No evidence for marked ethnic differences in accuracy of self-reported diabetes, hypertension, and hypercholesterolemia

Objective: To assess whether the accuracy of self-reported diabetes, hypertension, and hypercholesterolemia in high-risk groups differs according to ethnicity. Study Design and Setting: We analyzed data of 430 patients at high risk of cardiovascular disease from different ethnic origin, including Turkish, Surinamese, and Dutch. Risk factors based on self-reports were compared with data from medical records and with a gold standard based on clinical measurements. Proportions of concordance between self-reports and other methods and kappa statistics (kappa) were determined by ethnicity. Results: Concordance between self-reports and other data sources was highest in diabetes and lowest for hypercholesterolemia. Agreement of self-reports was substantial to almost perfect for diabetes (kappa: 0.84-0.76), substantial to moderate for hypertension (kappa: 0.63-0.51), and moderate for hypercholesterolemia (kappa: 0.55-0.48). There was no statistically significant association between ethnicity and concordance, except for self-reporting of diabetes among Surinamese vs. Dutch indigenous patients (odds ratio = 0.37; 95% confidence interval: 0.14-0.97). Conclusion: There are no marked ethnic differences in the accuracy of self-reports of diabetes, hypertension, and hypercholesterolemia in high-risk populations. Larger studies including multiple ethnic groups are needed to confirm these findings. (C) 2007 Elsevier Inc. All rights reserved.