Measuring timely geographical access to surgical care in India: a geospatial modelling study.

BACKGROUND: In 2015, the Lancet Commission on Global Surgery proposed timely access to surgical care facilities as an indicator of preparedness in surgical systems. Since then, several African and South American countries have mapped the population coverage for their surgical facilities. However, these estimates are missing for India, despite its large population, geographical scope, and sociodemographic diversity. We conducted a nationwide analysis of timely access to surgical care in India and point to rural-urban disparities. METHODS: We extracted a nationwide dataset of 20 802 geolocated surgical care facilities from the IndoHealMap project. We obtained accessibility motorised friction surface raster data cropped for India from the Malaria Atlas Project Explorer. We calculated the travel times from each grid cell (equivalent to 1 km2) in India to its nearest surgical facility under the optimal speed estimation by means of the Djikstra least-cost algorithm. District-level and state-level rural and urban populations were estimated through raster-based analysis via data from WorldPop, Urban-Rural Catchment Areas, and GADM version 3.6. The primary endpoint of the analysis was the proportion of population within 2 h of their nearest surgical care facility at national, state, and district levels. Wilcoxon tests adjusted for multiple comparisons (Holm-Bonferroni correction) were used to investigate rural-urban differences at a 5% significance level. FINDINGS: The motorised travel-times distribution is highly right-skewed, depicting that a large number of areas in India were within 2 h of their nearest surgical facility. At the national level, 99·2% of the rural population had timely access to surgical care facilities, compared with 99·8% of the urban population. However, less than 80% of the rural population in the northern and northeast regions had timely access to care. The rural regions had a significantly smaller proportion of residents with timely access to surgical care compared with their and urban counterparts at the district level (n=1299, effect size=0·57; p<0·001) and state level (n=71, effect size=0·51; p<0·001) analyses. INTERPRETATION: Our findings should be considered upper-bound estimates for geographical access because they assume readily available motorised transport and optimal travel speeds, which might not always be the case. These first-ever estimates can inform India's National Surgical, Obstetric, and Anesthesia Plan and which locations should be focused on for upcoming surgical facilities. FUNDING: No funding to declare.

Duke Scholars

Author Joao Ricardo Nickenig Vissoci Emergency Medicine