Metabolic And Cardiovascular Obesity Scores

Integrated Discrimination Improvement (IDI) Calculator

Clinical calculation support · Verify results before clinical use

Compare predicted risk between two models

Enter the mean predicted event probability for the event and non-event groups in each model.

MODEL COMPARISON

Patients who experienced the outcome

Use the average predicted probability among all patients with the observed event.

Baseline model mean risk (%) Mean predicted probability in event group

New model mean risk (%) Mean predicted probability in the same event group

Patients who did not experience the outcome

Use the average predicted probability among all patients without the observed event.

Baseline model mean risk (%) Mean predicted probability in non-event group

New model mean risk (%) Mean predicted probability in the same non-event group

How to interpret IDI

The integrated discrimination improvement (IDI) compares the discrimination slopes of a baseline risk model and a new model. The calculation uses average predicted probabilities in patients with and without the outcome, not the prediction for a single patient.

IDI = (Mean risk_new,event − Mean risk_{baseline,event})
− (Mean risk_{new,non-event} − Mean risk_{baseline,non-event})

Practical interpretation

IDI > 0: the new model increases the separation in mean predicted risk between event and non-event groups.
IDI = 0: no change in discrimination slope is seen from the entered means.
IDI < 0: the new model decreases the separation in mean predicted risk between groups.

IDI is a discrimination measure, not proof of clinical utility or better calibration. Interpret it alongside calibration, discrimination measures such as the C-statistic/AUC, uncertainty intervals, and the clinical context.

References

Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Statistics in Medicine. 2008;27(2):157–172. doi:10.1002/sim.2929

Pencina MJ, Fine JP, D’Agostino RB Sr. Discrimination slope and integrated discrimination improvement—properties, relationships and impact of calibration. Statistics in Medicine. 2017;36(28):4482–4490. doi:10.1002/sim.7139 · PubMed

Clinical and statistical note: This calculator computes IDI from user-supplied mean predicted probabilities and is intended for model evaluation, not individual-patient risk estimation. Use comparable outcome definitions, prediction horizons, and validation populations for both models. Statistical uncertainty and significance are not estimated by this tool.