Cost-Effectiveness and Resource Optimization Analysis of Machine Learning-Guided Screening Programs for Chronic Liver Disease in Primary Care Settings

• Ada John

  ladoke Akintola university of technology

  Author

Chronic liver disease (CLD) represents a growing global health burden, with most cases remaining undiagnosed until irreversible liver damage has occurred. Traditional screening approaches utilizing the Fibrosis-4 (FIB-4) index demonstrate limited sensitivity and poor completion rates in primary care settings. This study presents a comprehensive cost-effectiveness analysis of machine learning-guided screening programs for CLD detection in primary care environments. Using a state-transition decision-analytic model simulating cohorts of patients with metabolic dysfunction-associated steatotic liver disease (MASLD) and alcohol-related liver disease (ARLD), we evaluated four screening strategies: ML-based risk stratification with transient elastography (TE) confirmation, FIB-4 with TE, TE-only, and no systematic screening. The ML-based approach achieved superior predictive performance with 89.4% accuracy and an area under the receiver operating characteristic curve (AUC-ROC) of 0.818, outperforming FIB-4 across all metrics. At a willingness-to-pay threshold of $100,000 per quality-adjusted life-year (QALY), the ML-guided strategy demonstrated the highest probability of cost-effectiveness (>80%), yielding an incremental cost-effectiveness ratio (ICER) of $38,916/QALY compared to $72,502/QALY for FIB-4-based screening. The ML model identified 17.1% additional true positives while reducing false-negative diagnoses by 50%. These findings establish that ML-guided screening programs represent both clinically superior and economically viable alternatives to current practice, supporting widespread integration into primary care population health strategies.

• pdf

Copyright (c) 2026 Ada John (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.