Validation of questions designed for investigation of gastroenteritis

Background: Health departments routinely investigate cases of gastroenteritis through interviews to determine the source of infection. However, validation studies of dietary questionnaires typically focus on quantities consumed and don't assess questions designed to identify sources of foodborne illness. We aimed to assess the accuracy and reliability of information collected by surveys of food history recall for gastroenteritis investigations. Methods: A questionnaire was developed to investigate the sources of foodborne gastroenteritis in Australia, with questions on food exposures selected for validation. Fifty-five participants photographed all foods consumed and food receipts obtained during a seven-day observation period. These photographs were uploaded to an online survey or emailed to the researcher. Participants were contacted 14 days later for a telephone interview about foods consumed in the seven-day period. Questionnaire responses were compared to uploaded photographs. Kappa statistics (κ) and 95% confidence intervals were calculated. Sixty-two questions were assessed, including those targeting foods considered high-risk for foodborne gastroenteritis. Potential risk factors covered by these questions included: meats (poultry, beef, pork, and deli meats), the state of poultry purchased (raw versus precooked), and the number of meals eaten outside of the home. Results: Several questions targeting high-risk foods were found to have substantial-to-almost perfect agreement (κ ≥ 0.610) between what was eaten and what was reported by participants, with most questions showing at least a moderate level of agreement (κ = 0.410–0.600). Questions regarding exposure to different types of meat showed a high level of consistency. The only question with poor participant recall (κ < 0.000) was that relating to consumption of undercooked beef or veal. Conclusion: Several questions designed for investigation of gastroenteritis were found to provide at least a moderate level of accurate and reliable recall, even after a delay until interview. These questions are suitable for investigating sources of foodborne gastroenteritis