Objective: The objective was to evaluate the predictive validity of the American Dental Association's caries risk assessment (CRA) tool, adapted with permission, and used by the U.S. Department of Veterans Affairs dental services within their electronic dental record.
Methods: This analytic epidemiologic study with a retrospective longitudinal design included Veterans who had a minimum of three years of available data. The primary outcome was caries-related treatment during the twelve-month predictive period following the CRA category identification.
Results: The sample included 57,675 Veterans; 50.1% classified as low, 33.2% as moderate and 16.8% as high caries risk. During the twelve-month predictive period, both teeth/person and teeth with caries-related treatment rose sequentially from low to high CRA categories. However, poor sensitivity (0.34-0.58) and better specificity (0.53-0.78) values were observed. Similarly, better negative predictive values (0.72-0.79) compared to positive predictive values (0.28-0.34) were found. Adjusted logistic regression models showed that current treated caries (caries detected at the time of the CRA exam) was more strongly associated with future caries-related treatment than the subjective CRA determination.
Conclusions: The subjective CRA tool evaluated in this study is better at identifying patients at low risk of requiring future caries-related treatment versus those at higher risk. This makes it difficult to target the most caries susceptible patients with preventive measures. Furthermore, we found that the amount of caries treatment required at the time of the CRA exam had the strongest association with caries-related treatment during the subsequent twelve-month predictive period.
Clinical significance: The amount of caries-related treatment required at the time of the caries risk classification is the strongest predictor of future caries.
Keywords: Caries risk assessment; Dental caries; Statistics; Vulnerable populations.
Copyright © 2021. Published by Elsevier Ltd.