Revisiting indirect methods based on orphanhood data with micro-simulations

In countries lacking a comprehensive system of death registration, data on the survival of close relatives are routinely collected for estimating risks of dying in the adult population. One of the most widely used approach, the orphanhood method, consists in converting proportions of parents alive into probabilities of surviving through standard demographic models. Since the 1980s, HIV epidemics have introduced severe breaches in the assumptions underpinning this method. In this paper, a set of populations with substantial mortality from AIDS are generated with microsimulations. The extent of the biases introduced by HIV/AIDS in the mortality estimates conventionally obtained from reports on parental survival is evaluated. Procedures for adjusting theses biases are discussed. New coefficients to estimate adult mortality from data on orphanhood for the period prior to the scale-up of antiretroviral therapy are developed. The orphanhood technique can be used in settings heavily affected by HIV/AIDS. Countries lacking a comprehensive vital registration system should continue to ask survey and census respondents about the survival of their parents