The ready access provided by telemedicine benefits families and society but might increase total healthcare utilization with uncertain implications for costs. The objective of this study was to assess the net impact on healthcare utilization of introducing into inner-city childcare a telemedicine model designed to manage acute illness. A cohort study was done using comparable periods before and after introduction of telemedicine for all qualifying children (n = 112) using three innercity childcare centers. Because the utilization histories of these children differed in length, we chose child-months as the unit of analysis. Acute illness visits were ascertained for 1806 child-months among the 112 qualifying children. Following telemedicine startup, children's office and emergency department (ED) visits for illness fell by 1.73 and 0.20/child/year, respectively, replaced by telemedicine visits at 1.07/year. These observations could be misleading, however, because of the possibility of confounding factors. For example, the cohort aged during observation, and illness visits fall with age. Accordingly, in multivariate analysis we adjusted for season of the year, age, and within-child correlation. In this analysis, reduction in illness utilization overall tended toward an increase (rate ratio = 1.26, p = 0.13). The worst-case estimate (based on upper 95% confidence interval for rate ratio) for increase in illness utilization was 3.38 visits/child/year, and the most likely case was an increase of 1.26. Assuming (1) the worst-case effect (largest increase) on overall utilization and (2) reimbursement for ED, office, and telemedicine visits of 350 dollars, 45 dollars, and 45 dollars, respectively, the healthcare system would break even on telemedicine if it replaced 0.50 ED visits per child annually.