Survival following psychiatric diagnoses in early adulthood

Abstract

Aims: To establish the increased all-cause mortality risk after an inpatient episode of care with a diagnosis of a severe psychiatric disorder in young people.

Methods: The data included all psychiatric inpatient episodes for psychiatric diagnoses in Western Australia between 2005 and 2022 linked with the state death registry. Participants were only included if they turned 18 years of age between 2005 and 2016, so survival from first adult admission until the study end date could be compared with age-gender matched life tables.

Results: A total of 18,893 individuals had an admission with a primary or secondary diagnosis for a selected psychiatric diagnosis in the study period, across which time 485 died. Admission for substance use disorders presented the greatest risk of mortality, increasing the risk of death in early adulthood by more than three times (observed/expected = 3.07; 95% confidence interval = [2.76, 3.42]; p < 0.001), followed closely by bipolar disorders (observed/expected = 2.95; 95% confidence interval = [2.09, 4.03]; p < 0.001), while having any two or more comorbid disorders was associated with an increased death rate (observed/expected = 3.30; 95% confidence interval = [2.72, 3.97]; p < 0.001). The Kaplan-Meier curves also suggested that the proportionate increased risk of mortality remained relatively constant across the study period for all diagnoses.

Conclusion: Inpatient admission for psychiatric disorders increased the risk of all-cause mortality in early adulthood by between two and three times and the increased death rate did not substantively reduce over time. Effective long-term support services are needed to reduce the premature mortality observed among these young adults.

Keywords: Mortality; early adulthood; inpatient admission; psychiatric diagnosis.

Figure 1.

(a) Survival curve of all-cause mortality among those diagnosed with psychotic disorders in an inpatient episode of care versus the reference population (n = 3754, total years of life since DX = 25,375, observed deaths = 98, expected deaths = 46.5). (b) Survival curve of all-cause mortality among those diagnosed with SUD in an inpatient episode of care versus the reference population (n = 10,714, total years of life since SUD DX = 75,958, observed deaths = 345, expected deaths = 112.2). (c) Survival curve of all-cause mortality among those diagnosed with bipolar in an inpatient episode of care versus the reference population (n = 1658, total years of life since BPD DX = 10,588, observed deaths = 39, expected deaths = 13.2) and (d) Survival curve of all-cause mortality among those diagnosed with depressive disorders in an inpatient episode of care versus the reference population (n = 6454, total years of life since DEP DX = 42,905, observed deaths = 135, expected deaths = 58.9). Y-axis ranges from 100% to 90% survival.

Figure 2.

Survival curve of all-cause mortality among those diagnosed with two or more different disorders in two or more inpatient episode of care versus the reference population (n = 3190, total years of life since the first diagnosis DX = 26,364, observed deaths = 112, expected deaths = 33.9).