BEYOND BINARY: FUZZY PROPORTIONS FOR REAL-WORlD CLINICAL DECISIONS

This article introduces a fuzzy statistical framework for estimating proportions in biostatistical contexts where observations are imprecise or partially defined. Traditional methods assume binary outcomes, but clinical data often involve diagnostic ambiguity and graded assessments. We model such observations as trapezoidal fuzzy numbers and define a fuzzy proportion estimator based on fuzzy arithmetic. The method is applied to a real-world case study involving depression screening, where physicians express varying levels of diagnostic confidence. Results demonstrate that the fuzzy estimator preserves uncertainty and provides a more expressive summary than classical point estimates. The proposed approach maintains compatibility with traditional statistics while offering greater interpretability in the presence of vague or subjective data.