Spatial Analysis of Emergency Obstetric Care Services for Upgrading of Health Facilities: a Case Study of Homa Bay County.

Abstract

Maternal Mortality is when an expectant woman dies during pregnancy, childbirth, or at least within a month of giving birth because of pregnancy management. Incidental or accidental deaths do not constitute Maternal Mortality. Unpredictable Obstetric problems (infection, severe bleeding, hypertension, and obstructed labor) remains top causes of death and disability for women. However, the most reliable mitigation to pregnancy right/correct, timely and effective emergency obstetric care for respective complication remains unknown. Therefore, the objective of this study is to use Geospatial Techniques to support the selection of the existing public health facilities to be upgraded (to EmOC using Homa Bay County as a case study), to ensure better provision of Emergency Obstetric Care.In the methodology, the facilities that met the EmOC functions standard were identified, these were 9 (7 BEmOC and 2 CEmOC). Their catchments were then determined at 10km using Voronoi Polygons in QGIS 3.4. Catchment populations were extracted and 7 facilities of the 9 existing EmOC facilities were found to serve more population than recommended by WHO, i.e., 100,000 people per health facility. In order to determine which health facilities would be upgraded to offer EmOC services, a 2-hour buffer was created around each facility (walking and motorized scenarios). Then a set of inclusion criteria (>10km from the existing BEmOC facility, > 500m from a road, at least 100 people/km square) was ran at 21 combinations and the list of recommended facilities for upgrading was arrived at. There was a need to upgrade 4 facilities to meet the desired 13. However, the two health centers (Pala Masogo and Sena) automatically qualified because they are Level 3. After the criteria of selection, the other two facilities that qualified for upgrading were Godbura Dispensary and Ponge Dispensary (Mbita). Recommendations drawn from the findings were first, reporting in DHIS2 should be improved to ensure identification of EmOC signal functions at the health facility level is not a complex process. This would also ensure that there is data completeness and good quality. Secondly, the accessibility of BEmOC and CEmOC was determined using spatial analysis in Grasshopper, OSM, and QGIS. However, it can be achieved at once by use of Access MoD 5. Third, a simpler and more understandable feature picking method should be explored when choosing the best combination.