435 – Clinical Trial Design III
A Profile-Based Stratified Randomization and Its Application to a Double-Blind, Placebo-Controlled Clinical Trial
Andrew F. Magyar
Allergan, Inc.
Cornelia Haag-Molkenteller
Allergan, Inc.
Jihao Zhou
Allergan, Inc.
Stratified randomization is a common technique used in clinical trials to balance important baseline characteristics between treatment and control groups. However, as the number of stratification factors increases, the number of strata will become too large to handle in practice. A double-blind, placebo-controlled study (NCT01600716) was designed to investigate a compound for the treatment of urinary incontinence due to neurogenic detrusor overactivity in patients with multiple sclerosis, in which there are 4 stratification factors: age (< 55/>=55 years), baseline urinary incontinence episodes (< =9/>9), MS duration (< 12/>=12 years) and concurrent anticholinergic treatment use (Yes/No). To determine the strata for randomization, we proposed a profile-based stratification method, which used each subject's baseline profile determined using a propensity score. The parameters in the propensity score function was estimated based on the MS patients' data pooled from two similar previous Phase 3 clinical trials.