The contribution of financial hardship, socioeconomic position and physical health to mental health problems among welfare recipients

Recent changes to income support payments for single parents have prompted public debate about the financial adequacy of Newstart allowance and other welfare payments. Currently, the Newstart Allowance for a single person with no dependents is $501.00 per fortnight. The inadequacy of these payments has been highlighted by studies estimating that recipients of unemployment payments spend 122% of their income on daily living expenses, and 75% of Newstart recipients are reported to live in extreme poverty. Given that common psychiatric disorders occur more frequently among welfare recipients relative to the general population, and that financial hardship and socioeconomic disadvantage are key correlates of mental disorders, the types of welfare reforms recently introduced can have real implications for population health and wellbeing. We recently published a longitudinal analysis of the Household Income and Labour Dynamics in Australia (HILDA) study documenting how income support receipt and payment transitions predict changes in mental health over time. The key findings included: i) people receiving disability, unemployment or single parenting payments had poorer mental health than those never receiving welfare payments; ii) people reported greater decline in their mental health after transferring to, or during their time in, receipt of these payments; and iii) their poorer mental health was not entirely explained by covariates such as the experience of financial hardship and low income. Here, we adopt a different analytic approach to this data and employ a variation of the ‘explained fraction’ technique to investigate the specific role of financial hardship in explaining the association between mental health problems and income support. The purpose of our analysis was firstly to estimate the proportion of the increased risk of mental health problems amongst recipients of unemployment, disability and single parenting payments that could be uniquely attributed to their poor financial status; and secondly, to compare this to the proportion of increased risk that was uniquely explained by other factors such as socioeconomic position and physical function. The sample comprised 11,701 HILDA participants (49.4% men) who were of working age (mean baseline age=39) and provided up to nine years of longitudinal data. Mental health problems were defined by scores less than 50 on the 5-item mental health scale from the SF36. This outcome measure has previously been demonstrated to be a valid proxy for common psychiatric disorders such as depression and anxiety in epidemiological research. Time-varying and time-invariant covariates included age, sex, year, income support payment, marital status, socioeconomic position (education, work history, housing tenure, and parental occupation), lifestyle risks (smoking status and alcohol use), physical functioning and financial status (household equivalised income and financial hardship). Work history was defined by the proportion of time in employment since first leaving full-time education. Four mutually exclusive and dummy-coded categories of current income support payment were defined (unemployment, disability, single parent, other). To demonstrate the overall increased risk of mental health problems associated with income support, we first fit random effects logistic regression models adjusted for age, sex, year and payment type (none, unemployment, disability, parenting payment single, other) – Model 1. In a second step we conducted a series of analyses that added factors individually to the initial model, estimating the overall proportion of the increased risk of payment type that was explained by each factor (e.g. financial status, socioeconomic position, physical function), without accounting for the shared effects all other covariates (Model 2). In a final step, full multivariate adjusted models were tested, excluding (Model 3) and including (Model 4) the key factors of interest: socioeconomic position, financial status and physical function. The proportion of the increased risk of mental health problems that was uniquely explained by each factor was calculated by contrasting the Odds Ratios (OR) for income support status from models that excluded and included each factor. This was achieved for the respective factors by taking the difference between the ORs for each payment estimated in Model 4 from the ORs for each payment estimated in Model 3.