A. E. Giles1,2, A. G. Ramirez1,3, M. G. Shrime4,5 1Harvard School Of Public Health,Boston, MA, USA 2McMaster University,Surgery,Hamilton, ONTARIO, Canada 3University Of Virginia,Surgery,Charlottesville, VA, USA 4Harvard School Of Medicine,Program In Global Surgery And Social Change,Boston, MA, USA 5Massachusetts Eye And Ear Infirmary,Otolaryngology,Boston, MA, USA
Introduction:
Availability of surgery is gaining increasing importance in global health, yet few nationally representative surveys incorporate surgery-relevant indicators. We sought to derive a composite score that predicts surgery availability from existing population-level survey data, and validate it against known surgical data.
Methods:
The Demographic and Health Surveys Program Service Provision Assessment (SPA) survey from Malawi was used to construct a composite score. Sensitivity analysis was conducted to identify an appropriate weighting scheme. Validation was performed through re-creation of the composite score in Kenya’s Access, Bottlenecks, Costs, and Equity (ABCE) Project data and comparison of the score against actual facility surgical volume. Performance of the score was also compared to that of using cesarean section availability as the sole indicator, against an a priori set of surgical volumes as the basis of comparison.
Results:
Based on the sensitivity analysis, the final composite score was: 0.25[Caesarian Section] + 0.25[Physician Present] + 0.20[Anesthetist Present] + 0.15[Ketamine Available] + 0.15[Transfusion Capability]. A total of 52 facilities (of the 1,060 health care facilities) were identified as providing surgical care in Malawi: 4 central hospitals, 22 district hospitals, 22 community hospitals, and 4 urban clinics. Community hospitals displayed the widest variation in ability to provide surgery. The composite score correlated well with surgical volume when applied to the Kenyan data (beta 1,378, p<0.001). Using a cutoff of 50 or more operations annually to define a facility providing surgery, the score outperformed provision of caesarean section alone with a sensitivity of 97% and specificity of 92%, versus 84% and 95%, respectively (Figure 1).
Conclusion:
The composite surgery availability score is both sensitive and specific for predicting surgical service capability. Implications for adoption of such a score include standardized evaluation of population access to surgical services and monitoring progress over time at the subnational, national, and multinational levels. The proposed methodology may make available time-sensitive findings to inform relevant policy change and investment of resources for surgery as part of achieving universal health coverage.
