To count or not to count deaths: reranking effects in health distribution evaluation.

ABSTRACT

Populations' structures and sizes can be a result of healthcare policy decisions. We use a two-period theoretical framework and a dynamic microsimulation model to examine the consequences of this assertion on the appraisal of alternative health policy options. Results show that standard welfare-in-health measures are sensitive to changes in populations' sizes, in that taking into account the (virtual) existence of the dead can alter the ranking of policy options. Disregarding differences in the survivals induced by alternative policies can bias programmes' ranking in favour of less live-saving policies. The paper alerts on the risk of policy misranking by the use of ex-post cross-sectional analyses, neglecting deaths occurring in the past as well as counterfactual deaths in alternative policy scenarios.