A predictive model for significant periodontal disease progression: A large-scale cohort study

Area of Science:

• Periodontology
• Machine Learning in Healthcare
• Predictive Analytics

Background:

• Periodontitis progression is difficult to predict.
• Electronic health records (EHRs) offer a valuable data source.
• Identifying high-risk patients is crucial for timely intervention.

Purpose of the Study:

• To develop and validate a machine learning model for predicting significant periodontal disease progression.
• To identify key clinical and demographic predictors of disease advancement.

Main Methods:

• Retrospective cohort study of 4,117 patients with longitudinal periodontal data.
• Random Forest Classifier trained on baseline data (demographics, smoking, systemic conditions, periodontal measures).
• Primary outcome: significant progression defined as mean Clinical Attachment Level (CAL) worsening by ≥1 mm over ≥24 months.

Main Results:

• 28.0% of patients experienced significant progression over a mean follow-up of 34.7 months.
• The Random Forest model achieved an AUC-ROC of 0.82, accuracy of 81.6%, and recall of 79.2%.
• Key predictors identified: baseline mean CAL, smoking status, age, and diabetes.

Conclusions:

• Machine learning effectively predicts periodontal disease progression using real-world EHR data.
• Initial disease severity, smoking, age, and diabetes are critical risk factors.
• The model shows potential for personalizing periodontal treatment strategies.