NEHA October 2022 Journal of Environmental Health

October 2022 • Journal of Environmental Health 21 seemed to receive reviews that included more restaurant cleanliness keywords, both limited and full-service restaurants received similar frequencies of FBI keywords. Our study considered reviews only from the Houston area, and thus the results are not necessarily generalizable to other cities and markets. While we made a significant eort to ensure that the business analytics software analyzed keywords only within their proper context, it is possible that terms in atypical contexts might have deflated or inflated the recorded total frequency of keywords. For example, one particular restaurant with a restaurant cleanliness keyword as part of its name had to be excluded from text mining. Additionally, FBIs have incubation periods ranging from a few hours to 50 days, which can lead to the likelihood that some of the online reviewers might be misattributing their FBI experience to the restaurant they ate at most recently. The main intent of our study, however, was not to develop an exhaustive list of cleanliness and food safety issues within restaurants or to identify which specific restaurants contained the highest frequencies. Rather, our study sought to analyze trends within the food service market as a whole. Conclusion FBI cases are underreported due to several complexities such as long incubation periods of pathogens; lack of reporting to public health o†cials; and, in some instances, illness severity not warranting a healthcare visit. In the food service context, however, it is crucial to have a multipronged approach to investigate plausible causes of FBIs and design a proactive strategy to address these issues, especially because 61% of all FBI outbreaks are attributed to restaurants (CDC, 2018b). Hence, the goal of our study was to use text mining on a data set containing 231,381 online customer reviews from 954 restaurants in Houston, Texas. Our findings demonstrate statistically significant inverse correlations between the increased frequency of keywords in online reviews and customer satisfaction. Future research could investigate correlations between text mining using large data sets and the occurrence of FBI outbreaks in a hyperlocal setting to validate the e†cacy of these methodologies in real time. The results of our study can be used by public health practitioners to obtain customer perspectives on restaurant cleanliness and sanitation. Similarly, restaurateurs can determine what areas of cleanliness and sanitation are most important to customers to ensure repeat business and inform restaurant marketing strategies. Corresponding Author: Sujata A. Sirsat, Conrad N. Hilton College of Global Hospitality Leadership, University of Houston, 4450 University Drive, Houston, TX 77204-3028. Email: sasirsat@central.uh.edu. Average Restaurant Cleanliness Keyword Frequency by Restaurant Category 0 1 2 3 4 5 6 Fast Food Fast Casual Food Truck Moderate Midscale Upscale Fine Dining Average Keyword Frequency Restaurant Category Full Service Limited Service FIGURE 5 Average Foodborne Illness Keyword Frequency by Restaurant Category 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Fast Food Fast Casual Food Truck Moderate Midscale Upscale Fine Dining Average Keyword Frequency Restaurant Category Full Service Limited Service FIGURE 6 References on page 22

RkJQdWJsaXNoZXIy NTU5MTM=