| dc.description.abstract | Cost evaluation and cost control are the main concerns of any healthcare systems since national economies, regardless of their economic power, face constrained healthcare budgets while the demand for healthcare and health costs are ever-growing. Consequently, we need to understand how healthcare costs behave as a function of patients’ and health systems’ characteristics. However, although the positive association between education and health has been well documented, the causal relationship between years of schooling and healthcare costs is yet to be established. The aim of this thesis is to estimate, if any, this causal effect. To handle potential confounding and reverse causation problems in the education-healthcare costs relationship, this thesis applies a Two-sample Mendelian Randomisation (MR) approach on summary data from previous genome-wide association studies based on samples of European and white British ancestry. Given that the IV/MR-conditions discussed through this paper are fulfilled, this method might provide the causal effects of years of schooling on healthcare costs. This thesis finds that an extra year of completed schooling would reduce total healthcare costs by approximately £61 in 2019-GBP (£219 per std in years of schooling (3.6 years); CI £267 - £168). This main result remains comparable across methods meant to uncover potential violations of the MR-assumptions. However, only using summary data from previous studies pose some limitations to the analysis. Thus, these results should be treated with care. In addition, I conclude that, contrary to earlier concerns, a Two-sample MR with summary data can be used to perform exploratory analyses of economic outcomes. Genetic variants should not be dismissed as instruments beforehand on the argument that they won't induce enough variation to the exposure, However, methodologically solid research results would require using individual data, ideally containing information on siblings and/or family trios. There is a need to understand the biological pathways through which genetic instruments influence educational attainment. Without this knowledge it is difficult to make convincing arguments supporting the credibility of the exogenous condition, which cannot be tested otherwise. Finally, it seems of vital importance to rigorously establish which specific causal effects a MR-analysis identifies. | eng |