Screening 'Gum Disease or Problem' Status Using Molar Probing Depths: Model Development and Validation in NHANES...

Background/Objectives: A parsimonious multivariable model was engineered and externally validated using dental electronic health records to screen for a clinician-evaluated "gum disease or related problem" status. Designed for seamless electronic integration, a positive screening result triggers a targeted referral-and the model is well-suited for dental hygienist use- while providing clinical decision support for dentists during definitive diagnoses. This automated support standardises workflows, maximises team efficiency, and supports data-driven clinical decision-making. Methods: This cross-sectional study developed the model using the US National Health and Nutrition Examination Survey (NHANES) 2011-2012 cohort (n = 1732) and independently validated it on the 2013-2014 replication cycle (n = 1855) for the high-risk 30-50-year-old demographic. Survey-weighted logistic regression incorporated age, gender, and eight targeted molar probing depth (PD) measurements paired with binary indicators for tooth presence to predict whether an individual has a gum disease or related problem. Results: External validation demonstrated high parameter generalisability, yielding a pooled population-level AUC of 0.859 (95% CI: 0.828-0.885). At the prevalence-matched threshold of Pt = 0.367, the framework yielded a sensitivity of 78.39% (95% CI: 69.48-85.29%) and a specificity of 79.10% (95% CI: 74.55-83.04%). Demonstrating public health utility via Decision Curve Analysis, the model avoided approximately 39 unnecessary clinical examinations by a dentist per 100 individuals at a practical clinical decision threshold of 0.33, achieving a substantial pooled net benefit compared with a treat-all strategy without compromising case-detection metrics. Conclusions: This scalable screening model offers a highly practical pathway for gum disease and related problem interception. Optimised for seamless software integration, the framework facilitates rapid, hygienist-led risk stratification without increasing chairside burden, and delivers objective data streams to reinforce the dentist's definitive diagnostic workflow.