Use of public school immunization data to determine community-level immunization coverage.

OBJECTIVES To evaluate whether immunization data collected on a child's entry into kindergarten, i.e., Chicago Public School Immunization Data (PSID), was comparable to coverage levels determined by the National Immunization Survey (NIS) and to use these data to identify community areas with consistently low immunization coverage. METHODS The Chicago Department of Public Health obtained four years of PSID (2000-2003); these data included demographic information, home address, and immunization records. Coverage levels were determined in two ways: (1) one dose of measles-containing vaccine (MCV) and (2) four doses of diphtheria and tetanus toxoids and pertussis vaccine, three doses of poliovirus vaccine, and one dose of measles-containing vaccine (the 4:3:1 series), stratified by racial/ethnic group; these levels were compared to NIS estimates for the respective time periods. We used geographic information system software to illustrate variations in coverage levels between distinct community areas within Chicago. RESULTS Year 2000 MCV coverage levels determined from PSID closely approximated NIS estimates (84.6% vs. 87.2% +/- 4.6%, respectively). MCV coverage levels determined by race/ethnicity from PSID were within the 95% confidence intervals (CI) for all racial categories (white, 89.5% vs. 92.2% +/- 6.4%; black, 79.0% vs. 83.5% +/- 9.4%; Hispanic, 89.5% vs. 87.5% +/- 5.8%). Comparison of PSID and NIS 4:3:1 coverage levels revealed similar findings. For each study year, PSID identified 12 community areas with consistently low MCV coverage levels, i.e., < 80%. CONCLUSIONS PSID closely approximated NIS coverage estimates for MCV and 4:3:1 immunization. These methods can be used by state and city health departments to identify and direct resources to communities at greatest need.