Related Experiment Videos
Limitations of controlled augmentation trials in schizophrenia
R G Stern1, J Schmeidler, M Davidson
1Department of Veteran Affairs, F.D.R. Montrose VA Hospital, New York 10548, USA.
Biological Psychiatry
|July 15, 1997
Summary
Most schizophrenia augmentation trials lack sufficient patients and employ too many outcomes for reliable results. These studies often lack the statistical power to confirm treatment efficacy, making conclusions tentative.
Area of Science:
- Psychiatry
- Clinical Trials
- Pharmacology
Background:
- Add-on augmentation treatments are explored for schizophrenia management.
- Evaluating the empirical basis of these treatments requires rigorous trial design.
Purpose of the Study:
- To assess the experimental design of pharmacologic augmentation trials in schizophrenia.
- To compare trial designs against conventional statistical requirements.
Main Methods:
- A systematic search identified 13 double-blind, placebo-controlled, parallel, add-on neuroleptic augmentation trials from 1988-1992.
- Analysis focused on sample size, number of outcome measures, statistical power, and effect size.
Main Results:
- The average trial included 34.5 subjects and 25.0 outcome measures.
- Mean statistical power was low (.1-.4) for typical effect sizes.
- A 63% probability of significant findings by chance was observed.
- Sample sizes needed for adequate power (0.80) ranged from 58 to 216 subjects.
Conclusions:
- Most trials had insufficient sample sizes and excessive outcome measures for conclusive efficacy determination.
- Tentative conclusions are drawn from trials with fewer than 40-100 subjects or multiple hypotheses.