ABSTRACT.Context. An overarching national health goal of Healthy People 2010 is to eliminate disparities in leading health care indicators including immunizations. Disparities in US childhood immunization rates persist, with inner-city, black, and Hispanic children having lower rates. Although practice or clinic-based interventions, such as patient reminder/recall systems, have been found to improve immunization rates in specific settings, there is little evidence that those site-based interventions can reduce disparities in immunization rates at the community level.Objective. To assess the effect of a community-wide reminder, recall, and outreach (RRO) system for childhood immunizations on known disparities in immunization rates between inner-city versus suburban populations and among white, black, and Hispanic children within an entire county.Setting. Monroe County, New York (birth cohort: 10 000, total population: 750 000), which includes the city of Rochester. Three geographic regions within the county were compared: the inner city of Rochester, which contains the greatest concentration of poverty (among 2-yearold children, 64% have Medicaid); the rest of the city of Rochester (38% have Medicaid); and the suburbs of the county (8% have Medicaid).Interventions. An RRO system was implemented in 8 city practices in 1995 (covering 64% of inner-city children) and was expanded to 10 city practices by 1999 (covering 74% of inner-city children, 61% of rest-of-city children, and 9% of suburban children). The RRO intervention involved lay community-based outreach workers who were assigned to city practices to track immunization rates of all 0-to 2-year-olds, and to provide a staged intervention with increasing intensity depending on the degree to which children were behind in immunizations (tracking for all children, mail, or telephone reminders for most children, assistance with transportation or scheduling for some children, and home visits for 5% of children who were most behind in immunizations and who faced complex barriers).Study Participants. Three separate cohorts of 0-to 2-year-old children were assessed-those residing in the county in 1993, 1996, and 1999.Study Design. Immunization rates were measured for each geographic region in Monroe County at 3 time periods: before the implementation of a systematic RRO system (1993), during early phases of implementation of the RRO system (1996), and after implementation of the RRO system in 10 city practices (1999). Immunization rates were compared for children living in the 3 geographic regions, and for white, black, and Hispanic children.Immunization rates were measured by the same methodology in each of the 3 time periods. A denominator of children was obtained by merging patient lists from the practice files of most pediatric and family medicine practices in the county (covering 85% to 89% of county children). A random sample of children (>500 from the suburbs and >1200 from the city for each sampling period) was then selected for medical chart review at practices...