Measurement of Unmet Need for Contraception: A Counterfactual Approach

Abstract

Unmet need plays a critical role in reproductive health research, evaluation, and advocacy. Although conceptually straightforward, its estimation suffers from a number of methodological limitations, most notably its reliance on biased measures of women's stated fertility preferences. We propose a counterfactual-based approach to measuring unmet need at the population level. Using data from 56 countries, we calculate unmet need in a population as the difference between: (1) the observed contraceptive prevalence in the population; and (2) the calculated contraceptive prevalence in a subsample of women who are identified to be from "ideal" family planning environments. Women from "ideal" environments are selected on characteristics that signal their contraceptive autonomy and decision-making over family planning. We find significant differences between our approach and existing methods to calculating unmet need, and we observe variation across countries when comparing indicators. We argue that our indicator of unmet need is preferable to existing population-level indicators due to its independence from biases that are generated from the use of reported preference measures, the simplicity with which it can be derived, and its relevance for cross-country comparisons as well as context-specific analyses.

FIGURE 1

Current methodology for unmet need classification, DHS, from Bradley et al. (2012)

FIGURE 2

Estimated $iC{P}_{ct}$ by country NOTE: For countries with multiple survey rounds, country‐level $iCP$ are calculated by averaging survey‐round $iCP$ estimates.

FIGURE 3

Changes in $iC{P}_{ct}$ over time, by country

FIGURE 4

Kernel density plots, unmet need across definitions

FIGURE 5

Kernel density plots, difference between the new and old unmet need measure

FIGURE 6

Kernel density plots, matched unmet need across definitions