Risk assessment for canine periodontal disease using a hybrid causal Bayesian network.

Bayesian network accurately predicts canine periodontal disease risk using modifiable and non-modifiable factors.

This study presents a hybrid Bayesian network model for assessing canine periodontal disease risk, addressing the significant gap between disease prevalence and clinical detection in veterinary medicine. Researchers constructed a directed acyclic graph mapping causal relationships between risk factors, then developed a Bayesian network integrating data from 9.5 million electronic health records, 2,600 owner questionnaires, published studies, and expert opinion. The final model contained 19 nodes representing various risk factors and clinical indicators. The model effectively differentiated high-risk from low-risk breeds and captured associations with age, body size, head morphology, and dental hygiene practices. Clinical validation demonstrated strong predictive performance with ROC AUC values ranging from 0.583 to 0.962 across four independent datasets. Key findings showed baseline periodontitis probability of 12.4%, increasing to 47% with gingivitis presence. The bidirectional inference capability enables risk assessment using any combination of variables and can function for both probabilistic inference and causal prediction of intervention outcomes, providing a decision-support tool for clinical practice.

This summary was distilled by AI and may occasionally misinterpret data. Confirm critical details with the primary literature before clinical application.