Increasing the value from individual-level observational data: practical applications in health economics

Abstract

This thesis uses individual-level observational data to make new contributions to the literature. In total joint replacement, patient attributes associated with long-term gains from surgery are uncovered, while methodological contributions are made to the optimal timing of patient-reported outcomes, the identification of regression to the mean, and the estimation of an appropriate control group to improve the accuracy of economic evaluation. In congenital heart disease, a microsimulation model is developed to project the demographics of an at-risk cohort, and a costing study is completed to highlight the costs of uncertainty in anticoagulation treatments. Finally, the advantages of ‘big data’ techniques such as classification and regression trees (CART) for health economics analysis are show-cased in the CART modelling of general practitioner prescribing for cardiovascular disease, and the drivers of decision-making in health funding.