# A Machine-Learned Comorbidity Index

Source: [A Machine-Learned Comorbidity Index](<https://www.sota2.com/research/paper/6a322a39d33dfb23e783ec5b>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-18T05:00:06.089Z

Authors: Suleman Baloch, Kishlay Jha, Alberto M. Segre, Philip M. Polgreen, Bijaya Adhikari

Published: 2026-06-16T03:07:20.000Z

## Abstract

Traditional comorbidity scores (e.g., Charlson and Elixhauser) are widely used for risk adjustment and patient stratification, but they have two key limitations: (i) they are largely mortality-centric and do not align well with other clinical outcomes, and (ii) their linear, rule-based structure cannot capture nonlinear, outcome-specific risk relationships. We propose a Machine-Learned Comorbidity Index (MLCI) that maps diagnosis codes to a single scalar by maximizing the normalized Hilbert-Schmidt Independence Criterion (nHSIC) between the learned score and multiple clinical outcomes. MLCI captures nonlinear risk-outcome dependence and is supported by a theory that characterizes when a unified, informative admission-level ordering can be achieved across outcomes. Empirical results on multiple benchmark electronic health record (EHR) datasets show that MLCI outperforms strong baselines across multiple evaluation metrics.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [30-day Mortality Prediction (30M) on MIMIC-IV](<https://www.sota2.com/research/sota/30-day-mortality-prediction-30m-on-mimic-iv.md>): 30-day mortality prediction (30M); MIMIC IV; Mutual Information (MI): 0.07219.

- [30-day mortality prediction (30M) on MIMIC-III](<https://www.sota2.com/research/sota/30-day-mortality-prediction-30m-on-mimic-iii.md>): 30-day mortality prediction (30M); MIMIC-III; Mutual Information (MI) (x10^3): 77.77.

- [Clinical Outcome Prediction on MIMIC-IV](<https://www.sota2.com/research/sota/clinical-outcome-prediction-on-mimic-iv.md>): Clinical Outcome Prediction; MIMIC IV; Mortality (dCorr): 54.8.

- [Clinical Outcome Prediction on MIMIC-III](<https://www.sota2.com/research/sota/clinical-outcome-prediction-on-mimic-iii.md>): Clinical Outcome Prediction; MIMIC-III; Mortality (dCorr): 39.06.

- [In-hospital mortality prediction on MIMIC-IV (using MI)](<https://www.sota2.com/research/sota/in-hospital-mortality-prediction-on-mimic-iv-using-mi.md>): In-hospital mortality prediction (MORT); MIMIC IV; MI (x10^3): 74.22.

- [In-hospital mortality prediction (MORT) on MIMIC-III](<https://www.sota2.com/research/sota/in-hospital-mortality-prediction-mort-on-mimic-iii.md>): In-hospital mortality prediction (MORT); MIMIC-III; Mutual Information (MI): 0.08452.

- [Length of Stay Prediction (LOS) on MIMIC-IV (Mutual Information)](<https://www.sota2.com/research/sota/length-of-stay-prediction-los-on-mimic-iv-mutual-information.md>): Length of Stay Prediction (LOS); MIMIC IV; Mutual Information (MI): 0.1542.

- [Length of stay prediction (LOS) on MIMIC-III](<https://www.sota2.com/research/sota/length-of-stay-prediction-los-on-mimic-iii.md>): Length of Stay Prediction (LOS); MIMIC-III; Mutual Information (MI): 0.1645.

- [ICU Transfer Prediction on MIMIC-IV](<https://www.sota2.com/research/sota/icu-transfer-prediction-on-mimic-iv.md>): ICU transfer prediction (ICU); MIMIC IV; MI (x10^2): 19.01.

- [ICU Transfer Prediction on MIMIC-III](<https://www.sota2.com/research/sota/icu-transfer-prediction-on-mimic-iii.md>): ICU transfer prediction (ICU); MIMIC-III; MI (x10^3): 26.42.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2606.17450>)

[PDF](<https://arxiv.org/pdf/2606.17450.pdf>)