Access to care following injury in Northern Malawi, a comparison of travel time estimates between Geographic Information System and community household reports. Issue 5 (May 2022)

• Record Type: Journal Article
• Authors: Whitaker, John; Brunelli, Giulia; Van Boeckel, Thomas P.; Dube, Albert; Amoah, Abena S; Rickard, Rory F; Leather, Andrew J M; Davies, Justine
• Is Part Of: Injury. Volume 53:Issue 5(2022)
• Page Start: 1690
• Page End: 1698
• Publication Date: 2022-05
• Subjects: Wounds and injuries -- Healthcare systems -- Health services accessibility -- Geographic Information Systems -- Emergency Care; Wounds and injuries -- Surgery -- Periodicals; Accidents -- Periodicals; Wounds and Injuries -- surgery -- Periodicals; Lésions et blessures -- Chirurgie -- Périodiques; Electronic journals; Electronic journals; 617.1
• DOI: 10.1016/j.injury.2022.02.010
• Languages: English
• ISSNs: 0020-1383

Abstract

Highlights: Since the Lancet Commission on Global Surgery, analysis of population access to urgent surgical care, particularly using GIS, has expanded. Using GIS methodology in Northern Malawi, most of the studied population could not travel to a government secondary facility within two hours. Compared to community household-reports, GIS underestimated travel time, and correlated poorly with patient-reported travel time. This study is the first to validate GIS-estimated with community household-reported travel time, for injury care in any low-income country. GIS methods must be used with caution, accounting for potential error, when planning health system strengthening to improve injury care access. Abstract: Introduction: Injuries disproportionately impact low- and middle-income countries like Malawi. The Lancet Commission on Global Surgery's indicators include the population proportion accessing laparotomy and open fracture care, key trauma interventions, within two hours. The "Golden Hour" for receiving facility-based resuscitation also guides injury care system strengthening. Firstly, we estimated the proportion of the local population able to reach primary, secondary and tertiary facility care within two and one hours using Geographic Information System (GIS) analysis. Secondly, we compared community household-reported with GIS-estimated travel time. Methods: Using information from a Health and Demographic Surveillance Site (Karonga, Malawi) on road network, facility location, and local staff-estimated travel speeds, we used a GIS-generated friction surface to calculate the shortest travel time from all households to each facility serving the population. We surveyed community households who reported travel time to their preferred, closest, government secondary and tertiary facilities. For recently injured community members, time to reach facility care was recorded. To assess the relationship between community household-reported travel time and GIS-estimated travel time, we used linear regression to generate a proportionality constant. To assess associations and agreement between injured patient-reported and GIS-estimated travel time, we used Kendall rank and Cohen's kappa tests. Results: Using GIS, we estimated 79.1% of households could reach any secondary facility, 20.5% the government secondary facility, and 0% the government tertiary facility, within two hours. Only 28.2% could reach any secondary facility within one hour, 0% for the government secondary facility. Community household-reported travel time exceeded GIS-estimated travel time. The proportionality constant was 1.25 (95%CI 1.21–1.30) for the closest facility, 1.28 (95%CI 1.23–1.34) for the preferred facility, 1.45 (95%CI 1.33–1.58) for the government secondary facility, and 2.12 (95%CI 1.84–2.41) for tertiary care. Comparing injured patient-reported with GIS-estimated travel time, the correlation coefficient was 0.25 (SE 0.047) and Cohen's kappa was 0.15 (95%CI 0.078–0.23), suggesting poor agreement. Discussion: Most households couldn't reach government secondary care within recognised thresholds indicating poor temporal access. Since GIS-estimated travel time was shorter than community-reported travel time, the true proportion may be lower still. GIS derived estimates of population emergency care access in similar contexts should be interpreted accordingly.