Using a sample of retirees drawn from the US Health and Retirement Study (HRS), econometric techniques developed for dynamic panel data model with household effects, and a new and novel measure of health, we study several aspects of the impact of health and health change on the post-retirement evolution of wealth. We rely on item response theory to construct a broad and comprehensive measure of health status, and establish an indicator of poor health based upon it.  We then use this measure to study the incidence and persistence of poor health in retirement and its variation over demographic and economic groups of retirees, including those at various levels of prior wealth. We also use this measure to analyze the effect of adverse shocks to health status on overall wealth.   Our approach advances knowledge regarding the health related threats to economic independence for older citizens in several ways. First, we focus exclusively on a sample of individuals who have actually retired, whereas the vast majority of existing studies of resource adequacy utilize samples of older Americans made up of either pre-retirees or a mix of pre-retirees and retirees. Retirees are more constrained with respect to how to deal with health shocks than are non-retirees. They have less ability to alter work and earnings, and their health care coverage is likely to differ; many workers have employer sponsored coverage while fewer retired persons have such coverage. For these reasons research on retirees only will be better designed to address the question of the effect of health shocks on wealth of retirees than a sample defined by age. Second, we study longer term consequences of health shocks than other existing research.  This is important from a policy perspective as it is longer term or permanent changes in wealth that could be an important issue while temporary changes may not require policy attention. Third, the issue that we address is little studied. Policy reform can be improved with better understanding of the incidence of these risks, the groups of people for which incidence and/or consequences of the risks realization are the greatest, and the patterns by which resources change in response to health shocks over retirement years.The contributions noted above are enabled by our unique latent measure of true health status, which rests on a full estimation framework based on item response theory, thereby reducing the arbitrariness of prior estimates of the impacts of health shocks based on either several partial indicators of health status or ad hoc estimates of underlying health. Finally, we analyze the effect of shocks on a more comprehensive definition of wealth, including pension wealth; prior health shock studies have used more limited definitions of wealth, often excluding pension wealth. Hence, our findings will allow us to identify  those socio-economic groups for which negative health shocks represent a significant risk to retirement resources.