08 Pharmacoepidemiology

Pharmacoepidemiology
NG ICRI Pharmacoepidemiology
Outline of presentation
Definitions The importance of pharmacoepidemiology Studies on drug use Studies on drug effects Signal generation Risk quantification Hypothesis testing Problem solving Applications of pharmacoepidemiology
Definitions
Epidemiology-Study of distribution and determinants of disease in populations
Clinical epidemiology:How to critically evaluate medical literature and how to apply principles of epidemiology to clinical medicine
Pharmacoepidemiology: Definition - 1
Pharmacoepidemiology is the study of
Use of drugs Effect of drugs in large numbers of subjects.
Pharmacoepidemiology Definition 2
Application of the principles of epidemiology to drug use and drug effect in large numbers patients
Evolution of Pharmacoepidemiology as a science

1961 Thalidomide disaster Case reports of limb malformations in offspring of women. Treated with thalidomide Development of systems for Adverse Drug Reactions Monitoring.
Development of the science of Pharmacoepidemiology.
The importance of pharmacoepidemiology- the problem of ADRs

Account for 5% of all hospital admissions. Occur in 10-20% of hospital inpatients. Cause death in 0.1% of medical and 0.01% of surgical inpatients. Adversely effect patients quality of life.
Cause patients to lose confidence in their doctors.

Increase costs of patients care.
The importance of pharmacoepidemiology- the inadequacies of clinical trials For the investigation of certain drug events, models are not possible e.g. pregnancy. RCTs are often inadequate to answer questions on safety.
Aims of Pharmacoepidemiological studies
Signal generation
Risk quantification Hypothesis testing
Studies on drug use
Drug utilization studies

WHO definition: The marketing, distribution, prescription and use of drugs in a society with the resultant medical, social and economic consequences Quantitative or Qualitative (DUR or drug utilization reviews) DURs focus on specific drugs or class of drugs (4th generation cephalosporins, aminoglycosides)
The concepts of DDD and PDD

The defined daily dose (DDD) is the estimated average maintenance dose per day of a drug when used for its major indication (e.g aspirin)
Expressed as DDDs/1000 population In hospitals, DDDs/100 bed days, adjusted for occupancy
Prescribed daily dose -PDD

The prescribed daily dose or PDD is the average daily dose of a drug that has actually been prescribed Calculated from a representative sample of prescriptions
The need for drug utilization studies

Indication of pattern of use of drugs Early signals of irrational use of drugs
Allows comparisons between regions, countries

Interventions to improve drugs use Continuous quality improvement
Study of drug effects
Aims of Pharmacoepidemiological studies?

Signal generation
Risk quantification Hypothesis testing
What is a signal?
Reported information on a possible causal relationship between an adverse event and a drug, the relationship being unknown or incompletely documented previously. Usually more than one report is required to generate a signal depending upon the seriousness of the event and quality of the information.
Rawlins and Thompson classification -1991

Type A Predictable (Augmented) Dose-dependent Related to the drugs pharmacological properties. Diarrhea with antibiotic use. Type B Bizarre, unpredictable. Dose-independent (idiosyncratic). Unrelated
Hypersensitivity with penicillin.
Methods of signal generation
Spontaneous reporting - 1
A system by which practicing clinicians are encouraged to report any and all adverse events with a drug Reports complied at the National Centre Reports from National Centers are then sent to the Uppsala Monitoring Centre.
Spontaneous reporting:- Advantages

Early warning signals
Relatively inexpensive Do not interfere with clinical practice
Spontaneous reporting: Disadvantages

Information inadequate, incomplete, not verifiable Cause- effect difficult to establish Voluntary, thus under reporting Incidence and prevalence difficult to calculate
Prescription Event monitoring - 1

Used for both signal generation and risk quantification
Used to study large cohorts of drugs

5000 18,000 prescriptions from prescribers studied
Information:
All adverse events Death from any cause Hospitalization Fetal abnormalities Changes in laboratory values
Advantages: Calculation of incidence

Prescription event monitoring when?

New chemical entity
Predicted widespread use Identified but unquantified risks
Post-marketing surveillance
Carried out by the pharmaceutical industry A single cohort of 5000-10,000 patients studied
Follow up- months/years Data submitted to regulatory authorities
Record linkage
Monitoring drug use and effect via database research Databases provide accessible information on thousands of patients Each patient in the database has a unique identifier Examples of information- hospitalization, infections developed, death, birth defects, lab investigations, physician services Databases can also be linked Disadvantages: Accuracy of data, missing data, lack of data Advantages: Speed
Methods of Risk Quantification &

Hypothesis testing
What is risk?
The probability of developing an outcome, regardless of severity.
What is a hypothesis ?
A hypothesis is a supposition based on observation or reflection. Cigarette smoking causes lung cancer.
Use of O.C pills causes hypertension and thromboembolic phenomenon.

Epidemiological studies allow you to accept or reject a hypothesis.
Why measure risk postmarketing?

Risk quantification often requires large sample sizes Rule of 3 An AE of 1/3000 will require 9000 subjects to be studied Pre marketing studies usually pick up Type A adverse events (dose dependent)
Risk versus exposure

Single group risk [ cases/ total exposure], does not account for baseline risk.
Risk is always calculated against an another group- unexposed or an experimental exposure
Classification of risk - CIOMS

Very common 1 in 10 exposures
Common Uncommon Rare Very rare < 1 in 10 1 in 100 exposures < 1 in 100 1 in 1000 exposures < 1: 1000 1:10,000 exposures > 1 in 10,000 exposures
Measurement of risk
Absolute risk (AR) Absolute risk reduction (ARR) Relative risk (RR) Relative risk reduction (RRR) Odds ratio (OR) Number needed to treat (NNT)
Problem - 1
In a randomized trial, investigators compared mortality rates in patients with bleeding oesophageal varices treated either by endoscopic ligation or endoscopic sclerotherapy. After a mean follow up of 10 months, 18/64 pts treated with ligation died, 29/65 pts treated with sclerotherapy died.
(1) Absolute risk (AR)

Simplest measure of association Absolute risk of dying in the ligation group is 18/64 or 28% Absolute risk of dying in the sclerotherapy group is 29/65 or 45%
Measurement of risk
(2) Absolute risk reduction (ARR)

Establishes a relation between the two absolute risks Also called as Risk difference (RD) Tells us what proportion of pts are spared the outcome if they receive the experimental therapy (rather than conventional therapy) ARR = 0.445-0.281 = 0.165 = 16.5%
Measurement of risk
(3) The concept of relative risk

RR or risk ratio Defined as the ratio of the risk rate in the exposed population versus unexposed or control versus experimental population
Contingency table/ 2x2 table

An Did not develop Total outcome an outcome Exposed to drug/test a b a+b
Not exposed to c drug/control

RR = a = P1 a+b
d
c = P2 c+d
c+d
P1 P2
P1= 0.446 P2= 0.281 P1/P2 = 0. 63 Thus the risk of death in subjects is 2/3rds that with ligation
Or Subjects receiving sclerotherapy were 1.58 times or one and a half times as likely to die as compared to ligation treated subjects
Measurement of risk
(4) Relative Risk Reduction- RRR

Another measure of assessing effectiveness of treatment What is the proportion of baseline risk that is removed by the experimental therapy ARR / Baseline risk
Relative risk reduction (RRR)

ARR = 0.446 - 0.281 Baseline risk = 0.446 (sclerotherapy group)
RRR = 0.165/0.446 = 0.369 Or ligation decreases the risk of death by 37%

Measurement of risk
(5 ) Odds Ratio
Measures the strength of association between an exposure and outcome
Calculation of odds ratio

Odds in the exposed/odds in the unexposed Odds in the exposed: a/a+b = a b/a+b b Odds in the unexposed:c/c+d = d/c+d
c d
Contingency table/ 2x2 table

An outcome Did not develop Total an outcome
Exposed to drug/test Not exposed to drug/control OR =
a+b
c+d
ad/bc
Calculation of odds ratio

Death No Death Total
Exposed to ligation
18 a
46 b
64 a+b
Exposed to sclerotherapy
29 c
36 d
65 c+d
OR ad/bc
Odds Ratio
Odds of dying in the ligation group is 0.39
Odds of dying in the sclerotherapy group is 0.80

Odds ratio = 0.39/0.80 = 0.48 Inference: Pts treated with ligation have half as likely to die as compared to sclerotherapy
Or, 0.8/0.39, pts treated with sclerotherapy as twice as likely to die as compared to ligation treated subjects
Measurement of risk
(6) Number needed to treat

Another way to express the impact of the new treatment
Allows comparison between treatments

NNT = 1/ARR For every 17 pts treated with ligation, 1 pt will be saved
For 100 pts, 28 pts will die with ligation

For 100 pts, 44.6 pts will die with sclerotherapy
Case control and cohort studies

Case control
Cohort
Proceeds from effect to cause.

Starts with the disease.
Proceeds from cause to effect.

Starts with the risk factor.
Fewer subjects. Larger number of subjects. Quicker results.
Longer follow up.

Relative inexpensive.
Expensive.
Levels of Evidence - 1
Level 1- RCTs, direct comparisons of drugs within the same class, rather than with placebos, for effect on important treatment outcomes
Level 2: 1) RCTs, direct comparisons of drugs within the same class, but on validated surrogate outcomes or 2) Comparisons of active agents with placebos on clinically important outcomes or validated surrogate outcomes
Levels of evidence- 2
Level 3: Placebo controlled trials where outcomes are restricted to unvalidated surrogate markers
Level 4: Non randomized studies (case control, cohort studies)
Applications of Pharmacoepidemiology
Estimation/quantitation of risk Patient counseling
Formulation of public health/policy decisions

Formulation of therapeutic guidelines and discovery of new indications Pharmacoeconomic decision making
Estimation of risks of drug use

The most common application
Use of case-control and cohort studies E.g., Clozapine and agranulocytosis and neutropenia (1:5000 versus 1:200)
Formulation of public-health policy decisions
Is the risk associated with the drugs too high?

Does the labeling need to be changed?
Is there too much of inappropriate prescribing?

Patient counseling
Termination of pregnancy versus continuation of pregnancy when the risk of malformation is low/high
Formulation of therapeutic decision making
Effectiveness and safety of drugs used in groups of people not included in Phase III trials (elderly, pediatrics)
E.g, Use of ciprofloxacin for typhoid fever in children
Summary
Application of principles of epidemiology to the study of the use and effects of drugs in large numbers of patients Deals with signal generation, risk quantification and hypothesis generation Applications in risk quantification, patient counseling, therapeutic, regulatory and economic decision making

08 Pharmacoepidemiology

Uploaded by

Copyright:

Available Formats

08 Pharmacoepidemiology

Uploaded by

Document Information

Original Description:

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

08 Pharmacoepidemiology

Uploaded by

Copyright:

Available Formats

Pharmacoepidemiology

Evolution of Pharmacoepidemiology as a science

Development of the science of Pharmacoepidemiology.

The importance of pharmacoepidemiology- the problem of ADRs

Cause patients to lose confidence in their doctors.

Aims of Pharmacoepidemiological studies

Studies on drug use

Drug utilization studies

The concepts of DDD and PDD

Prescribed daily dose -PDD

The need for drug utilization studies

Allows comparisons between regions, countries

Study of drug effects

Aims of Pharmacoepidemiological studies?

Rawlins and Thompson classification -1991

Hypersensitivity with penicillin.

Methods of signal generation

Spontaneous reporting:- Advantages

Spontaneous reporting: Disadvantages

Prescription Event monitoring - 1

Used to study large cohorts of drugs

Advantages: Calculation of incidence

Prescription event monitoring when?

Follow up- months/years Data submitted to regulatory authorities

Methods of Risk Quantification &

The probability of developing an outcome, regardless of severity.

Use of O.C pills causes hypertension and thromboembolic phenomenon.

Why measure risk postmarketing?

Risk versus exposure

Classification of risk - CIOMS

(1) Absolute risk (AR)

(2) Absolute risk reduction (ARR)

(3) The concept of relative risk

Contingency table/ 2x2 table

Not exposed to c drug/control

(4) Relative Risk Reduction- RRR

Relative risk reduction (RRR)

RRR = 0.165/0.446 = 0.369 Or ligation decreases the risk of death by 37%

Measures the strength of association between an exposure and outcome

Calculation of odds ratio

Contingency table/ 2x2 table

Exposed to drug/test Not exposed to drug/control OR =

Calculation of odds ratio

Odds of dying in the sclerotherapy group is 0.80

(6) Number needed to treat

Allows comparison between treatments

For 100 pts, 28 pts will die with ligation

Case control and cohort studies

Proceeds from effect to cause.

Proceeds from cause to effect.

Fewer subjects. Larger number of subjects. Quicker results.

Longer follow up.

Level 4: Non randomized studies (case control, cohort studies)

Formulation of public health/policy decisions

Estimation of risks of drug use

Formulation of public-health policy decisions

Is the risk associated with the drugs too high?

Is there too much of inappropriate prescribing?

Formulation of therapeutic decision making