Abstract
BACKGROUND AND HYPOTHESIS: Recent research showed that young people who presented to hospital with self-harm in Finland had a significantly elevated risk of later psychosis. We investigated the prospective relationship between hospital presentation for self-harm and risk of psychosis in an unprecedentedly large national Swedish cohort. STUDY DESIGN: We used inpatient and outpatient healthcare registers to identify all individuals born between 1981 and 1993 who were alive and living in Sweden on their 12th birthday and who presented to hospital one or more times with self-harm. We compared them with a matched cohort, followed up for up to 20 years, and compared the cumulative incidence of psychotic disorders. Furthermore, we examined whether the strength of the relationship between hospital presentation for self-harm and later psychosis changed over time by examining for cohort effects. STUDY RESULTS: In total, 28 908 (2.0%) individuals presented to hospital with self-harm without prior psychosis diagnosis during the follow-up. For individuals who presented to hospital with self-harm, the cumulative incidence of diagnosed psychosis was 20.7% at 20 years follow-up (hazard radio = 13.9, 95% CI 13.3-14.6, P-value <5 × 10-308). There was no evidence of a dilution of the effect over time: while the incidence of hospital self-harm presentation increased, this did not result in an attenuation over time of the strength of the relationship between hospital self-harm presentation and subsequent psychosis. CONCLUSIONS: Individuals who present to hospital with self-harm in their teens and 20s represent an important risk group for psychosis prediction and prevention.