Severe mental illness (SMI) is associated with increased stroke risk, but little is known about how SMI relates to stroke prognosis and receipt of acute care.

To determine the association between SMI and stroke outcomes and receipt of process-of-care quality indicators (such as timely admission to stroke unit).

We conducted a cohort study using routinely collected linked data-sets, including adults with a first hospital admission for stroke in Scotland during 1991–2014, with process-of-care quality indicator data available from 2010. We identified pre-existing schizophrenia, bipolar disorder and major depression from hospital records. We used logistic regression to evaluate 30-day, 1-year and 5-year mortality and receipt of process-of-care quality indicators by pre-existing SMI, adjusting for sociodemographic and clinical factors. We used Cox regression to evaluate further stroke and vascular events (stroke and myocardial infarction).

Among 228 699 patients who had had a stroke, 1186 (0.5%), 859 (0.4%), 7308 (3.2%) had schizophrenia, bipolar disorder and major depression, respectively. Overall, median follow-up was 2.6 years. Compared with adults without a record of mental illness, 30-day mortality was higher for schizophrenia (adjusted odds ratio (aOR) = 1.33, 95% CI 1.16–1.52), bipolar disorder (aOR = 1.37, 95% CI 1.18–1.60) and major depression (aOR = 1.11, 95% CI 1.05–1.18). Each disorder was also associated with marked increased risk of 1-year and 5-year mortality and further stroke and vascular events. There were no clear differences in receipt of process-of-care quality indicators.

Pre-existing SMI was associated with higher risks of mortality and further vascular events. Urgent action is needed to better understand and address the reasons for these disparities.