Data Challenge: From resistance trajectories to lives lost — a SPIDAAR-anchored multi-dataset framework for AMR stewardship and R&D prioritization

Antimicrobial resistance (AMR) kills more than a million people each year, with the heaviest burden in low- and middle-income countries. The Vivli AMR Register contains rich antibiotic-susceptibility surveillance data, but most prior analyses have stopped at resistance trends without bridging to the question that matters most for patients and clinicians: how many lives are being lost to AMR, and which choices within the existing antibiotic arsenal could reduce that toll? This research uses three Vivli datasets — SPIDAAR (the only register containing patient mortality and length-of-stay outcomes), Pfizer ATLAS, and Merck SMART — together with the Global AMR R&D Hub funding database, to build a unified analytical framework that connects resistance trajectories to patient outcomes and to global investment. Help improve patient outcomes: Using SPIDAAR's mortality and length-of-stay data from Ghana, Kenya, Malawi, and Uganda, the analysis estimates the additional risk of death associated with resistant versus susceptible bacterial infections. This translates abstract resistance percentages into lives-lost estimates that clinicians and policymakers can act on. Strengthen stewardship: A facility-level "what if" simulator, built on the existing open-source AMR Sentinel platform, lets clinical teams compare empiric antibiotic choices against their own local antibiogram. The tool is parameterized so hospitals can re-run it with their own breakpoints, ward-mix, and antibiotic-cost data. Inform public health practice: A Bayesian projection of resistance trajectories through 2030 across countries, pathogens, and antibiotics identifies the country–pathogen combinations where AMR-attributable mortality is growing fastest, informing where surveillance, diagnostic, and stewardship investment is most urgent. Strengthen health systems: An R&D Mismatch Index ranks projected mortality burden against current per-pathogen research and development investment, flagging where global funding is least aligned with where it would save the most lives. All outputs — analysis code, projections, and the stewardship simulator — will be released openly under an Apache 2.0 license, designed so catchment-region researchers and clinicians can refine the framework with their own local parameters before drawing any country-specific conclusions.