Prevalence and predictors of comorbid diabetes among newly diagnosed Hispanic cancer patients in Connecticut

Background: U.S. Hispanics are at elevated risk of diabetes, and comorbid diabetes has been associated with poorer prognosis among patients with certain cancers. This study examined prevalence and predictors of comorbid diabetes in newly diagnosed Hispanic cancer patients in Connecticut. Methods: For all Hispanic patients diagnosed (at age 10+ years) with cancer in 2002–2003 and reported to the population-based Connecticut cancer registry, information from hospital medical records in the state was sought for comorbid diabetes (i.e., in any hospital admission 12 months prior to cancer diagnosis) and its treatment, body weight and height, preferred or primary language, and type of health insurance. Multiple logistic regression (MLR) models were used to examine predictors of presence of comorbid diabetes. Results: Comorbid diabetes was found for 192 (25.1%) of 765 patients, and was statistically significantly associated with preferred language (30.3% of 535 with non-English vs. 15.1% of 166 with English), age, and body mass index in a MLR model. Type of diabetes treatment was found for 92% of the 192 patients with diabetes. Conclusions: The high prevalence of comorbid diabetes among newly diagnosed Hispanic cancer patients, especially those without English as their preferred language, should be an impetus to further research on quality of diabetes care and survival of these patients.