An Implementation Study of Interventions to Promote Safe Motherhood in Jimma Zone Ethiopia

A three-arm cluster trial design will be used to measure the impact of the intervention packages in three districts (Gomma, Seka Chekorsa, Kersa) in Jimma Zone, Ethiopia. Twenty-four clusters with 160 individuals each are required to detect a 17% change in the primary outcome (proportion of facility-based births) with 80% power,assuming a cluster autocorrelation of 0.8 and an intra-cluster correlation coefficient of 0.1. Primary health care units (PHCU) which consist of a health centre and several community-based health posts will serve as trial clusters. All PHCUs with maternity waiting areas (MWAs) constructed will be eligible for selection. Eligible women will be randomly selected from PHCU catchment areas stratified by MWA functionality and health centre basic emergency obstetric care (BEmOC) capacity. This is to ensure balanced distribution of poorly functioning MWAs and facilities with diminished BEmOC capacity between trial arms. Cross-sectional household surveys will be conducted with eligible women to collect information on socio-demographics, knowledge, attitudes and practices regarding maternal health services, reproductive history and maternal health service utilization. Surveys will be administered by trained interviewers on tablet computers programmed using Open Data Kit at baseline prior to intervention roll out and at endline. Multilevel regression models will be used to quantify the effect of the intervention packages on outcomes of interest. Random effect terms for PHCUs will be included to account for the clustered nature of the data.