Assessing the sustainability of the Nigerian urban reproductive health initiative facility-level programming: longitudinal analysis of service quality

Table 7 Random effect model results predicting the effect of overall quality and program component scores on the number of new family planning acceptors in the last year per reproductive health staff member, MLE facility surveys from Kaduna and Ilorin, Nigeria

	Number of new acceptors per reproductive health staff member
	Model 1		Model 2
	Marginal	95% CI	Marginal	95% CI
Overall quality index	1.06	−0.55; 2.67	na
Training index	na		5.32	−8.06; 18.70
Method provision index	na		2.24	−0.30; 4.78
Stock-out (30 days) index	na		−2.61	−4.70; −0.52*
IEC index	na		0.15	−1.60; 1.89
Outreach index	na		−1.10	−3.33; 1.13
Ilorin (ref: Kaduna)	0.71	−10.82; 12.23	−0.18	−11.98; 11.62
Survey wave
Baseline – 2011 (ref)	Ref		ref
Endline – 2014	10.81	−0.11; 21.73	10.15	−0.93; 21.24
Sustainability – 2017	19.45	8.83; 30.08***	18.05	6.54; 29.55**
NURHI- supported facility (ref: not NURHI)	16.88	4.28; 29.49**	14.69	2.39; 26.98*
Private facility (ref: public)	−11.34	−19.35; −3.33**	−14.48	−22.37; −6.59***
Ilorin and Endline (2014) interaction	−7.42	−22.58; 7.73	−8.35	−24.08; 7.38
Ilorin and sustainability (2017) interaction	−20.57	−35.42; −5.72**	−18.54	−34.36; − 2.71*
NURHI facility and sustainability (2017) interaction	1.81	−13.22; 16.83	1.61	−13.99; 17.20
Error Components:
Sigma mu	9.66	5.35; 17.43***	3.24	+
Sigma epsilon	29.88	27.42; 32.56***	30.90	28.84; 33.10***
Number of observations		416	416
Number of unique facilities		168	168

na – not applicable for specific model; excludes outliers. + The maximum likelihood estimator did not report a standard error in this case indicating a large standard error and an insignificant effect for the time invariant error. * p ≤ 0.05; ** p ≤ 0.01; ***p ≤ 0.001

ISSN: 1472-6963