BACKGROUND: Texas has experienced outbreaks of Cyclosporiasis annually for the last 3 years. Case-control studies are often employed to study large-scale outbreaks, however they often require a large amount of time and resources. The case-chaos study design is a supplemental analytic study that has the potential to reduce the time and resources needed. This method may be used to help identify vehicles associated with illness, and guide the outbreak response without utilizing the same level of time and resources required of traditional analytic study designs.
METHODS: Data was obtained using the National Electronic Disease Surveillance System (NEDSS), and consisted of cases and food history from outbreaks that occurred in 2013 - 2015. A case-chaos study was conducted for outbreaks within each year separately. Cases were matched to controls with a 1:3 ratio. Case-chaos results were compared to Food Net values.
RESULTS: Fifty six food items were analyzed in 2013. Among those with Cyclosporiasis, odds of consuming cilantro (the implicated food item) were fairly modest at 2.41 (95% CI: 1.62, 3.59). Twenty one food items were analyzed in 2014. Among those with Cyclosporiasis, odds for consuming cilantro were greater than all other food items (OR: 14.78; 95% CI: 5.77, 28.70). The odds ratio for cilantro in 2014 was higher compared to other food items using Food Net (OR: 11.80; 95% CI: 7.51, 18.54). In 2015, the same twenty one food items were analyzed. Among those with Cyclosporiasis, odds for consuming cilantro was not among the highest (OR: 1.81; 95% CI: 0.326, 10.15). However, the highest odds was given by the food item “pico de gallo or salsa” which was often made using cilantro as reported by cases (OR 148.38; 95% CI: 0.109, 201698.48).
CONCLUSIONS: Case-chaos was able to identify the implicated food item in 2014, and was consistent with food items identified by Food Net. This study design was also a great indicator in the 2015 investigations leading us to revisit the previously implicated food item, cilantro, although indirectly. A relationship between prevalence of food items among cases and odds ratios of food items were observed in case-chaos studies for all three years. However, further use and evaluation of case-chaos studies are needed to determine its utility as a tool for outbreak investigations.