The economic burden of breast cancer is a popular topic within economic and public health research. Denmark's extensive registry data systems allow for causal inference on this topic and measurement of outcomes longitudinally. While previous studies used matching methods to assess the economic effects of breast cancer, few have incorporated detailed economic and socio-demographic registry data for more refined matching, nor have they evaluated the impact of modeling methods on the measured results.
This project evaluates the impact of matching and estimation choices on measuring the economic effect of breast cancer in Denmark. I will show the utility of exact matching using a broad set of matching variables, and the effect that important modeling assumptions have on measured income losses.