W H Helderman1, J Mulder, M A van'T Hof, G J Truin. 1. WHO Collaborating Centre for Oral Health Care Planning and Future Scenarios, University of Nijmegen, Nijmegen, The Netherlands. w.vanpalenstein@dent.kun.nl
Abstract
OBJECTIVE: This study aimed to validate a caries prediction method, the Dentoprog-Method, which was developed on Swiss children. METHOD: A Dutch child population was used for validation. The diagnosis of caries, age of the children and the caries experience were slightly different from the population from which the Dentoprog-Method was developed. With the Dentoprog-Method, Dutch children were ranked in increasing order of high caries risk probability. The measure of prediction was expressed in sensitivity and specificity, which were calculated at 10% risk probability intervals for each class of "high" caries increment and for different forecast periods. ROC curves presenting sensitivity as a function of (100-specificity) were employed to summarize the obtained information. The area under the ROC curve was used as a measure of predictive accuracy. RESULTS: The area under the ROC curves for various caries increment classes of 7.5-year-old and 9.5-year-old Dutch children was in the range of the area under the ROC curves of Swiss children. CONCLUSION: This analysis indicated that the Dentoprog-Method when applied on a sample of Dutch children was robust enough to overcome small differences in caries diagnosis, age and caries experience.
OBJECTIVE: This study aimed to validate a caries prediction method, the Dentoprog-Method, which was developed on Swiss children. METHOD: A Dutch child population was used for validation. The diagnosis of caries, age of the children and the caries experience were slightly different from the population from which the Dentoprog-Method was developed. With the Dentoprog-Method, Dutch children were ranked in increasing order of high caries risk probability. The measure of prediction was expressed in sensitivity and specificity, which were calculated at 10% risk probability intervals for each class of "high" caries increment and for different forecast periods. ROC curves presenting sensitivity as a function of (100-specificity) were employed to summarize the obtained information. The area under the ROC curve was used as a measure of predictive accuracy. RESULTS: The area under the ROC curves for various caries increment classes of 7.5-year-old and 9.5-year-old Dutch children was in the range of the area under the ROC curves of Swiss children. CONCLUSION: This analysis indicated that the Dentoprog-Method when applied on a sample of Dutch children was robust enough to overcome small differences in caries diagnosis, age and caries experience.