpen:
van de meeste ingewikkelde queries in Access tot het gezamenlijk werken aan
de General Discussion in het weekend. Daarnaast heb je voor een professor
ongevenaarde sociale eigenschappen, van een vette roddel tot het meeleven
met mijn moeder. Niets is voor jou teveel, je hebt een ongelofelijke berg werk te
verzetten en ik zie je er nooit moedeloos van worden.
Jacobijn Gussekloo is mijn andere promotor: jij bent de huisarts-professor van de
afdeling Public health en Eerstelijnsgeneeskunde van het Leids Universitair Medisch
Centrum. Jij hebt na het eerste jaar het stokje overgenomen van Pim Assendelft
die zich meer ging richten op de preventieve huisartsgeneeskunde. Als geen ander
weet jij hoe je succesvol oud kan worden. Jouw heldere en nuchtere kijk op de
manuscripten ben ik steeds meer gaan waarderen. Niet zelden raakte jij als lezer de
weg kwijt in de eerste versies van een manuscript. Je kon dan duidelijk aangegeven
hoe ik een navigatiesysteem kon opzetten om de lezer met succes naar de conclusie
te begeleiden. Ook jij was heel motiverend met je begeleidend commentaar (Mooi
werk) en ook in de eindfase stimuleerde je me om nog even vol te houden, omdat
ik er bijna was. Ook ben ik je heel dankbaar voor de samenwerking met de mensen
van jouw afdeling.
Adrianne Faber, collega-apotheker en onderzoeker bij het SIR. Jij was mijn copromotor en wat voor n. Allereerst ben jij een zeer warm persoon die zich meer
kan inleven dan wie dan ook en openstaat voor alle zielenroerselen. In het begin
heb je heel hard meegeholpen met het onderzoek naar het farmacotherapeutisch
200
D a n k wo o rd
| C hapter 6.3
thuisconsult (FTC) en hebben we samen nascholing gegeven. Ook bij de

medicatiebeoordeling van de medicatierollen was jij de controller. Tijdens en na
jouw 1e zwangerschap werd ik zelfstandiger in het onderzoek doordat je andere
taken kreeg, maar ook dat was uiteindelijk een goede ontwikkeling. Jij bent er altijd
als dat moet en ik hoop nog lang met je samen te werken.
Wendy den Elzen, onderzoeker bij Public health en Eerstelijnsgeneeskunde, ik ben
je zeer dankbaar voor je co-auteurschap en voor de hulp bij de statistiek en vooral
door het learning by doing zodat ik zelf lineair mixed modelling kon uitvoeren in
SPSS. Daarnaast was het fijn om met jouw zeer vrolijke persoonlijkheid te mogen
samen te werken en je enthousiaste e-mails te lezen. Jean-Marie Riemens-Louisse,
je hebt een goede start gemaakt met de systematische review: ook al is het geen
Cochrane review geworden, het is toch gepubliceerd! Lynette Bermingham, it was a
pleasure for me as a pharmacist to cooperate with you as a general practitioner from
the UK on our systematic review on collaboration between GP and pharmacists.
We certainly collaborated well! Our weekly meetings were enjoyable. Suddenly, you
had to return to the UK. I still hope that I can meet you in the near future. Svetlana
Belitser, statisticus van de universiteit Utrecht, ook jou wil heel erg bedanken voor
de statistische analyses bij de systematische review, speciaal voor het maken van de
hele mooie plaatjes in R.
De dames van het secretariaat van de afdeling Farmaco-epidemiologie en Klinische
Farmacologie van de Universiteit Utrecht (Anja, Ineke en Suzanne) wil ik hartelijk
danken voor alle secretarile ondersteuning en het altijd vinden van een plekje voor
mij als dagjesmens.
Alle professoren van de beoordelingscommissie, Toine Egberts, Peter de Smet, Niek
de Wit, Marcel Olde Rikkert en Theo Raynor, wil ik hartelijk bedanken voor het
goedkeuren van mijn proefschrift en hun positieve commentaar.
Het al genoemde onderzoek naar het FTC was de reden om te starten met dit
promotietraject. Het was verreweg het meest intensieve onderzoek en naast
Adrianne was Paulien Schul, nu van BENU Apotheken (voorheen LLOYDS
apotheken), mijn steun en toeverlaat. Paulien, jij had het zelf al heel druk en
veranderde van functie tijdens de rit, maar bleef tijd houden om het FTC te
cordineren. Sterker nog, jij bent in heel wat apotheken persoonlijk geweest
om te helpen met de data-invoer en het opvragen van de laboratoriumwaarden.
Zonder jou was het complexe FTC-onderzoek nooit een succes geworden. Met
201
alle deelnemende FTC-apothekers heb ik gedurende 2,5 jaar zeer intensief contact
gehad en ik wil jullie van harte bedanken: Ellemieke Boerstoel, Esther Coes, Ellen
van Briel, Eva de Groot, Paul Jansema, Mirjam Meerburg, Monique Martens-van
Motman, Martine Uiterwijk-Winkel, Bart Vincken en Mieke Waardenburg. Ook
Anil Sewratan, Youssef Amararouchi en Marcelle Broshuis wil ik bedanken voor
het uitvoeren van de follow-up en aanleveren van data uit verschillende apotheken.
Het FTC-onderzoek was niet mogelijk geweest zonder de financile ondersteuning
van de sponsoren KNMP, Achmea, Menzis en Astra Zeneca.
Tijdens het promotietraject ontdekte ik gaandeweg dat het wat sneller ging als (nog)
meer mensen gingen meedenken en -helpen. Allereerst was daar het onderzoek
naar de kennis en therapietrouw van de medicatierol in de SIR Masterclass
farmaceutisch praktijkonderzoek 20102011. Ik wil daarom Laurens Biesma, Hinke
Dorhout, Rene Dull, Eva de Groot, Rik van der Meer en Ruud Timmerman heel
hartelijk bedanken voor de vruchtbare samenwerking bij het uitdenken van de
onderzoeksopzet en voor de uitvoering van het onderzoek in jullie apotheken. Ik
ben natuurlijk reuze trots dat de huidige voorzitter van de KNMP heeft meegedaan
aan dit onderzoek. Ook de twee niet-Masterclass apothekers, Karin ten Kate en
Inge Stollman-Truijen, wil ik bedanken voor hun medewerking aan dit onderzoek.
Maar niet alleen apothekers hebben me geholpen. Bij de SIR dienden zich heel wat
mensen aan die stage wilden lopen. Reeds genoemd, Youssef Amarouchi, student
farmacie, heeft meegeholpen met FTC-onderzoek. Zo ook Cynthia Oosterlee,
studente farmakunde, die een half jaar lang de 1e fase van het Masterclassonderzoek mede heeft uitgevoerd en geanalyseerd. Buiten SIR had ik nog nooit
iemand ontmoet die zo handig was met Access als jij. Ook toen jij al klaar was
met je stage heb je me nog geholpen met de gegevensverwerking. De 2e fase van
het onderzoek werd uitgevoerd door Glenn Stolk, student farmacie. Glenn, jij
hebt heel Nederland doorgereisd met het openbaar vervoer om de interviews
in de verschillende apotheken af te nemen. Ook voor de vragenlijst voor de
cognitieve functie draaide jij je hand niet om. Met je verslag heb je de 1e aanzet
gegeven voor een internationale publicatie en ben je 2e auteur. Niet te vergeten,
Sanne Verdoorn, een zeer gemotiveerde studente farmacie met grote interesse
voor medicatiebeoordeling. Sanne, jij had een excellent verslag gemaakt over
de medicatiebeoordeling bij de apotheken van Connecting Care. Ook heb je
meegedacht over de 1e app medicatiebeoordeling voor de iPad. Samen hebben we
jouw onderzoek uitgewerkt tot het STOPP-START-manuscript. Sanne, ik hoop van
harte dat jijzelf ook promotieonderzoek gaat doen.
202
D a n k wo o rd
| C hapter 6.3
Mark Naunton, you are an Australian pharmacist/lecturer married with a Dutch

pharmacist. I was very pleased to meet you at the station of Assen, exact one day
before you went back to Tasmania. I just decided that I was going to conduct the
Dutch variant of the Australian Home Medicines Review. You told me everything
about the home visits in Australia and the forms youre using. I am very delighted
and proud that you will give a lecture at my dissertation day.
Het farmacotherapeutisch thuisconsult beleefde in 2007 een vliegende start met de
tv-uitzending Kies beter op RTL. Ik wil de heer en mevrouw Mingelen en mevrouw
Lens hiervoor hartelijk bedanken voor het filmen van een medicatiegesprek bij hen
thuis. Het geeft een heel goed gevoel dat de afsluiting van het promotietraject ook
weer bij de familie Mingelen thuis plaatsvond, resulterend in de foto van meneer
op de voorkant. Ook de heer Verzijden en meneer en mevrouw Kromhout wil ik
bedanken voor de bereidheid om op de foto en de film te gaan voor respectievelijk
het Pharmaceutisch Weekblad en het KNMP-congres. Tenslotte wil ik alle mensen
bedanken waarmee we een medicatiegesprek hebben gehouden.
Graag wil ik ook de apothekers van de 6 apotheken bedanken die deelnamen
aan het onderzoek naar de medicatiebeoordeling bij de medicatierol. Delia Titre,
destijds studente Farmacie, wil ik bedanken voor de hulp bij de dataclassificatie.
Daarbij wil ik ook Apotheek Voorzorg, in het bijzonder Jan de Groot, en de
KNMP bedanken voor het sponsoren van dit onderzoek. Ook de apothekers
van de 13 Connecting Care apotheken wil ik bedanken voor al hun uitgevoerde
medicatiebeoordelingen en medewerking bij aanleveren medicatiegegevens bij het
STOPP-START-onderzoek. Speciaal gaat mijn dank uit naar apothekers Paul van
Bakel en Bert Mouthaan van Connecting Care voor het beschikbaar stellen van
Sanne Verdoorn voor het STOPP-START-onderzoek en Marita van Herpen van van
Herpen Kwaliteitsmanagement voor alle cordinatie en secretarile ondersteuning.
Francis te Nijenhuis, de ontwerper van mijn proefschrift. We kennen elkaar
nog maar een paar maanden. Ik wil je bedanken voor je conscintieuze werk en
kritische instelling, maar vooral ook voor de enorme hoeveelheid energie en werk
die je erin hebt zitten. Heel bijzonder vond ik ook je inlevingsvermogen in mijn
privomstandigheden en het meedenken hoe we dit konden laten terugkomen in
het proefschrift.
Tijdens het promotietraject kwam ik ook andere kandidaat-promovendi tegen.
Twee daarvan wil ik hier naar voren halen. Als eerste Anne Leendertse. Jij begon in
203
2007 ook met een onderzoek naar medicatiebeoordeling nadat je al een succesvol
HARM-onderzoek had afgerond. Naast de vele gesprekken over de invulling van
medicatiebeoordeling en het onderzoek heb je mij ook gevraagd mee te denken
in de ontwikkeling van de multidisciplinaire KKCZ-richtlijn Polyfarmacie bij
ouderen, die in 2012 is verschenen. De deelname aan deze werkgroep heeft mij
mede genspireerd bij het uitvoeren van de systematische literatuurbeoordeling
en daar wil ik ook Rob van Marum, geriater, en Monique Verduyn, apotheker
bij NHG, voor bedanken. Het was leuk om gezamenlijk een prijs (ZonMw Parel)
te ontvangen voor deze richtlijn. Anne, veel succes met je innoverende project
rondom de apotheker in de huisartsenpraktijk. Ten tweede, Annemieke Floor,
collega-apotheker van het SIR. Onze onderzoekstrajecten liepen parallel en
we spraken elkaar het meest op maandag op de universiteit in het Sebastiaan
Wendt-gebouw. We hebben veel uitgewisseld over de drukte en de balans tussen
onderzoek en andere werkzaamheden. Ik vind het bijzonder hoe je zon intensief
promotietraject ook nog eens combineert met je functie als voorganger bij het
Apostolisch Genootschap. Als onze promoties achter de rug zijn, hoop ik nog vele
jaren met je samen te werken.
Een aantal mensen van het SIR zijn al genoemd, maar hier werken nog veel meer
leuke en inspirerende mensen. Ik wil jullie ook in het zonnetje zetten voor het
meeleven met mijn onderzoek en danken voor de gezellige gesprekken die we
hebben. Seline, bedankt voor alle organisatieactiviteiten rondom mijn promotiedag.
Ook de hulpmedewerkers, bezorgers, apothekers-assistenten en apothekers van
apotheek Stevenshof wil ik hier graag op de voorgrond zetten. Wat een bijzondere
en fijne club mensen bij elkaar. Alle lof en dank voor jullie flexibiliteit en geduld
en voor het meeleven met mijn onderzoek waardoor ik steeds minder in beeld
was in de apotheek. Niet te vergeten, de huisartsenpraktijk Stevenshof en Zaaijer
waar we als apotheekteam zo intensief mee samenwerken. Leonie Hulst, Peter van
Hartingsveldt en Bram Mertens, ik wil jullie graag bedanken voor de invulling van
de apothekerstaken in de afgelopen jaren.
Dan is er nog een apotheker die ontbreekt in dit rijtje: Anne-Margreeth Krijger.
Dat heeft een reden, want jij bent namelijk paranimf. We werken bijna even lang bij
apotheek Stevenshof en delen een bijzondere eigenschap: een wel heel apart gevoel
voor humor. Wat hebben we veel gelachen bij het recepten nakijken. Tijdens de
eerste jaren van mijn onderzoek was jij een van de expert reviewers van SIR en heb
je een hele grote bijdrage geleverd aan mijn onderzoeken. Jouw Dr House review
zal ik niet gauw vergeten. Dan die andere paranimf, Duco Smit, een van mijn grote
204
D a n k wo o rd
| C hapter 6.3
vrienden. We kennen elkaar al uit de studententijd uit Groningen en kwamen elkaar

toevallig weer tegen in Haarlem. We hebben veel samen hardgelopen, gevolleybald,
getennist, biertjes gedronken en bovenal goede gesprekken gevoerd. Tegenwoordig
ben je een 2e carrire begonnen als drummer van de fenomenale rock coverband
Flow. Omdat ik geen instrument kan spelen, had ik de eer daarvan de manager te
worden en hoop ik die taak de komende jaren wat serieuzer in te vullen.
Het ontspannende element komt steeds nadrukkelijker naar voren aan het eind van
dit dankwoord. Daar wil ik mijn vrienden en vriendinnen van volleybal, tennis,
school, studententijd en mijn vriendenclub uit Meppel voor bedanken.
Nu de finish bijna in zicht is denk ik weer terug aan de marathon van Rotterdam
in 2009. Wat was het fijn om mijn vrouw en oudste zoon als eerste te zien na het
passeren van de finish. Ik wil mijn gezin bedanken voor het geduld en ondersteuning
bij dit promotietraject, vooral in die laatste intensieve fase. Ik heb er zin in om weer
veel leuke dingen met ons vieren te gaan doen.
Aan het eind van dit promotietraject is mijn moeder overleden. Ik wil graag mijn
zus en haar familie bedanken voor het in huis nemen van mijn moeder gedurende
haar laatste 5 maanden. Ik heb vele treinreizen gemaakt richting Assen zodat ik
onderweg kon schrijven aan het proefschrift en mijn moeder kon bezoeken. Lieve
ma, jij was een hele warme moeder waarmee het altijd gezellig was en je hebt me
veel vrijheid gegeven: Onderzoekt alles en behoudt het goede.
205
Br i ef aan huis ar ts van mijn vad er

Geachte huisarts,
Ik schrijf u deze brief als zoon van de heer F. Kwint. Ik ben werkzaam
als apotheker in een gezondheidscentrum, waar apothekers een
herhaalreceptenlijn bemannen en dagelijks overleg hebben met de
huisartsen over deze recepten. In deze hoedanigheid hou ik mij bezig met
medicatieanalyse en voorstellen ter verbetering.
Ik maak mij grote zorgen over de conditie van mijn vader. Naar mijn idee
heeft hij zeer ernstig hartfalen, omdat hij in rust zeer vermoeid is en steeds
de neiging heeft in slaap te vallen. Daarnaast houdt hij zeer veel vocht vast in
het hele lichaam, maar mogelijk ook in zijn gezicht gezien zijn dikke lippen,
wangen en oogleden. Hij lijkt ook een dikke tong te hebben, omdat hij zeer
moeilijk praat.
Daarnaast heeft hij zelf het idee hallucinaties te hebben op het moment dat hij
zijn ogen weer opendoet. Zo zag hij bijvoorbeeld een vogel de kamer binnen
vliegen. Het kan echter ook te maken hebben met zijn bovenste ooglid dat
voor zijn oog hangt waardoor hij tijdelijk even een vlekje heeft.
Mijn vader zet zelf de medicatie uit in een weekcassette. Hij gebruikt de
volgende medicatie:
Acetylsalicylzuur 80 mg 1x daags 1
Bisoprolol 2,5 mg
1x daags 1
Atorvastatine 20 mg
1x daags 1
Furosemide 40 mg
1x daags 1 (i.p.v. 2x daags 1)
Ramipril 5 mg
2x daags 1 (i.p.v. 2x daags 1,5)
Esomeprazol 40 mg
1x daags 1
Momenteel gebruikt hij doxycycline in verband met een luchtweginfectie.
Mijn vader is onlangs bij de cardioloog geweest en deze heeft hem geadviseerd
furosemide 40 mg 2x daags 1 te gebruiken en ramipril 5 mg van 2x daags 1
206
D a n k wo o rd
| C hapter 6.3
te verhogen naar 2x daags 1,5. Hij hoefde pas na 6 maanden terug te komen.
Mijn vader had sindsdien last van hallucinaties en maakte zich ook zorgen
over het in slaap vallen. Hij las in de bijsluiter van ramipril dat de maximale
dosering 10 mg is en heeft daarom besloten dit weer te verlagen naar 2x daags
5 mg. Ook heeft hij zelfstandig de dosering van furosemide verlaagd naar 1 x
daags 40 mg.
Ramipril
Wat betreft de dosering van ramipril ben ik het met mijn vader eens, dat 15 mg
per dag, zeker voor hem een erg hoge dosis is. Bij een verminderde nierfunctie
(CL 20-50 ml/min) mag officieel maximaal 5 mg ramipril worden gegeven. Ik
weet dat mijn vader een verminderde nierfunctie heeft, maar weet niet of deze
de laatste maanden nog bepaald is. Verder is bij ACE-remmers in zeldzame
gevallen (<1/1000) angioneurotisch oedeem beschreven met zwelling van het
aangezicht, tong en larynx. Ik kan moeilijk het onderscheid maken tussen
oedeem door vochtretentie en angioneurotisch oedeem, maar mijn vader
heeft zeker vergelijkbare symptomen. Daarnaast heeft hij al tijden last van
huiduitslag op de armen, wat sinds de dosisverhoging is uitgebreid naar de
romp volgens mijn vader. Mogelijk is dus sprake van een overgevoeligheid.
Gezien de belangrijke waarde van ACE-remmers bij hartfalen is het moeilijk
deze te stoppen om na te gaan of sprake is van overgevoeligheid. Wellicht zou
toch een andere ACE-remmer (enalapril) of een angiotensine-II-antagonist,
zoals losartan, geprobeerd kunnen worden, uiteraard in overleg met de
cardioloog.
Furosemide
Ik denk dat het heel onverstandig is van mijn vader maar 1 tablet furosemide
40 mg te slikken in plaats van twee. Ik vraag mij zelfs af of 2x daags 40 mg
voldoende is gezien de mate van oedeem. Na overleg met een collegaapotheker ML Bouvy, die gepromoveerd is op het gebied van hartfalen, zijn er
wellicht nog de volgende opties:
t een injectie met furosemide, omdat bij zeer veel oedeem furosemide oraal
vaak minder goed wordt opgenomen door de darmwand;
207
t tijdelijk hydrochloorthiazide erbij (daarbij moeten natrium en kalium

spiegels gecontroleerd worden);
t een ander lisdiureticum, zoals bumetanide. Bumetanide 1 mg komt overeen
met furosemide 40 mg. Bumetanide kan zo nodig in dosering opgevoerd
worden tot bijvoorbeeld 5 mg.
Mijn collega-apotheker Bouvy zei ook nog dat de moeilijke spraak een gevolg
kan zijn van lage natriumspiegels. Ook is een mogelijkheid dat hij een (lichte)
TIA heeft gehad.
Het is geenszins mijn bedoeling mij te bemoeien met uw voorschrijfbeleid. Als
apotheker probeer ik ideen aan te dragen, waardoor de conditie van mijn
vader mogelijk kan verbeteren. Ik hoop, dat u iets met deze informatie kunt
en dat deze in overeenstemming is met uw eigen perceptie.
Met vriendelijke groet,

HF Kwint, apotheker
Apotheek Stevenshof, Leiden
208
Aff iliations dur ing the conduc tance of the res earch
Ly nette B ern ing h am

Department of Public Health and Primary Care, Leiden University Medical Center,
Leiden, The Netherlands.
Marcel L. B ouv y
Division of Pharmacoepidemiology & Clinical Pharmacology, Utrecht Institute for
Pharmaceutical Sciences (UIPS), Faculty of Science, Utrecht University, Utrecht,
The Netherlands.
SIR Institute for Pharmacy Practice and Policy, Leiden, The Netherlands.
Wendy P. J. d en E l z en
Adri ann e Fab er
SIR Institute for Pharmacy Practice and Policy, Leiden, The Netherlands.
Ja c obij n Gu ss ek l o o
Gl enn Stol k
The Netherlands.
209
S anne Verd o or n
The Netherlands.
210
Internat i on a l publ i c ati ons pres ente d i n t h i s t h e si s

Kwint HF, Faber A, Gussekloo J, Bouvy ML.
Effects of medication review on drug-related problems in patients using automated
drug-dispensing systems: a pragmatic randomized controlled study.
Drugs Aging 2011;28:305-14.
Kwint HF, Faber A, Gussekloo J, Bouvy ML.
The contribution of patient interviews to the identification of drug-related
problems in home medication review.
J Clin Pharm Ther 2012;37:674-80.
Kwint HF, Bermingham L, Faber A, Gussekloo J, Bouvy ML.
The relationship between the extent of collaboration of general practitioners and
pharmacists and the implementation of recommendations arising from medication
review: a systematic review.
Drugs Aging 2013;30:91-102.
O t her publi c ati ons rel ate d to t h e t h e si s ( i n dutch )

Kwint HF, Luchtman T, Krijger-Dijkema JM, Eekhof J, De Kanter J.
Arts en apotheker houden samen toezicht. Controle van herhaalreceptuur op
geneesmiddelgerelateerde problemen.
Pharm Weekbl 2003;138(49):1732-7.
van Haarlem N.
Apotheker Lloyds komt aan huis.
Pharm Weekbl 2006;141(24):808-1.
211
Bos E.
Trek open die medicijnkast.
Pharm Weekbl 2006;141(50):1588-91.
Bos E.
Wordt het farmacotherapeutische thuisconsult succesvol? Alle medicatie op tafel
graag.
Pharm Weekbl 2009;144(27):24-25.
van Wijck F.
Populariteit Baxterrol vraagt om goede leidraad.
Pharm Weekbl 2011;146(21).
de Graaf L.
FPZ moet zich nestelen in de wetenschap. Meer promoties nodig van openbaar
apothekers.
Pharm Weekbl 2012;147(23).
Faber A.
Medicatiereview aanbevolen bij geautomatiseerd distributiesysteem.
PW Wetenschappelijk Platform. 2012;6:e1201.
Faber A.
Bijdrage farmacotherapeutische anamnese aan medicatiebeoordeling.
PW Wetenschappelijk Platform 2013;7:e1301.
212
Henk-Frans Kwint was born on 17 January

1970 in Meppel, the Netherlands. He is
married and has two sons. He studied
pharmacy in the city of Groningen since
1988 where he obtained his pharmacy degree
in 1996. Subsequently, he started working
for the academic community pharmacy
Stevenshof in Leiden in 1996 and still is. In
2001 he became managing pharmacist in this
pharmacy.
Since 1997 he made his first steps in
pharmacy practice research as a tutor in the
SIR Masterclass organised by SIR Institute for Pharmacy Practice and Policy, which
is affiliated with the academic pharmacy Stevenshof. In 2007, he started his PhD
research on medication review and multidose drug dispensing at the Division of
Pharmacoepidemiology and Clinical Pharmacology of the Utrecht Institute for
Pharmaceutical Sciences (UIPS) of Utrecht University.
He was a member of the Special Interest Group asthma/COPD of the Royal
Dutch Association for the Advancement of Pharmacy (KNMP) from 1998 till
2000. Furthermore, he was a regional pharmacist coordinator for forming local
networks between general practitioners and community pharmacists in the region
of Rijnland-Midden-Holland from 1997 till 2001. In 2010 he became member
of the working group who developed the Dutch multidisciplinary guideline
Polypharmacy in the elderly. He was the principal author of the chapter describing
the clinical medication review process.
213

Kwint PDF

Uploaded by

Copyright:

Available Formats

Kwint PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kwint PDF

Uploaded by

Copyright:

Available Formats

Photography:

Cover design and

Patricia Nauta, Leiden, and Ruth Veneboer, s-Hertogenbosch

CIP-gegevens Koninklijke Bibliotheek, Den Haag

Prof.dr. M.L. Bouvy

The relationship between the extent of collaboration

Multidose drug dispensing

3.1 Medication adherence and knowledge of older

3.2 Effects of medication review on drug-related

Chapter 4 Home Medication Review

4.2 The contribution of patient interviews to the

4.3 Completeness of medication reviews provided by

4.4 Evaluation of medication reviews in older patients

Chapter 5 General discussion

Chapter 6 From Summary to About the author

6.4 List of co-authors

6.5 List of publications

6.6 About the author

medication management) and medication reviews (aimed at appropriateness of the

65 years and 5 drugs

mismanagement (e.g. adherence and practical issues such as difficulties with

interview).22 The development of the Medicines Use Review (MUR) was an

Medicines Reviews (HMRs) in Australia are structured collaborative reviews

Fi g u re 3 The medication review process according to Dutch multidisciplinary

DRP = drug-related problem

misused.25 As explicit criteria for inappropriate prescribing, STOPP (Screening

restricted. It is recommended that the agreed actions are written on an information

In Chapter 3.1 we compared the self-reported medication adherence and knowledge

Stichting Farmaceutische Kengetallen (SFK). Polyfarmacie bij een op de tien

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

as was inadequate clinical training of the pharmacist.13 Without access to medical

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

clinical trial (RCT), pharmacist and GP involved, community-dwelling patients in

Ke y eleme nts of the inte r ve ntion

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

Potentially relevant publications

RCTs = randomized controlled trials

Q u a l ity ass essment of stu d i es

Quality assessment of studies

Not clear Yes

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

Home-based MR. Care

Nurse met pts at home

(A) 141 (0.5)

Description of included studies: effect on clinical, intermediate and process outcomes

MAI ( ); no. of PIMs

Home visit and MR by

Clinic-based MR, then

MRs without pt interview

Description of included studies: effect on clinical, intermediate and process outcomes

No. of daily units of

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

One clinical pharmacist

Clinical MR (one health

Description of included studies: effect on clinical, intermediate and process outcomes

G P/ p h a r m acist co lla b o rat io n in m edic ation re v iew

Sorensen et al.27 (2004)

Sellors et al.26 (2003)

Volume et al.29 (2001)