Jill Wechsler
|
Can larger clinical trials detect more safety problems before a drug comes to market and thus avoid serious adverse events
in larger subject populations? Such an approach may be worth the time and money involved, according to some research analysts.
So far, efforts to address drug safety issues have focused primarily on strengthening postmarketing assessment and adverse
event monitoring. But expanded preapproval studies in some cases may be able to detect problems before a new therapy comes
to market and thus avoid serious adverse events later on.
The downside of proposals that the Food and Drug Administration require more preapproval safety data is that this could delay
the approval of needed treatments. Such an approach, moreover, could boost the cost of drug development and of resulting medicines.
An alternative strategy is to limit access to a new drug for a period of time (one to four years) after it comes on the market
while the sponsor conducts postmarketing studies and monitors utilization very closely. Those patients willing to assume more
risk may agree to use the newly approved therapy during this trial or "conditional" period, but most sponsors oppose the uncertainty
inherent in such an approach.
More focus on safetyIn calculating the pros and cons of larger and longer trials, the nature of the drug, the seriousness of the condition being
treated, and the availability of alternative therapies all effect decisions on optimal clinical trial size and scope. These
issues are considered in a recent analysis of the costs and benefits of larger preapproval clinical trials and their capacity
for reducing postapproval injuries and deaths. Researchers at the Duke University School of Medicine headed by associate professor
Shelby Reed propose a new research framework that can better identify those situations where larger clinical trials may reveal
higher risks of adverse events and thus avoid safety problems, particularly in certain target populations.
The Push for More Subgroups in Trials
|
In a paper published online in August by Health Affairs [ http://content.healthaffairs.org/cgi/content/abstract/27/5/w360/], which can only be accessed by subscribers or on a pay-per-view basis, the authors present a model for calculating whether
increasing the size of a clinical trial can improve the odds of detecting an adverse drug event (ADE) among those taking the
test drug. The number of subjects in a trial and the background frequency of a certain ADE in the general population can determine
the study's odds ratio.
In looking at COX-2 inhibitors as an example, the analysts calculate that a clinical trial with 2000 subjects in each treatment
arm would be likely to detect ADEs associated with the test drug 76% of the time. Doubling the size of the study to 4000 subjects
in each treatment arm would increase detection rates to 96%. When plugging in the cost of running the larger clinical trial,
the researchers found that boosting enrollment to 4000 subjects per arm is cost effective for avoiding safety problems: the
incremental cost is only about $27,000 per life-year saved. However, a further increase to 8000 subjects per arm would provide
little additional detection power and is not worth the additional expense.
Based on this model, Reed and colleagues might want to raise the size of premarket clinical safety databases recommended by
the International Conference on Harmonization (ICH). For medications to treat chronic conditions, ICH advises sponsors to
present data on 1500 patients, including 300 to 600 treated for at least six months and 100 for one year. Although the ICH
policy leaves room to increase those numbers where there are safety concerns for similar compounds, the Duke analysts consider
these recommended sample sizes much too small to characterize many ADEs. They propose that study sponsors and regulatory authorities
identify potential ADEs of interest for any given drug and the odds ratio at which the drug's side effects might outweigh
the benefits of the therapy.
