Abstract TP194: Internet-based Control Recruitment for a Case-Control Study of Major Risk Factors for Stroke in Korea: A Failure and Lessons Learned from the Experience

Abstract

This study aimed to estimate the population attributable risks (PARs) of major risk factors for stroke in Korea through an internet-based case-control study and test the feasibility and validity of this approach. Control subjects were recruited by a web-based questionnaire from April 2008 to September 2009. Cases were consecutive stroke patients hospitalized to during the same period. Each control was matched with 2 cases for age and sex. The presence of hypertension, diabetes, hypercholesterolemia, smoking, atrial fibrillation, coronary heart disease, history of stroke, family history of stroke and obesity were investigated. Age-standardized prevalence, adjusted odd ratios (ORs) and PARs for 9 risk factors for stroke were calculated. One thousand forty one control subjects were matched with 2,082 cases. Adjusted ORs for stroke were 2.80 for hypertension, 3.56 for diabetes, 3.65 for smoking, 3.95 for history of stroke, and 1.25 for obesity. Risk of atrial fibrillation, coronary heart disease, and family history of stroke were not significant, while hypercholesterolemia showed inverse correlation. PARs for stroke were 16.3% for hypertension, 9.4% for diabetes, 23.6% for smoking, 5.84% for stroke history, and 4.18% for obesity. Comparisons of education level and religion between the control group and the national census data for validating its representativeness revealed significant discrepancy. In conclusion, internet-based control recruitment for a case-control study of stroke risk factors may be feasible; however, careful approaches for reducing selection and reporting biases are essential.