Data Challenge: SSI-Predict: A Clinical Decision Support Tool for Surgical Antimicrobial Prophylaxis in Low- and Middle-Income Countries

Surgical site infections (SSIs) are the most prevalent healthcare-associated infections in low- and middle-income countries (LMICs), affecting up to 23% of surgical patients, more than double the rate in high-income countries(WHO, 2018). In Sub-Saharan Africa alone, SSIs account for up to 38% of all surgically-related nosocomial infections, with Tanzania reporting rates of 19–24% in district and tertiary hospitals(Fehr J, et al. 2006). Despite this burden, resistance data to guide perioperative antimicrobial prophylaxis, the single largest driver of antibiotic use in hospitals globally, remains critically absent in LMIC surgical settings. In Ghana, 55% of E. coli and 77% of K. pneumoniae SSI isolates were ESBL-producing, with 51% of gram-negative SSI isolates resistant to gentamicin, the backbone of affordable LMIC surgical prophylaxis . In East Africa, ceftriaxone, the most frequently prescribed prophylaxis agent, shows resistance rates of up to 69%, yet most hospitals lack the capacity to perform cultures to detect this. Meanwhile, inappropriate surgical antimicrobial prophylaxis in LMICs is associated with increased mortality, re-admission rates, and length of hospital stay((Sanders T et al., 2022, Berhe F. et al., 2025). We propose SSI-Predict, the first clinical decision support tool designed specifically to guide surgical antimicrobial prophylaxis in LMICs. Using SSI-associated isolates extracted from the Pfizer ATLAS datasets, supplemented by SPIDAAR outcome data, we will: 1. Characterise 18-year resistance trajectories for key SSI pathogens stratified by phenotype (MRSA/MSSA/ESBL); 2. Build the first LMIC SSI-specific antibiogram; 3. Deploy a machine-learning phenotype prediction model as web-based, offline-capable tool outputting WHO AWaRe-guided prophylaxis recommendations. We will additionally integrate Global AMR R&D Hub investment data to quantify the gap between SSI AMR burden and surgical AMR research funding in LMICs. SSI-Predict is proposed precisely for resource-limited settings, where prophylaxis is decided empirically, laboratory capacity is limited, and the cost of a wrong antibiotic choice is a patient's life.