- Ordinal logistic regression was performed to calculate odds ratios (OR) and 95% confidence intervals (CI) with adjustment for gender, age, occupation, marital status, income, and general health status at the survey. Changes in general health status (i.e., objective variable) was coded as improve=1, no change=0, and deteriorate=−1
- †Coded as deteriorate=1, no change=0, and improve=−1
- ††Coded as increase=1, no change=0, and decrease=−1