Treatment-resistant late-life depression prevalence and clinical/sociodemographic correlates: An electronic health records study.

BACKGROUND

Treatment-resistant late-life depression (TRLLD) remains a clinical challenge. Our aim was to identify demographic and clinical factors associated with TRLLD.

METHODS

A large naturalistic retrospective cohort study was conducted using mental health records from south London. Patients were selected at first recorded depression diagnosis after the age of 60. We defined TRLLD when ≥3 antidepressants were prescribed across their mental health record or a natural language processing algorithm indicated depression was described as treatment-resistant. We collected demographic and clinical characteristics around the first depressive episode diagnosis at age ≥ 60 and used multivariate logistic regression models to investigate factors associated with TRLLD.

RESULTS

Of 8171 patients with late-life depression, 1443 (17.7 %) had TRLLD. Amongst those with severe, psychotic or recurrent depression the prevalence of TRLLD was 30.7 %, 31.4 % and 27.0 % respectively. Female sex, recurrent, severe, or psychotic depression and higher self-harm risk were associated with higher odds of TRLLD after adjustment for demographics (Adjusted Odds ratio (aOR): 1.36, 2.05, 2.34, 2.21 and 1.30 respectively). Older age, Black ethnicity, cognitive impairment and difficulties in activities of daily living were associated with a lower likelihood of TRLLD (aOR: 0.97, 0.65, 0.71 and 0.76 respectively). Several physical conditions examined were associated with increased odds for TRLLD, the strongest association was with hypertension (aOR: 2.20).

CONCLUSIONS

This is the first large-scale study examining the socio-demographic data as well as psychiatric and physical comorbidities in TRLLD without known dementia. Prevention and management of multi-morbidity should be considered in primary or secondary prevention of treatment-resistant depression.