Geographic analysis of immunization patterns in Michigan using MCIR data

Introduction: Information about concentrations of underimmunized children is useful for improving immunization programs. The Michigan Care Improvement Registry (MCIR) contains immunization, address, and geographic coordinate information for most of Michigan's children, which can be used to map areas of low immunization coverage. Materials / Methods: MCIR supplied records for children aged 19-35 months. Density mapping (ArcView 9) and cluster analysis (SaTScan 7) were applied to each of 7 recommended childhood vaccines. The unit of analysis was the child (completely unimmunized child vs. partially/completely immunized child) or the immunization (immunization given vs. not given, where >1 immunization is recommended; e.g., with DTaP, each child counts 4 times, once per recommended DTaP immunization). Results (Genesee County): MCIR provided records on 9327 children; 89% had geographic coordinates. Children without coordinates were more likely to be unimmunized (OR ≈ 2). The percentage of unimmunized children ranged from 15% (varicella) to 3% (hepatitis B). Density mapping displayed great detail; Northwestern Flint had higher densities of unimmunized children (and children in general). Significant DTaP underimmunization clusters covered Flint and northern Genesee county; in contrast, PCV7 underimmunization clusters covered parts of Flint and southern Genesee county. Additional geographic analysis is in progress. Discussion: Density mapping and cluster analysis complement each other. Density mapping shows where many underimmunized children are near each other. Cluster analysis detects areas where more underimmunization exists than expected from population density. This information can be used to improve immunization coverage or outbreak response; Genesee County is using it to plan immunization outreach efforts.