NBER Working Papers by Donald S. Shepard

Deciding whether to fund a given health program involves both statistical and ethical issues. Traditional statistical methods of measuring program effectiveness may give misleading results unless careful attention is paid to the question of population heterogeneity. Even within particular age and sex categories, members of a population typically differ in both their mortality rate and the extent to which they would benefit from a given medical intervention. It may or may not be possible to identify the risk factors (e. g., weight, smoking behavior) that explain these differences. If an intervention confers unequal benefit on different risk groups, it will change their mixture within the population over time. If those helped most are those at greatest risk, a "traditional assessment" will o...