Table 4

Unadjusted and adjusted individual-level analysis showing odds of reliable and estimated child and maternal deaths and disability

			Control	Intervention	Random stratum effects		Fixed stratum effects
					Unadjusted	Adjusted	Unadjusted	Adjusted
					OR (p value; 95% CI)	OR (p value; 95% CI)	OR (p value; 95% CI)	OR (p value; 95% CI)
Child deaths*			n	n
Follow-up data only	Reliable	Died	116	80
	Reliable	Alive	2116	2107	0·67 (0·15; 0·39 to 1·16)	0·70 (0·18; 0·43 to 1·18)	0·70 (0·034; 0·50 to 0·97)	0·76 (0·075; 0·55 to 1·03)
	Total estimated	Died	132	87
	Total estimated	Alive	2966	2730	0·71 (0·15; 0·44 to 1·14)	0·75 (0·17; 0·50 to 1·14)	0·68 (0·007; 0·51 to 0·90)	0·72 (0·026; 0·54 to 0·96)
Maternal deaths (per individual mother)
Follow-up data only	Reliable	Died	68	52
	Reliable	Alive	2246	2155	0·75 (0·26; 0·48 to 1·17)	0·83 (0·42; 0·53 to 1·31)	0·79 (0·28;0·54 to 1·15)	0·87 (0·47; 0·61 to 1·26)
	Total estimated	Died	80	63
	Total estimated	Alive	3056	2791	0·83 (0·45; 0·56 to 1·24)	0·91 (0·67; 0·61 to 1·38)	0·85 (0·44; 0·61 to 1·20)	0·93 (0·69; 0·67 to 1·31)
Child disability
	Screened	Positive	181	132
		Negative	1935	1974	0·66 (0·10; 0·40 to 1·09)	0·64 (0·085; 0·39 to 1·06)	0·65 (0·021; 0·46 to 0·94)	0·63 (0·013; 0·44 to 0·91)

All models take account of clustering of women within village development committees and also the pairing of clusters (strata) using either ‘random stratum effects’ or ‘fixed stratum effects’. Adjusted OR are adjusted for baseline asset score, maternal literacy, maternal age and main household occupation using propensity scoring.
Child death rate excludes stillbirths and,and deaths in infants up to 4 weeks of age.