The information-seeking behaviour of doctors: a review

Blackwell Publishing Ltd

of the evidence
Karen Davies, Department of Information Science, Loughborough University,
Loughborough, Leicestershire, UK

Abstract
This paper provides a narrative review of the available literature from the past
10 years (1996–2006) that focus on the information seeking behaviour of
doctors. The review considers the literature in three sub-themes: Theme 1, the
Information Needs of Doctors includes information need, frequency of doc-
tors’ questions and types of information needs; Theme 2, Information Seeking
by Doctors embraces pattern of information resource use, time spent searching,
barriers to information searching and information searching skills; Theme 3,
Information Sources Utilized by Doctors comprises the number of sources uti-
lized, comparison of information sources consulted, computer usage, ranking
of information resources, printed resource use, personal digital assistant (PDA)
use, electronic database use and the Internet. The review is wide ranging. It would
seem that the traditional methods of face-to-face communication and use of
hard-copy evidence still prevail amongst qualified medical staff in the clinical
setting. The use of new technologies embracing the new digital age in informa-
tion provision may influence this in the future. However, for now, it would seem
that there is still research to be undertaken to uncover the most effective meth-
ods of encouraging clinicians to use the best evidence in everyday practice.

• the use of electronic information resources by

Introduction
doctors;
This narrative review is concerned with doctors’ • where medical librarians can impact on doctors’
need for information, specifically in the clinical information-seeking behaviour.
area. The aim of this paper is to review the literature Since the seminal work of Covell et al.1 on the
on the information needs, literature searching and information needs of doctors there has been research
resources utilized by doctors. This review will in this area using a variety of methodologies from
examine the past 10 years, from 1996 to 2006, observation to interviews and self reports. These
which have seen the development of user-friendly have commonly identified a clinical information
electronic resources, particularly with regard to need, but the frequency of questions raised has
Internet search engines. varied. Linked to this information need is the doc-
The key objectives are to report from the litera- tors’ willingness and ability to personally search
ture (published as journal articles): for the answer. Clinical information needs may have
• the research findings from relevant articles; tight deadlines or may occur outside the hours a
• the clinical information needs of doctors; library is staffed so a degree of end-user searching
should be anticipated. The final element researched
Correspondence: Karen Davies, Department of Information Science, is the resources themselves. To ensure 24/7 access,
FK029, Holywell Park, Garendon Wing, Loughborough University,
Loughborough, Leicestershire LE11 3TU, UK. E-mail: the most viable resources are electronic. However,
K.S.Davies@lboro.ac.uk the range of resources available is huge and the

© 2007 The author

78 Journal compilation © 2007 Health Libraries Group. Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 79

challenge is in resource selection as much as skills The review is not restricted to the UK. The pro-
in using the resource itself. fessional role or job title of the doctor is dictated
The three elements: information need, literature by the country of study. This paper will refer to
searching and resources, are all interlinked and the ‘doctor’, which in this context will include all
should not be considered in isolation. If there is no medical and surgical staff, including general prac-
expressed information need, there is no driving titioners (GPs). Studies that focused on medical
force to undertake a literature search. Similarly, a students or other healthcare staff, qualified or not,
lack of access or lack of search skills negatively were excluded. Studies were included if doctors
affects the use of electronic resources to answer and other healthcare staff were involved, but the
information needs. The actual mechanics and doctors results were reported separately.
skills of literature searching are extensive and will Studies were only included that contained the
not be covered in this article. full descriptive and analytical reporting of the research.
Qualitative studies were excluded. Grey literature
was also not considered in this review, as only
Methodology
research published in journal articles was included.
In identifying studies (specifically journal articles) Table 12–35 provides a comparison of the studies
to be included in this narrative review, the following selected for this review. The actual date for data
databases were searched: ,  (1996 – collection, the country where the research was
2006) and :     undertaken, the research methodology utilized,
. The search strategies used keywords the doctors’ professional role and the number of
and was limited to English language articles. The active participants in the research are detailed.
search strategies were developed for the three separate
themes and limited to the title and abstract:
Information needs of doctors
• Theme 1, the Information Needs of Doctors—
information AND need AND (physician OR
Information need
doctor OR clinician).
• Theme 2, Information Seeking by Doctors— There are various types of need. The most recognizable
(information OR literature) AND (search OR is ‘articulated’ or expressed need, which is an
seek) AND (physician OR doctor OR clinician). actual recognized need that demands an answer.36
• Theme 3, Information Sources Utilized by The need may be acknowledged, but no attempt
Doctors—(resource OR database) AND (phy- can be made to answer it. Unperceived or unexpressed
sician OR doctor OR clinician). needs are not recognized as information needs.37
Additionally, the reference lists of articles were In other words, the doctor is unaware that there is
searched. Research completed and published up a gap in their knowledge.
to 2006 was included in this review and only pub- It is assumed, by librarians and information
lished data was considered. This time restriction specialists, that doctors have information needs
does mean that research prior to 1996 has been associated with clinical work. However, as with
excluded from this review. any profession, the need varies amongst the popu-
The criteria for inclusion was that the focus of lation and is not homogenous. Little work has
the paper was either information needs, literature been carried out to quantify this need in the UK.
searching or resources used by qualified (post- The existing research and associated literature is
registration) doctors. Studies were excluded that heavily USA based. Whilst the information need
looked at patient administration and management, expressed by US clinicians may be the same as UK
such as electronic patient records. Studies were clinicians, the clinical setting is not. The National
also excluded that focused on continuing education, Health Service (NHS) is a unique environment
clinical research trials and teaching. and very different to that of private or insurance-
Studies were considered from both the primary based practice. This is probably more applicable
and secondary care sector, so encompassing an in the different interfaces between primary and
holistic approach to health care. secondary care.

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 80
Table 1 Comparison of studies selected for this review

Information-seeking behaviour, Karen Davies

Date of data Number of active
Study Country collection Research method Doctor’s professional role particpants

Ali, 20022 Australia 10/94–03/95 Questionnaire Hospital specialists and registrars 98

Alper et al., 20053 USA 01/04–04/04 Clinicians searched for answers Primary care clinicians 52
Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94

to clinical questions
Andrews et al., 20054 USA Not stated Questionnaire Primary care practitioners 59
Arroll et al., 20025 New Zealand 11/99–02/00 Pre-questionnaire and direct Family physicians 50
observation
Barrie and Ward, 19976 Australia Not stated Doctors recorded their own General practitioners 27
questions. Followed up with
a semistructure interview
Bennett et al., 20047 USA 12/02–01/03 Questionnaire (faxed) Physicians 3347
Bergus and Emerson, 20058 USA 5-year study Analysed clinical questions Family medicine residents 49
(dates not specified)
Bryant, 20049 UK 01/99–04/99 In-depth interviews and General practitioners 58
discussion groups
Chambliss and Conley, 199610 USA 02/95–07/95 Post-clinical session Family physicians 9
interviews to gather
unanswered questions
Cheng, 200411 Hong Kong 1999 Questionnaire, interviews and Hospital clinicians 50 interviews,
randomized control study 254 questionnaires,
800 randomized
control studies
Cogdill et al., 200012 USA 1996 Interviews Community-based primary 15
care physicians
Crowley et al., 200313 USA 07/00–04/01 Computer system collected Internal medicine residents Not stated
clinical questions
Cullen, 200214 New Zealand 01/01–08/01 Questionnaire, followed General practitioners 294 questionnaires,
up with interviews 20 interviews
D’Alessandro et al., 200415 USA 11/99–05/00 Control group V intervention General paediatric 52
and 12/00–05/01 (received 10 min training). residents (hospital)
Telephone interview
using critical incident technique
Ebell and White, 200316 USA Not stated Exit interviews and physicians Physicians 27 exit interviews,
© 2007 The author

self-reporting cards (from five specialisms) 35 self-report

Ely et al., 200517 USA Not stated Observation, interviews between Generalist physicians 48
patient visits and general interview
Ely et al., 199918 USA 04/96–12/97 Observation—recorded Family doctors 103
clinician’s questions
Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
© 2007 The author

Table 1 Continued

Date of data Number of active

Study Country collection Research method Doctor’s professional role particpants

Green et al., 200019 USA Not stated Interviewed after each patient visit Primary care internal 64
to identify unanswered questions medicine residents
Jerome et al., 200120 USA 04/97–12/97 Librarians entered clinical Hospital healthcare Not stated
questions on to a database staff and patients
Kim et al., 200521 USA Not stated Questionnaire and Office based general 20
semistructured interviews paediatricians
Lappa, 200522 Greece Not stated Questionnaire Emergency-care clinicians 35
Magrabi et al., 200523 Australia 10/02–11/02 Clinicians used online evidence General practitioners 227
system to answer their questions
McAlister et al., 199924 Canada 06/97–10/97 Questionnaire General internists 294
Oliveri et al., 200425 Denmark 07/00 Questionnaire Hospital doctors 225
Ram and Wellington, 200226 UK 12/01–03/02 Questionnaire General practitioners 86
Ramos et al., 200327 USA Not stated Direct observation and self reporting Faculty and residents 38
Sackett and Strauss, 199828 UK 04/97–05/97 Descriptive feasibility study of Medical doctors and students 17
use of evidence and questionnaire
Schilling et al., 200529 USA 10/01–06/02 Attempted to answer questions and Internal medicine residents 43
then completed a questionnaire
Schwartz et al., 200330 USA 04/01–07/01 Physicians recorded their Clinical faculty in an 3

Information-seeking behaviour, Karen Davies

questions and how they urban residency
attempted to answer them
Scott et al., 200031 Australia 10/98–01/99 Questionnaire on EBM and Consultant physicians 111
literature searching in adult medicine
Swinglehurst et al., 200132 UK Not stated A clinical informationist General practitioners and 34
answered questions nurse practitioners
Veness et al., 200333 Australia/ 12/00 Questionnaire to determine Radiation oncologists 191
New Zealand views about EBM and use and registrars
of EBM resources
Wilson et al., 200334 UK 07/01–10/01 Questionnaire on NHS Net and General practice staff 670
resources (paper and electronic) (managers, nurses and doctors)
Wong and Veness, 200535 Australia/ 2003 Questionnaire to determine Radiation oncology registrars 69
New Zealand/ views about EBM and use of
Singapore EBM resources

81
82 Information-seeking behaviour, Karen Davies

Table 2 Types of information required by doctor’s professional role

Drug therapy/ Treatment/

Diagnosis information Epidemiology therapy

General practitioners23 40% 35%

General practitioners32 23% 60%
Primary care physicians3 22% 49%
Primary care physicians12 —no student present 23% 29% 19%
Primary care physicians12 —student present 33% 20% 18%
Family physicians5 33% 39%
Family physicians10 29% 15% 37%
Family medicine residents8 43% 36%
Primary care internal medicine residents19 27% 38%
Internal medicine residents13 22% 53%
Internal medicine residents29 19% 43%
Hospital staff20 53%
Hospital clinicians11 46%*
Hospital specialists and registrars2 10% 27% 58%
Physicians (from five specialisms)16 —self report 23% 50%
Physicians (from five specialisms)16 — 29% 34% 14%
post-patient visit Interview
Clinical faculty30 13% 50%

*Includes drug therapy.

To fully comprehend the information-seeking the course of their careers. Therefore, doctors need
behaviour of doctors, it is important to under- to be able to refresh their memories and update
stand their information needs. Information needs their knowledge efficiently and effectively.
are the driving force behind literature searching.
Doctors have diverse information needs. Most
Types of information need
of the studies reviewed in this article focus on
clinical information needs rather than collecting Table 2 illustrates the main types of clinical
data on full range of clinician’s information needs. information required by medical staff from
The clinical information need is relatively easily selected research, organized by doctor’s professional
categorized and covers the main role of qualified role. There are limitations to this table. Responses
medical staff, namely treating the patient. that are not comparable with other studies have
An issue that can negatively affect information not been included and only those reported in three
need is information overload. Doctors can start with or more studies have been included. Responses
too little information and quickly become inundated, under 10% have also not been included. Therefore,
often with irrelevant or unreliable information. not all the responses are shown for each study.
Advances in information dissemination, such as Comparison of the results from these studies is
electronic newsletters, push technologies (informa- hindered by the different definition of terms, varied
tion sent to the recipient without it being requested) subjects, range of settings and diverse methods of
and RSS feeds have increased the amount of infor- data collection. However, basic analysis suggests that
mation received by medical professionals. In fact, the top categories of information need treatment
the hardest task now is to actually locate the infor- or therapy (average 38%), diagnosis (average 24%)
mation required from the flood of information and drug therapy/information (average 11%).
received. This is a particular issue in the medical field, The fifteen research studies reviewed used more
where the biomedical research information doubles than forty different definitions to analyse their
every 20 years.38 It would be virtually impossible data. However, over 90% of the studies used
for medical staff to retain all the knowledge ‘treatment’ or ‘therapy’ and more than 80% utilized
required to treat all the patients they examine over ‘diagnosis’.

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 83

Table 3 Comparison on background and foreground questions

Background questions Foreground questions

Cheng11 33% consultants 67% consultants

57% senior medical officers 43% senior medical officers
45% medical officers 55% medical officers
Green et al.19 28% 66%

Two research articles analysed the questions

posed according to whether they were foreground
(directly related to the patient, such as diagnosis,
treatment and prognosis) or background (general
information on a condition or disease) in nature.11,19
The foreground question usually contains at least
three out of the four PICO39 elements: patient or
problem; intervention; comparison with inter-
vention; and outcome. Table 3 compares the back-
ground and foreground questions posed by doctors
in research by Cheng11 and Green et al.19 The
more senior the medical professional the more
likely the question will be foreground. This is Figure 1 The average number of questions asked per patient visit
understandable, as their background knowledge
would have developed through experience. of research into clinical information needs of doc-
One research study divided the questions raised tors would not allow this information to be moni-
into three specific areas. Medical fact, such as the tored in the long term to determine if it ultimately
reported side-effects of a prescribed drug, accounted was utilized for patient care.
for 50%5 of the questions. Medical opinion, for
example patient management, answered 30%.5 The
Frequency of information need
remainder, 20%,5 related to non-medical informa-
tion. This illustrated that the majority of the ques- The literature was considered to identify the
tions (80%) were related to patient care. Two research average number of questions asked by clinicians in
articles identified that 98%40 and 81%15 of ques- the clinical setting. Figure 1 illustrates the range of
tions related to patient care. The ‘evidence cart’ questions per patient visit from selected research.
survey in the UK found that 81% of the evidence The research showed that the frequency of doc-
sought affected diagnostic and/or treatment deci- tors information needs varied, ranging from 0.16
sions.28 Clinicians in New South Wales, Australia, to 1.27 questions per patient. The range may be
used Clinical Information Access Program (CIAP) because of the different medical specialisms involved
75% of the time for clinical reasons, such as con- or to the different research methodologies utilized in
firming a diagnosis, developing a treatment plan the studies, which does hinder direct comparison of
or selecting a diagnostic test.41 The purpose of the results. From the data available it is not possible
searches by doctors in a study (after decentralized to determine definitive reasons for the varied results.
 was introduced) was only for patient care Observation was the chosen methodology for
10% of the time, research 30% and for both (research one research study that resulted in higher than
and patient care) 53% of the time.42 Unfortunately average questions—1.27 per patient visit.27 How-
research studies are unlikely to take into account ever, two other observational studies had results of
information obtained for other reasons (such as 0.45 questions per patient18 and 0.29 questions.5
education or general research) that may ultimately The different results between the two groups may
be utilized in patient care management. The timescale be because of the clinical setting, as the work by

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
84 Information-seeking behaviour, Karen Davies

Ramos et al.27 was in a university based clinic, profession may in fact produce similar results. The
which are usually teaching settings. USA-based studies are often conducted within
Even within the same environment, when the office practices, but these are not necessarily
circumstances altered, so did the number of ques- general practitioners as known in the UK. Moreover,
tions generated. The doctors asked fewer ques- a wide variety of specialists work in these environ-
tions when students were present—0.29 per patient ments, such as cardiologists, gastroenterologists,
compared with 0.42 when alone.12 endocrinologists, pulmonary disease specialists
The doctors’ perceived need for information and infectious disease specialists.1 There has also
may not reflect their actual need. In one study the been research conducted outside the USA in
perceived need was lower than their actual need. hospital based environments.2,11,22,25,28,31
Interviews after the patient visit identified more
questions than self reporting by doctors.16 How-
Information seeking by doctors
ever, Arroll et al.5 found the reverse to be the case
as doctors overestimated the number of questions Whether or not information needs are pursued
by threefold. depends on the individual’s ability to identify their
The average number of questions raised per need and then to express the need in terms that are
patient visit from the research articles reviewed in searchable by themselves or a third party.36
this article is 0.47. A study by Ely et al.17 counted There is a distinct difference between looking up
the number of questions raised by doctors in a a fact and literature searching. For example,
half-day observation period (4 h). On average checking the dosage for a prescribing drug is an
doctors asked 5.5 questions per half-day. It was information need of the doctor and can easily be
noticeable that the number of patients seen per located in the British National Formulary (BNF)
session was not quantified in the article, but this so does not require an extensive search of the liter-
figure would appear to be much higher that the ature. This is often not differentiated in research
other studies reviewed in this review. on doctors information needs.
Gorman et al.43 undertook a study to determine
if the information seeking behaviour of primary
Patterns of information resource use according to
care medical staff in rural areas were different from
purpose of search
those in non-rural areas. The results showed there
were no statistically significant differences in terms Evaluating the complexity of questions asked by
of the number of questions asked, the number of doctors suggests that 60% of the questions are simple
questions pursued and the number of questions (that is, one concept) such as the dose of a drug.5
answered. This reinforced the earlier work by This type of question doctors should be comfortable
Dorsch44 who reached a similar conclusion. answering with a minimal amount of training on
Despite the variance of number of questions computers and resource selection. Therefore, to
raised, these research articles prove there is an answer these simple concept questions, user educa-
information need generated by patient visits. tion should focus on resources that can be used
effectively with minimal training.46 In the UK this
could include training on how to access and search
Differences according to type of healthcare
resources such as the BNF and the Cochrane Library.
setting
A study focused on whether clinical questions
Thirteen of the studies focus on the primary care could be answered by evidence-based medicine
sector. Six of these studies are from New Zealand, (EBM) resources.47 Librarians searched ,
Australia and the UK.5,6,9,14,23,32 It is acknowledged the     
that these studies may not be relevant for other and HealthGate Clinical Guidelines to answer
professional groups of doctors. However, it should both complex and general care management ques-
not be assumed that any professional group is tions. The results did not reflect well on the EBM
cohesive enough to be a distinct unique group.45 resources. Only 20% of the complex clinical ques-
Therefore, research in other specialisms in the tions were answered, and just under half of the

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 85

general care management queries using the EBM

Time required to search effectively
resources. No answers were found for 40% of the
complex clinical questions and 30% of the general Some information is particularly time sensitive,
care management. However, the primary literature especially if it influences patient care. Research
resources answered 95% of the complex clinical indicates there are two speeds in which information
questions and 85% of the general care management. is required. The first is immediate; the ‘bottom
line’ required when dealing with a patient, followed
by the less-immediate morning rounds, case
Time spent searching
consultations and discussions with colleagues.50
The key barrier to information searching is time. In a study by Barrie and Ward,6 61% of the
In one study, 60% felt time was an issue.19 Cogdill questions were answered during the consultation.
et al.12 found that 40% of questions were not pursued Sackett and Strauss28 found that 90% of the
because of a lack of time. It is difficult to estimate searches on patient management were successful
the time that an information search will take. in finding the information in the time available on
Searching for information 15 years ago involved rounds. A 2003 study reported answers were sought
trawling through Index Medicus and browsing the to 74% of the questions immediately, finding
journal shelves. Since the introduction of electronic satisfactory answers to 87% of them.27 In another
databases, the time spent searching for the evidence study, of the questions pursued, 25% were on the
has declined dramatically, as users do not have to same day that the question was asked.19 This
travel to the library to look at physical books and figure is much lower than the other studies.
journals, but can access electronic resources remotely.48 The success of the searches conducted by
However, it is important to combine a suitable doctors is important as this really determines if
search strategy with the identification and searching time and effort has been successfully utilized. Resi-
of relevant information resources. dents managed to answer 89% of their questions
In a study by Krahn et al.,49 an informaticist (in in a study by Schilling et al.29 and, in another study,
this case a specialist who searches and critically doctors answered 80% of the questions pursued.18
appraises literature) spent an average of 53 min to Cheng11 found that 73% of the patient-care problems
answer surgery questions. Two studies had lower were solved. A survey by Ely et al.17 found that, of
figures, with an average of 14.4 min15 and a mean the 55% of questions pursued, 72% were actually
of 15 min.19 However, when considering resources answered. One issue of concern is that, when infor-
used at the point of care, the median search time mation is not located, it is then assumed not to
was lower, at between 5 and 10 min.30 The amount exist. However, it may actually exist, but the search
of time taken to answer these questions is longer strategy utilized was ineffective in retrieving the
than the time usually devoted to a single patient answer. This can leave the doctor with a mistaken
appointment. In the NHS, GP’s patient appoint- view of the contents and coverage of the database.
ments are generally acknowledged to be between 7 One study considered the efficiency of the searches.
and 10 min long. This type of information search- These were regarded as very good or good 18% of
ing is not practical in the clinical setting with the the time, fair 52% and poor 30% of the time.31
patient present. In a survey of residents, 66% of The suggestion that clinicians only have time to
their searches took only 2 min, whilst the maxi- complete ‘quick and dirty’ searches because of time
mum for a search was 15 min.27 The study of Ely constraints is worrying. This comment arises because
et al.18 study found that doctors spent on average speed is often the most important aspect in search-
less than 2 min pursuing an answer. The research ing. This is where the filtered evidence resources
utilizing an ‘evidence cart’ found that resources come to the fore. The Cochrane Library provides
had to be accessible in between 10.2 and 25.4 s the best evidence (systematic reviews) in a usable
for them to be practical at the bedside.28 In and understandable format. Whereas 
practice, if the search takes more than 2 min it will contains information that has not been evaluated.51
not produce information suitable for that patient The evaluation of the ‘raw’ information from 
consultation. can be time consuming for busy clinicians.

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
86 Information-seeking behaviour, Karen Davies

Table 4 Barriers to literature searching

Issues with IT or Limited Lack of interest

Study Lack of time online resources search skills Cost or urgency

Scott et al.31 74% 43% 41%

Green et al.19 60% 7%
Andrews et al.4 76% 22% 25% 33%
Ely et al.17 19% 8% 15%
Lappa22 74% 80%+

found that question formulation did not improve

Perceived competence and confidence
as they progressed in their clinical training.8 This
Holtum47 suggests that librarians have downplayed lack of skill is compounded when the doctors
their skills and not highlighted the complexity of believe the search failed because of a lack of relevant
literature searching, giving health professionals a evidence, rather than realizing the problem was the
false perception of their skill levels. errors in their search strategy.53 Retrieving incomplete
information and the failure to locate any useful
information may discourage doctors from searching
Barriers to information searching
for evidence again. Ely et al.17 found that the two
Table 4 compares the various barriers to literature key reasons why doctors failed to pursue answers
searching categorized in research studies. Only was the doubt that an answer existed (25%) and
responses that occurred in two or more studies referring a patient rather than undertaking a
have been included and responses below 5% have literature search (22%).
been excluded. A lack of training or issues with the technology
The most popular response in the studies iden- is often raised by doctors. The lack of training was
tified issues with the ‘online resources’ or ‘infor- the most common barrier to using the Internet
mation technology (IT)’ in general. Two-thirds of (74%) and databases (62%) according to a UK
the studies mentioned a ‘lack of time’ and half the study.54 However, in a study by D’Alessandro
‘limited search skills’ of the doctors. et al.55 on accessing the digital health sciences
Responses that occurred only once included: library, only 34% of the participants had taken up
• 66% lacked basic IT skills;22 the opportunity for on-site training. A study by
• 49% mentioned irrelevant material;31 Erickson and Warner56 indicated that 60% of the
• 25% felt that the answer did not exist or would residents questioned would be interested in
not affect care.17 improving their  skills, but none had taken
advantage of existing programmes offered by the
library staff. However, a study of end users found
Information searching skills
that, whilst education was important, practice was
There is so much information available that poor the biggest reason behind improvement in the quality
searching or navigating difficulties will hinder the of searching.57 Conversely, literature searching skills
retrieval of the required evidence.7 Information decline if not used regularly.58 A critical mass of
professionals realize that converting the clinical searches is required to maintain searcher compet-
question to a searchable strategy for an information ence. A survey of Australian GPs found that they
resource is a crucial skill, but is difficult to master. conducted on average 8.7 searches per month.23
Inexperienced searchers have been found to use Research on the use of CIAP in Australia found
inappropriate search terms, wrong or inappropriate that 60% of users sought evidence more than twice
databases, incorrectly spelt terms, inappropriate a month and 30% at least once a week.41 In a sur-
connectors, and brand drug names rather than vey of GPs, 45% expected to use an information
generic.52 A study of resident doctors disappointingly source to find the latest evidence a minimum of

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 87

once a fortnight.59 Therefore, it is possible for The first is if the doctor considers the patient’s
doctors to maintain their information skills at a problem to be urgent. The second is the doctor’s
reasonable level through regular searching. belief that an answer exists and it can be located to
To be effective literature searchers, doctors must assist in the clinical decision-making process.60
possess more than basic search skills. One study A study by D’Alessandro et al.15 found that on
identified the  search strategies that end- average 1.2 resources were utilized to answer one
users utilized, namely keywords (94%), applied question. Potter and Rotert61 suggests that end-
‘limits’ (46%), combined terms with ‘AND’ (39%) users choose one or two electronic resources and
and medical subject headings (MeSH) 11% of the take the time to learn how to search them quickly
time.14 Decentralizing  increased the and effectively.
number of searches and the doctor’s knowledge of
the database.42 The search strategies included ‘by
Comparison of information sources consulted
author’ (40%) and using ‘MeSH’ (60%). Concepts
such as explosion (55%), and focus (55%) were Table 5 compares the information sources utilized
also utilized. These results represent an encourag- by doctors in percentages from selected research.
ingly high use of specialist search skills. Only responses that are comparable with other
studies have been included.
The two information sources used most fre-
Information sources utilized by doctors
quently (averaged from the studies specified) are
There are a vast range of resources that can be text books (39%), followed by ‘humans’ (25%).
utilized by doctors, from journal articles to Computers were used on average by 13%, but the
government reports to original research. In fact, in highest percentage use, 53%, occurred in the latest
a time pressurized environment, such as on the published research in 2005.29
patient ward, it would seem probable that there are The heading of ‘humans’ covers all those a
almost too many potential information sources for doctor may contact. The participants of the Green
doctors to locate information efficiently. Technology et al. study19 consulted with attending doctors
develops rapidly, so this section focuses on research (17%), fellow residents (5%) and speciality con-
published from 2000 onwards. sultants (3%).

Number of sources utilized Computer usage

There are two main reasons why a question is In the Ely et al. study of 1992,62 only 2% of the
pursued, especially beyond one information source. family doctors had access to personal computers

Table 5 Information sources utilized by doctors

Study Textbooks Journal articles Humans Computer search Drug sources

Barrie and Ward6 14% 40% 43%

McAlister et al.24 45% 73% 61%
Ely et al.18 36%* 36% 25%
Cogdill et al.12 69% 27%
Green et al.19 31% 30% 25% 5%
Ramos et al.27—Resident 50% 49%
Ramos et al.27—Faculty 70% 24%
D’Alessandro et al.15—Control 48% 27% 15%
D’Alessandro et al.15—Intervention 36% 30% 21%
Ely et al.17 39% 6% 18% 16%
Schilling et al.29 5% 33% 1% 53%

*Includes all printed information—textbooks, journals and information on posters.

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
88 Information-seeking behaviour, Karen Davies

Table 6 Ranking of information sources utilized by doctors

Study Text sources Humans Personal library Electronic resources Internet

Arroll et al.5 1st 2nd 3rd

Cullen et al.14 1st 2nd 3rd 4th
Bryant et al.9 3rd 1st 2nd
Oliveri et al.25 1st 2nd 3rd 3rd
Bennet et al.7 1st 3rd
Kim et al.21 2nd 1st 3rd 3rd

in their offices, so their lack of computer usage is practitioners by Andrews et al.,4 print and inter-
not surprising. By the next study of Ely et al.18 personal resources were more likely to be used than
(using data from 1996 to 1997), 13% used online sources. Printed drug resources were used
computers. This figure had risen to 16% by the several times a week or daily by 61% of respondents,
study published in 2005.17 and medical textbooks by 58%. A study of British
GPs highlighted their preference for referring to
paper publications rather than the electronic version
Ranking of information resources
of that resource.34 The resources utilized most
Table 6 ranks the information sources utilized by often were the British Medical Journal (93%), Drug
doctors in order of their importance. Only those and Therapeutics Bulletin (92%), National Institute
sources mentioned by two or more studies are for Clinical Excellence (NICE) guidance (82%),
included. Bandolier (69%) and Clinical Evidence (65%).
Disappointingly, but not surprisingly, two-thirds
of the studies ranked text sources first. In most of
Personal digital assistant (PDA) use
the studies, the category of ‘humans’ is the second
choice. Clinicians often find it quicker to ask a The ‘evidence cart’ experiment discovered that
colleague for advice.63 However, a study of clinical house staff liked having access to high quality
communications in a Bristol hospital highlighted evidence, but were more concerned with being able
that many of the phone calls were in fact unsuccess- to access the information within 30 s.28 The
ful attempts to locate the correct person.64 The downside with the evidence cart was its bulk, but
success rate of call events (telephone and paging) new technologies such as PDAs are more user-
was 74%. However, one-quarter of all call events friendly in the clinical setting.65 There are still
were attempts to identify the name of an individual barriers to hand-helds, which include their reliability,
with a specific role. Medical colleagues accounted size, security, limited memory and battery life.66
for 42% of the calls medical staff received. However, a study in a clinical setting in the UK
found that hand-held technology is an effective
tool to aid evidence-based medicine, as the devices
Printed resource use
held sufficient information to be relevant, accessible
Ramos et al.27 discovered that US residents were and effective at the point of care.67
more likely to use a pocket reference or consult A British study found that doctors preferred
another person (usually an attending or specialty UK resources to US ones. The BNF was on the
consultant) to answer their questions. The faculty PDAs in Phase II of a project. This was utilised by
were more likely to consult texts (pocket reference more users compared with the US drug informa-
and comprehensive texts). These results are similar tion source from Phase I.67
to the study by Arroll et al.5 where the doctors’
sources for most answers were books, followed by
Electronic database use
colleagues. Only 5% of questions were answered
using a computer, and most of these were simple, Three research studies did show substantial
not complex, questions. In a study of family computer usage. In the study of Schilling et al.,29

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 89

Table 7 Respondents answers to specific questions over the life of the study

Immediate Three months

Questions Pre-intervention post-intervention post-intervention

Sometimes or often searched MEDLINE 13% 36% 65%

to assist in clinical decision making
Often or very often incorporated information 13% 15% 30%
from a search into their clinical decision making

the electronic sources utilized were  (44%), occurred during this study. One of the major bar-
 (42%), Internet search engines (5%), riers when searching the primary literature sources
 (3%) and the Cochrane Library such as  is that the questions generated by
(2%). The Cullen study14 found doctors accessed medical clinicians rarely match those answered by
 (70%), clinical guidelines websites (47%) health researchers.69
and Cochrane Library (38%). Another study by
Scott et al.31 found the evidence source used most
The Internet
frequently was  (76%), followed by Cochrane
Library (11%), and then Best Evidence (7%). The Internet as an information resource for
In a survey by Ramos et al.,27 EBM resources doctors has been evaluated by several studies. A
such as Patient-Oriented Evidence that Matters postal survey of British GPs found that 82% used
(POEMs), Clinical Evidence and the Cochrane the Internet, but fewer (53%) used electronic
Library were rarely used by the residents. In biomedical databases.54 A New Zealand study
London, a study was undertaken to determine the use reported that 48.6% of family practitioners used
of the Cochrane Library by GPs.26 Only 33% of the Internet for clinical information.14 The
the GPs had used the Cochrane Library and most respondents in this study were confused about the
(85%) of these had only used it occasionally. A difference between search engines and databases.
disappointing 52% had never heard of the Cochrane  and PubMed were listed as ‘other search
Library. A later survey of UK general practice staff engines’ by 6% of respondents. Several medical sites
found that 11% were unaware of the Cochrane such as the British Medical Journal and Cochrane
Library.34 However, awareness and use of Cochrane Library were also listed as search engines.
Library is increasing. A postal survey in 2000 of The Internet is a useful tool to disseminate med-
Australian and New Zealand radiation oncologists ical research evidence and guidance. Government
and registrars determined their awareness and use departments, such as the Department of Health
of electronic resources.33 These respondents were and organizations (for example, NICE and the
reasonably aware of the Cochrane Library, with Royal Colleges) use the Internet to enable access to
18% unaware of the resource. However, the figure full-text documents and guidelines. Electronic
for not knowing about     databases and other resources are accessible via
   () was extremely the National Library for Health (http://www.
high (61%), especially as this has been part of the library.nhs.uk/Default.aspx?ref=at), either avail-
Cochrane Collaboration since 1996. A survey of just able to all for free or to NHS staff via Athens
the radiation oncologist registrars was repeated in authentication passwords.
2003. Ninety-four per cent were aware of the Cochr- Unfamiliarity with the computerized resources
ane Library and 49% had used it. This is a notable means searching would take even longer, as a suit-
increase from the 2000 data.35 Awareness of  had able resource has to be located first.5 The Informa-
risen from 19 to 50% and usage from none to 15%. tion for Health Strategy70 outlined the importance
In a study by Byrnes et al.68 participants were of developing ‘desktop’ facilities for clinicians.
taught how to use  (and access it via The National Electronic Library for Health, now
PubMed). Table 7 illustrates the changes that the National Library for Health (http://www.

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
90 Information-seeking behaviour, Karen Davies

library.nhs.uk/Default.aspx?ref=at) links to all the doctor. Basic analysis suggests that the top three
resources clinicians require to practice medicine categories are treatment or therapy, followed by
and contains best-practice guidance (National diagnosis and then, a distant third, by drug therapy/
Service Frameworks and the Guidelines Finder). information. When teaching information skills to
doctors, librarians can draw attention to the different
resources and different search skills. This will enable
Discussion
the doctors to limit the search to specified areas.
This review focuses on the clinical information The research articles that considered foreground
needs of doctors. Librarians can provide doctors and background questions11,19 highlight a distinc-
with the skills to answer their ‘expressed’ or identified tion between two types of information need. Fore-
needs.26 Needs that remain unexpressed or are not ground focus on the patient and background on
identified by doctors cannot be answered. the condition or disease. Background questions
In the available literature, the studies that have are usually a simple ‘look-up’ retrieval of informa-
been undertaken to determine the information tion. Foreground questions are more complex and
needs of doctors utilized many different research this is where teaching doctors the four elements of
methodologies. Each methodology utilized has PICO59 is highly beneficial.
potential limitations associated with it that can Figure 1 illustrates the frequency with which
impact on the results. doctors asked questions derived from patient care.
This narrative review cannot compare like with This ranged from 0.16 to 1.27 questions per patient.
like as the studies employ such a variety of method- This is a substantial range and may be as a result
ologies. No systematic research has been under- of the different medical specialisms involved or the
taken or reported in the literature. different research methodologies utilized. All of
Observation creates an unnatural environment the research published has identified clinical infor-
and the observed may react differently to their mation needs. Therefore, it is important that all
normal working mode. Interviewing doctors may doctors are encouraged and given opportunities to
actually encourage questions as it allows them to develop basic information searching skills, or
revisit the patient encounter and reflect back, thus alternatively are signposted to those who can find
distortion of the data collected may occur. A ques- the information for them.
tionnaire is also problematic as it relies on the A study by Arroll et al.5 identified that 60% of the
recall of the respondent and their interpretation of questions asked by doctors are simple one-concept
the question posed. Self-reporting requires doc- queries. This suggests that raising the awareness of
tors to note responses during a busy clinical ses- doctors to the different resources available to them
sion, which may not be 100% complete because of (and providing basic instruction on their use) should
time constraints and other distractions. The most enable them to answer these simple questions.
effective study combines more than one method. It is difficult to estimate how long a specific
Covell et al.1 used both doctor self-reported ques- information search will take. However, ensuring
tionnaires and interviews (post-patient and post- that doctors have the necessary skills to effectively
clinic). Cheng11 utilized posted questionnaires, search should minimize the time spent locating rel-
interviews and a randomized controlled study. evant information. This is particularly important
Another problem with these studies is that most as, in certain time sensitive situations, doctors do
involved small sample sizes and therefore general- not have time to contact the library (or the ques-
ization to a larger population is problematic. The tion occurs outside library opening hours) and so
study of Ely et al.18 is the only one involving more must undertake the search for evidence themselves.
than a hundred doctors, and so can be acknow- According to a UK study,34 lack of training was
ledged as statistically robust. The studies of Ely the most common barrier to using the Internet
et al.18 and Ebell and White16 are the only two that and electronic databases. An American study56
counted more than a thousand patient encounters. indicated that over half the doctors were interested
Table 2 illustrates the types of information in improving their  skills, but none had
required (organized) by the professional role of the taken advantage of existing education programmes

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 91

offered by the library staff. Promoting the sessions range of electronic resources available, no current
offered by librarians is vital. The Internet and resource was comprehensive enough or efficient
Intranet are useful tools, but require the doctor to enough for the clinicians.71 This is a major issue as
seek out the information. More pro-active meas- time constraints mean clinicians would prefer to
ures can be considered; for example, direct mar- utilize one authoritative source, rather than search
keting, leaflets in staff rooms and displays in several information sources. The high percentage
prominent areas such as the staff canteen. The (53%) of respondents using computers in the latest
onus still remains on the doctors to approach the research of Schilling et al. (2005)29 may be the start
library to actually book and attend information of an increased usage in IT as access and usability
skills training. There is no compulsion to develop issues are being addressed. However, research is
their skills and, until that is the case, it is unlikely minimal so it is too early to suggest that this is an
that all doctors will develop adequate information- upwards trend in IT use by doctors. More research
searching skills. is required in this area.
From the studies reported in Table 6, two-thirds The confusion expressed by doctors about search
of the doctors who responded ranked text sources engines (suggesting the British Medical Journal
first. However, these are not always the optimal and Cochrane Library were search engines)14
sources for evidence-based practice because of the raises concerns. Doctors must be aware of the
likelihood that they may not be up to date. This is validity and integrity of the websites they access.
not an issue if the information required is back- General web search engines will locate sites that do
ground or general information that can be retrieved not provide authentic healthcare information.
from textbooks. The real issue is when searching The doctor should double check and validate all
for information that can change rapidly, such as information from the Web, which can be time-
diagnostic tests and treatments, especially drugs. consuming and is not always straightforward. This
Critically Appraised Topics (CATs) are a more tendency to use Internet search engines is not
evidence-based option, but are only the appraisal limited to doctors, as research among higher educa-
of one article. If this is not the most suitable article tion students (JUBILEE and JUSTEIS) illustrates
available on that particular subject then the value that, from the users perspective, search engines are
of that CAT is minimal. Printed resources were the preferred method for locating information.72
found to be preferable to the same electronic Whilst the Internet is an excellent dissemination
resource by British GPs.34 The doctors ranked tool, doctors need to be taught by librarians how
‘humans’ as their second choice. The ease of use to evaluate websites. This could be a specific course
and interaction with others means that for the on the Internet or added to existing courses deliv-
doctor these are ideal social and informal methods ered by librarians on information skills.
for gaining information. Unfortunately, these The changes in the use of  after doctors
responses may not be based on any reliable evi- were taught how to use it,68 illustrates the importance
dence. This information is more likely to be some- of librarians providing information skills training
one’s opinion, which is actually ranked at the sessions. The use of the Cochrane Library in the
bottom of the evidence hierarchy. This suggests studies of 2000,31 200214 and 200529 is disappoint-
that, in the real world, doctors are not yet embrac- ingly low. In the context of minimizing searching
ing the use of computerized resources. and appraisal time, this is disturbing as the infor-
From the research of Ely et al.62 in 1992 to their mation in the Cochrane Library has already been
200517 article, the increase in use of computers critically appraised. The poor uptake of Cochrane
illustrates an increasing awareness and usage of IT, Library suggests librarians can raise doctors’ aware-
possibly linked with the development of more ness of this resource.
user-friendly database search facilities. It is not
surprising that UK doctors preferred UK drug
Conclusion
information sources to American sources on their
PDAs.67 In the clinical setting, the UK informa- The recognized scholarly knowledge base continues
tion sources would be more relevant. Despite the to be one of the sources used by clinicians. The

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
92 Information-seeking behaviour, Karen Davies

other options of face-to-face communication, hard 7 Bennett, N. L., Casebeer, L. L., Kristofco, R. E. & Strasser,
copy and telephone calls still prevail in the clinical S. M. Physicians’ Internet information-seeking behaviours.
The Journal of Continuing Education in the Health
setting. The adoption and uptake of new technologies
Professions 2004, 24, 31–8.
within the clinical setting may influence this in the 8 Bergus, G. R. & Emerson, M. Family medicine residents
near future. However, for now, it would seem that do not ask better-formulated clinical questions as they
there is still research to be undertaken to uncover advance in their training. Family Medicine 2005, 37, 486–
the best way(s) to encourage clinicians to use the 90.
9 Bryant, S. L. The information needs and information
best evidence in everyday practice.
seeking behaviour of family doctors. Health Information
and Libraries Journal 2004, 21, 84–93.
10 Chambliss, M. L. & Conley, J. Answering clinical
Key Messages
questions. Journal of Family Practice 1996, 43, 140–4.
Implications for Practice 11 Cheng, G. Y. T. A study of clinical questions posed by
hospital clinicians. Bulletin of the Medical Library
• Doctors can develop skills to undertake Association 2004, 92, 445–58.
literature searching. 12 Cogdill, K. W., Friedman, C. P., Jenkins, C. G., Mays, B. E.
& Sharp, M. C. Information needs and information seeking
• Librarians need to highlight pre-appraised
in community medical education. Academic Medicine 2000,
resources such as the Cochrane Library. 75, 484–6.
• Librarians need to provide information 13 Crowley, S. D., Owens, T. A., Schardt, C. M., Wardell, S. I.,
literacy training. Peterson, J., Garrison, S. & Keitz, S. A. A web-based
compendium of clinical questions and medical evidence to
Implications for Policy educate internal medicine residents. Academic Medicine
2003, 90, 370–9.
• Doctors do have a need for information in the
14 Cullen, R. J. In search of evidence: Family practitioners’
clinical setting. use of the Internet for clinical information. Journal of the
• Electronic resources remain an under-utilized Medical Library Association 2002, 90, 370–9.
information source in the NHS. 15 D’Alessandro, D. M., Krieter, C. D. & Peterson, M. W.
• Use of evidence-based medicine needs to be An evaluation of information-seeking behaviours of
general pediatricians. Pediatrics 2004, 113, 64–9.
championed by librarians.
16 Ebell, M. H. & White, L. What is the best way to gather
clinical questions from physicians? Journal of the Medical
References Library Association 2003, 91, 366–8.
17 Ely, J. W., Osheroff, J. A., Chambliss, M. L., Ebell, M. H.
1 Covell, D. G., Uman, G. C. & Manning, P. R. Information & Rosenbaum, M. E. Answering physicians’ clinical
needs in office practice. Are they being met? Annals of questions: obstacles and potential solutions. Journal of the
Internal Medicine 1985, 103, 596–9. American Medical Informatics Association 2005, 12, 217–
2 Ali, I. Library provided information and clinical decision 24.
making: a study of two hospitals in Canberra. Australian 18 Ely, J. W., Osheroff, J. A., Ebell, M. H., Bergus, G. R.,
Academic and Research Libraries 2000, 31, 30–45. Levy, B. T., Chambliss, M. L. & Evans, E. R. Analysis of
3 Alper, B. S., White, D. S. & Bin, G. Physicians answer more questions asked by family doctors regarding patient care.
clinical questions and change clinical decisions more often British Medical Journal 1999, 319, 358–61.
with synthesized evidence: a randomized trial in primary 19 Green, M. L., Ciampi, M. A. & Ellis, P. J. Residents’
care. Annals of Family Medicine 2005, 3, 507–13. medical information needs in clinic: are they being met?
4 Andrews, J. E., Pearce, K. A., Ireson, C. & Love, M. M. American Journal of Medicine 2000, 109, 218–23.
Information-seeking behaviours of practitioners in a 20 Jerome, R. N., Giuse, N. B., Gish, K. W., Sathe, N. A. &
primary care practice-based research network (PBRN). Dietrich, M. S. Information needs of clinical teams:
Journal of the Medical Library Association 2005, 93, 206– analysis of questions received by the clinical informatics
12. consult service. Bulletin of the Medical Library Association
5 Arroll, B., Pandit, S., Kerins, D., Tracey, J. & Kerse, N. Use 2001, 89, 177–85.
of information sources among New Zealand family 21 Kim, G. R., Bartlett, E. I. Jr. & Lehmann, H. P.
physicians with high access to computers. Journal of Family Information resource preferences by general paediatricians
Practice Online 2002, 51. Available from: http:// in office settings: a qualitative study. BMC Medical
www.jfponline.com (accessed 10 February 2006). Informatics and Decision Making 2005, 5. Available from:
6 Barrie, A. R. & Ward, A. M. Questioning behaviour in http://www.biomedcentral.com/content/pdf/1472-6947-5-
general practice: a pragmatic study. British Medical Journal 34.pdf (accessed 17 May 2006).
1997, 315, 1512–5. 22 Lappa, E. Undertaking an information-needs analysis of

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
 Information-seeking behaviour, Karen Davies 93

the emergency-care physician to inform the role of 38 Wyatt, J. C. & Sullivan, F. ABC of health informatics-
the clinical librarian: a Greek perspective. Health keeping up: learning in the workplace. British Medical
Information and Libraries Journal 2005, 22, 124–32. Journal 2005, 331, 1129–32.
23 Magrabi, F., Coiera, E. W., Westbrook, J. I., Gosling, A. S. 39 Blenkinsopp, J. Appraising the literature. In: Booth, A. &
& Vickland, V. General practitioners’ use of online Walton, G. (eds) Exploiting Knowledge in Health Services.
evidence during consultations. International Journal of London: Facet Publishing, 2004: 204–13.
Medical Informatics 2005, 74, 1–12. 40 Shelstad, K. R. Information retrieval patterns and
24 McAlister, F. A., Graham, I., Karr, G. W. & Laupacis, A. needs among practising general surgeons: a statewide
Evidence-based medicine and the practising clinician. experience. Bulletin of the Medical Library Association
Journal of General Internal Medicine 1999, 14, 236–42. 1996, 84, 490–7.
25 Oliveri, R. S., Gluud, C. & Willie-Jørgensen, P. A. Hospital 41 Westbrook, J. I., Gosling, A. S. & Coiera, E. Do clinicians
doctors’ self-rated skills in and use of evidence-based use online evidence to support patient care? A study of
medicine: a questionnaire survey. Journal of Evaluation in 55 000 clinicians. Journal of the American Medical
Clinical Practice 2004, 10, 219–26. Informatics Association 2004, 11, 113–20.
26 Ram, F. S. F. & Wellington, S. R. General practitioners 42 Darmoni, S. J., Benichou, J., Thirion, B., Hellot, M. F. &
use of the Cochrane Library in London. Primary Care Fuss, J. A study comparing centralized CD-ROM and
Respiratory Journal 2002, 11, 123–4. decentralized intranet access to . Bulletin of the
27 Ramos, K., Linscheid, R. & Schafer, S. Real-time Medical Library Association 2000, 88, 152–6.
information-seeking behaviour of residency physicians. 43 Gorman, P. N., Yao, P. & Seshadri, V. Finding the answers
Family Medicine 2003, 35, 257–60. in primary care: information seeking by rural and
28 Sackett, D. L. & Strauss, S. E. Finding and applying non-rural clinicians. MEDINFO 2004, 11, 1133–7.
evidence during clinical rounds: the ‘evidence cart’. Journal 44 Dorsch, J. L. Information needs of rural health
of the American Medical Association 1998, 280, 1336–8. professionals: a review of the literature. Bulletin of the
29 Schilling, L. M., Steiner, J. F., Lundahl, K. & Anderson, Medical Library Association 2000, 88, 346–54.
R. J. Residents’ patient-specific clinical questions: 45 Booth, A. Selecting Appropriate Resources. In: Booth, A. &
opportunities for evidence-based learning. Walton, G. (eds) Managing, Knowledge in Health Services.
Academic Medicine 2005, 80, 51–6. London: Library Association, 2000: 206–21.
30 Schwartz, K., Northrup, J., Israel, N., Crowell, K., 46 Holtum, E. A. Librarians, clinicians, evidence-based
Lauder, N. & Neale, A. V. Use of online evidence-based medicine, and the division of labour. Bulletin of the Medical
resources at the point of care. Family Medicine 2003, 35, Library Association 1999, 87, 404–7.
251–6. 47 Koonce, T. Y., Giuse, N. B. & Todd, P. Evidence-based
31 Scott, I., Heyworth, R. & Fairweather, P. The use of databases versus primary medical literature: an in-house
evidence-based medicine in the practice of consultant investigation on their optimal use. Bulletin of the Medical
physicians. Results of a questionnaire survey. Australian Library Association 2004, 92, 407–11.
and New Zealand Journal of Medicine 2000, 30, 319–26. 48 Line, M. B. Libraries and information services in 25 tears’
32 Swinglehurst, D. A., Pierce, M. & Fuller, J. C. A. A clinical time: a British perspective. In: Lancaster, F. W. (ed.) Libraries
informaticist to support primary care decision making. and the Future: Essays on the Library in the Twenty-First
Quality in Health Care 2001, 10, 245–9. Century. New York: Haworth Press, 1993: 73–83.
33 Veness, M., Rikard-Bell, G. & Ward, J. Views of Australian 49 Krahn, J., Sauerland, S., Rixen, D., Gregor, S., Bouillon, B.
and New Zealand radiation oncologists and registrars & Neugebauer, E. A. M. Applying evidence-based surgery
about evidence-based medicine and their access to in daily clinical routine: a feasibility study. Archives of
Internet-based sources of evidence. Australasian Radiology Orthopaedic and Trauma Surgery 2006, 126, 88–92.
2003, 47, 409–15. 50 Muir Gray, J. A. & De Lusignam, S. National electronic
34 Wilson, P., Glanville, J. & Watt, I. Access to the Library for Health (NeLH). British Medical Journal 1999,
online evidence base in general practice: a survey of the 319, 1476–9.
Northern and Yorkshire Region. Health Information and 51 Grandage, K. K., Slawson, D. C. & Shaughnessy, A. F.
Libraries Journal 2003, 20, 172–8. When less is more: a practical approach to searching for
35 Wong, K. & Veness, M. J. Internet access, awareness and evidence-based answers. Journal of the Medical Library
utilization of web-based evidence: a survey of Australian, Association 2002, 90, 298–304.
New Zealand and Singaporean radiation oncology 52 Verhoeven, A. A. H., Boerma, E. J. & Meyboom-de Jong,
registrars. Asia-Pacific Journal of Clinical Oncology 2005, 1, B. Use of information sources by family physicians:
120–7. a literature survey. Bulletin of the Medical Library
36 Lor, P. J. The Information Needs of the General Practitioner: Association 1995, 83, 85–90.
A Theoretical Model Based on an Analysis of the General 53 Sanders, S. & Del Mer, C. Clever searching for evidence:
Practitioner’s Cognitive Environment. University of South new search filters can help to find the needle in the
Africa, 1979. haystack. British Medical Journal 2005, 330, 1162–3.
37 Buckland, M. K. Library Services in Theory and Context, 54 Doney, L., Barlow, H. & West, J. Use of libraries and
2nd edn. Oxford: Pergamon, 1988. electronic information resources by primary care staff:

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94
94 Information-seeking behaviour, Karen Davies

outcomes from a survey. Health Information and Libraries hospital setting: an observational study. British Medical
Journal 2005, 22, 182–8. Journal 1998, 316, 673–6.
55 D’Alessandro, D. M., D’Alessandro, M. P., Galvin, J. R., 65 Straus, S. E. Bringing evidence to the point of care.
Kash, J. B., Wakefield, D. S. & Erkonen, W. E. Barriers to Evidence-Based Medicine 1999, 4, 70–1.
rural physician use of a digital health sciences library. 66 McAlearney, A. S., Schweikhart, S. B. & Medow, M. A.
Bulletin of the Medical Library Association 1998, 86, 583– Doctors’ experience with handheld computers in clinical
93. practice: qualitative study. British Medical Journal 2004,
56 Erickson, S. & Warner, E. R. The impact of an individual 328, 1162–5.
tutorial session on  use among obstetrics and 67 Honeybourne, C., Sutton, S. & Ward, L. Knowledge in the
gynaecology residents in an academic training programme: palm of your hands: PDAs in the clinical setting. Health
a randomized trial. Medical Education 1998, 32, Information and Libraries Journal 2006, 23, 51–9.
269–73. 68 Byrnes, J. A., Kulick, T. A. & Schwartz, D. G.
57 McKibbon, K. A. & Walker-Dilks, C. J. The quality and Information-seeking behaviour changes in
impact of  searches performed by end-users. community-based teaching practices. Journal of the
Health Libraries Review 1995, 12, 191–200. Medical Library Association 2004, 92, 334–40.
58 Davidoff, F. & Florance, V. The informationist: a new 69 Booth, A. Turning research priorities into answerable
health profession? Annals of Internal Medicine 2000, 132, questions. Health Information and Libraries Journal 2001,
1085–6. 18, 130–2.
59 Tovey, D. & Godlee, F. General practitioners say that 70 NHS Executive Information for Health. An Information
evidence-based information is changing practice. British Strategy for the Modern NHS 1998–2005.
Medical Journal 2004, 329, 1043. A National Strategy for Local Implementation. 1998.
60 Thompson, M. L. Characteristics of information resources Available from: http://www.medgraphics.cam.ac.uk/
preferred by primary care physicians. Bulletin of the Downloads/IPU_0006.PDF (accessed 20 July 2006).
Medical Library Association 1997, 85, 187–92. 71 Houston, T. K. Clinical questions: new research on barriers
61 Potter, B. & Rotert, E. Making evidence-based practice a and solutions to finding information. John Hopkins
reality. Wisconsin Medical Journal 2005, 104, 22–4. Advanced Studies in Medicine 2005, 5, 153–4.
62 Ely, J. W., Burch, R. J. & Vinson, D. C. The information 72 Griffiths, J. R. & Brophy, P. Student searching behaviour in
needs of family physicians: case-specific clinical questions. the JISC Information Environment. Ariadne 2002, 33.
The Journal of Family Practice 1992, 35, 265–9. Available from: http://www.ariadne.ac.uk/issue33/edner/
63 McKibbon, K. A. Evidence-based practice. Bulletin of the intro.html (accessed 30 March 2007).
Medical Library Association 1998, 86, 396–401.
64 Coiera, E. & Tombs, V. Communication behaviours in a Received 9 February 2007; Accepted 4 April 2007

© 2007 The author

Journal compilation © Health Libraries Group 2007 Health Information and Libraries Journal, 24, pp.78–94