부산대학교 도서관

Background: People with severe mental ill health often die prematurely, and suicide risk is known to be particularly elevated after discharge from inpatient mental health care. Few studies have investigated the risk of suicide alongside other causes of death after discharge and how risk differs from the general population. It is also unclear how post-discharge mortality risk compares to the risk associated with poor mental health more generally. Although early follow-up and support from health services has been recommended there are noticeable gaps in research into primary care provision at this time. Methods: Routinely collected data from the Clinical Practice Research Datalink (CPRD) in England from 2001 to 2019 was used to develop three interlinked studies to address these gaps. Outcomes included all-cause mortality, natural and external causes of death, suicide, accidental death, alcohol specific and drug related death. Two studies used matched cohort designs comparing a discharged cohort to a) working-age and older adults in the general population and b) adults with severe mental illness (SMI) being treated in the community. The cumulative incidence (absolute risk) of each outcome at one year was calculated and relative risks estimated using stratified Cox proportional hazards models. The final study utilised a nested case-control design. Conditional logistic regression was used to assess variation in a range of primary care consultation and clinical management variables between discharged patients who died by suicide and those still alive 1 year after discharge. Results: The risk of death for patients in the year after discharge was higher compared to the general population and those with SMI in the community for all causes examined. The risk of suicide was greatly elevated in the first three months (HR 191.1, 95% CI 125.0-292.0 in working-age adults vs. the general population). Older adults were also more vulnerable to dying by natural causes in the first three months. The risk of suicide in the discharged compared to community SMI groups (HR 11.6, 95% CI 6.4-20.9) was not explained by recent self-harm. Most discharged patients who died by suicide consulted with their GP at least once. Those who died were less likely to consult in the first two weeks, but more likely in the week before death, were prescribed more psychotropic medication and were more likely to be readmitted. Hospital discharge information was not identified for most patients. Conclusions: This work has highlighted the enduring risk of suicide and death by other causes in people recently discharged from hospital not only compared to the general population but also to those with SMI in the community. It identified additional risks of death by natural causes in older adults in the first few months. Therapeutic inpatient care and follow-up after discharge, including supporting physical health, is needed. It has shown that discharged patients who die by suicide do access primary care providing opportunities for intervention. The role of primary care in post-discharge guidance should be more clearly defined.