Table 2

Performance measures for the prediction models

Definition 1: HbA1c ≥48 mmol/mol (6.5%)Definition 2: glucose-lowering treatment initiation
AUCt (%)Brier score (%)IPAAUCt (%)Brier score (%)IPA
Main model72.7 (71.2–74.3)10.7 (8.8–12.6)18.279.4 (77.7–81.0)7.5 (5.9–9.1)17.1
Minimum model68.2 (66.7–69.7)12.8 (10.7–14.8)2.879.8 (78.1–81.4)8.6 (6.8–10.5)4.6
  • The models were validated (using the validation sample) for both definitions of type 2 diabetes. High AUCt values indicate good discrimination. Low Brier scores indicate good calibration. High IPA indicates good average performance.

  • The main model for HbA1c ≥48 mmol/mol (6.5%) included baseline HbA1c, age, sex, body mass index (BMI), treated hypertension, pre-existing pancreatic disease, absence of cancer, unhealthy diet, doctor’s advice to lose weight or change dietary habits, self-reported lack of anyone to talk to, and good self-rated health. The main model for glucose-lowering treatment initiation included baseline HbA1c, age, sex, BMI, and doctor’s advice to lose weight or change dietary habits.

  • AUCt, time-dependent area under the curve; IPA, index of prediction accuracy.