INTRODUCTION/INTENDED USE

The following data comprise exposure frequencies for sporadic culture-confirmed Escherichia coli O157 cases that occurred in Minnesota from 2009 through 2015. Cases were interviewed with a detailed questionnaire about exposures that occurred in the 7 days prior to their illness onset (http://mnfoodsafetycoe.umn.edu/wp-content/uploads/2014/02/Salmonella-and-STEC-Interview-Form.pdf). Epidemiologists can use these frequencies to estimate the background exposure rate in the population for a particular food item or other exposure and then use this estimate in a binomial model comparison which is available in Epi Info 7 under StatCalc or as an excel spreadsheet at (https://public.health.oregon.gov/DiseasesConditions/CommunicableDisease/Outbreaks/Gastroenteritis/Pages/Outbreak-Investigation-Tools.aspx#binomial). This can be useful to quickly evaluate potential hypotheses during the hypothesis generation phase of an investigation prior to conducting a more resourceintensive study method like a case-control study. Cases that were not interviewed, cases who reported international travel within the week prior to illness onset, and outbreak-associated cases, regardless of the transmission route, were not included in these data. Epidemiologists should carefully consider any potential biases arising from the use of E. coli O157 case exposure information to estimate background population exposure rates. Using E. coli O157 cases to estimate the background exposure rate in the population will likely introduce bias away from the null hypothesis when evaluating exposures that are risk factors for E. coli O157. For example, the percentage of sporadic E. coli O157 cases that were exposed to a known E. coli O157 risk factor such as ground beef is potentially higher than the true background rate of exposure to ground beef in the population. Epidemiologists should also consider if the food item under evaluation has seasonal consumption patterns, or if consumption of the food varies by gender or age. The questionnaire administered to E. coli O157:H7 cases was modified in 2011. Food exposure questions that were added to the interview questionnaire are indicated by an asterisk (*). Food exposures questions that were removed from the questionnaire in 2011 are indicated by a double asterisk (**). Another potential source for background exposure data is the FoodNet Population Survey Atlas of Exposures, 2006-2007 (http://www.cdc.gov/foodnet/surveys/FoodNetExposureAtlas0607_508.pdf). For more specific exposure food frequencies to match the demographics of your current investigation, please contact Joshua.rounds@state.mn.us