Are antidepressants worth taking

The short version: antidepressants produce a real but modest benefit for some people, while causing significant side effects in many, and the evidence base is weaker than most patients are told.

The largest real-world study of depression treatment, STAR*D, gives the most honest picture. In step one, about 37 in 100 people achieved remission on a single antidepressant. With each additional treatment attempt, the remission rate dropped: 31 in 100 at step two, then 14 in 100, then 13 in 100 [6]. Meanwhile, the share of people who stopped because they could not tolerate the drug rose from 16 in 100 at step one to 30 in 100 by step four [6]. That is a pattern of diminishing returns and accumulating intolerance, not a story of reliable help.

The side effects are not minor. In one pooled analysis of 4,768 patients on duloxetine, about 43 in 100 reported at least one side effect, compared to 31 in 100 on placebo [1]. Sexual dysfunction was three times more common on the drug than on placebo [1]. In a head-to-head comparison, escitalopram caused treatment-emergent sexual dysfunction in about 49 in 100 patients, duloxetine in about 33 in 100, and placebo in about 17 in 100 [5]. Another study found that about 32 in 100 patients starting an antidepressant had an adverse drug reaction, mostly gastrointestinal [3]. The side effects are not rare; they are the norm.

The efficacy data is also muddier than the headlines suggest. A 2021 analysis found that side effects themselves inflate depression rating scores, meaning the true drug effect is smaller than the published numbers show [1]. When you remove the items that overlap with side effects, the drug-placebo difference shrinks [1]. The published effect sizes, around Cohen's d of 0.75 for SSRIs and 0.68 for SNRIs [2], are modest and shrink further when unpublished negative trials are accounted for.

Psychotherapy alone or combined with medication outperformed medication alone in a 24-month trial. Long-term psychodynamic psychotherapy produced a 74 in 100 remission rate, fluoxetine alone produced 22 in 100, and the combination produced 65 in 100 [4]. The number needed to treat for psychotherapy versus fluoxetine was 3, meaning one extra remission for every three people who choose therapy over pills [4].

The evidence does not support the idea that antidepressants are a straightforward solution. For about 3 in 10 people, the first drug helps without intolerable side effects. For the rest, the path involves switching drugs, adding drugs, or accepting side effects for a modest gain. The most effective approach, psychotherapy, is the one least offered and least covered by insurance.

My call: antidepressants are worth trying for moderate to severe depression, but only with full knowledge that the benefit is modest, side effects are common, and psychotherapy is likely more effective and should be pursued first or alongside. Confidence: moderate.