Table 5. Primary analysis: Newborn mortality in intervention areas compared to comparison areas in 2007 and 2013 and adjusted for baseline.
Neonatal Mortality Rate per 1,000 Live Births | 2007 (5 districts) | 2013 (5/6 districts) | ||||||
---|---|---|---|---|---|---|---|---|
Intervention wards | Comparison wards | OR~ (95% CI) | p | Intervention wards | Comparison wards | OR~ (95% CI) | p | |
Live births | 33,553 | 30,603 | 21,898 /24,381 | 21,085/23,307 | ||||
Deaths, days 0–27 | 1,140 | 1,035 | 661/749 | 616/679 | ||||
Neonatal mortality rate per 1,000 live births (95% CI) | 35.1 (33.1–37.2) | 34.9 (32.8–37.1) | 1.0 (0.9–1.1) | 0.830 | 31.0 (28.8–33.5)/ 31.6 (29.5–34.0) | 30.0 (27.7–32.5)/ 29.9 (27.8–32.3) | 1.0 (0.9–1.2)/1.1 (0.9–1.2) | 0.547/0.339 |
Effect adjusted for baseline mortality and population size within the wards and division | 1.0 (0.9–1.2) | 0.779 |
~We used multilevel logistic regression to compute ORs specifying the ward and subvillage level. The intracluster correlation coefficients were 5.6% (95% CI 3.5%–8.8%) for the subvillage nested within the ward. The respective value for the ward level was 0.5% (95% CI 0.1%–2.7%).