Collaborative Development of a Departmental Database in an Urban, Low-Resource Hospital to Explore Predictors of Maternal Mortality and to Propose Targeted Health Service Research Interventions

Background:

Korle Bu Teaching Hospital (KBTH) in Accra, Ghana conducts ~10,000 deliveries per year with a MMR of 1063 deaths in 100,000 live births. A majority (28%) of maternal deaths are attributable to hypertensive complications for which effective clinical interventions exist; however, barriers to their implementation (like service delay) impede positive clinical outcomes. Creation of an onsite database of obstetrical events and service delivery indicators allows tracking and correction of the high rate of mortality.

Methods:

Through a twinning partnership with two teaching institutions, retrospective data of 20,331 deliveries at KBTH was used to build a multiple logistic regression model exploring the relationship between maternal mortality and predictor variables of appropriate care before reaching the facility (timely administration of magnesium) and appropriate care at the facility (delay in care due to screening, delay in care due to treatment) for women with hypertension or hypertensive complications of pregnancy.

Results:

We generated a model that will inform subsequent prospective data collection, as well as contribute to the implementation of interventions to improve service delivery. New data will be integrated every 6 months and leadership will meet to plan, monitor and evaluate evidence-based interventions in real-time.  Lessons learned include establishing a flexible shared work plan that can redirect research focus as necessary. Establishing regular communication schedules will hold both partners responsible for assigned tasks.

Conclusions:

This implementation science research may ultimately strengthen human and resource capacity and provide a model to other academic institutions focusing on sustainably improving health outcomes in low-resource facilities.