Factors Associated With Hospital Presentation of Self-Harm Among Older Canadians in Long-Term Care: A 12-Year Cohort Study

Abstract

Objectives: This study aimed to examine the incidence of, and factors associated with, hospital presentation for self-harm among older Canadians in long-term care (LTC).

Design: Retrospective cohort study.

Setting and Participants: The LTC data were collected using Resident Assessment Instrument–Minimum Data Set (RAI-MDS) and Resident Assessment Instrument–Home Care (RAI-HC), and linked to the Discharge Abstract Database (DAD) with hospital records of self-harm diagnosis. Adults aged 60+ at first assessment between April 1, 2003, and March 31, 2015, were included.

Methods: Adjusted hazard ratios (HRs) of self-harm for potentially relevant factors, including demographic, clinical, and psychosocial characteristics, were calculated using Fine & Gray competing risk models.

Results: Records were collated of 465,870 people in long-term care facilities (LTCF), and 773,855 people receiving home care (HC). Self-harm incidence per 100,000 person-years was 20.76 [95% confidence interval (CI) 20.31–25.40] for LTCF and 46.64 (44.24–49.12) for HC. In LTCF, the strongest risks were younger age (60–74 years vs 90+: HR, 6.00; 95% CI, 3.24–11.12), psychiatric disorders (bipolar disorder: 3.46; 2.32–5.16; schizophrenia: 2.31; 1.47–3.62; depression: 2.29; 1.80–2.92), daily severe pain (2.01; 1.30–3.11), and daily tobacco consumption (1.78; 1.29–2.45). For those receiving HC, the strongest risk factors were younger age (60–74 years vs 90+: 2.54; 1.97–3.28), psychiatric disorders (2.20; 1.93–2.50), daily tobacco consumption (2.08; 1.81–2.39), and frequent falls (1.98; 1.46–2.68). All model interactions between setting and factors were significant.

Conclusions and Implications: There was lower incidence of hospital presentation for self-harm for LTCF residents than HC recipients. We found sizable risks of self-harm associated with several modifiable risk factors, some of which can be directly addressed by better treatment and care (psychiatric disorders and pain), whereas others require through more complex interventions that target underlying factors and causes (tobacco and falls). The findings highlight a need for setting- and risk-specific prevention strategies to address self-harm in the older populations.