Monday, August 11, 2008

DCRI researchers weigh benefits of smaller versus larger drug-safety databases
By Kelly Winget, DCRI Communications

Several prescription drugs recently had to be withdrawn from the market due to adverse drug events that a percentage of patients experienced. This has heightened the debate on whether to increase safety and surveillance studies once a drug has been approved and is publicly available, or to have larger clinical trials focusing on safety before the drug is approved to better identify and possibly prevent adverse events.

DCRI researchers, led by Shelby Reed, Ph.D., created a hypothetical model to evaluate the number of adverse events that could potentially be avoided once the drug is approved. The goal was to see if using a larger patient database for pre-approval trial would be a cost-effective way to prevent adverse events. The results were published in the August 5 issue of Health Affairs.

When researchers are planning clinical trials, they typically base their patient sample size on what is needed to determine if the drug is effective. For medications that are designed for chronic but not life-threatening conditions, the standard recommendation is to gather data on 1,500 patients about the drug's safety. Among those patients, the International Conference on Harmonization recommends that between 300 and 600 patients were treated with the drug for at least six months and 100 were treated for at least a year.

DCRI researchers note that a patient database this size might not have enough statistical power to identify adverse events that occur in as many as one in 100 patients.

The hypothetical model created by DCRI researchers analyzed how many adverse events could be avoided, comparing larger versus smaller databases while also reviewing how cost-effective having larger databases might be. For the model, researchers assumed 10 million patients would be treated with the hypothetical drug, and that 75,000 patients might have an adverse drug event.

With a pre-approval safety database of 2,000 patients, approximately 57,000 adverse events could be detected and avoided. Using a database of 4,000 patients, the number increases to approximately 72,000 adverse events that could be avoided.

According to the model, when smaller patient databases have enough statistical data to detect an increased risk for adverse events, larger databases would just increase the cost and duration of the trial unnecessarily.

In scenarios when a smaller database does not have enough statistical information to detect adverse events in a reliable way, large databases offer significant benefits to future patients by helping to avoid adverse events. This could mean the drug is not approved or it is not approved for a large patient population.

For serious adverse events, the model indicated that there could be scenarios where both the smaller and larger patient databases did not have enough information to detect the adverse events. There are estimates that serious adverse events were not detected in approximately half the drugs on the market until they were made available to a large patient population.

The researchers note that conducting safety studies after a drug has been released to the market is necessary, even if pre-approval trials are required to use larger patient databases. The reason is that clinical trials can't reveal all the risks associated with real-world clinical use and clinical trials can't offer information on the long-term effects of a drug before it has been approved.

Researchers recommend standardizing how safety data is reporter so clinicians and patients can better evaluate and compare the potential risks of drugs.

The other DCRI researchers involved with this study include Kevin Anstrom, Ph.D., Damon Seils, Rob Califf, MD, and Kevin Schulman, MD.

Click here to read the full article.