38 Chapter 2 inequalities in mental health treatment outcomes. Existing evidence was limited to studies reporting inequalities in access to mental health care or disparities in treatment outcomes observed in randomised controlled trial [3, 15, 17, 18, 29]. We addressed one notable challenge in the observational literature, which is the unavailability of data on mental health need combined with information on treatment intensity, treatment outcomes, and income. We overcame this limitation by using detailed measures of diagnosis and baseline severity recorded by mental health care professionals. Additionally, our study has the advantage of covering the entire patient population in Dutch specialist mental health care and exploring patterns across diagnoses. Our findings contribute to the policy debate on the accessibility of specialist mental health services by emphasising the need for early detection and targeting to prevent lowincome individuals from first accessing specialist mental health care in a state of lower functioning. Furthermore, our results raise three hypotheses relevant to policy. First, that individuals of a low socioeconomic status require more treatment minutes and courses of treatment to obtain equivalent outcomes compared with higher income people. This scenario could be the case because of low treatment adherence or a higher likelihood to discontinue care [30-32]. Financial barriers might prevent low-income individuals from receiving the appropriate amount of treatment. Previous studies show the relevance of out-of-pocket payments in limiting access to care in the Netherlands [33-35], but there is little evidence about effects of financial barriers on the amount of care used. Nevertheless, we hypothesise that the deductible is more relevant in an individual’s decision to start rather than to continue care, because its amount is easily surpassed by the initial expenditures of starting a record, and individuals are aware of their low ability to control expenditures during the treatment. Second, that the effectiveness of the treatment provided is lower for patients with lower incomes [20, 21], suggesting complementarities between treatment, personality traits, and social environment [36]. A third hypothesis is that low-income individuals receive a lower quality of care. Future research should test these hypotheses, because each has different policy implications. Our study has several limitations. First, we focused on the first observed treatment record in our data. This method does not rule out that patients received treatment before the lookback period. Second, there are ongoing debates on the clinical appropriateness of using GAF [25, 37]. However, in our study, each patient was evaluated by the same medical professional over time, which addresses some of the concerns in the literature. Furthermore, together with diagnosis, GAF is the most relevant indicator of severity and has been widely used internationally. A related limitation is that the mandatory recording of GAF in Dutch specialist mental health treatment records stopped in 2014, resulting in more missing data towards the end of our study period (Table A2). Nevertheless, our results and conclusions did not change when we only used data for the period when GAF recording was compulsory (Table A16). We studied the amount of treatment minutes provided, which is only one aspect of quality of care, and we measured additional
RkJQdWJsaXNoZXIy MjY0ODMw