In this study we attempted to systematically utilize the varying gender and age distributions of foodborne outbreaks, combined with knowledge of Danish age and gender specific food intakes from the survey "The Danish National Survey of Diet and Physical Activity 2003-2008 " (DANSDA), to risk rank food groups and obtain an indication of the outbreak source. The study included 12 Danish foodborne outbreaks with a known outbreak source. For each of the 12 outbreaks and for the Danish population, the theoretical proportion of exposed persons for each of 214 food groups was calculated. The theoretical exposure for each food item in each outbreak and in the Danish population was compared using three different methods (preference measures). The preference measures comprised the difference (Præfdiff) and the quotient (Præfkvot) between exposure in the outbreak and the population and Præfdiff compared to the maximum difference for the food group (Præfpot). In addition, the similarity between the gender- and age distribution of outbreak cases was compared with the distribution by gender and age of the part of the Danish population which the DANSDAfound exposed to each food group. The similarity between the two gender- and age distributions were calculated for each food group as a Proportional Similarity Index (PSI). For each outbreak, the food groups were ranked according to the size of the three preference measures (Præfdiff, Præfkvot og Præfpot) as well as the PSI, and the ranking of selected indicator food groups for the outbreak source was compared between methods. Large differences in food preferences were found for patients in outbreaks with different food source, and strong similarities were observed between food preferences for patients in outbreaks with the same or similar food sources. The study identifies Præfpot as the best preference measure. An indicator food group for the outbreak source could be found among the 25 highest-ranked Præfpot in 11 of the 12 outbreaks. The probability of this outcome of the investigation to be a result of pure chance is less than 1 per thousand. The outbreak, in which an indicator food group for the outbreak source was not in the top-25 for Præfpot, was a sub-outbreak comprising the “tail” of cases in the end of a larger outbreak. For the preference measures Præfdiff, Præfkvot and for PSI, indicator food groups for the outbreaks source were only ranked in the top-25 of respectively, 9, 5 and 3 of the 12 outbreaks. Calculation and ranking of Præfpot in this study was Excel-based and is fast and easily conducted, once the underlying Excel-data sheets have been established. Ranking of Præfpot for food groups was assessed to be a useful supplement to the investigation of foodborne outbreaks. There is a need for further studies to determine how and when the food preference ranking is best used, and the method can probably be refined by adding seasonal and geographical variation in food intake into the ranking of food groups.
|Place of Publication||Søborg|
|Number of pages||84|
|Publication status||Published - 2015|