Totality of Clinical Trial Data Considered in FDA Decision Making

By TCT Daily Staff

Meeting the primary endpoint in a clinical trial is important, but it is not the only factor considered by the FDA for approval of an investigational device, according to Ashley B. Boam, MS.

Boam, chief of the Interventional Cardiology Devices Branch of the FDA’s Office of Device Evaluation in its Center for Devices and Radiological Health, outlined on Tuesday the elements the FDA looks for in clinical trial design, data reporting, and analysis.

"Some people think that it’s really pass-fail: If you meet the endpoint, all systems are go; if you miss the primary endpoint, approval is not possible," Boam said.

"We really look at the totality of the data, the individual components or a composite, as well as the safety issues. At the end of the day we want an accurate representation of all the results in the labeling."

Boam said the FDA’s review of clinical trial submissions is always comprehensive, but its focus may vary depending on the reason for the trial, whether it was a proof-of-principle study, a pivotal trial for device approval, a postmarketing study, or a study of a new device iteration.

Critical elements in trial protocol

Trial protocols must contain certain critical elements to allow for proper interpretation and evaluation, Boam said.

Most important, outcome measures must be meaningful and reproducibly measurable.

Hypotheses must be clearly expressed in words as well as mathematical expressions. If superiority is to be shown, differences must be clinically significant, not just statistically significant. If noninferiority is the endpoint, appropriate margins must be specified.

In addition, the follow-up schedule must be reasonable to allow evaluation of the outcomes of interest and must be workable in the real world of clinical practice.

Patient selection

Patient selection must strike a balance; on the one hand, broad enough to ensure that the study results are generalizable to the patient population and, on the other hand, narrow enough to facilitate interpretation of results.

If patients are enrolled at multiple sites, demographics and other factors that could influence the study outcome must be considered, Boam said.

In statistical analysis planning, appropriate tests to evaluate the study hypotheses are needed, and adjustments for covariate differences between treatment and control groups must be considered.

Boam advised the use of independent data safety monitoring and clinical events committees. She said the FDA pays close attention to informed consent instruments, and these should be thoughtfully formulated.

She urged investigators to ask the FDA for help with trial design if they run into questions.

"FDA is here to partner with you, to help investigators and companies as they put together study protocols," she said.

Disclosures:

（source：www.tctmd.com）