Relationship between pharmacotherapy for major depressive disorder and healthcare utilisation in British Columbia, Canada: a retrospective population-based cohort.

Objectives: To describe the population that meets the criteria for major depressive disorder (MDD) in British Columbia (BC), compare patterns of healthcare utilisation between those with MDD who are and are not prescribed pharmacotherapy, and assess these relationships in models that control for potential confounding variables.

Design: We used a population cross-sectional study design among a cohort of individuals living with MDD and examined the relationship between pharmacotherapy and healthcare utilisation between 2019 and 2020 using linked billing and administrative data.

Setting: This study identified individuals with MDD using a validated case definition of International Classification of Diseases (ICD) codes in BC, Canada.

Participants: The final study cohort included 549 029 adult participants who met the MDD case definition.

Explanatory Variable: Explanatory variable was the use of prescription antidepressant medication during the study period, based on BC PharmaNet data.

Covariates: Covariates include sociodemographic characteristics (age, sex, urban/rural residence, neighbourhood income quintile and comorbidities).

Primary Outcome Measure: Primary outcome measure was healthcare utilisation (outpatient physician visits, emergency department (ED) visits and hospitalisations).

Results: We stratified our analysis based on whether study participants were classified as 'recently incident' or 'actively prevalent'. The odds ratio (OR) for health service utilisation between the pharmacotherapy group and the non-pharmacotherapy group for individuals who were recently incident was 8.14 (95% CI 7.40, 8.95) for outpatient physician visits, 1.04 (95% CI 1.02, 1.07) for ED visits and 1.05 (95% CI 1.00, 1.10) for hospitalisations, after adjusting for comorbidities and other sociodemographic variables in our regression analyses, whereas for the actively prevalent group the ORs were 7.57 (7.27, 8.49), 0.91 (0.89, 0.92) and 1.00 (0.98, 1.02), respectively.

Conclusion: This study examined the association of pharmacotherapy on healthcare utilisation for adults with MDD in BC. The study revealed higher outpatient physician visits for the pharmacotherapy group and no major association for inpatient visits. For ED visits, recently incident individuals on pharmacotherapy had slightly higher odds of having an ED visit, whereas individuals who were actively prevalent and on pharmacotherapy had a slight decrease in odds. This may suggest a protective effect of pharmacotherapy against a utilisation of resource-intensive healthcare services.