Citation

Milner A, Disney G, Byars S, King TL, Kavanagh AM, Aitken Z. Soc. Psychiatry Psychiatr. Epidemiol. 2020; ePub(ePub): ePub.

Affiliation

Disability and Health Unit, Melbourne School Population and Global Health, Centre for Health Equity, The University of Melbourne, Victoria, 3010, Australia. zoe.aitken@unimelb.edu.au.

Copyright

(Copyright © 2020, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s00127-020-01844-6

PMID

32055895

Abstract

PURPOSE: We aimed to understand how much of the gender difference in mental health service use could be due to the joint mediation of employment, behavioural and material factors, social support and mental health need.

METHODS: We used data from employed individuals aged 18-65 years who participated in the 2015-2017 waves of the Household, Income and Labour Dynamics in Australia survey. The exposure (male, female) and confounders were measured in 2015, mediators in 2016 and the outcome-whether a person had seen a mental health professional in the previous year-was measured in 2017. We estimated natural mediation effects using weighted counterfactual predictions from a logistic regression model.

RESULTS: Men were less likely to see a mental health care provider than women. The total causal effect on the risk difference scale was  - 0.045 (95% CI  - 0.056,  - 0,034). The counterfactual of men taking the mediator values of women explained 28% (95% CI 1.7%, 54%) of the total effect, with the natural direct effect estimated to represent an absolute risk difference of  - 0.033 (95% CI  - 0.048,  - 0.018) and the natural indirect effect  - 0.012 (95% CI  - 0.022,  - 0.0027).

CONCLUSION: Gendered differences in the use of mental health services could be reduced by addressing inequalities in health, employment, material and behavioural factors, and social support.

Language: en

Keywords

Causal mediation analysis; Counterfactual; Gender; Mental health service use