Validity of self-reported health plan information in a population-based health survey

To validate information on private health insurance coverage in a population-based study.

Respondents to the Massachusetts Behavioral Risk Factor Surveillance System were asked the name of their health plan company (affiliation) and specific brand of insurance (product), the duration in which they belonged to the plan, and demographic and health-related data. Information on plan affiliation and product was used to classify individuals on type of coverage. At the end of the survey, respondents with health insurance were asked to retrieve their health plan cards, and to read detailed information from the cards. Self-reported data were compared with information from the cards.

Self-reported information on health plan affiliation agreed with plan cards for 93 percent of individuals, while agreement was 79 percent for health plan product. Among health maintenance organization (HMO) participants, 93 percent correctly reported being in an HMO (sensitivity), whereas 76 percent of respondents in a non-HMO plan correctly self-reported (specificity). Individuals with higher levels of income, those with a primary care doctor, and those in a health plan for at least 1 year had higher agreement. Higher validity was associated with poor physical health and recent cancer screening.

Self-reported data on health plan affiliation and product have good validity in a population-based sample of adults. While agreement differs according to specific respondent characteristics, these differences do not appear substantial.