DYNAMICS OF HEALTH INSURANCE COVERAGE Suk-fong S. Tang, Beth K. Yudkowsky Practice and Research, American Academy of Pediatrics, Elk Grove Village, IL

BACKGROUND: Health insurance data from national surveys are often used to guide policy makers in expanding coverage for the low-income population. However, estimates derived from these data sources often do not adequately describe the magnitude of the uninsured population as surveys typically do not record the duration of health insurance coverage during the survey period. Since health insurance coverage is indicated for even the shortest episode of coverage, estimates based on these data include only those lacking coverage for the full year and exclude those who are uninsured part or most of the year.

OBJECTIVE: To demonstrate the pervasiveness of gaps in health insurance coverage among children and young adults and show how operational definitions of the "uninsured" can change the estimate of the uninsured population.

DESIGN/METHODS: Using monthly health insurance coverage data from the 1998 Full Year Population Characteristics File of the Medical Expenditure Panel Survey, 3 measures were constructed to reflect (1) the proportion without coverage throughout the calendar year, (2) the proportion uninsured in an average month, and (3) the proportion lacking coverage at any time during the year, among children and young adults through age 21.

RESULTS: 10.8M persons through age 21, or 12.2%, were uninsured throughout 1998. In an average month, the estimate rose to 15.9M, or 18.3%. Finally, the estimate of the uninsured reached 23.0M, or 26.0%, when all persons uninsured for any period of time during 1998 were included. About 53.8M, or 60.8%, of this group had at least some episode of employment-based coverage in 1998, but 4.1M (7.6%) lost that coverage during the year. 61.1% of those who lost employment-based coverage became uninsured for 3 months or longer. An estimated 20.4M, or 22.9%, had Medicaid coverage in 1998, but 4.5M(22.1%) lost that coverage during the year. 57.4% of those disenrolled became uninsured for 3 months or longer subsequent to disenrollment.

CONCLUSIONS: How the uninsured is defined has serious financial and programmatic implications as policy makers consider the size of this population when allocating resources. While nearly 11M children and young adults were uninsured for all of 1998, an additional 12M had gaps in their insurance coverage that typically exceeded 3 months. This latter group presents a sizable challenge to providing health care for the needy, and therefore must not be undercounted or overlooked.