الفهرس 
ContentsOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
The aim of the present study was to develop a multivariate caries prediction model that describes the relationship of the probability of high caries risk status to the risk factors and background characteristics. To meet this aim, a sample was selected using a multistage stratified random technique. The study originally included 1024 children, but 210 children were lost to follow up so that 814 first graders were followed from their first year in school until they reached third grade. At baseline, data were collected by means of a clinical oral examination, questionnaire and bacteriologic sampling. At the final examination, only the clinical examination was repeated to determine the amount of change in caries.
Statistical modeling was performed using logistic regression analysis for permanent caries and multiple regression analysis for caries in the primary dentition. In both cases modeling was done using caries increment and fmal caries experience as response variables. All explanatory variables were used after excluding unreliable and correlated variables. Different levels of significance were selected in every case to allow the choice of a reasonable number of explanatory variables that can predict the outcome as accurately as possible.