Importance: The association between income and mental health has long been a question of interest. Nationwide register data provide means to examine trends and patterns of these associations.
Objectives: To compare income-specific trends in the incidence rates of first psychiatric hospital admissions and to evaluate whether an income gradient exists in the incidence rates at all levels of household income.
Design, setting, and participants: This population-based open cohort study used linked registry data from nationwide Finnish Hospital Discharge and Statistics Finland population registers to determine annual incidence rates of first psychiatric hospital admissions. All Finnish citizens (N = 6 258 033) living in the country at any time from January 1, 1996, through December 31, 2014, contributed to 96 184 614 person-years at risk of first inpatient treatment for mental disorders. The analyses were conducted from August 1, 2018, through September 30, 2019.
Exposures: Equivalized disposable income, sex, age group, reduction in income decile in the previous 3 years, urbanicity, educational level, and living alone status.
Main outcomes and measures: Annual percentage changes in the age-standardized incidence rates and incidence rate ratios (IRRs).
Results: Altogether, 186 082 first psychiatric inpatient treatment episodes occurred (93 431 [50.2%] men), with overall age-standardized incidence rates per 1000 person-years varying from 1.59 (95% CI, 1.56-1.63) in 2014 to 2.11 (95% CI, 2.07-2.15) in 2008. In the highest income deciles, a continuous mean decrease per year of 3.71% (95% CI, 2.82%-4.59%) in men and 0.91% (95% CI, 0.01%-1.80%) in women occurred throughout the study period, in contrast to the lowest deciles, where the trends first increased (1.31% [95% CI, 0.62%-2.01%] increase in men from 1996 to 2007 and 5.61% [95% CI, 2.36%-8.96%] increase in women from 1996 to 2001). In the adult population, an income gradient was observed at all levels of household income: the lower the income decile, the higher the adjusted IRRs compared with the highest decile. The IRRs in the lowest decile varied from 2.94 (95% CI, 2.78-3.11) to 4.46 (95% CI, 4.17-4.76). In other age groups, the gradient did not persist at the highest income deciles. Diagnosis-specific income gradient was steepest in schizophrenia and related psychotic disorders, with estimated IRRs of the lowest income decile of 5.89 (95% CI, 5.77-6.02).
Conclusions and relevance: In this cohort study, clear negative income gradient in the incidence rates of first hospital-treated mental disorders was observed in the adult population of Finland. These findings suggest that reduction in the use of inpatient care has not taken place equally between different income groups.