evidencelast verified 2026-09-19

Does psychiatric medication keep you out of hospital, or put you there?

Both claims get made with total confidence, and the studies support neither as a slogan. for schizophrenia and bipolar disorder, the largest studies ever done tie staying on medication to fewer hospital stays. a randomized trial and a twenty year follow-up say lifelong use is not settled. starting an antidepressant carries two specific, measured risks that depend on who you are. and the question people most want answered, what happens to someone with no diagnosis who is handed a prescription, has barely been studied. this page sets out what was measured, in each direction, with the limits of each study beside it.

nothing here is a reason to start, stop or change a medication. several of these studies show that stopping suddenly and alone is the most dangerous option of all. if you are in crisis in the US, call or text 988.

22.4%
of schizophrenia hospital stays are followed by a readmission within 30 days; 15.0% for mood disorders (AHRQ, 2012)
0.53
hazard ratio for rehospitalization on clozapine against no antipsychotic, same person on and off (29,823 people, Sweden)
2.83
hazard ratio for mania in the 3 months after starting an antidepressant alone in bipolar disorder
72.7%
of visits where an antidepressant was prescribed recorded no psychiatric diagnosis (US, 2007)

the trap in every one of these numbers. people are put on medication when they are most unwell. so people on medication are hospitalized more than people who are not, whether the drug helps, harms or does nothing. researchers call it confounding by indication. the studies below are sorted by how well they get around it: randomized trials first in weight, then studies that compare each person with themselves, then everything else.

1. how often people come back

the federal hospital database counted 847,000 stays for mood disorders and 383,000 for schizophrenia in 2012. within thirty days, 15.0 percent of the mood disorder stays and 22.4 percent of the schizophrenia stays were followed by another admission for any cause. for every other kind of illness it was 15.4 percent. people on Medicare or Medicaid came back 40 to 75 percent more often than people with private insurance.

over years, the picture is starker. in national registers, 43.7 percent of people with schizophrenia in Sweden and 54.0 percent of people hospitalized for bipolar disorder in Finland were back at least once. in the general population it is rare: 1.3 percent of US adults had an inpatient mental health stay in the past year, against 16.7 percent who took a medication for their mental health. most people on these drugs never see the inside of a psychiatric ward.

2. the evidence that medication keeps people out

schizophrenia, Sweden, 29,823 people. each person was compared with themselves, on a drug and off it. against no antipsychotic, the risk of rehospitalization was about half on clozapine (hazard ratio 0.53) and on long-acting injections (0.51 to 0.58). it was barely lower on oral quetiapine (0.91). read the byline too: co-authors included employees of Janssen-Cilag, which makes the injection that ranked first, and the lead author reports fees from most makers of antipsychotics.

bipolar disorder, Finland, 18,018 people. same design. lithium was tied to the lowest risk of psychiatric rehospitalization among mood stabilizers (0.67) and the lowest risk of admission for any cause (0.71). the most prescribed drug, quetiapine, managed 0.92. the best record belongs to a generic salt nobody markets; the weakest to the drug most often chosen. that is the marketing story seen from the outcomes side.

stopping, Finland, 8,719 first admissions, up to twenty years. against people who stayed on an antipsychotic, the risk of rehospitalization or death was 1.63 times higher for those who stopped at discharge and 7.28 times higher for those who stopped after five years. the risk of death was 174 to 214 percent higher in non-users and early stoppers. the authors: the risk after stopping “does not decrease as a function of time during the first 8 years of illness”. people who stop are not a random sample. some stop because they are well, many because they are unwell, homeless or using substances.

3. the evidence that lifelong is not settled

the only randomized trial with long follow-up. Dutch patients whose first psychosis had been in remission for six months were randomly assigned to guided dose reduction or to maintenance. the reduction group relapsed more in the first two years. at seven years, 40.4 percent of them had recovered, meaning few symptoms and a working life, against 17.6 percent on maintenance. 103 people, open label, not yet replicated at that length. it applies to people already well, tapered by clinicians.

twenty years in Chicago. among 70 people with schizophrenia followed six times over two decades, “patients with schizophrenia not prescribed antipsychotics had significantly better work functioning” from year four onward. nobody was randomized, and those who came off and stayed off may simply have had a milder illness. read beside the Finnish data it suggests both things can be true: the drugs prevent relapse, and a subgroup does well without them over decades. nobody can yet tell in advance who is in it.

4. the measured risks of starting an antidepressant

mania. in 3,240 Swedes with bipolar disorder, each compared with themselves, starting an antidepressant alone was followed by a hazard ratio for mania of 2.83 in the first three months. with a mood stabilizer it was 0.79. a third were given the antidepressant alone despite a known diagnosis. across 109 trials, mania or hypomania occurred in 12.5 percent of people on antidepressants and 7.5 percent without, and the relative increase was larger in people diagnosed with depression, some of whom have a bipolar illness nobody has seen yet. if you have ever had days of little sleep with racing thoughts and unusual energy, or a close relative has bipolar disorder, say so before you start.

suicidal behavior, by age. the FDA pooled 372 placebo-controlled trials with 99,231 adults. the odds ratio for suicidal behavior on an antidepressant was 2.30 under age 25, 0.87 from 25 to 64, and 0.06 at 65 and over. these are randomized trials, so under 25 the drug accounts for the difference. in absolute terms the events were rare.

and the timing, which cuts the other way. across 82,285 antidepressant starts in one health plan, suicide attempts serious enough for a hospital stay ran at about one per thousand starts in the following six months. they were “highest in the month before starting antidepressant treatment and declined progressively after starting medication.” people start at their worst moment. a hospital stay soon after a first prescription is weak evidence that the prescription caused it.

one small study says antidepressants raise readmissions. in Zurich, 45 inpatients given antidepressants were matched to 45 who were not. within a year 35.6 percent of the first group were rehospitalized against 22.2 percent. the authors say their findings “raise the possibility” that antidepressants impair recovery. ninety people, no randomization, never replicated at scale. it is here because the question is fair and almost untested.

5. starting with no diagnosis, and never leaving

this is the question with the least evidence and the most people in it. by 2007, 72.7 percent of US doctor visits at which an antidepressant was prescribed recorded no psychiatric diagnosis, up from 59.5 percent in 1996. in Baltimore, where the same adults were interviewed four times over twenty four years, 69 percent of antidepressant users had never met criteria for major depression and 38 percent had never met criteria for any of five common mood or anxiety disorders. the approval trials enrolled only people who met full criteria. for the rest there is no trial evidence of benefit, and the side effects and the difficulty of stopping are the same.

nobody has followed that group forward to count hospital stays. we looked and did not find it. what has been measured is whether people leave. from 1999 to 2010 antidepressant use rose from 6.5 to 10.4 percent of US adults, and the entire rise was use lasting two years or more (3.0 to 6.9 percent). short-term use fell slightly. the increase was confined to patients of general medical providers. the door in stayed the same size. the door out got smaller. when long-term users did try to leave under supervision, in the ANTLER trial, 56 percent relapsed within a year against 39 percent who stayed on, and withdrawal is easy to mistake for relapse. see how many adults take these drugs and how long each was tested.

what to take from it

Questions people ask

How often do people go back to hospital after a psychiatric stay?

In the federal hospital data for 2012, 15.0% of stays for mood disorders and 22.4% of stays for schizophrenia were followed by a readmission for any cause within 30 days; the figure for all other conditions was 15.4%. Over years, not days, national register studies found 43.7% of people with schizophrenia in Sweden and 54.0% of people with bipolar disorder in Finland were rehospitalized at least once.

Does starting a psychiatric medication raise the risk of being hospitalized?

Not as a general rule, and for some conditions the evidence points the other way. People are usually started on medication when they are at their worst, so they are hospitalized more than people who are not, whatever the drug does. In 82,285 antidepressant starts, suicide attempts leading to hospital were highest in the month before the first prescription and fell afterwards. The measured exceptions are specific: an antidepressant alone in bipolar disorder nearly tripled the rate of mania, and in randomized trials antidepressants roughly doubled suicidal behavior in people under 25.

Do antipsychotics prevent rehospitalization?

In the largest study, 29,823 people with schizophrenia in Sweden were each compared with themselves on and off medication. The risk of psychiatric rehospitalization was about half on clozapine (hazard ratio 0.53) or a long-acting injection (0.51 to 0.58), and barely lower on oral quetiapine (0.91). Co-authors included employees of the maker of the top-ranked injection. A Finnish 20-year study found stopping was followed by more rehospitalization and death. A randomized Dutch trial found the opposite for recovery: 40.4% recovered at seven years after early dose reduction against 17.6% on maintenance.

What happens to people who start an antidepressant with no diagnosis?

It is common and almost unstudied. By 2007, 72.7% of US doctor visits where an antidepressant was prescribed recorded no psychiatric diagnosis. In a Baltimore community study, 38% of antidepressant users had never met criteria for any of five common mood or anxiety disorders. Nobody has followed that group forward to count later hospital stays. What is known is that people stay on: the entire rise in US antidepressant use from 1999 to 2010 was use lasting two years or more.

Is it safe to stop?

Not on your own, and not quickly. In the Finnish study the danger after stopping an antipsychotic did not fade with time. In the ANTLER trial of long-term antidepressant users, 56% relapsed within a year of a supervised taper against 39% who stayed on, and withdrawal symptoms can be mistaken for relapse. The Dutch trial that found better recovery with lower doses reduced them step by step under clinical supervision, in people who had already been well for six months.

Sources

  1. Heslin KC, Weiss AJ. Hospital Readmissions Involving Psychiatric Disorders, 2012. HCUP Statistical Brief #189. AHRQ, May 2015. https://hcup-us.ahrq.gov/reports/statbriefs/sb189-Hospital-Readmissions-Psychiatric-Disorders-2012.jsp
  2. Tiihonen J, Mittendorfer-Rutz E, Majak M, et al. Real-World Effectiveness of Antipsychotic Treatments in a Nationwide Cohort of 29 823 Patients With Schizophrenia. JAMA Psychiatry 2017;74(7):686–693. https://doi.org/10.1001/jamapsychiatry.2017.1322 read our summary of this study →
  3. Lähteenvuo M, Tanskanen A, Taipale H, et al. Real-world Effectiveness of Pharmacologic Treatments for the Prevention of Rehospitalization in a Finnish Nationwide Cohort of Patients With Bipolar Disorder. JAMA Psychiatry 2018;75(4):347–355. https://doi.org/10.1001/jamapsychiatry.2017.4711
  4. Tiihonen J, Tanskanen A, Taipale H. 20-Year Nationwide Follow-Up Study on Discontinuation of Antipsychotic Treatment in First-Episode Schizophrenia. Am J Psychiatry 2018;175(8):765–773. https://doi.org/10.1176/appi.ajp.2018.17091001
  5. Wunderink L, Nieboer RM, Wiersma D, Sytema S, Nienhuis FJ. Recovery in remitted first-episode psychosis at 7 years of follow-up of an early dose reduction/discontinuation or maintenance treatment strategy. JAMA Psychiatry 2013;70(9):913–920. https://doi.org/10.1001/jamapsychiatry.2013.19
  6. Harrow M, Jobe TH, Faull RN, Yang J. A 20-Year multi-followup longitudinal study assessing whether antipsychotic medications contribute to work functioning in schizophrenia. Psychiatry Res 2017;256:267–274. https://doi.org/10.1016/j.psychres.2017.06.069
  7. Viktorin A, Lichtenstein P, Thase ME, et al. The risk of switch to mania in patients with bipolar disorder during treatment with an antidepressant alone and in combination with a mood stabilizer. Am J Psychiatry 2014;171(10):1067–1073. https://doi.org/10.1176/appi.ajp.2014.13111501
  8. Tondo L, Vázquez G, Baldessarini RJ. Mania associated with antidepressant treatment: comprehensive meta-analytic review. Acta Psychiatr Scand 2010;121(6):404–414. https://doi.org/10.1111/j.1600-0447.2009.01514.x
  9. Stone M, Laughren T, Jones ML, et al. Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. BMJ 2009;339:b2880. https://doi.org/10.1136/bmj.b2880
  10. Simon GE, Savarino J, Operskalski B, Wang PS. Suicide risk during antidepressant treatment. Am J Psychiatry 2006;163(1):41–47. https://doi.org/10.1176/appi.ajp.163.1.41
  11. Hengartner MP, Passalacqua S, Andreae A, et al. Antidepressant Use During Acute Inpatient Care Is Associated With an Increased Risk of Psychiatric Rehospitalisation Over a 12-Month Follow-Up After Discharge. Front Psychiatry 2019;10:79. https://doi.org/10.3389/fpsyt.2019.00079
  12. Mojtabai R, Olfson M. Proportion of antidepressants prescribed without a psychiatric diagnosis is growing. Health Aff 2011;30(8):1434–1442. https://doi.org/10.1377/hlthaff.2010.1024
  13. Takayanagi Y, Spira AP, Bienvenu OJ, et al. Antidepressant use and lifetime history of mental disorders in a community sample: results from the Baltimore Epidemiologic Catchment Area Study. J Clin Psychiatry 2015;76(1):40–44. https://doi.org/10.4088/JCP.13m08824
  14. Mojtabai R, Olfson M. National trends in long-term use of antidepressant medications: results from the U.S. National Health and Nutrition Examination Survey. J Clin Psychiatry 2014;75(2):169–177. https://doi.org/10.4088/JCP.13m08443
  15. Lewis G, Marston L, Duffy L, et al. Maintenance or Discontinuation of Antidepressants in Primary Care (ANTLER). N Engl J Med 2021;385(14):1257–1267. https://doi.org/10.1056/NEJMoa2106356 read our summary of this study →
  16. SAMHSA. Key Substance Use and Mental Health Indicators in the United States: Results from the 2024 National Survey on Drug Use and Health. https://www.samhsa.gov/data/sites/default/files/reports/rpt56287/2024-nsduh-annual-national-report.pdf

adam