Background Internationally there is a large treatment gap for common mental disorders and disparities across population groups but little understanding about whether people occupying multiple advantaged/disadvantaged statuses access mental health support differentially. This study examines discrimination and mental health service use (MHSU) at the intersections of social statuses.

We hypothesised i) greater past-year discrimination among disadvantaged social statuses; ii) different patterns of MHSU/treatment among single, compared to multiple social status groups; and after accounting for need: iii) less MHSU/treatment among multiply disadvantaged than more advantaged social status groups; and, iv) less MHSU/treatment among those reporting discrimination.

Methods English population-based data came from the 2014 Adult Psychiatric Morbidity Survey. Latent Class Analysis (LCA) was used to define intersectional social status groups. Multivariate logistic regression models were estimated to examine associations between MHSU/treatment and i) single social statuses and ii) latent classes of social status by gender, adjusting for confounders and need (physical/mental illness), followed by exposure to past year discrimination.

Results Five-class LCA solutions were selected for men and women, characterised as: 1) retired White British; 2) employed migrants, 3) economically inactive migrants 4) employed White British, lower education/social class; and 5) employed White British, high education/social class. Discrimination was more common among disadvantaged groups, and patterns differed across single and intersectional analyses. After adjustments, MHSU/treatment was elevated among females (OR 1.88, 95% CI: 1.49–2.06) and sexual minorities (OR 1.63: 1.06–2.51) but lower among Black ethnic groups (OR 0.28: 0.14–0.56). Adjustments for discrimination attenuated associations for sexual minorities and Black respondents. Intersectionally, findings were similar by gender except the retired White British group, for whom MHSU/treatment was significantly elevated among women (OR 1.96: 1.32–2.90) but not men. Among women and men respectively, compared to the most advantaged group, greater odds of MHSU/treatment were found for ‘employed migrants’ (OR 2.50: 1.71–3.67; OR 2.56: 1.77–3.71), ‘economically inactive migrants’ (OR 4.47: 3.00–6.40; OR 4.60: 3.10–6.83) and ‘employed White British, low social class/education’ (OR 1.91: 1.32–2.79; OR 1.95: 1.43–2.63). Adjustments for discrimination had little influence.

Conclusion Accounting for need, MHSU/treatment disparities are apparent but differ when considering single, or multiple social statuses. Single status analyses mask discrepancies observed intersectionally, while intersectional data-driven analyses miss inequities by minority statuses which do not distinguish latent classes. For some groups, discrimination may elevate, rather than inhibit MHSU/treatment. To better inform policy and practice, research should incorporate multiple and mixed-methods approaches to identify complexities of social stratification processes.