OBJETIVE: To examine the relationship between access to health care and undiagnosed diabetes among the high-risk, vulnerable population in the border region between the United States of America and Mexico. METHODS: Using survey and fasting plasma glucose data from Phase I of the U.S.-Mexico Border Diabetes Prevention and Control Project (February 2001 to October 2002), this epidemiological study identified 178 adults 18-64 years old with undiagnosed diabetes, 326 with diagnosed diabetes, and 2 966 without diabetes. Access to health care among that sample (n = 3 470), was assessed by type of health insurance coverage (including "none"), number of health care visits over the past year, routine pattern of health care utilization, and country of residence. RESULTS: People with diabetes who had no insurance and no place to go for routine health care were more likely to be undiagnosed than those with insurance and a place for routine health care (odds ratio [OR] 2.6, 95% confidence interval [CI] 1.0-6.6, and OR 4.5, 95% CI 1.4-14.1, respectively). When stratified by country, the survey data showed that on the U.S. side of the border there were more people with undiagnosed diabetes if they were 1) uninsured versus the insured (28.9%, 95% CI 11.5%-46.3%, versus 9.1%, 95% CI 1.5%-16.7%, respectively) and if they 2) had made no visits or 1-3 visits to a health care facility in the past year versus had made > 4 visits (40.8%, 95% CI 19.6%-62.0%, and 23.4%, 95% CI 9.9%-36.9%, respectively, versus 2.4%, 95% CI -0.9%-5.7%) (all, P < 0.05). No similar pattern was found in Mexico. CONCLUSIONS: Limited access to health care-especially not having health insurance and/or not having a place to receive routine health services-was significantly associated with undiagnosed diabetes in the U.S.-Mexico border region.