Relative risk and risk factors of all-cause, natural-cause and external-cause mortality after psychiatric hospitalisation: a population-based electronic health-record study in Hong Kong.
Discharge from psychiatric hospitalisation is associated with increased mortality risk, but risk estimates across fine-grained post-discharge intervals and identified risk factors are lacking.
We conducted a population-based study utilising a medical health record database of public healthcare services in Hong Kong. We identified all individuals aged ≥12 years who had been discharged from psychiatric inpatient units between January-01-2006 and December-31-2021. Individuals without history of psychiatric hospitalisation attending primary care clinics during the same period served as comparison group (i.e., primary-care clinics [PCC] cohort). Survival analyses estimated cumulative incidences and hazard ratios (HRs) for all-cause, natural-cause and external-cause mortality across discrete intervals from ≤1 month to >10 years post-discharge. Risk factors for excess mortality 1-year and 10-year post-discharge were evaluated.
The discharged cohort comprised 114,395 patients with 975,229 controls. Within the first-year post-discharge, within the deceased patient group, natural causes and suicide-related mortality accounted for 65.8% and 18.8% of deaths in the discharged cohort, and for 87.8% and 1.6% of deaths in the PCC cohort. The relative proportion of all deaths due to suicide in the discharged cohort decreased from 18.8% at 1-year to 11.7% at 10-year, while the relative proportion of all deaths from natural causes increased from 65.8% at 1-year to 73.3% at 10-year. Respiratory diseases (year 1 = 30.6%, year 10 = 35.0%) and cardiovascular diseases (year 1 = 10.7%, year 10 = 11.0%) were consistently leading causes of natural deaths. Relative all-cause mortality risk peaked within 1-month post-discharge (HR = 10.31 [95% confidence interval (CI) = 9.36-11.36]), remaining high at 1-year (HR = 3.40 [95% CI = 3.17-3.65]) and 10-year (HR = 2.61 [95% CI = 2.52-2.69]). External-cause mortality, particularly suicide-related mortality, showed extremely high risk during 1-month (HR = 78.54 [95% CI = 52.64-117.18]), and remained high at 1-year (HR = 42.40 [95% CI = 30.02-59.87]) and 10-year (HR = 10.29 [95% CI = 8.75-12.10]). Depression diagnosis at discharge conferred the highest short-term suicide risk, while substance/alcohol use disorders showed the highest long-term any-cause mortality risk. Risk factors for premature mortality within 1-year and 10-year post-discharge included male sex, old age, not receiving social/disability allowance, greater physical morbidity burden, self-harm history, and poor discharge psychiatric status. Ethnicity data were not available in the database.
The immediate post-discharge phase is a critical window for targeted and diagnosis-specific intervention, including psychiatric/social/financial support. Future research should prioritise the development of robust risk prediction models that incorporate key risk factors identified to facilitate personalised care and improves post-discharge outcomes for high-risk individuals.
Health and Medical Research Fund.
We conducted a population-based study utilising a medical health record database of public healthcare services in Hong Kong. We identified all individuals aged ≥12 years who had been discharged from psychiatric inpatient units between January-01-2006 and December-31-2021. Individuals without history of psychiatric hospitalisation attending primary care clinics during the same period served as comparison group (i.e., primary-care clinics [PCC] cohort). Survival analyses estimated cumulative incidences and hazard ratios (HRs) for all-cause, natural-cause and external-cause mortality across discrete intervals from ≤1 month to >10 years post-discharge. Risk factors for excess mortality 1-year and 10-year post-discharge were evaluated.
The discharged cohort comprised 114,395 patients with 975,229 controls. Within the first-year post-discharge, within the deceased patient group, natural causes and suicide-related mortality accounted for 65.8% and 18.8% of deaths in the discharged cohort, and for 87.8% and 1.6% of deaths in the PCC cohort. The relative proportion of all deaths due to suicide in the discharged cohort decreased from 18.8% at 1-year to 11.7% at 10-year, while the relative proportion of all deaths from natural causes increased from 65.8% at 1-year to 73.3% at 10-year. Respiratory diseases (year 1 = 30.6%, year 10 = 35.0%) and cardiovascular diseases (year 1 = 10.7%, year 10 = 11.0%) were consistently leading causes of natural deaths. Relative all-cause mortality risk peaked within 1-month post-discharge (HR = 10.31 [95% confidence interval (CI) = 9.36-11.36]), remaining high at 1-year (HR = 3.40 [95% CI = 3.17-3.65]) and 10-year (HR = 2.61 [95% CI = 2.52-2.69]). External-cause mortality, particularly suicide-related mortality, showed extremely high risk during 1-month (HR = 78.54 [95% CI = 52.64-117.18]), and remained high at 1-year (HR = 42.40 [95% CI = 30.02-59.87]) and 10-year (HR = 10.29 [95% CI = 8.75-12.10]). Depression diagnosis at discharge conferred the highest short-term suicide risk, while substance/alcohol use disorders showed the highest long-term any-cause mortality risk. Risk factors for premature mortality within 1-year and 10-year post-discharge included male sex, old age, not receiving social/disability allowance, greater physical morbidity burden, self-harm history, and poor discharge psychiatric status. Ethnicity data were not available in the database.
The immediate post-discharge phase is a critical window for targeted and diagnosis-specific intervention, including psychiatric/social/financial support. Future research should prioritise the development of robust risk prediction models that incorporate key risk factors identified to facilitate personalised care and improves post-discharge outcomes for high-risk individuals.
Health and Medical Research Fund.
Authors
Chan Chan, Lui Lui, Choo Choo, Chu Chu, Chung Chung, Wong Wong, So So, Sham Sham, Solmi Solmi, Correll Correll, Chang Chang
View on Pubmed