Our objectives were to explore health insurance status and insurance type, adjusted for self-reported and perceived health variables, as determinants of having and using a usual care provider in the Women's Health Initiative (WHI) Observational Study (OS). This analysis describes insurance status in a large, diverse group of older women and tests the hypothesis that insurance was a key predictor of their access to healthcare in the mid-1990s. Multiple logistic regression analysis was used to evaluate determinants of having visited a usual healthcare provider within the proceeding 12 months, using cross-sectional information provided by a population-based cohort of 55,278 postmenopausal women. Five percent of women younger than 65 years and 0.2% of women 65 or older in the OS cohort lacked health insurance. Among the 31,684 women, aged 50-64 years, Hispanic women and those with fewer years of education and lower household income and who were current smokers were less likely, and those lacking insurance were the least likely, to have seen their healthcare provider within the preceding year. Among 23,594 women, aged 65-79 years, African American and Hispanic women and those with lower household income, and Medicare only and those who were current smokers, were less likely to have seen their healthcare provider within the preceding year. In both age groups, women with chronic medical conditions and poorer perceived health scores and those with prepaid insurance were more likely to have seen their healthcare provider. In the WHI OS, both health (self-reported and perceived) and type of health insurance remained independently associated with having visited a usual healthcare provider after multivariate adjustment for one another as well as for pertinent sociodemographic characteristics.

BACKGROUND: Amid current changes in health care access across the United States, the importance of health insurance status and insurance type relative to demographic, actual, and perceived health variables as determinants of screening for breast, colorectal, and cervical cancer is uncertain. This analysis evaluates the hypothesis that health insurance independently predicts cancer screening in the Women's Health Initia tive Observational Study cohort. METHODS: Questionnaire data from 55,278 women en rolled in the Women's Health Initiative Observational Study between September 1994 and February 1997 were analyzed by multiple logistic regression to identify predictors of self-reported mammography within 2 years, Pap smear within 3 years, and stool guaiac or flexible sigmoidoscopy within 5 years. RESULTS: Positive determinants of reporting cancer screening were age, ethnic origin, household income, educational level, family history of cancer, having a usual care provider, time since last provider visit, and insurance status and type. Smoking, diabetes, and, among older women, prior cardiovascular events were negative determinants of cancer screening. Among women younger than 65, lacking health insurance or having fee-for-service insurance was strongly associated with failure to report cancer screening, independently of having or using a usual care provider and of demographics, self-perceived health, and health characteristics. Among women 65 and older, those with Medicare alone were less likely, whereas those with Medicare + prepaid insurance were more likely, to report cancer screening. CONCLUSIONS: In the Women's Health Initiative Obser vational Study, a large, diverse group of older women, health insurance type and status were among the most important determinants of cancer screening indepen dent of demographics, chronic health conditions, and self-perceived health characteristics.