Table 4.
Associated factors for poor glycaemic control using multiple logistic regression analysis.
| Variable | Regression coefficient | Adjusted odds ratioa (95% CI) | Wald statistic | p-valuea |
|---|---|---|---|---|
| Social support score | 0.06 | 1.06 (1.03,1.10) | 11.80 | 0.001 |
| Employment status | ||||
| Employed | 1 | |||
| Unemployed | −0.78 | 0.46 (0.22,0.95) | 4.46 | 0.035 |
| Pensioner | −1.26 | 0.28 (0.13,0.61) | 10.48 | 0.001 |
| DM interfere ADL | ||||
| No | 1 | |||
| Yes | 1.16 | 3.18 (1.17,8.70) | 5.11 | 0.024 |
Note: ADL = activities of daily living, CI = confidence interval, DM = diabetes mellitus.
a
Forward LR multiple logistic regression model. No multicollinearity problem or significant interaction terms. Overall, the Hosmer–Lemeshow test (p = 0.321), correctly classified percentage of 76.3%, and area under the receiver operating characteristic curve of 69.3% indicates that the model was fit.