Predicting foodborne outbreaks at restaurants: Results of a risk assessment model

This study examined relationships between risk factors and foodborne outbreaks at restaurants. Two approaches were used: 1) historical cohort and 2)case control. The historical cohort design was used to test relationships between outbreaks and prior violations among 19,753 restaurants during a one-year period. A case control design was used to test relationships between outbreaks and the organizational culture, foods served, and number of customers. A case was defined as a restaurant with a foodborne outbreak during a one-year study period. In the case control study, five inspectors surveyed managers from 186 “case” restaurants and 277 “control” restaurants, randomly selected and stratified by number of seats and risk rating. Using multivariate analysis, outbreaks were significantly associated with buffets (OR=3.14, 95%CI=1.54-6.64), higher price entrées (OR=1.64, 95% CI=1.29-1.83), larger number of customers (OR=1.24, 95% CI=1.14-1.36), larger number of seats (OR=2.32, 95% CI=2.05-2.63), and sanitizing violations (OR=1.98, 95% CI=1.03-3.50). Other factors associated with outbreaks included cooking/reheating violations, contamination violations, full wait service, off-site catering, food prepared for multi-day use, larger number of foodhandlers, menu size, ownership type, receiving imported seafood, receiving/pooling eggs, and serving raw meats or produce. A structural model based on number of customers, foodhandling violations, and raw food served was developed for estimating the risk of foodborne outbreaks at restaurants. The results suggest that structural equation modeling is an effective tool for predicting foodborne outbreaks at restaurants. Utilizing a risk-based model would allow health departments to focus more inspections towards restaurants that pose the greatest risk to public health.