Among the predictive models, the best clinical model based on age, gender, physical activity, diabetes, systolic blood pressure, antihypertensive medication, and HDL cholesterol correctly classified 86.4% of patients in the general population who did and did not progress to end-stage renal disease.

The albumin-to-creatinine ratio performed better, correctly identifying 89.3%. Estimated GFR alone was better yet (93.3%).

Combining the two marginally improved prediction overall (93.6%), but substantially improved the more important true-positive rate across all potential screening populations.