Determining maximal achievable effect sizes of antidepressant therapies in placebo‐controlled trials

• Published in: Acta psychiatrica Scandinavica 2021-09, Vol.144 (3), p.300-309
• Main Authors: Hieronymus, Fredrik, Lisinski, Alexander, Hieronymus, Magnus, Näslund, Jakob, Eriksson, Elias, Dinesen Østergaard, Søren
• Format: Article
• Language: English
• Subjects: adults; Antidepressants; antidepressive agents; Clinical trials; depression; depression rating-scale; difference; efficacy; hamilton depression; major depression; Mental depression; minimal clinically important; minimal clinically important difference; Neurologi; Neurology; outpatients; Patients; placebo effect; Placebos; Psychiatry; psychotherapies; Psykiatri; Remission; Serotonin uptake inhibitors; themselves
• ISSN: 0001-690X; 1600-0447
• DOI: 10.1111/acps.13340

Objective Antidepressants outperform placebo with an effect size of around 0.30. It has been suggested that effect sizes as high as 0.875 are necessary for a minimal clinically important difference. Whether such effect sizes are achievable in placebo‐controlled trials is unknown. Therefore, we aimed to assess what effect sizes are theoretically achievable in placebo‐controlled trials of antidepressants. Methods Patient‐level analyses comparing Hamilton Depression Rating Scale (HDRS‐17) outcomes for simulated antidepressant therapies to placebo‐treated participants (n = 2201) from clinical trials of selective serotonin reuptake inhibitors. Results An optimally effective antidepressant, where all treated participants achieve HDRS‐17 scores comparable to those displayed by healthy volunteers (remission‐type model), had a maximum effect size of 1.75, with a mean difference of 11.6 points on the HDRS‐17. In simulations where patients received an additional 50% symptom reduction over that obtained with placebo (improvement‐type model), the maximum effect size was 1.08 with a mean HDRS‐17 difference of 7.2. When adjusting for normal rates of treatment discontinuation, maximum effect sizes were 1.10 (remission‐type model) and 0.76 (improvement‐type model) with HDRS‐17 mean differences of 8.8 and 5.6, respectively. Conclusions Three methodological issues (i) a large and variable placebo response, (ii) a high rate of dropout and (iii) HDRS‐17‐ratings significantly larger than zero in healthy volunteers, reduce the degree of treatment‐placebo separation achievable in depression trials. Assuming that those who discontinue treatment have only partial response, even a highly effective antidepressant would have difficulties surpassing such effect size cut‐offs as have been suggested to signify a minimal clinically important difference.