Early escape option helps balance needs of patients, investigators

Scientists designing crossover clinical trials, particularly those involving rare diseases, often must strike a delicate balance between maximizing the study’s potential for discovery and minimizing patient exposure to ineffective treatment.

The Division of Biostatistics and Epidemiology partners with nearly all clinical and translational research programs at Cincinnati Children’s to help them design studies and manage statistical analysis. Its faculty members also conduct direct studies to ensure and enhance the strength and innovation of research.

In a recent example of this work, a study published in July 2014 in Contemporary Clinical Trials, explored findings from previous work on treatments for the inflammatory disorder familial Mediterranean fever (FMF). The study’s co-authors included first author was Bin Huang, PhD, from Biostatistics and Epidemiology; senior author Philip Hashkes, MD, from the Shaare Zedek Medical Center in Israel; and colleagues from the Division of Rheumatology at Cincinnati Children’s.

The team compared study power and dropout rates among three types of crossover trial designs: traditional without early escape, early escape as a patient choice, and early escape based on accepted protocol.

Both early-escape approaches were preferable to the traditional method. Researchers found that supporting exits based on patient choice reduced the dropout rate as much as 29 percent. However, the patient choice approach also increased patient exposure to the less effective treatment arm, which suggests that investigators must be mindful of the trade-offs between study design methods.

Citation

Huang B, Giannini EH, Lovell DJ, Ding L, Liu Y, Hashkes PJ. Maximizing study power and minimizing patient exposure to ineffective treatment. Contemp Clin Trials. 2014;38(2): 204-212.

Lead Researcher:

A photo of Bin Huang, PhD. 
Bin Huang, PhD

Expected time on placebo was calculated under the assumption of a Poisson distribution following the formula presented in Appendix A, with parameter settings similar to the familial Mediterranean fever (FMF) tria, i.e. a=1,A=5.Y axis shows the expected days on placebo; X axis corresponds to the rate of attack. The solid line corresponds to results of early escape per protocol; dashed line corresponds to results of early escape per patient choice. Under the traditional crossover design without escape, the number of days on placebo treatment is 180 days.
Click on image to view caption.