What is Pharmacoepidemiology? Flashcards
Definition of pharmacoepidemiology
The study of the use of and the effect of drugs in large numbers of people
Pharmacology - FOCUS of inquiry
Epidemiology - METHODS of inquiry
Definition of pharmacology
The study of the effects of drugs
Definition of clinical pharmacology
The study of the effects of drugs in humans
Central principle of clinical pharmacology
Therapy should be individualized, or tailored, to the needs of the specific patient at hand
Individualization of therapy
Requires the determination of a risk/benefit ratio specific to the patient at hand
Definition of epidemiology
The study of the distribution and determinants of diseases in populations
Pharmacoepidemiology falls within epidemiology
Subdivisions of epidemiology
A study of infectious diseases in populations
The study of chronic diseases
Major application of pharmacoepidemiology principles
After drug marketing
Pure Food and Drug Act (1906)
In response to excessive adulteration and misbranding of the food and drug at the time
No requirement for proof of efficacy or safety of marketed drug
Federal government was allowed to remove drug that was adulterated or misbranded
Food, Drug, and Cosmetic Act (1938)
Preclinical toxicity testing was required and manufacturers were required to gather clinical data about drug
No proof of efficacy was required
Phase I testing
Determine safety and dosage
Phase II testing
Evaluate efficacy, look for side effects
Phase III testing
Confirm efficacy, monitor adverse reactions for long term use
Phase IV testing
Additional post-market testing
Where pharmacoepidemiology is emphasized
Problems with clinical trials
Expensive; small; often drugs are compared against placebo; exclude elderly, children, pregnant women, patients with important comorbidities; may be unethical; not timely
HOWEVER, clinical trials provide gold standard evidence of drug effects
FDA Amendment Act of 2007
FDA has the right to require post-marketing studies to be completed
Study design: Most causal to least causal
RCT
Cohort
Case-control
Analyses of secular trends
Case series
Case reports
Case Reports
A report of an event in a single patient
Useful for generating hypotheses
Simple and inexpensive
Case Series
Collections of patients all of whom have either a single exposure or single outcome
NO control group = NO hypothesis testing
Useful to quantify an ADE and ensure particular ADE(s) are not happening in population larger than that studied prior to drug marketing
Analyses of Secular Trends (ecological studies)
Examine trends in an exposure that is a presumed cause and trends in a disease that is a presumed effect and test whether the trends coincide
Trends examined over either over time or across boundaries
LACK INDIVIDUAL DATA: Ecological fallacy, no control for confounding variables
Case-Control Study
Compare cases (with outcome) to controls (without outcome) to look for differences in antecedent exposure(s)
Useful to study multiple exposures and uncommon diseases
Potential for bias in control selection and exposure determination
Measure of association = OR
Cohort Study
A study that identifies a cohort of subjects and follows them over time to determine outcome
Loss to follow-up greater concern
Randomized Control Trial
Experimental trials: A study in which the investigator controls the exposure received by each participant
RCT: a study where participants are randomly assigned between exposure and control groups
ONLY DESIGN THAT CONTROLS FOR UNKNOWN CONFOUNDERS
Pragmatic Clinical Trials
A study in which the investigator tests the effectiveness of an intervention under “real-world” conditions
Not generalizable
Evidence paradox
Evidence Paradox
In order to inform clinicians on how to treat patients, we need real-life patients
People with comorbidities
People who forget to take their medication
Relative Risk
Measure of association
Based on incident data
Cohort, RCTs
Ratio of risk
a/(a+b) / c/(c+d)
Odds Ratio
Measure of association
Based on prevalent data
Case-control, cross sectional
Ratio of odds
a/c / b/d = ad/bc
P-value of significance
< 0.05
Significant CI
Not including the null
Study results
- No association: Exposure and effect are independent
- Artificial association (spurious): Either chance (unsystemic variation) or bias (systemic variation)
- Indirect association (confounded)
- Causal association (direct or true)
Types of error
- Random: Tested for and fixed statistically
- Bias: Avoided with good study design
- Confounding: Adjusted for when recognized
Biases
- Information bias: Interview bias (probing more thoroughly some subjects more than others) or recall bias (more than just a faulty memory)
- Selection bias: When controls don’t represent the population that produced the cases or differential loss to follow-up
Confounder
A variable related to both the exposure and the outcome
NOT be on the path from exposure to outcome
Confounding occurs when such a variable is distributed unequally between groups
