Abstract
This paper estimates the short-run causal effect of tumor detection and treatment on psychosocial well-being, work and income. Tumor detection can be considered as a random event, so that we can compare individuals' average outcomes in the year of diagnosis with the year before. We argue for using panel data estimation techniques that enable us to control for observed and unobserved information intrinsic to the individual and time constants. We use data of a national representative panel in the Netherlands that includes health survey information and data on work, education, and income between 2007 and 2012. Our findings show differences in the psychosocial dysfunction of men and women in response to tumor detection and treatment. Women, not men, are decreasingly likely to participate in the labor force as a result of malignant tumor detection, while no significant effects are found on her personal or household income. We also demonstrate that fixed effects panel data models are superior to matching techniques.