Conference proceeding
Counterfactual Optimization of Combinatorial Treatments Using Inverse Classification
Proceedings (IEEE International Conference on Healthcare Informatics. Online), pp.1394-1396
06/01/2026
DOI: 10.1109/ICHI69079.2026.00217
Abstract
Optimizing multi-drug regimens from observational data is challenging due to combinatorial explosion: k binary treatments yield 2 k possible combinations. We present an inverse classification framework using neural networks that avoids explicit interaction modeling by learning a predictive function across the treatment space. Applied to 124,031 post-AMI Medicare patients optimizing 7 medication classes (128 combinations) to minimize 2-year mortality, the neural network achieved AUC 0.793. Individualized recommendations differed from observed treatments in 98% of cases, with 1.8% of patients showing predicted mortality reductions exceeding 10 percentage points. Predicted benefits were largest among vulnerable subgroups including older patients, females, minorities, and those with higher comorbidity burden.
Details
- Title: Subtitle
- Counterfactual Optimization of Combinatorial Treatments Using Inverse Classification
- Creators
- Sulyun LeeW. Nick Street - University of IowaAaron Sporrer - University of IowaBarry L. Carter - University of IowaLinnea A. Polgreen - University of Iowa
- Resource Type
- Conference proceeding
- Publication Details
- Proceedings (IEEE International Conference on Healthcare Informatics. Online), pp.1394-1396
- DOI
- 10.1109/ICHI69079.2026.00217
- eISSN
- 2575-2634
- Publisher
- IEEE
- Language
- English
- Date published
- 06/01/2026
- Academic Unit
- Bus Admin College; Nursing; Pharmacy Practice and Science; Computer Science; Business Analytics
- Record Identifier
- 9985219921602771
Metrics
1 Record Views