Pilot validation of a digital tool for predicting the risk of temporomandibular disorders: integration of Delphi and ICONIX

Introduction: Temporomandibular disorders (TMD) are prevalent conditions that often remain underdiagnosed due to limited access to standardized tools. Objective: To conduct a preliminary evaluation of the diagnostic performance of a web-based screening application for TMD, developed as a digital health tool aimed at early detection and support clinical decision-making. Methods: A cross-sectional pilot study was conducted between March and May 2024 at the Department of Clinical Surgery, School of Stomatology, Benemérita Autonomous University of Puebla (Mexico). A convenience sample of 18 undergraduate students (>18 years) completed a 10-item self-administered web questionnaire, designed using the Delphi method and weighed through the ICONIX modeling approach. Clinical evaluation was performed blinded to the application results, using the Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) as the reference standard. Statistical analysis included descriptive statistics, Student’s t-test, Cohen’s kappa coefficient, and ROC curve analysis. Results: The web application demonstrated promising diagnostic performance, correctly identifying 12 of 13 clinically confirmed TMD cases. Sensitivity was 92.3%, specificity 60%, positive predictive value 85.7%, and overall, diagnostic accuracy 83.3%. The area under the ROC curve (AUC) was 0.762, with a Youden index of 0.52. Cohen’s kappa coefficient was 0.56, indicating a moderate level of agreement between digital classification and clinical diagnosis. No significant differences were observed in mandibular functional variables between groups with and without TMD. Conclusion: This preliminary pilot study suggests that the web-based application is a feasible and moderately accurate tool for syndromic risk stratification of TMD. While the results are encouraging, the small sample size and limited population generalizability highlight the need for future studies with larger and more diverse samples to validate its diagnostic performance.