Related Experiment Videos
A method to quantify rater bias in antidepressant trials
E Petkova1, F M Quitkin, P J McGrath
1Department of Biostatistics and the Department of Therapeutics, Columbia University College of Physicians and Surgeons, New York State Psychiatric Institute, New York, NY 10032, USA.
Summary
Rater bias in antidepressant clinical trials may slightly inflate drug efficacy. However, this study found bias insufficient to invalidate conclusions, as patient-reported outcomes still showed significant drug effectiveness compared to placebo.
Area of Science:
- Clinical Psychology
- Psychopharmacology
- Biostatistics
Background:
- Concerns exist regarding incomplete blinding in antidepressant drug trials.
- Rater bias potentially inflates perceived drug efficacy, questioning antidepressant effectiveness.
Purpose of the Study:
- To quantify rater bias in clinical trials of antidepressant efficacy.
- To assess the contribution of bias to claims of antidepressant drug effectiveness.
Main Methods:
- Derived patient-based effect size using Symptom Check List (SCL-90) depression scores.
- Contrasted patient-based effect size with clinician-based effect size to indirectly measure bias.
- Assessed patient prodrug bias likelihood using active placebo response rates.
Main Results:
- Patient-derived effect sizes were smaller than clinician-derived effect sizes, suggesting potential rater bias.
- Quantified bias was insufficient to invalidate original conclusions based on clinician ratings.
- Patient self-ratings showed significant differences between drugs and placebo, with non-overlapping confidence intervals.
Conclusions:
- While some clinician ratings may be biased, the extent appears insufficient to alter conclusions on antidepressant efficacy in this trial.
- The proposed method offers an indirect measure of rater bias by comparing patient and clinician-derived effect sizes.
- Further application of this approach in other trials is needed to establish generalizability.