Older adults who present with self-harm are significantly more likely to die from unnatural causes during the first year following self-harm incidence, per cohort data published in The Lancet.

Investigators abstracted data from the UK Clinical Practice Research Datalink, which contains anonymized patient records from 674 general practices around the United Kingdom. The data the Research Datalink captures is broadly representative of the national population per age, gender, and ethnicity. The present study cohort comprised adults 65 and older with a self-harm episode recorded during 2001 to 2014. Self-harm episodes were identified by Read Codes from patients’ electronic primary care health records, including information captured during a general practitioner consultation or inpatient hospital care. 

Self-harm inclusion criteria comprised both milder forms of non-suicidal self-injury to near-fatal attempted suicide. Researchers first calculated annual incidence of self-harm in older individuals, then monitored patients for clinical outcomes following self-harm. Psychiatric referrals, prescription of psychotropic medication, psychiatric diagnoses, and physical illness diagnoses following self-harm were captured. Using Cox regression in a matched cohort of older adults without self-harm, researchers also estimated cause-specific mortality risks.

A total of 4124 older adults were identified as having an incidence self-harm episode, of which 3327 (80.7%) involved ingestion of drugs, 234 (5.7%) self-cutting, 118 (2.9%) were by other means, and 435 (10.6%) with no specified method. The overall incidence was 4.1 per 10,000 person-years, with stable gender-specific incidences observed across the 13-year study period. Compared with the reference group of adults age 65 to 74, incidence of self-harm increased by 31% in adults age 75 to 84 (rate ratio 1.31; 95% CI, 1.23-1.41) and by 76% in individuals 85 and older (rate ratio 1.76; 95% CI, 1.62-1.91).  After self-harm, 335 of 2854 adults (11.7%) were referred to mental health services, 1692 (59.3%) were prescribed an antidepressant and 336 (11.8%) were prescribed a tricyclic antidepressant. Women were referred more often to mental health services than men (13.1% vs 9.7%; P =.005).

Related Articles

In a subsequent Cox regression analysis, individuals from the initial analyses (n=2454) were matched with individuals from a comparison cohort (n=48,921). Compared with the comparison cohort, a prior diagnosis of mental illness was twice as prevalent in the self-harm cohort (prevalence ratio 2.10; 95% CI, 2.03-2.17). In a similar fashion, the prevalence of a previous physical health condition was 20% higher in the self-harm cohort compared with the comparison cohort (prevalence ratio 1.20; 95% CI, 1.17-1.23). Adults from the self-harm cohort died from unnatural causes an estimated 20 times more frequently than the comparison cohort during the first year. In addition, a significantly elevated risk for suicide (hazard ratio 145.4; 95% CI, 53.9-392.3) was observed in the self-harm cohort compared with the comparison cohort.

These data highlight the significant burden of mental and physical illness experienced by older adults who present with self-injury. As these individuals are also at significantly elevated risk for suicide and death from unnatural causes, rigorous screening for depressive symptoms, suicide ideation, and self-harm behavior is crucial. Healthcare professionals should also be aware of the elevated risk for self-harm in older individuals with existing physical or mental illness.


Morgan C, Webb RT, Carr MJ, et al. Self-harm in a primary care cohort of older people: incidence, clinical management, and risk of suicide and other causes of death [published online October 15, 2018]. Lancet. doi:10.1016/S2215-0366(18)30348-1