Canine Periodontal Disease Risk Assessment with a 19-Node Bayesian Network

A hybrid Bayesian network model for assessing canine periodontal disease risk, integrating data from 9.5 million electronic health records, 2,600 owner questionnaires, previous studies, and expert elicitation. The model comprises 19 nodes with 101 states and over 33,200 conditional probabilities, validated across four independent datasets. It was authored by Ciaran O’Flynn and last updated in April 2026.

Use Cases

• Predicting periodontal disease probability in dogs based on breed, age, and morphology.
• Evaluating the impact of modifiable factors like dental hygiene on disease risk.
• Validating causal relationships between clinical indicators such as biofilm presence and gingivitis.
• Supporting clinical decision-making through bidirectional probabilistic and causal inference.

Strengths

• Model integrates a large volume of data from 9.5 million electronic health records and 2,600 questionnaires.
• The network structure is explicitly defined with 19 nodes, 101 states, and over 33,200 conditional probabilities.
• Validation performance is reported with ROC AUC values ranging from 0.583 to 0.962 across four datasets.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The primary data file is a DOCX document, which may require parsing to extract structured data.

Provenance

Source

figshare

Collection Method

Constructed from electronic health records, questionnaires, previous studies, and expert elicitation.

Time Range

null

Freshness

Last updated 2026-04-23 04:21:40; freshness should be verified.

Geography

null