Abstracto

When should the sequential parallel comparison design be used in clinical trials?

Lee Baer, Anastasia Ivanova

Randomized clinical trials in many conditions such as mental illness, pain, addiction, allergy and asthma are subject to moderate-to-high rates of placebo response. This reduces the effect size for the comparison of the active treatment to the placebo group, hence reducing the efficiency of studies. Since attempts to resolve this problem using a placebo lead-in design have not been successful, in 2003 the sequential parallel comparison design (SPCD) was developed. In the SPCD, the study is conducted in two treatment stages, with the efficacy analysis including (a) all subjects in stage 1 and (b) stage 2 data from stage 1 placebo nonresponders. The relevant data from the two stages are pooled to compute an overall p-value. This article reviews the situations in which the SPCD is preferable to the standard parallel design as well as the circumstances in which SPCD is not a good choice.