This study investigated the association between caries status and sealant need at a prior survey and subsequent sealant use in a Medicaid program. Clinical data from a 1986-87 statewide epidemiological survey (N = 8026) representative of North Carolina (NC) schoolchildren (grades K-12) were linked with all NC Medicaid dental claims submitted during 1987-92, yielding 570 children in the survey who had at least one dental visit during 1987-1992. From the 570, 390 children were included: 71 who received sealants (S) and 319 who received non-sealant care (NS). Children were excluded based on age, having preexisting sealants, or having no sealant-eligible molars or premolars. S and NS were compared on baseline dfs, DMFS, and sealant need, controlling for the patient's age, number of visits, and the provider's propensity to seal. At all ages, NS was twice as likely to have had prior dfs or DMFS (OR = 2.04, 95% CI = 1.15, 3.70). The association between sealant receipt and prior sealant need varied by age. At 6 to 11 years, S and NS had equal likelihood of sealant need (OR = 1.41, 95% CI = 0.62, 3.18). At 12 to 15 years, NS had a greater likelihood of sealant need (OR = 6.82, 95% CI = 1.60, 29.08). Caries-free status was associated with subsequent sealant receipt. Prior sealant need caused variability in dentists' decisions, depending on the child's age and past caries experience. Sealants were used infrequently by most providers and for a minority of patients. These findings are important for the Medicaid program and for future non-randomized studies of sealant effectiveness.