Adverse reactions: Published data
When considering information about ADRs in published studies, don't simply accept the data that you find. Remember to assess the relative validity of different sources of information.
Clinical trials
These give a helpful indication of common reactions with a guide to incidence, but:
- The comparatively small patient populations involved may mean that serious, but uncommon, reactions are not discovered.
- The patients involved in clinical trials may be highly selective and not representative of the real patient population. Often the elderly and children are excluded, for
example.
- The procedures for ADR monitoring and reporting in clinical trials are variable in their rigour.
Case reports
A case report can raise suspicion about an ADR, and may give helpful advice about how a specific situation was managed in practice. However, it is often insufficient by itself to establish a causal association because:
- A single case report linking a reaction to a drug could be just coincidence, or due to a number of confounding influences.
- There is no comparison group not exposed to the drug to allow for a quantitative estimate of risk.
- There is no reliable denominator of drug-exposed patients from which an incidence rate can be estimated.
Epidemiological studies
Given the limitations of clinical trials and case reports, post-marketing epidemiological studies generally provide the best source of quantitative information on ADRs. These fall into two broad categories:
- Follow-up (cohort) studies. Groups defined by exposure status to a particular drug are followed and subsequent events recorded and compared. For example 750 patients on leflunomide are scrutinised for adverse events.
- Case-control studies. Groups defined by their outcomes are enrolled and prior drug exposures ascertained and compared. For example, 4,000 patients with upper GI bleeding may be investigated in an attempt to identify factors that may have caused it.
