Income inequalities beyond access to mental health care: a Dutch nationwide record-linkage cohort study of baseline disease severity, treatment intensity, and mental health outcomes.

Background: Existing literature shows low and unequal access to mental health treatment globally, resulting in policy efforts to promote access for vulnerable groups. Yet, there is little evidence about how inequalities develop once individuals start treatment. The greater use of mental health care among individuals with low income, such as in the Dutch system, might be driven by differences in need and might not necessarily lead to better treatment outcomes. In this study, we aimed to examine income inequalities in four stages of the mental health treatment pathway while adjusting for need.

Methods: We constructed a nationwide retrospective cohort study, examining all patients aged older than 18 years with a first specialist mental health treatment record in the Netherlands between 2011 and 2016, excluding those who did not receive any treatment minutes. We linked patient-level data from treatment records to administrative data on income, demographics from municipal registries, and health insurance claims. We used multivariate models to estimate adjusted associations between household income quintile (standardised for household size) and outcomes characterising four stages of mental health treatment: severity at baseline assessment based on the Global Assessment of Functioning (GAF) score, treatment minutes received, functional improvement by the end of the initial record, and additional treatment in a subsequent record. Estimates were adjusted for patient need (97 categories of primary diagnosis and severity at baseline assessment measured by GAF) and demographic covariates.

Findings: Our study population consisted of 951 530 adults with a first specialist mental health treatment record in the Netherlands between Jan 1, 2011, and Dec 31, 2016. Patients in our cohort were on average aged 45·0 years (range 19-107) and mostly female (529 859 [55·7%] women and 421 671 [44·3%] men; no ethnicity data were available). First, we found that patients with the lowest income had the greatest initial therapist-assessed disease severity (5·545 GAF points), which was 0·353 GAF points (95% CI 0·347-0·360) lower than those in the highest income quintile. Second, we found that the negative association between income and treatment minutes was reversed once we adjusted for diagnosis and severity at baseline, with patients with the lowest income receiving 1·8% fewer treatment minutes (95% CI 1·1-2·4) than those in the highest quintile. Third, those in the highest income quintile were 17·3 percentage points (95% CI 17·0-17·6) more likely to have functional improvements by the end of the initial record, compared with 25·8% of patients with an improvement in the lowest income quintile. Fourth, while 35·7% of patients in the lowest income quintile received additional treatment in a subsequent record, this was only 3·0 percentage points (95% CI 2·7-3·3) lower for those in the highest quintile. None of these patterns were explained by diagnosis, severity at baseline, or treatment minutes received.

Interpretation: Disparities favourable to patients with a higher income persist through the different stages of mental health treatment. These differences highlight the limitations of solely focusing on improving access to care to reduce the mental health gap. Our findings call for a better understanding of the role of social environment and quality of care as complementary mechanisms explaining inequalities during mental health treatment.

Funding: Erasmus Initiative Smarter Choices for Better Health (Erasmus University Rotterdam), European Union's Horizon 2020, and Nederlandse Organisatie voor Wetenschappelijk Onderzoek (Dutch Research Council).

Translation: For the Dutch translation of the abstract see Supplementary Materials section.