PT - JOURNAL ARTICLE AU - Robert J Magnani AU - John Ross AU - Jessica Williamson AU - Michelle Weinberger TI - Can Family Planning Service Statistics Be Used to Track Population-Level Outcomes? AID - 10.9745/GHSP-D-17-00341 DP - 2018 Mar 21 TA - Global Health: Science and Practice PG - 93--102 VI - 6 IP - 1 4099 - http://www.ghspjournal.org/content/6/1/93.short 4100 - http://www.ghspjournal.org/content/6/1/93.full SO - GLOB HEALTH SCI PRACT2018 Mar 21; 6 AB - Estimates of the modern contraceptive prevalence rate (mCPR), a population-level indicator, that are derived directly from family planning service statistics lack sufficient accuracy to serve as stand-alone substitutes for survey-based estimates. However, data on contraceptive commodities distributed to clients, family planning service visits, and current users tend to track trends in mCPR fairly accurately and, when combined with survey data in new tools, can be used to approximate the annual mCPR in the absence of annual surveys.The need for annual family planning program tracking data under the Family Planning 2020 (FP2020) initiative has contributed to renewed interest in family planning service statistics as a potential data source for annual estimates of the modern contraceptive prevalence rate (mCPR). We sought to assess (1) how well a set of commonly recorded data elements in routine service statistics systems could, with some fairly simple adjustments, track key population-level outcome indicators, and (2) whether some data elements performed better than others. We used data from 22 countries in Africa and Asia to analyze 3 data elements collected from service statistics: (1) number of contraceptive commodities distributed to clients, (2) number of family planning service visits, and (3) number of current contraceptive users. Data quality was assessed via analysis of mean square errors, using the United Nations Population Division World Contraceptive Use annual mCPR estimates as the “gold standard.” We also examined the magnitude of several components of measurement error: (1) variance, (2) level bias, and (3) slope (or trend) bias. Our results indicate modest levels of tracking error for data on commodities to clients (7%) and service visits (10%), and somewhat higher error rates for data on current users (19%). Variance and slope bias were relatively small for all data elements. Level bias was by far the largest contributor to tracking error. Paired comparisons of data elements in countries that collected at least 2 of the 3 data elements indicated a modest advantage of data on commodities to clients. None of the data elements considered was sufficiently accurate to be used to produce reliable stand-alone annual estimates of mCPR. However, the relatively low levels of variance and slope bias indicate that trends calculated from these 3 data elements can be productively used in conjunction with the Family Planning Estimation Tool (FPET) currently used to produce annual mCPR tracking estimates for FP2020.