Data analysis
Bivariate analyses (chi-square tests) were used to examine differences between the men and women. Logistic regression analyses were then used to determine the factors significantly associated with health services utilization. The dependent variables in the logistic regression models included physician visits and the use of traditional Korean medicine. The independent variables were age, gender, education, marital status, income and the presence of chronic
conditions. To avoid over-fitting the model, only those variables found to be significant on univariate analysis were included in the final regression model. All tests were conducted at the 5% level of significance. The percentage and the odds ratio (OR) were reported with a 95% confidence interval (CI).