Till receipts - a new approach for investigating outbreaks? Evaluation during a large Salmonella Enteritidis phage type 14b outbreak in a north west London takeaway restaurant, September 2009

Selecting suitable controls for outbreak investigations is often difficult and if done inappropriately will lead to biased inferences. Till receipts and other sales records are frequently available on food premises, but their applicability has not been fully explored. Using data from an investigation into a Salmonella outbreak affecting 66 individuals exposed in a London takeaway restaurant, this study aimed to evaluate the use of till receipts to assess associations between sales and illness. Cases identified through local case-finding were subjected to a standardised exposure questionnaire. Till receipts over the time period when cases arose were analysed. Estimated food exposures from sales were compared to case reported exposures and till receipts analysis showed strong association between illness and consumption of rotisserie chicken (odds ratio (OR): 2.75; confidence interval (CI): 1.7-4.5). Chicken sales immediately prior to food consumption for cases were compared to two control periods in an ecological case-crossover design. On average there was an estimated increase of 3.7 (CI: 2.2-5.2) extra chickens sold in the hour immediately prior to the consumption in the cases (p< 0.0001) and the risk of becoming ill at busy times increased by 5% with each additional chicken quarter sold per hour (OR: 1.05; CI: 1.03-1.08). Microbiological and environmental investigations revealed Salmonella Enteritidis phage type (PT) 14b in all available cases' stool samples, two environmental samples and leftover chicken from the takeaway. The feasibility of this novel approach to obtain exposure information in the population at risk has been demonstrated, and its limitations are discussed. Further validation is required, comparing results with those in a concurrent classic case-control study.