Incorporating genomic virulence in a quantitative microbial risk assessment to assess the public health impact of alternative microbiological criteria for Salmonella

We developed a QMRA model to assess changes in Non-Typhoidal Salmonella (NTS) illnesses under scenarios of microbial criteria (MC) based on proportion tested ( TP ) of production lots (combos), concentration rejection thresholds per gram ( CT ), differential targeting of the “higher” (HV) and “lower” (LV) virulence serovars, and concentration reduction (e.g. kill steps). MC scenarios were compared against a baseline without diversion of product positive for NTS. The QMRA incorporates a novel epidemiologically-validated genomic method to group serovars by virulence, and corresponding dose-response functions reflecting differences in HV and LV infectivity. HV serovars were S. Typhimurium, S. 1,4, ( Risk Assessment and Analytics Staff 2019 ),12:i:-, S. Enteritidis, S . Newport, S. Dublin, S . Paratyphi B, S . Muenchen, S. Agona, and S . Litchfield. Illness reductions >25% can be achieved for TP >50%, and CT of 1 or 10 CFU/g. Mean illness reduction of 71.0% (95%PI=61.6-76.7%) is predicted with TP =100% at CT =10 CFU/g. At >50% TP , illness reductions were similar for CT =1 vs CT =10 while more combos were diverted (e.g., 5.1e-3% vs 1.1e-3% of combos, respectively under 100% TP ). MC targeting only HV achieved significant illness reductions, but with fewer diverted combos. For example, applying the CT =10 CFU/g only to HV with TP =100% led to a mean reduction of 46% (34.9-56.4%) and 1.1e-5 of combos diverted. MC targeting higher virulence and/or higher concentrations of NTS can achieve large public health gains. Optimal MC should also consider feasibility and cost of product diversions, speed and precision of assays to accurately identify high concentration combos, and individual plant characteristics that affect NTS concentrations and virulence profiles.