Statistical Method to Analyze Data from Delayed Start Design in Parkinson’s Disease
View Presentation View Presentation
*Ohidul I Siddiqui, CDER, FDA 

Keywords: delayed start design, missing data, parkinson's disease

Development of a drug that could potentially modify the course of Parkinson’s disease is extremely challenging. A design that has been proposed for such a claim is delayed start design. In this design, a comparison is made between the patients who are early randomized to treatment (early start) vs. the patients who are initially randomized to placebo are given active treatment after certain weeks (delayed start). It is hypothesized that if the early start treatment group has lower mean clinical score compared to the mean score of the delayed start treatment group at the end of study period, the drug has disease modifying benefit.

In delayed start design, missing data due to dropouts are unavoidable. The missing data at the end of delayed start period are often imputed using LOCF approach. Clinical trial simulations are conducted to evaluate the statistical properties (including typ