Transcript here - Boston University

Research Activities at The
Boston Collaborative Drug
Surveillance Program; Boston
University School of Public
Health
Susan Jick
Professor of Epidemiology
Director of the Boston Collaborative Drug Surveillance Program
Primary Objectives of the BCDSP

Post-marketing drug safety studies

Disease epidemiology to support drug safety concerns

Expertise in large electronic databases
What kinds of safety Questions Get
Asked?

Generally interested in rare or uncommon adverse events

Difficult to study using clinical trials

Difficult to collect information on large enough patient populations

Case-control

Cohort

Large databases have made it possible to study large cohorts of drug users

We can now study drugs in relation to rare outcomes
Early Post-Marketing Safety

What is post-marketing safety?

Generally observational research on marketed drugs

Prior to the 1960’s there were no formal drug safety studies

All data prior to the late 1970’s was collected by hand

First use of automated data was initiated around 1978

Group Health Cooperative

Around 300,000 members
This is where I came in

I first started working with data from GHC in the early 1980’s

While getting my MPH at BU SPH

Later went on to get my doctorate at BU

Started teaching the Drug Epidemiology class while getting my
doctorate

And the rest is history….
Later Developments

Other resources became available in the 1980’s

Research capabilities increased dramatically

Large cohorts of drug exposed persons could be identified

Rare events could be studied

Databases with covariate information became available
The Clinical Practice Research Datalink

Started in 1988


Data on millions of patients
Became available to the public in the 2000’s

GP data from the UK

All diagnoses

Drugs

Clinical details

Patient characteristics
Great Potential

Research capabilities have increased dramatically because of available data

Our research has changed correspondingly

So has the potential for error and poor research

Understanding research methods and data resource has never been more
important
What are the big challenges?
Major methodologic concerns

Confounding

By age, sex, calendar time, geography

And by indication

Appropriate and accurate exposure information and definition

Poor quality data

Data errors

Need for validation

Is information complete?

Ability to test assumptions
Examples of Drug Safety Studies at the BCDSP

MMR vaccine and autism

Oral contraceptives and venous thromboembolism

Vioxx and MI

Accutane and depression

Hormone replacement and breast cancer

Antidepressant drugs and suicide
What could go wrong?


What is confounding by indication?

People take drugs for a reason

Is that reason, or indication for use, associated with the outcome of interest?
In antidepressant users can you evaluate whether or not the drug increases
the risk of depression?

How do we control for it?

Can one always control for it?
A glimpse of our research