Use of single case analyses in off‐on‐off‐on trials in affective illness: A demonstration of the efficacy of nimodipine

Abstract

AbstractThis paper 1) demonstrates the efficacy of nimodipine in the treatment of a patient with rapid‐cycling bipolar disorder using an intensive single subject design, and 2) evaluates three statistical techniques that are useful in the analysis of single case data. The patient is a 42‐year‐old bipolar II female refractory to multiple medication trials who underwent double‐blind treatment with placebo (B), nimodipine (A), and verapamil (C) in a B‐A‐B‐A‐C‐A design. The use of statistics in single subject studies has been the subject of controversy. One problem with single subject data is that data points are often correlated with data points from adjacent time periods (i.e., nonindependent), and therefore such data have been considered inappropriate for statistics that assume independence of error terms. One statistical technique we examined was a t‐test in the context of repeated on‐offdrug trials suggested by Gentile, Roden, and Klein (1972), in which all data points from treatment phases are aggregated and then compared to aggregated data from baseline or placebo phases. We compared this t‐test to a variant of meta‐analysis in which each phase is statistically compared to its adjacent phases and the t‐values from those comparisons are summed and converted to a z‐score. As illustrated with this patient, the t‐test method tended to be more conservative than the z‐score method in instances where there is an overall time trend for improvement or deterioration. The Chassen technique also provides an alternative for analysis of data with time trends. The 2 method was more conservative than the single aggregated t‐test if the direction of predicted change on a given measure (i.e., improvement during an active drug phase and deterioration during a placebo phase) did not occur as expected at each medication transition. These statistical techniques, in combination with the sequential off‐on‐off‐on clinical trial design, provide strong evidence for the mood‐stabilizing effect of nimodipine in this rapid‐cycling BP II female. Depression 2:259–271 (1994/1995). © 1995 Wiley‐Liss, Inc.