Are Antidepressants Overprescribed? The Question the Field Keeps Dodging

Antidepressants are now so culturally embedded that they appear in dating profile bios, and nobody finds that alarming enough.

The Normalization Problem

There is a particular kind of clinical failure that hides behind the language of progress. When more than 14% of Australians are taking antidepressants, one of the highest rates recorded anywhere in the world, the instinct inside the system is to frame this as a success story: more people getting help, stigma declining, access improving. That framing is not just incomplete. It is a way of not asking the harder question.

For young women especially, SSRIs (selective serotonin reuptake inhibitors) have undergone a cultural transformation that should make clinicians uncomfortable. A medication class with a real but specific evidence base has become something closer to a wellness product, referenced casually in group chats, joked about on social media, hashtagged under TikToks about Lexapro. The cultural normalization of antidepressant use is not the same thing as appropriate clinical use, and conflating the two has consequences for real patients who will eventually need to come off these drugs without anyone having planned for that moment.

The question worth asking is not whether SSRIs work. They do, in specific, well-defined populations, for specific, well-defined indications. The question is whether the rate at which they are being prescribed reflects anything resembling that specificity, or whether prescribing has simply decoupled from clinical appropriateness altogether.

What the Data Actually Shows

The numbers do not support complacency. A study led by Eduard Vieta and Diego Hidalgo-Mazzei, reviewing data from 947,698 people who attended primary care services in Catalonia between 2010 and 2019, found that antidepressant prescribing increased by 404% over that decade, while diagnoses of the conditions that actually warrant antidepressant treatment increased by just 49%. That is not a rounding error. That is an eightfold divergence between prescribing rate and clinical indication, published in the Journal of Psychiatry and Mental Health, and it deserves to be taken seriously rather than explained away.

In the BMJ, consultant psychiatrist Richard Braithwaite cited Mitchell and colleagues' meta-analysis to make a point that still has not landed with sufficient force in primary care: that overdiagnosis of depression is massive, and that overprescribing follows overdiagnosis as a direct consequence (BMJ 2014;348:g1436). The diagnostic threshold for depression has drifted so far from its original clinical anchoring that normal human suffering; grief, situational distress, the grinding exhaustion of financial precarity, is now routinely captured inside it and medicated accordingly.

That last point is the one psychiatry would rather not confront directly. Prescribing an SSRI as a stopgap for inaccessible psychotherapy or unaddressed social determinants of health is not a clinical decision. It is a rationing decision dressed up as one. Dr. Mark Horowitz has said plainly that it is not acceptable to prescribe antidepressants as a stopgap in the absence of accessible, affordable services; and he is right, but saying it plainly apparently remains controversial.

What This Looks Like in the Consulting Room

The clinical reality of routine SSRI prescribing rarely resembles the controlled trial populations on which the efficacy data is based. Major depressive disorder trials typically enroll patients with moderate-to-severe, persistent depressive episodes. The patient sitting in a GP's office after a relationship breakdown, a redundancy, or six months of housing instability does not match that profile; but receives the same prescription, often with minimal psychoeducation, no defined treatment duration, and no exit plan discussed at the point of initiation.

What follows is predictable and well-documented. The patient stabilizes, or improves because the situational stressor resolves, or tolerates the side effects well enough not to complain. Months become years. The original indication (which may never have met formal diagnostic criteria) is never revisited. When the patient eventually asks about stopping, they are told to taper over two to four weeks, which is not a taper by any pharmacologically defensible definition. Antidepressant discontinuation syndrome emerges. The patient and clinician alike interpret it as relapse. The prescription is reinstated. Duration of use extends further.

This is not an edge case. This is the modal patient experience for a substantial proportion of the millions of people currently taking SSRIs in high-income countries. The Catalonia data makes it structural: the majority of people on these drugs are on them for reasons that were never rigorously indicated in the first place, which means they are also bearing the entire burden of the adverse effect profile (sexual dysfunction, emotional blunting, weight changes, and the not-small risk of post-SSRI sexual dysfunction (PSSD) persisting after discontinuation) without the clinical justification that would make that risk-benefit calculation defensible.

A Framework for Doing This Better

The first requirement is honesty at the point of prescribing. Patients starting an SSRI should understand three things that current practice rarely communicates: what the actual evidence base for their specific presentation looks like, what the known adverse effects include (including the ones the field has historically minimized), and what the plan is for eventually discontinuing the medication. An exit strategy is not optional clinical information. It is part of informed consent.

The second requirement is periodic reassessment that is genuine rather than perfunctory. Prescribers should be asking, at regular intervals, whether the original indication still holds, whether the patient has derived the expected benefit, and whether the risk-benefit calculus has shifted. For patients whose original presentation was situational or subthreshold, the presumption at the one-year or two-year mark should favor discontinuation, not continuation.

When deprescribing is indicated, the method matters enormously. The Horowitz and Taylor hyperbolic tapering model, published in The Lancet Psychiatry and operationalized in the Maudsley Deprescribing Guidelines, establishes that dose reductions should be made in proportion to the current dose rather than in fixed milligram steps, because serotonin transporter occupancy follows a hyperbolic rather than linear relationship with dose. The clinical implication is that the final reductions, from low doses toward zero, produce disproportionately large changes in receptor occupancy and carry the highest risk of withdrawal. Standard two-to-four-week tapers are not designed around this pharmacology. They are designed around clinical convenience, and patients pay for that with their symptoms.

The Maudsley Deprescribing Guidelines recommend tapering periods measured in months to years for patients on long-term SSRIs, with reductions of no more than 10% of the current dose per step, and adequate stabilization time between reductions. This is what the evidence supports. It is not what most patients receive.