Objectives: This study describes determinants of dropout in a three-year community intervention trial of the effects of sugar-substituted chewing gums on caries progression rates.
Methods: A total of 602 children aged 9-14 years from 28 school classes in five secondary schools in Kaunas, Lithuania, were given a clinical and radiographic baseline caries examination. The schools were assigned randomly to one of the following interventions: sorbitol/carbamide gum, sorbitol gum, xylitol gum, placebo gum, or no gum. Children in the four intervention schools were asked to chew at least five pieces of chewing gum per day, preferably after meals. The children were reexamined clinically after one, two, and three years of study, and radiographically after three years.
Results: A total of 33 children (6%) had dropped out before the one-year clinical examination, an additional 29 children (5%) dropped out before the two-year examination, and a further 108 children (18%) dropped out before the final three-year clinical examination. A total of 230 children (39%) were not available or refused to participate in the three-year radiographic examination. Analyses using random effect logit models showed that, irrespective of time of follow-up, most of the cluster variation in dropout was related to school classes within the primary randomization units, the schools. The most important predictors of individual dropout were age and baseline caries experience, whereas sex was not associated with dropout.
Conclusions: The results show that a community intervention trial of chewing gums carried out among schoolchildren is subject to cluster effects. Dropout was not primarily related to the randomization units themselves, i.e. the schools, but rather to subclusters of classes within the schools. These findings should be considered when designing community intervention trials and practical preventive programs among schoolchildren.