The basis for clinicians' caries risk grouping in children

Despite the growing emphasis on targeting care to high-risk groups, little is known about the factors used by clinicians to designate risk. This study investigates the degree to which factors measured in a typical dental health survey are used by clinicians in assigning children to dental disease risk groups. A random sample of 9690 children aged 5-15 years was selected from the South Australian School Dental Service. Dentists or dental therapists judged each child as low-, medium-, or high-risk for dental disease. Clinicians recorded caries experience (DMFS/dmfs), and children's parents completed a questionnaire on dental behaviors and socioeconomic status (SES). Two binary logistic models were fitted using the risk grouping as the outcome variable, one comparing low- with moderate-risk and the other comparing moderate- with high-risk. Sixty percent of children were judged as moderate-risk, 27% as low-risk, and only 13% as high-risk. In the logistic models, proximal DMFS/dmfs were stronger predictors of assignment to the higher-risk groups than were factors indicating past occlusal caries, while factors describing caries on buccal or lingual surfaces appeared unimportant. Untreated lesions on permanent and primary teeth were among the strongest predictors of assignment to the higher-risk groups. Other significant factors (P < 0.05) were: exposure to professionally applied fluoride and sealants, country of birth, frequency of toothbrushing, and exposure to fluoridated water. No SES factors reached significance. The models explained nearly one-half the variation in the risk predictions. Clinical markers of past caries experience explained the greatest variation in the judgments, showing that clinicians base their risk predictions largely on children's past disease. The three types of surfaces contribute unevenly to the judgments and unrestored caries was the largest contributor to the decision.