Table 3.
The AUC for Various Models in Prediction of Nonadherence to PSOR
| Model | Description | AUC (95% CI) |
|---|---|---|
| 1 | Basea | 0.64 (0.61–0.67) |
| 2 | Plus DR severity | 0.69 (0.66–0.72) |
| 3 | Plus DME | 0.68 (0.64–0.71) |
| 4 | Plus glaucoma | 0.68 (0.65–0.71) |
| 5 | Plus AMD | 0.67 (0.63–0.70) |
| 6 | Plus cataract | 0.68 (0.65–0.71) |
| 7 | Plus other eye condition | 0.68 (0.65–0.71) |
| 8 | Plus all six eye conditionsb | 0.84 (0.81–0.87) |
| 9 | Plus two or more eye conditions | 0.70 (0.67–0.73) |
| 10 | Plus all six eye conditions and two or more eye conditions | 0.84 (0.81–0.87) |
| 11 | Conditional inference regression tree | 0.84 (0.81–0.86) |
a
Modified Poisson model adjusted for age, sex, BMI, SBP, triglycerides, and worse eye presenting VA.
b
The AUC for model 8 was significantly higher than models 1–7, and 9.
AUCs for all models was significantly higher than the base model.