Understanding unmet need: history, theory, and measurement.

Abstract

During the past two decades, estimates of unmet need have become an influential measure for assessing population policies and programs. This article recounts the evolution of the concept of unmet need, describes how demographic survey data have been used to generate estimates of its prevalence, and tests the sensitivity of these estimates to various assumptions in the unmet need algorithm. The algorithm uses a complex set of assumptions to identify women: who are sexually active, who are infecund, whose most recent pregnancy was unwanted, who wish to postpone their next birth, and who are postpartum amenorrheic. The sensitivity tests suggest that defensible alternative criteria for identifying four out of five of these subgroups of women would increase the estimated prevalence of unmet need. The exception is identification of married women who are sexually active; more accurate measurement of this subgroup would reduce the estimated prevalence of unmet need in most settings.