Does Context Affect SCHIP Disenrollment? Findings from a Multilevel Analysis

Jane E. Miller, Rutgers University
Julie A. Phillips, Rutgers University

The State Children's Health Insurance Program (SCHIP) was enacted in 1997 to improve coverage of low-income children. Most existing analyses of SCHIP disenrollment ignore variation in time since enrollment, producing biased estimates of disenrollment rates. Moreover, few studies report variation in rates by demographic, socioeconomic or geographic characteristics. We estimate survival models within a multilevel model, using family level data on over 24,000 families enrolled in NJ KidCare and county characteristics from the Area Resource File, Census, and NJ FamilyCare physician roster. Disenrollment was markedly higher in plans involving cost-sharing, particularly among black families. Families with infants or with several children enrolled were less likely to disenroll. These findings were robust to inclusion of a wide variety of county-level demographic, socioeconomic, programmatic and insurance characteristics. Of these, only geographic density of NJ KidCare physicians and population density were significantly related to the overall risk of disenrollment.