Antidepressants versus placebo

The authors selected published FDA trials on antidepressants to meta analyze, they found no average efficacy for mild to moderate depression compared to placebo.

The cited NCHS paper shows antidepressant use by 11% of the population over 12, exacerbated by unethical papers like this. It unequivocally promotes antidepressants for severe depression though the cited “Emperor’s New Drugs…” investigates breaking of blind conditions as one cause.

The cited paper by Turner et al implies antidepressant trials are encouraged by journals towards showing positive results, the project paper tries to reduce this bias by its methodology but then gives a misleading result by failing to include or mention any of these withheld trials in its results. Turner et al also found 31% of FDA trials of antidepressants were not published implying 94% were positive when only 51% were.

The cited paper by Montcrieff and Kirsch implies artificial cut off points between remission and non-remission are unethically used to amplify small effects. The project paper ethically uses continuous results on the HRSD scale, it also implies some trials use disproportionately more patients in the severe versus moderate category cutoff to avoid unfavorable placebo comparisons.

The cited paper by Dunlop and Baja implies pressure to unethically eliminate placebo trials with antidepressants ostensibly to protect patients. The project paper confronts this ethically by concluding on average attrition doesn’t affect its results, implying patients are not harmed by substituting placebos.

The authors fail to point out that severely depressed patients don’t give informed consent to trial antidepressants little better than placebos, nor to have the results distorted to deceive many uninformed doctor’s patients with placebo effects. When the public receives placebo medications the trial participants become uninformed accomplices to fraud. The authors ethically mention that declining responses from placebos don’t mean antidepressants are working, disclosed in the cited paper “Initial Severity…” by Kirsch et al.

The HRSD scale also measures subjective feelings rather than just disease symptoms, this incentivizes trials testing for drugs people like rather than treating disease. Montcrieff and Cohen show alcohol or sedatives improve HRSD scores but cannot be ethically promoted as antidepressants. The project paper critically fails to explain or reference how or even if antidepressants work while endorsing them, providing correlation without explaining its nature or causation.

Drowsiness and gastrointestinal upsets mean patients often detect the antidepressants, the project paper then biases towards higher HRSD scores from these severely depressed patients’ expectations. The project paper ignores bias from shorter trials artificially cutting off lower scores from later disappointments or relapse. Sometimes patients previously responsive to antidepressants are then selected for other trials, another ignored bias.

References

N Engl J Med. 2008 Jan 17;358(3):252-60. doi: 10.1056/NEJMsa065779.Selective publication of antidepressant trials and its influence on apparent efficacy.Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R.
Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, et al. (2008) Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration. PLoS Med 5(2): e45. doi:10.1371/journal.pmed.0050045
Efficacy of antidepressants in adults BMJ 2005; 331 doi: http://dx.doi.org/10.1136/bmj.331.7509.155 (Published 14 July 2005)
NCHS Data Brief ■ No. 76 ■ October 2011 Antid