Developing a Pediatric Database to Facilitate Dental Caries Risk Assessment

Grayson Cole
University of Minnesota 
Dental School

Grayson Cole1, Neel Shimpi1; Rajesh Koralkar2; Ingrid Glurich1, Amit Acharya1,2
1Institute for Oral and Systemic Health, 2Biomedical Informatics Research Center

Research area:  Institute for Oral and Systemic Health

Background: High prevalence of dental caries is an established health issue among the pediatric population with immediate and long-term impacts on oral health. Previous studies have indicated that factors including fluoride, lead, and smoke exposure; dietary habits; and socioeconomic status correlate with future incidence of dental caries. This study sought to evaluate and expand the spectrum of risk factors for developing dental caries in primary dentition.

Methods: Eligible patients (18 month well child visit between January 2003 and November 2009; ≥1 dental visit between age 3 and 8 years) were identified from Marshfield Clinic’s Data warehouse. Unstructured data were manually abstracted from 1,700 handwritten 18 month visit ink-over forms and were combined with variables extracted from structured fields found within the data warehouse. Those that were not relevant to analyses of dental caries, not sufficiently diverse, or with ≥50% missing values were excluded from statistical analysis. Risk factors achieving statistical significance by the chi square test and t-test were used in logistic regression to develop a risk-assessment model.

Results: Among 1700 records abstracted, 86% (n=1461) included oral evaluation; comprising 902 cases (primary dentition caries found and/or restored) and 559 controls (no caries). Median age at first dental visit was 5 years. Of 49 variables identified, 19 met inclusion criteria for regression modeling. Healthy head circumference, fluoride in water, use of an interpreter, and past missed medical appointments were associated (p < 0.05) with higher incidence of primary dentition caries.  However, correlation of fluoride was inconclusive due to a higher missing data rate. 

Conclusions: The variables which achieved significance in this study were not previously identified as risk factors for caries in primary dentition. Validation of these candidate risk factors is required. A more standardized and structured ink-over form could contribute to a more universal tool to support care delivery.