Background
Ethiopia has one of the highest maternal mortality in the world. Institutional delivery is the key intervention in reducing maternal mortality and complications. However, the uptake of the service has remained low and the factors which contribute to this low uptake appear to vary widely. Our study aims to determine the magnitude and identify factors affecting delivery at health institution in two districts in Ethiopia.

Methods
A community based cross sectional household survey was conducted from January to February 2012 in 12 randomly selected villages of Wukro and Butajera districts in the northern and south central parts of Ethiopia, respectively. Data were collected using a pretested questionnaire from 4949 women who delivered in the two years preceding the survey.

Results
One in four women delivered the index child at a health facility. Among women who delivered at health facility, 16.1% deliveries were in government hospitals and 7.8% were in health centers. The factors that significantly affected institutional delivery in this study were district in which the women lived (AOR: 2.21, 95% CI: 1.28, 3.82), women age at interview (AOR: 1.96, 95% CI: 1.05, 3.62), women’s education (AOR: 3.53, 95% CI: 1.22, 10.20), wealth status (AOR: 16.82, 95% CI: 7.96, 35.54), women’s occupation (AOR: 1.50, 95% CI: 1.01, 2.24), antenatal care (4+) use (AOR: 1.77, 95% CI: 1.42, 2.20), and number of pregnancies (AOR: 0.25, 95% CI: 0.18,0.35). We found that women who were autonomous in decision making about place of delivery were less likely to deliver in health facility (AOR: 0.38, 95% CI: 0.23,0.63).

Conclusions
Institutional delivery is still low in the Ethiopia. The most important factors that determine use of institutional delivery appear to be women education and household economic status.