2025, Vol. 10, Issue 12, Part B

Oral cancer is a major public health issue in Assam, largely due to the prevalent use of areca nut, tobacco, smoking and alcohol. This study uses binary logistic regression to examine factors influencing the prognosis of 1000 oral cancer patients. Patient status (either “Death” or “Alive”) is the dependent variable, with factors categorized as Personal (age, gender, behavioural factors, education, family income, food habits and locality [rural / urban]) and Clinical (cancer-directed treatment [CDT], follow-up, topography sites, stage groups, treatment given prior to registration [TGPR], metabolic risk factors [MRF] [over weight, raised BP and raised BG]). The model correctly predicts 99.0% when selected variables are included. Significant factors include Behavioural factors, Family Income, CDT, Follow up, Stage groups, TGPR, MRF and Locality (Rural / Urban). The non-significant factors include age, gender, education, food habits and topography sites. The study concludes that more consumption of alcohol, not assigned CDT, TGPR, higher MRF and the patients of urban areas increase the risk of oral cancer and the others have no significant impact. The model fits the data and demonstrates strong predictive accuracy.