The purpose of this study was to determine: (1) whether preschool-age patients who utilize the emergency department (ED) are undervaccinated compared with patients having the same primary care provider and (2) whether reducing missed vaccination opportunities in the primary care office can potentially reduce the differences in undervaccination between the groups. This retrospective cohort study involved two groups: 583 ED patients, aged 4 to 48 months, who had primary care providers; and 583 control subjects randomly selected from primary care sites and matched according to date of birth and primary care site. The major outcome variable was the point prevalence of undervaccination, defined as more than 60 days past due for a vaccine at the time of the ED visit, and for control subjects, at the time of their matched patient's ED visit. Demographic variables, vaccination history, presence of chronic illness, and office utilization history were abstracted from office charts. The mean age of all patients was 20.0 months. Emergency department patients were more likely to be boys (61% vs 50%) and had more chronic illness, but did not differ racially from those in the control group. Primary care sites included a hospital-based clinic (n = 137), neighborhood health centers (n = 172), and private practices (n = 274). The undervaccination rates by primary provider type were for (1) hospital clinic ED patients 21.1%, control subjects 19.7%; (2) neighborhood health center ED patients 29.1%, control subjects 22.7%; and (3) private practice ED patients 26.6%, control subjects 14.9%. Overall, the odds ratio of ED patients' being undervaccinated compared with control subjects was 1.8 (95% confidence interval 1.3 to 2.5).(ABSTRACT TRUNCATED AT 250 WORDS)