do two-thirds of homeless people have a mental illness, and two-thirds a substance use disorder?
“about two-thirds of the homeless population has a current diagnosable mental illness and about the same percentage have a current substance use disorder.”
a guest, a psychiatrist and senior scientist at CAMH said on PsychEd: Educational Psychiatry Podcast, 2026-09-01, at 07:00. hear it in context →
the episode: PsychEd Episode 78: Mental Health and Homelessness with Dr. Vicky Stergiopoulos (at 07:00). we link the show's own page; we do not re-host audio.
first, the part that holds
the first half holds. the largest global meta-analysis puts current mental disorder among people experiencing homelessness at 67%, which is two-thirds, and a high-income-country meta-analysis puts it higher, at 76%.
what we found when we went and looked
we read two meta-analyses. a 2024 jama psychiatry review of 85 studies (48,414 people) found current prevalence of any mental health disorder of 67% (95% ci 55-77), so the mental illness figure checks out. the same review found current substance use disorder of 44% (95% ci 30-59), and 56% over a lifetime, not two-thirds. a 2021 plos medicine review of high-income countries found current alcohol use disorders at 36.7% and drug use disorders at 21.7%. the only figure near two-thirds that we found was 61.6%, from a canadian cohort that enrolled only homeless adults who already had a mental illness.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| about two-thirds of homeless people have a current diagnosable mental illness | confirmed | current prevalence of any mental health disorder was 67% (95% ci 55-77) across 85 studies worldwide (2024), and 76.2% (95% ci 64.0-86.6) across 8 studies in high-income countries (2021). heterogeneity is very high (i² 88-98%), and most studies come from the us, canada and germany. [1][2] |
| about the same share, roughly two-thirds, have a current substance use disorder | overstated | the pooled current substance use disorder figure is 44% (95% ci 30-59). even the upper bound falls short of two-thirds, and the lifetime figure is 56% (95% ci 43-68). in high-income countries the pooled figures are 36.7% for alcohol use disorders and 21.7% for drug use disorders. these two overlap, so they cannot simply be added together. [1] |
| a figure near two-thirds for substance dependence exists in canadian homeless data | misattributed | 61.6% (977 of 1,585) met interview criteria for alcohol and/or substance dependence in the past 12 months. but every participant had to have a mental disorder to enrol, so this is a rate among homeless people with mental illness, not among homeless people in general. [3] |
| the mental illness figure and the substance use figure describe two separate conditions | unverifiable | the 2024 review calls substance use disorder 'the most prevalent mental health disorder' but never says in its methods whether substance use disorder counts toward the 67% any-disorder figure. the 2021 review defines any current disorder as dsm axis i, and the dsm axis system placed substance-related disorders on axis i. so the two-thirds figure probably includes addiction rather than sitting alongside it, but neither paper says so outright. [2] |
both directions
for the speaker: the mental illness figure matches the 2024 global meta-analysis almost exactly (67% current), and the high-income-country estimate is higher still (76.2%). substance use disorder is the single most common diagnosis in both reviews. its lifetime prevalence (56%) and the 61.6% past-year dependence rate in a canadian multi-site cohort both come close to the higher number. the 61.6% cohort is the kind of population a toronto clinician is likely to see.
against: the claim says 'current'. on current substance use disorder, the pooled estimate is 44%, and even the top of its confidence interval (59%) is below two-thirds. the high-income review's separate alcohol (36.7%) and drug (21.7%) figures also do not reach two-thirds. the near-two-thirds canadian figure comes from a sample that had to have a mental illness to enrol, which inflates substance dependence compared with homeless people in general. also, if the 67% mental illness figure already counts substance use disorder, then presenting the two numbers side by side suggests a larger total burden of separate conditions than the data show.
what this cannot say
pooled estimates rest on very different studies (i² 88-99%). only 16 studies fed the 2024 current any-disorder estimate, and only 8 fed the 2021 one. the authors say sampling misses the 'hidden homeless', and that studies relying on health care records probably undercount. the reviews mix shelter, street and service-based samples, different instruments and different time windows. results vary by city and region, so a local figure could be higher than the pooled one. the speaker named no source, so we cannot rule out a local survey we did not find. we did not read every table in the 2024 review, and the fetched text never says whether substance use disorder is counted inside its 67%.
sources, each one fetched and read
- The prevalence of mental disorders among homeless people in high-income countries: an updated systematic review and meta-regression analysis. PLOS Medicine, 2021, 18(8):e1003750 What we read there: any current mental disorder 76.2% (95% ci 64.0-86.6, 8 studies, i² 88%); alcohol use disorders 36.7% (95% ci 27.7-46.2, 29 studies); drug use disorders 21.7% (95% ci 13.1-31.7, 23 studies); schizophrenia spectrum 12.4%; major depression 12.6%; personality disorders 25.4% Fetched 2026-10-22 (HTTP 200). https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1003750
- same review, full text on PubMed Central (PMC8423293) What we read there: defines 'any current mental disorder (Axis I disorders in the Diagnostic and Statistical Manual of Mental Disorders [DSM]'; 'The high burden of substance use disorders and schizophrenia spectrum disorders need particular attention in service development.' Fetched 2026-10-22 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC8423293/
- comorbidity of alcohol and/or substance dependence and major psychiatric diagnoses among homeless adults with mental illness, At Home/Chez Soi baseline. Frontiers in Psychiatry, 2018, doi:10.3389/fpsyt.2018.00701 What we read there: 977 of 1,585 participants (61.6%) met interview criteria for alcohol and/or substance dependence in the previous 12 months; a diagnosed mental disorder was required for enrolment Fetched 2026-10-22 (HTTP 200). https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2018.00701/full
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam