A study of 1,070 adults discharged from acute psychiatric units in Alberta found that emergency-department use before admission and unstable housing were among the clearest markers of a psychiatric emergency visit during the following year.
Forty percent of participants had at least one psychiatric ED visit for mental-health reasons within 12 months of discharge. The participants had been treated in Edmonton, Calgary or Grande Prairie between March 2022 and February 2024.
The results point to the importance of understanding a person’s circumstances after discharge, not just their diagnosis or experience during the hospital stay. They do not, however, show that changing housing or employment would itself prevent a future emergency visit.
What the Alberta study found
People who had visited a psychiatric ED in the six months before their admission had 3.72 times the odds of a psychiatric ED visit in the 12 months after discharge. The 95% confidence interval was 1.82 to 7.63.
Unstable housing was also associated with higher odds of a return visit. Participants who were living in a shelter, couch-surfing or living on the street had 2.60 times the odds of a return psychiatric ED visit compared with homeowners. The confidence interval was 1.29 to 5.22.
Unemployment was associated with 1.57 times the odds of a return visit compared with employment. Renters had an odds ratio of 1.48 compared with homeowners, but that estimate was not statistically significant. Students and retirees did not differ from employed participants in the analysis.
These are odds ratios, not a person’s individual probability of returning to the ED. The study identified patterns across groups; it could not determine what would happen to any one person.
Clinical and personal factors were not significant in the full model
Once 14 predictor groups were modeled together, depression, anxiety, resilience, well-being, diagnosis and satisfaction with the hospital stay were not significant predictors. Their odds ratios were close to 1.
Most participants—89%—said they were satisfied or very satisfied with their hospital stay. Participants who were dissatisfied had an odds ratio of 1.33 for a return visit, but the confidence interval, 0.82 to 2.13, included no clear difference.
The full analysis included 929 participants and explained only about 8.5% to 11.4% of the variation in who returned to the ED. That limited explanatory power means the factors should be viewed as risk markers rather than reliable forecasts.
Why housing may matter for follow-up
A person leaving psychiatric hospital care may face practical barriers that are not captured by a diagnosis or a measure of hospital satisfaction. Housing instability can make it harder to maintain a routine, store medication, receive messages or attend planned follow-up. The Alberta analysis did not measure all of these circumstances, so it cannot establish which pathways were most important.
Other potentially relevant factors—including substance use, medication adherence, social support, physical illness and contact with an outpatient clinician—were not included in the model. Only 66 participants were unstably housed, and the relatively wide confidence interval means the housing estimate should be interpreted cautiously.
Related findings from other settings underline the concern without proving the same pattern applies everywhere. In Ontario, 27.2% of people who were homeless at psychiatric discharge visited an ED within 30 days, compared with 11.6% of housed patients. In New South Wales, 37% of 35,056 people discharged from a first psychiatric admission visited an ED at least once within two years, while 40% were readmitted. These studies used different populations and time periods.
What the study can—and cannot—say about support
The Alberta participants came from Text4Support, a larger trial evaluating daily automated text messages, either alone or combined with peer support, after psychiatric hospitalization. The messages covered coping strategies, hope, medication reminders and seeking help early.
The study of predictors was not designed to test whether those interventions reduced 12-month ED visits. In that analysis, usual care and texts plus peer support each had an odds ratio of 0.79 compared with texts alone, with confidence intervals crossing 1.
A separate 2025 effectiveness analysis reported fewer readmissions with text messaging, with or without peer support, than with usual care, with a mean difference of 0.26 admissions. It also reported that texts alone reduced time in hospital by an average of 7.28 days. Readmissions, hospital time and ED visits were different outcomes measured over different time windows, so those results should not be treated as interchangeable.
What readers should take from the findings
- A previous psychiatric ED visit was the strongest association reported in the Alberta analysis.
- Unstable housing and unemployment marked groups with higher odds of another visit, but the study did not prove that housing or employment changes would prevent one.
- The analysis did not support using diagnosis, depression, anxiety or hospital satisfaction alone to predict who would return.
- The model left most of the variation in future ED visits unexplained.
- The sample required a mobile phone and the ability to read English texts, which may have left out some people facing the greatest barriers to care.
For services planning discharge support, the findings suggest that prior emergency use and a person’s housing situation deserve attention alongside clinical information. They do not provide an individual forecast or establish how much providing housing at discharge would change future emergency-department use.
AI tools were used to assist with the preparation of this article.
Leave a Reply