Cost-Aware Active Feature Acquisition for Differential Diagnosis under Realistic Clinical Availability Constraints
- 1Technion
Abstract
Active Feature Acquisition (AFA) adaptively selects which diagnostic test to order next and offers a route to reduce unnecessary laboratory testing in acute care. Existing clinical AFA evaluations, however, assume every feature can be retrieved on demand and split data at the visit level, both of which inflate apparent performance. We re-evaluate cost-aware AFA under constraints designed to reflect deployment. From MIMIC-IV we constructed a cohort of 64,766 acute admissions (39,884 patients; 21 conditions; 55 features in 30 test panels) with a patient-level split, a 12-hour decision cutoff, and a per-patient availability mask from what was actually measured, and priced panels using the 2026 Medicare fee schedule under panel-level billing. We evaluated EIG-Cost, which scores each panel by Monte-Carlo Expected Information Gain penalised by its dollar cost, against eight published methods across budgets 30−60 over five patient-level resamples. At a $30 budget, EIG-Cost achieved the highest macro-F1 (0.188, 95% CI [0.185, 0.191]) at the lowest cost ($17.28), exceeding the strongest baseline in all five resamples (p < 0.001; Cohens d = 4.0), and led at every budget. Three of the eight methods collapsed to a vitals-only baseline (macro-F1{approx} 0.040), acquiring nothing even at higher budgets, a genuine failure to adapt to availability rather than a budget limitation. Despite modest absolute accuracy, EIG-Costs probabilities were well-calibrated (expected calibration error 0.048). Under realistic availability constraints, clinical AFA is substantially harder than full-availability benchmarks imply, several published methods fail outright, and cost-aware information-gain scoring is a robust choice in this harder setting.
Related
- Epidemiology of Syphilis, Diagnosis, Differential Diagnosis and Treatment: Up-to-Date Knowledge
- Rare Subungual Fibroepithelial Polyp With Keratosis: Differential Diagnosis
- The Diagnosis of Diabetic Striatopathy Requires Typical MRI Features and the Exclusion of All Potential Differential Diagnoses
- Case Report: Severe persistent transaminitis during pregnancy: a complex differential diagnosis involving suspected toxoplasmosis, possible drug-induced liver injury, and intrahepatic cholestasis
- AVMD-ST: attention-variational mode decomposition based spectrogram transformer for differential diagnosis of asthma and COPD
BibTeX
@misc{bingham2026costaware,
title = {{Cost-Aware Active Feature Acquisition for Differential Diagnosis under Realistic Clinical Availability Constraints}},
author = {J. C. Bingham and N. Arussy},
howpublished = {medRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.30.26361745},
url = {https://www.medrxiv.org/content/10.64898/2026.08.30.26361745v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Cost-Aware Active Feature Acquisition for Differential Diagnosis under Realistic Clinical Availability Constraints
AU - J. C. Bingham
AU - N. Arussy
PY - 2026
DA - 2026/08/30
PB - medRxiv
DO - 10.64898/2026.08.30.26361745
UR - https://www.medrxiv.org/content/10.64898/2026.08.30.26361745v1
AB - Active Feature Acquisition (AFA) adaptively selects which diagnostic test to order next and offers a route to reduce unnecessary laboratory testing in acute care. Existing clinical AFA evaluations, however, assume every feature can be retrieved on demand and split data at the visit level, both of which inflate apparent performance. We re-evaluate cost-aware AFA under constraints designed to reflect deployment. From MIMIC-IV we constructed a cohort of 64,766 acute admissions (39,884 patients; 21 conditions; 55 features in 30 test panels) with a patient-level split, a 12-hour decision cutoff, and a per-patient availability mask from what was actually measured, and priced panels using the 2026 Medicare fee schedule under panel-level billing. We evaluated EIG-Cost, which scores each panel by Monte-Carlo Expected Information Gain penalised by its dollar cost, against eight published methods across budgets 30−60 over five patient-level resamples. At a $30 budget, EIG-Cost achieved the highest macro-F1 (0.188, 95% CI [0.185, 0.191]) at the lowest cost ($17.28), exceeding the strongest baseline in all five resamples (p < 0.001; Cohens d = 4.0), and led at every budget. Three of the eight methods collapsed to a vitals-only baseline (macro-F1{approx} 0.040), acquiring nothing even at higher budgets, a genuine failure to adapt to availability rather than a budget limitation. Despite modest absolute accuracy, EIG-Costs probabilities were well-calibrated (expected calibration error 0.048). Under realistic availability constraints, clinical AFA is substantially harder than full-availability benchmarks imply, several published methods fail outright, and cost-aware information-gain scoring is a robust choice in this harder setting.
ER -