Text version
Do antidepressants work well enough to justify their use given increasing concern about the frequency of severe withdrawal symptoms that can occur when antidepressants are stopped?
Hi! Jim Phelps here for the Psychopharmacology Institute. In the last Quick Take, I reviewed the PANDA study, the largest study of antidepressants not sponsored by a pharmaceutical company. In that study, sertraline was not superior to placebo for depression symptoms in a large primary care sample, but there were some positive findings. Anxiety improved, at least statistically. And overall, multiple meta-analyses have found antidepressants statistically superior to placebo for major depression, with an effect size of 0.3. How big a benefit is this? An effect size of 0.3 may be statistically significant. But is it clinically significant?
Here’s a paper that may help with that question. Drs. Hengartner and Ploderl present the minimal clinically important differences for common measures of depression. Remember the question is whether antidepressants are associated with clinically meaningful improvement. Take the Hamilton Depression Rating Scale (HAM-D). The common version has 17 items, each scored 0–4 by a clinician with a maximum of 51 points. A total score of 0–7 is generally regarded as not depressed, 8–16 as mild depression, 17–23 as moderate depression, and over 24 severe depression. So, how big a decrease in HAM-D scores would you actually notice when a patient comes back to see you in 4 to 6 weeks after starting an antidepressant? How big a change would be big enough to suggest continuing that antidepressant? That change is referred to as the minimal clinically important difference (MICD). That’s the smallest difference in a score that would mandate a change in patient management.
According to Hengartner and Ploderl, that MICD for the HAM-D is 3–5 points. You’ll see the implications of that number in a moment when we compare it to the average decrease in HAM-D scores seen in randomized trials of antidepressants. The point of Hengartner and Ploderl’s paper is to present the data behind that minimal clinically important difference. Where did that figure of 3–5 points come from? Was it fairly established? A very brief look then at their analysis. Estimates of the MCID had been produced several different ways, each somewhat different in how clinicians and patients’ impressions are weighted, whether the comparison is between patients or within a patient, meaning change over time, and whether statistical distributions, like the standard deviations of the measure, for example, are incorporated. And yet, interestingly, and it’s Hengartner and Ploderl’s main point, all the different methods produce very similar MCID values. For the HAM-D, for example, it’s that 3–5 point range I mentioned a moment ago.
Which brings us to the punchline, the real reason why Hengartner and Ploderl wrote this paper in the first place. If the MCID on the HAM-D is 3–5 points, how big a decrease can be expected from an antidepressant on average? The answer is that the meta-analytically derived mean HAM-D reduction associated with antidepressants is 2 points. By this analysis then, antidepressants do not, on average, produce meaningful clinically important differences—not that this will resolve the controversy about the benefit–risk ratio of antidepressants. But anyone who prescribes antidepressants should be following this debate. And if you’re struggling with Hengartner and Ploderl’s very negative conclusion regarding the efficacy of antidepressants, make sure to understand this paper, which is linked here at the Psychopharmacology Institute. It’s not the final word, but it’s an important one.
Abstract
The efficacy of antidepressants in the acute treatment of moderate-to-severe depression remains a controversial issue. The minimal important difference (MID) is relevant to judge the clinical significance of treatment effects. In this analysis paper, we discuss estimates of the MID for common depression outcome measures. For the Hamilton Depression Rating Scale 17-item Version (HDRS-17), according to both anchor-based and distribution-based approaches, MID estimates range from 3 to 8 points, and the most accurate values are likely between 3 and 5 points. For the 6-item version (HDRS-6), MID estimates range between 2 and 4 points. For both the Montgomery-Åsberg Depression Rating Scale (MADRS) and the Beck Depression Inventory II (BDI-II), MID estimates range between 3 and 9 points, with estimates of 3-6 points likely being the most accurate. Quality of life appears to be more important to patients than core depression symptoms. We thus also evaluated the Short-Form 36 (SF-36) mental component score, a popular mental-health-related quality of life measure. Its MID estimate is likely about 5 points. By contrast, the average treatment effects of antidepressants on the HDRS-17, HDRS-6, MADRS, BDI-II and SF-36 are 2 points, 1.5 points, 3 points, 2 points and 3-5 points, respectively. In conclusion, the efficacy of antidepressants in the acute treatment of moderate-to-severe depression consistently fails to exceed the lower bound of the MID estimates for common depression outcome measures. The clinical significance of antidepressants thus remains uncertain and we call for more research on quality of life measures, which are the patients’ most valued outcome domains.
Download PDF and other files
Reference
Hengartner, M. P., & Plöderl, M. (2022). Estimates of the minimal important difference to evaluate the clinical significance of antidepressants in the acute treatment of moderate-to-severe depression. BMJ Evidence-Based Medicine, 27(2), 69-73.
Related References
- Hengartner, M. P., & Plöderl, M. (2018). Statistically significant antidepressant-placebo differences on subjective symptom-rating scales do not prove that the drugs work: effect size and method bias matter!. Frontiers in Psychiatry, 9, 517.
- Hengartner, M. P., Plöderl, M., Braillon, A., Jakobsen, J. C., & Gluud, C. (2020). Sertraline in primary care: comments on the PANDA trial. The Lancet Psychiatry, 7(1), 17.
- Lewis, G., Duffy, L., Ades, A., Amos, R., Araya, R., Brabyn, S., … & Lewis, G. (2019). The clinical effectiveness of sertraline in primary care and the role of depression severity and duration (PANDA): a pragmatic, double-blind, placebo-controlled randomised trial. The Lancet Psychiatry, 6(11), 903-914.
