Use of a spatial scan statistic to identify clusters of births occurring outside Ghanaian health facilities for targeted intervention. Issue 2 (November 2016)

• Record Type: Journal Article
• Authors: Bosomprah, Samuel; Dotse‐Gborgbortsi, Winfred; Aboagye, Patrick; Matthews, Zoe
• Is Part Of: International journal of gynaecology and obstetrics. Volume 135:Issue 2(2016:Nov.)
• Page Start: 221
• Page End: 224
• Publication Date: 2016-11
• Subjects: Circular spatial window -- Clusters -- District health information system -- Geographic information system -- Home births -- Spatial scan statistic; Gynecology -- Periodicals; Obstetrics -- Periodicals; Electronic journals; 618
• DOI: 10.1016/j.ijgo.2016.04.016
• Languages: English
• ISSNs: 0020-7292

Abstract

Abstract: Objective: To identify and evaluate clusters of births that occurred outside health facilities in Ghana for targeted intervention. Methods: A retrospective study was conducted using a convenience sample of live births registered in Ghanaian health facilities from January 1 to December 31, 2014. Data were extracted from the district health information system. A spatial scan statistic was used to investigate clusters of home births through a discrete Poisson probability model. Scanning with a circular spatial window was conducted only for clusters with high rates of such deliveries. The district was used as the geographic unit of analysis. The likelihood P value was estimated using Monte Carlo simulations. Results: Ten statistically significant clusters with a high rate of home birth were identified. The relative risks ranged from 1.43 ("least likely" cluster; P = 0.001) to 1.95 ("most likely" cluster; P = 0.001). The relative risks of the top five "most likely" clusters ranged from 1.68 to 1.95; these clusters were located in Ashanti, Brong Ahafo, and the Western, Eastern, and Greater regions of Accra. Conclusion: Health facility records, geospatial techniques, and geographic information systems provided locally relevant information to assist policy makers in delivering targeted interventions to small geographic areas.