219 General Discussion opportunity to combine research and policy in facing the unequal distribution of poor - and also positive - mental health. The findings of this thesis reinforce the call to address mental health and mental health care inequalities. Attention is required by challenges we have known and studied for a long time, such as financial barriers and communitybased support for those with severe conditions. But there are also challenges just being uncovered, such as the persistence of inequalities along the mental health treatment pathway. Policy-wise, there is a “white canvas” to design strategies targeted at mental health inequalities. Research can have a role in informing this strategy, which should be a collaborative effort between academic disciplines and non-academic stakeholders. To do so, future research should focus on identifying what interventions, programs and policies best work to reduce mental health inequalities, causally framing research questions in a consequential way. Exposures of interest include those designed to focus on disparities but also population-wide policies that might (un)intendedly have distributional effects. These exposures are often unsuitable for randomisation, demanding an essential contribution of retrospective and prospective quasi-experimental work. The experience gathered in this thesis suggest that maximising the role of quasi-experiments requires going beyond the current academic approach of publishing single-paper evaluations that often trade-off policy relevance by the accomplishment of innumerous robustness checks. The uptake of quasi-experimental evidence can be increased by having these studies as part of a broader research strategy complemented with other sources of information. Examples of such approaches exist, for instance, in the work conducted by the World Bank under the concept of impact evaluation [32]. Furthermore, researchers should acknowledge the continuum of evidence towards causality and be willing to produce evidence that is positioned along that continuum. Methodologically, this includes developing better approaches to quantitatively estimate the impact of potential bias in the results (as the field of economic evaluation has been doing with the uncertainty of modelling estimates for health technology assessment). There is also room to strengthen the understanding of the potential bias through qualitative evidence in mixed methods studies. Institutional incentives would also need to change to maximise the potential of quasi-experimental work. This includes recognising the role of policyrelevant output alongside peer-reviewed publications. Last, and on a different domain, improving (mental health) inequalities research does also demand a systematic and widespread collection of detailed information on SES in contexts where this is not yet done, namely clinical settings. Researchers should have a pivotal role in setting up these data collection efforts, so that the evaluation of distributional effects does not have to be limited to the existence of cohort studies or linked administrative data infrastructures (as the one that could be used for this thesis). Policy has to do its part too to achieve a successful collaboration and strategy to close the mental health gap. Systematic data collection will rarely be sustainable solely through research efforts. Embedding this practice in the busy schedules of practitioners, whether 7
RkJQdWJsaXNoZXIy MjY0ODMw