“Dropouts” from a six-year longitudinal study of adolescent caries

Objectives: A radiographic study of approximal caries and caries-associated factors in adolescents was performed in a group, initially aged 14y, and followed-up for 6 years. It aimed to investigate subjects that dropped out of the investigation and the effect this had on the overall results. Methods: A random sample of 148 subjects living in: a) fishing; b) rural, and c) urban communities was investigated for approximal caries, using bitewing radiographs, and caries-risk factors using a structured questionnaire (Árnadóttir et al., Community Dent Oral Epidemiol 1998). The sample was followed up at the ages of 16y and 20y using the same methods. Results: The mean increment in approximal caries from 14-16y was 2.4 (s.d. 4.8; n=123) lesions or 1.2 lesions/ subject/ year and was 4.46 (s.d. 5.7; n=53) lesions or 1.1 lesions/ subject/ year by 20y (t=0.2; n.s.). The proportions of subjects caries-free at 14y,16y, 20y, were 29.5%, 20% and 11.6% respectively. Dropouts from this study occurred mostly after 16 y when many subjects changed residence, often due to further education. Analysis of subjects dropping out of the study showed that they were least from the rural community and most from the fishing community (47.4%; chi2 8.091; p<0.02). Furthermore, those dropping out had a history of lower attendance at their dentist (chi2 10.861; p<0.02) and had a higher frequency of sugar consumption (chi2 8.14; p<0.05). Dropouts showed a trend to having a higher caries prevalence and incidence of approximal caries by the age of 16y. Conclusions: It appears that subjects with a high-risk behaviour or residence in a community with high caries prevalence were more likely to drop out of the study. Although there are many advantages in conducting longitudinal studies of caries, the conclusions may become biased because of the distribution of subjects becoming increasingly skewed during the study period.