Test properties of 5 models for predicting complications in diabetes (validation dataset)*

Type of complication
Micro- and macrovascularInfectious diseaseMetabolic
*Sens = sensitivity; Spec = specificity. Diagnostic terms defined in glossary. Data on specificity, +LR, and –LR provided by author.
†The “best” models for predicting complications included predictors from the following categories: patient demographics, previous diagnoses of complications, metabolic measurements, medication, and healthcare utilisation measures.
‡Cut point of patients with the highest 30% of predicted risk scores.
Best models†‡72%73%2.680.3872%71%2.490.3983%70%2.790.24
Previous events72%72%2.570.3944%86%3.100.6533%99%24.60.68
3 variables‡71%73%2.630.4068%71%2.350.4575%70% 2.520.35
Risk score†74%70%2.470.3767%71%2.320.4682%71% 2.860.25
HbA1c concentration‡31%70%1.040.9838%70%1.280.8859%70% 1.970.59