Skip to main content

Table 5 Area under the receiver operating characteristics (ROC) curve of the case-mix based logistic regression models to identify patients located above the 99th percentile of health spending

From: Predictive risk modelling in the Spanish population: a cross-sectional study

  Demographic variables only ACG-PM CRG DCG-HCC
Out of the box AUC 95% CI AUC 95% CI AUC 95% CI AUC 95% CI
Variables:         
A&S + dx + rx + cost percentiles (age < 65)    0,852 (0.847 - 0.857)     
A&S + dx + rx + cost percentiles (age 65+)    0,779 (0.774 - 0.783)     
A&S + dx      0,814 (0.811 - 0.818) 0,858 (0.8556 - 0.861)
A&S + dx + rx + cost        0,882 (0.880 - 0.885)
Recalibration         
Variables:         
A&S (Age and sex) 0,811 (0.809 - 0.814)       
A&S + Only Dx    0,882 (0.880 - 0.884) 0,829 (0.826 - 0.832) 0,884 (0.882 - 0.886)
A&S + Only Rx    0,862 (0.859 - 0.864)    0,865 (0.863 - 0.868)
A&S + Dx + Rx    0,889 (0.887 - 0.891)    0,892 (0.890 - 0.894)
A&S + Dx + Rx + cost percentile    0,897 (0.895 - 0.899) 0,869* (0.866 - 0.871) 0,899 (0.897 - 0.901)
A&S + Dx + Rx + cost percentile + DI    0,897 (0.895 - 0.899) 0,869* (0.866 - 0.872) 0,899 (0.897 - 0.902)
  1. *The CRG models do not include prescriptions.
  2. Age groups: 14–17; 18–34; 35–44; 45–54; 55–64; 65–69; 70–74; 75–79; 80–84; 85 + .
  3. Cost groups in percentiles: 0–1; 2–10; 11–25; 26–50; 51–75; 76–90; 91–93; 94–95; 96–97; 98 + .
  4. Dx diagnoses, Rx prescriptions, DI Deprivation index.