Impact of a Community-based Smartphone Intervention on Maternal Health Service Utilization in Rural Tanzania

Improving access to safe facility-based delivery (FBD) is a challenge in rural Tanzania, where 50% of women deliver at home without skilled medical assistance. Community health workers (CHWs) may improve women’s demand for, and uptake of FBD services; however CHW performance can be suboptimal without supportive supervision and job aids. This dissertation addresses key questions in maternal health services research and contributes to the evidence base for mobile health (“mHealth”), for which rigorous impact studies in low-income countries remain scarce. This study employed a cluster-randomized, controlled, mixed-methods trial design to evaluate the impact of a smartphone-based intervention (SP+) designed to assist CHWs with data collection, prenatal education delivery, gestational danger sign identification, and referral on women clients’ utilization of FBD. Pairs of CHWs in 32 villages were cluster-randomized to receive training on either SP+ or standard, paper-based protocols for use during household visits with clients. The main outcome (delivery location) was ascertained via postnatal household surveys with 572 randomly selected women. In-depth qualitative interviews with 60 CHWs and 14 healthcare professionals, and focus group discussions with 56 women clients were conducted to explore perceptions of CHW performance, quality of care, and SP+ implementation strength. SP+ was associated with increased FBD: after adjusted analyses, the odds of FBD in intervention villages were two times greater than the odds in control villages (OR=1.95; p=0.02). A key underlying mechanism was increased household visit frequency by smartphone-assisted CHWs; these CHWs reported higher job satisfaction compared to peers in the control group. Qualitative findings suggest that SP+ led to perceived improvements in data management, communication, decision-making support, emergency response, enhanced social status and credibility, and perceived health system improvements among clients. However, concerns regarding privacy and data security were raised. Through triangulation of methods and data sources, I demonstrate how factors influencing women’s use of FBD in this context are multidimensional and thus require multi-level maternal health programs and policies. I argue that while SP support for CHWs can be efficacious in targeting increased FBD, prevailing health system weaknesses including a lack of formal support for CHWs limits the potential impact of such strategies.Ph.D