Introduction to Pharmacoepidemiology

Question 1

Pharmacoepidemiology:

Answer 1

study of the use, risks, and benefits of drugs in populations

Question 2

Pharmacovigilance:

Answer 2

continual monitoring for unwanted effects and other safety-
related aspects of marketed drugs

Question 3

Comparative effectiveness research (CER):

Answer 3

determining what therapeutic intervention (not just drug products) works best for a given disorder in patients likely to be seen in clinical practice

Question 4

Pragmatic research:

Answer 4

studies (often using randomization) that often test small
practical changes that could have an impact on health outcomes

Question 5

Experimental

Answer 5

RCTs (active treatment, usual care, pragmatic, others)

Question 6

Nonexperimental (Observational)

Answer 6

• case-control
• cohort
• others
• pharmacoepidemiologic and pharmacovigilance studies are primarily observational

Question 7

Pharmacoepidemiology

Answer 7

Study of the use, risks, and benefits of drugs in populations
* “the study of the utilization and effects of drugs in large numbers of people”
* pharmacology /pharmacotherapy + epidemiology
* studies to provide an estimate of probability of beneficial effects in
populations, or probability of adverse effects in populations
* application of epidemiological methods to pharmacological issues

Question 8

Applications of pharmacoepidemiology

Answer 8

Supplement information from premarketing studies
Identify new information not available from premarketing studies
General contributions

Question 9

Supplement information from premarketing studies

Answer 9

better quantify ADRs and beneficial effects
* higher precision
* can include populations not well represented in premarketing trials * can study effects of other drugs/disease states
* can study effects relative to other drugs for same indication

Question 10

Identify new information not available from premarketing studies

Answer 10

• previously undetected ADRs/beneficial effects: uncommon effects, delayed effects
• patterns of drug utilization
• effects of varied doses
• economic impact of drug use

Question 11

General contributions

Answer 11

• reassurance of drug safety
• ethical and legal obligations

Question 12

Data sources for pharmacoepidemiology

Answer 12

• Adverse drug reaction reports
• Medical claims data
• Electronic medical records (EMR)
• Richer, more robust data sources and increased computational abilities have made more rigorous pharmepi. studies possible
• Indiana Network for Patient Care (INPC)

Question 13

Medical claims data

Answer 13

• private and/or government medical/Rx drug insurance providers
• some collected and sold by third party vendors e.g., “TRUVEN”
• typically contains diagnostic, procedure, lab, prescription codes with basic patient information

Question 14

Electronic medical records (EMR)

Answer 14

• institutional/health system
• country wide (e.g. General Practice Research Database [GPRD] in UK)
• other (e.g., health information exchanges)

Question 15

Indiana Network for Patient Care (INPC)

Answer 15

> 100 separate healthcare entities providing data which includes:
* major hospitals, health networks, and insurance providers
* data on >18 million patients
* >10 billion clinical observations
* >951 million encounter records
* >147 million mineable text
* prescription drug data
* percent of residents who have touched the INPCR has grown to approximately two thirds of Indiana’s population.

Question 16

Limitations of observational studies in pharmepi.

Answer 16

bias and confounding

Question 17

Bias and Confounding

Answer 17

• Bias: systematic deviation from the truth that distorts the results of research
• Confounding: relationship between treatment and response (or exposure and disease) is actually attributable to another variable (the “counfounder”); confounder is independently related to BOTH the exposure and the outcome

Question 18

Information bias

Answer 18

• bias related to information regarding exposure or outcome
• includes measurement and/or classification error

Question 19

Detection bias

Answer 19

• specific outcome is diagnosed preferentially in subjects exposed to the agent
• e.g., may be more likely to look for an AE in someone who is exposed to a drug

Question 20

Confounding by indication

Answer 20

• indication for a drug or severity of disease predicts the use of the drug
• e.g., ACEI’s in preventing MI in patients with hypertension
• hypertensive patients with comorbidities (e.g., DM) may be more likely to get ACEI vs those without
• e.g., COXIB’s and GI bleeds

Question 21

Selection bias

Answer 21

• bias related to procedures used to select subjects/influence study participation
• due to systematic differences in characteristics between those who are selected for the study and those who are not

Question 22

Referral bias

Answer 22

reason for encounter is related to drug treatment (e.g., when the use of the drug contributes
to the diagnostic process)

Question 23

Protopathic bias

Answer 23

• exposure of interest is used unknowingly to treat an adverse event related to outcome/agent is used for early manifestation of a disease that has not yet been diagnosed
• e.g., an antipsychotic may be started to treat delirium, but the drug may have anticholinergic effects that contribute to delirium

Question 24

Prevalence bias

Answer 24

prevalent cases rather than new (incident) cases are selected

Question 25

Confounding by Indication

Answer 25

• occurs when the risk of an event is related to the indication for
  medication use but not the use of the medication itself
• appears when the reason of prescription is associated with the outcome of interest.

Question 26

Protopathic Bias “reverse causality”

Answer 26

• occurs if a particular treatment was started, stopped or otherwise changed because of the baseline manifestation caused by a disease or other outcome event
• in pharmacoepidemiology, protopathic bias occurs when the drug is initiated in response to the first symptoms of the disease which is, at this point, undiagnosed

Question 27

Time related biases

Answer 27

Lag time
* e.g. proton pump inhibitors and fracture risk

Question 28

“Immortal Time Bias”

Answer 28

• period of follow-up when, due to the exposure definition, the outcome
  being studied could never occur
• first described in heart transplant studies
• survival time in patients who received heart transplant vs. those on transplant list that did not receive one
• survival time in both groups calculated from the time they were added to transplant list
• survival time of those receiving a transplant (mean 111 days) was significantly longer than that of those accepted for a transplant who did not receive one (mean 74 days)
• problem?: * if a patient died before receiving a transplant, by default, they were in the non-
  transplant cohort
• patients in the transplant cohort were “immortal” in the time between being put on the list until the time of their transplant

Question 29

“Immortal Time Bias”: beta-blocker post-MI

Answer 29

• treated group: all subjects with 2 or more beta-blocker Rx; all others
  were untreated group
• followed from discharge
• observed significantly lower rates of death in the treated group (even with low doses)

Question 30

Pharmacovigilance

Answer 30

• Continual monitoring for unwanted effects and other safety-related aspects of marketed drugs
• Detection, evaluation, understanding, prevention of adverse drug reactions
• “the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other possible drug-related problems”
• Historically has involved collection of spontaneous reports of drug related morbidity or mortality
• Wider use of observational data conducted across multiple databases, development of large networks of observational databases
• post-marketing surveillance, signal detection, surveillance, data mining * often involves regulatory authorities, industry
• Rationale:
• premarketing studies exclude patients who have complicated
  comorbidities
• premarketing studies typically only determine the efficacy of a specific drug versus a placebo
• statistical power of a premarketing study is too weak to detect infrequent adverse events.

Question 31

Number of postmarketing/post approval safety programs/systems

Answer 31

• FDA Adverse Event Reporting System (FAERS) –> houses postmarketing adverse event reports received by FDA
• receives MedWatch reports
• FDA Sentinel System
• monitors safety of FDA regulated products
• data submitted by a number of sources (clinical, academic, claims, others)
• FDA VARES (vaccine adverse event reporting system)

Question 32

Comparative effectiveness research

Answer 32

the conduct and synthesis of research comparing the benefits and harms of different interventions and strategies to prevent, diagnose, treat and monitor health conditions in “real world” settings
The purpose of CER is to assist consumers, clinicians, purchasers, and policy makers to make informed decisions that will improve health care at both the individual and population level
“Research that identifies what clinical and public health interventions work best for improving health”
Determining what therapeutic intervention works best for a given disorder in patients likely to be seen in clinical practice
Generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat and monitor a clinical condition, or to improve the delivery of care.
* interventions should be compared on the basis of a health- related outcome measure

Question 33

CER Multiple study designs

Answer 33

• RCTs with active treatment arms
• observational studies
• other

Question 34

Efficacy vs. Effectiveness

Answer 34

• Efficacy: whether a drug (or other treatment) has the ability to bring about a given intended effect in controlled settings
• translates to: “in an ideal world, could a drug (treatment) achieve it’s intended effects?”
• Effectiveness: whether, in real-world patients and settings, a treatment, in fact achieves it’s desired effect

Question 35

Goals of CER

Answer 35

• to inform decisions on interventions or approaches to health care in real-world settings with regard to their intended and unintended outcomes that are relevant to patients
• to put new treatment into proper perspective in relation to older treatment
• identify patients who are more/less likely to respond to a given intervention than others
• overcome external validity problems with traditional RCTs

Question 36

Pragmatic Research

Answer 36

studies (often using randomization) that often test practical
changes that could have a big impact on health outcomes
* Pragmatic RCT: “a randomized clinical trial with one or more pragmatic elements”
* intended to overcome the limitations of traditional RCTs in order to answer CER questions
* include real world patients from diverse backgrounds
* beside intervention, other aspects of care are controlled by clinician
* include providers from diverse settings
* comparators are those used in clinical practice
* outcomes matter to clinicians and patients

Question 37

Pragmatic clinical trials Hybrid design

Answer 37

between RCT and routine care
Measure effectiveness in clinical practice
* “randomized observational design”
* typically aims to help clinician decide between new intervention and current standard of care
* takes place in routine clinical settings
* advantages of randomization with increased external validity