Can Family Planning Service Statistics Be Used to Track Population-Level Outcomes?

Figures & Tables

TABLE 1.

Countries Included in the Analyses, by Availability of Service Statistics Data Elements

Countries With Service Statistics Available (N=22)	Years of Data Available
Countries With Service Statistics Available (N=22)	3–4	5–6	≥7
Commodities data only (n=10)	4	1	5
Commodities and visits data (n=4)
Commodities	2	1	1
Visits	1	2	1
Commodities and users data (n=5)
Commodities		2	3
Users	1	3	1
Visits and users data (n=1)
Visits			1
Users			1
Commodities, visits, and users data (n=2)
Commodities		2
Visits		2
Users		1	1

TABLE 2.

Median MSE Results Across Countries Providing at Least 1 Service Statistics Data Element

MSE Component	Commodities to Clients Data (21 Countries)	Service Visits Data (7 Countries)	Current Users Data (8 Countries)
RMSE	0.0695	0.0961	0.1864
Variance	0.0014	0.0002	0.0012
Level bias	0.3114	0.3452	0.5774
Slope bias	0.0114	0.0069	0.0104

TABLE 3.

Relationship Between Service Statistics Estimates and Survey-Based Estimates of mCPR

Relationship	Commodities to Clients Data (21 Countries)	Service Visits Data (7 Countries)	Current Users Data (8 Countries)
Service statistics estimates always < survey estimates	14	5	3
Service statistics estimates always > survey estimates	5	1	4
Varies by year	2	1	1

TABLE 4.

Pairwise Comparisons of Median MSE Across Countries Providing at Least 2 Service Statistics Data Elements

MSE Component	Commodities vs. Visits (6 Countries)		Commodities vs. Users (7 Countries)		Visits vs. Users (3 Countries)
MSE Component	Commodities Data	Visits Data	Commodities Data	Users Data	Visits Data	Users Data
RMSE	0.0859	0.1182	0.1019	0.2216	0.2163	0.2115
Variance	0.0009	0.0001	0.0010	0.0013	0.0003	0.0008
Level bias	0.4054	0.4483	0.3103	0.5664	0.6015	0.5904
Slope bias	0.0029	0.0112	0.0161	0.0079	0.0158	0.0129

Abbreviations: MSE, mean square error; RMSE, root mean square error.
Note: Data from the 2 countries that collect all 3 types of service statistics are included in this table.