In recent years a strategy of selective, symptom-based intervention of carious primary teeth has been developed amongst some British general dental practitioners. Practice-based studies appear to provide evidence that policies of restoration of symptomless carious primary teeth do not confer any significant benefits above those associated with non-restorative care. However, results from these studies contrast with those of many clinical trials and prospective studies of primary molar restorations. In the current investigation, cohort study data from 5,168 carious primary molar teeth from 2,654 British children aged 4-5 years at baseline, augmented with Dental Practice Board treatment data, was utilised to assess the effect of restorative treatment on the likelihood of carious teeth subsequently progressing to either exfoliation or extraction. The effect of demographic and tooth level covariates on the fate of these teeth was also assessed. Multivariate multilevel parametric survival models were applied to the analysis of the carious-exfoliation and carious-extraction transitions to which the teeth were subject, assuming an underlying data hierarchy with teeth nested within individuals. Time of occurrence of caries affected survival experience, with teeth in which caries occurred later in life being associated with higher survival rates to extraction. Amongst filled teeth, later fillings were also associated with higher survival rates to extraction. Demographic and tooth level variables had a limited effect on survival experience. Treatment was found to be significantly associated with survival with respect to extraction, with survival rates of over 80% at 14 years, double those of untreated teeth.