Measuring contraceptive prevalence among women who are at risk of pregnancy.

Objectives: The contraceptive prevalence rate (CPR) is generally reported among in-union women ages 15-49. Here, union status and age serve as proxies for exposure to the risk of pregnancy. As a result of changing dynamics, age and union status proxies may be insufficient for determining the rate of contraceptive use among women at risk of pregnancy. Our objectives are to define a measure of contraceptive use among women at risk, to measure contraceptive use among such women and to compare this rate with conventional CPR.

Study Design: Using data from the United States 2011-2013 National Survey of Family Growth (NSFG), we explore self-reported data on contraceptive use, sexual recency, pregnancy status and fecundity to develop an alternative CPR (ACPR) measure, contraceptive prevalence among women at risk of pregnancy. After defining and measuring ACPR, we compare ACPR and conventional CPR estimates using NSFG and Demographic and Health Survey data from 48 surveys completed from 2000 to 2015 in low- and middle-income countries.

Results: For measuring ACPR, it is best to limit sexual activity recency component to "four weeks" to minimize underreport of coital-specific methods. It is best to limit the contraceptive use component to "current use" rather than "use at last sex" to minimize underreport of permanent methods. In the United States, 86% of women at risk of pregnancy are currently using contraception.

Conclusion: Women at risk of pregnancy report higher levels of contraceptive use than the conventional CPR indicates.

Implications Statement: Development of ACPR exposed some potential family planning measurement weaknesses. Specifically, because CPR is based on report of "current use," our analyses show that CPR may underreport coital-specific methods. As a result, CPR estimates may be somewhat deflated, and unmet need estimates may be somewhat inflated.