Application of fuzzy consensus for oral pre-cancer and cancer susceptibility assessment

Health questionnaire data assessment conventionally relies upon statistical analysis in understanding disease susceptibility using discrete numbers and fails to reflect physician’s perspectives and missing narratives in data, which play subtle roles in disease prediction. In addressing such limitations, the present study applies fuzzy consensus in oral health and habit questionnaire data for a selected Indian population in the context of assessing susceptibility to oral pre-cancer and cancer. Methodically collected data were initially divided into age based small subgroups and fuzzy membership function was assigned to each. The methodology further proposed the susceptibility to oral precancers (viz. leukoplakia, oral submucous fibrosis) and squamous cell carcinoma in patients considering a fuzzy rulebase through If-Then rules with certain conditions. Incorporation of similarity measures using the Jaccard index was used during conversion into the linguistic output of fuzzy set to predict the disease outcome in a more accurate manner and associated condition of the relevant features. It is also expected that this analytical approach will be effective in devising strategies for policy making through real-life questionnaire data handling.