Correlates Between Colorectal Cancer Screening and Diabetes Mellitus Among Residents of Ohio Appalachia

Abstract

Introduction: Colorectal cancer (CRC) is the third mostly commonly diagnosed cancer and leading cause of cancer mortality in the United States. Notably, Appalachian regions have higher CRC incidence and mortality rates and lower screening rates than non-Appalachian regions. Previous investigation has shown that having a primary care provider, patient-provider communication, and social influence are correlated with CRC screening patterns among Appalachian Ohio residents. However, it is not established whether these factors differ among residents with diabetes mellitus (DM), an independent risk factor of CRC, as compared to those without. Aim: To characterize CRC screening behaviors and barriers among patients with and without DM. Methods: We conducted interviews with and examined medical record data from Appalachian Ohio residents 50-75 years of age from September 2009 to September 2013. We identified correlates of being within CRC screening guidelines using multivariable logistic regression. Multiple imputation using the fully conditional specification method was used to impute both income and the CRC screening outcome where missing. In order to determine whether there were demographic factors that differentially predicted CRC screening in DM vs. non-DM participants, we fit multivariable logistic regression models that included an interaction between the factor and DM for each of the imputed datasets. Demographic predictors were considered separately. Estimates were combined using SAS PROC MIANALYZE (v9.3, SAS Institute, Cary, NC). Results: There were 4,401 participants included in the analysis. Among the 3,771 participants whose CRC screening status was known, 590 (15.6%) had DM and 1,359 (36.0%) were within CRC screening guidelines. Using the multiple imputations, participants with DM who asked their doctor to order a screening test (OR=5.05, 95% CI: 2.47-10.34) had higher odds of screening than those without DM (OR=2.43, 95% CI: 1.92-3.07, interaction p=0.04). Similarly, divorced, widowed or separated participants with DM had higher odds of CRC screening compared to never married participants (OR=1.98, 95% CI: 1.17-3.37) than those without DM (OR=1.20, 95% CI: 0.76- 1.89, interaction p=0.02). However, overall, there was no significant difference in the CRC screening rate between patients with DM and those without (OR=0.94, 95% CI: 0.79-1.09). Conclusion: Adherence to CRC screening guidelines is low among Appalachian Ohio residents. Although the association between DM and CRC risk is well-established, there was no significant difference in the screening rate among residents with DM as compared to those without DM. However, this study underscores the importance of targeting patient-provider communication in strategies to improve CRC screening rates.