The study forming the basis of this thesis was designed to examine the feasibility of a partnership between health visitors and dentists to access pre-school children for data collection which could allow identification of those children at risk of developing dental caries. The purpose of the investigation was twofold: 1) to investigate the feasibility of a partnership between dentists and health visitors (existing health services personnel) to access pre-school children in order to collect dental, microbiological, health behaviour and socio-demographic data at ages 1,2,3 and 4-years and 2) to develop a novel caries risk assessment model (using such data) for the identification of 4-year old pre-school children at high risk of developing dental caries. The basis for the investigation was a prospective 4-year longitudinal study of consented children from age 1- to 4-years inclusive. This was the first, large scale longitudinal study of pre-school children to involve a consented, but nonexclusive, population cohort. The cohort comprised all those children born and resident in Dundee between 1 April 1993 and 31 March 1994 for whom written consent was obtained by the child's health visitor at 8-months of age (n = 1683). Health visitors obtained microbiological (saliva sampling) and sociodemographic (parental and health visitor questionnaires) data in partnership with a study dentist collecting dental data (dental examination).The results of the study suggest that health visitors could, within their daily caseload of duties, both access the majority of pre-school children and,independently, collect caries risk assessmendt ata relating to these children with sustained diligence over a 4-year period. Risk model development was carried out using both logistic regression and CHAID (Chi-squared Automatic Interaction Detector) analyses. Data collected at age 1-year was used to predict caries at age 4-years. This resulted in the development of the Dundee Caries Risk Model (DCRM) (sensitivity 69% and specificity 60%) (n = 784). The key predictive factors in this model were type of housing, use of a feeder cup and use of vitamins. It might have been assumed that microbiological factors would be of significance. However, they were not found to be sufficiently predictive for incorporation into the model. This reduces the cost and increases the simplicity of the risk model.Development of the DCRM may facilitate preventive care being targeted towards those at risk of developing dental caries in order to prevent overt manifestation caries in pre-school populations.