A newly published study protocol outlines the design and evaluation of Hepatitis C Emergency Department, or HepC-EnD, a machine learning-based clinical decision support tool intended to establish targeted and sustainable hepatitis C virus screening within emergency department workflows. Described by researchers in Implementation Science Communications, the project seeks to evaluate how algorithmic risk stratification can integrate directly into routine clinical practice to address long-standing gaps in viral detection.
Hepatitis C virus infection currently affects more than 2.4 million individuals across the United States, with approximately one-third of those infected remaining unaware of their condition. Even though highly curative direct-acting antiviral treatments are available, low rates of screening and treatment uptake continue to present major obstacles to national viral elimination efforts.
These persistent detection barriers prompted the US Centers for Disease Control and Prevention in 2020 to expand its screening recommendations to advise universal testing for all adults. Hospital emergency departments offer a key setting for screening initiatives because they frequently care for patient populations facing elevated risks of infection. Nevertheless, existing nontargeted emergency department testing strategies require substantial operational resources and may lack long-term sustainability.
To establish a more sustainable approach, the HepC-EnD initiative is structured around an implementation study guided by the Exploration, Preparation, Implementation, Sustainment framework alongside clinical decision support Five Rights principles. Under the project’s first aim, researchers are developing and validating machine learning risk prediction algorithms using electronic health record data obtained from OneFL+. These predictive models incorporate sociodemographic characteristics, clinical variables, and social determinants of health.
The second phase of the project focuses on translating these algorithms into an electronic health record-integrated clinical decision support tool that alerts emergency department healthcare professionals when a patient’s calculated risk score surpasses a prespecified threshold. To optimize usability and clinical integration, investigators will conduct stakeholder meetings and semi-structured interviews with emergency clinicians to guide iterative refinements of the alert system.
Under the third aim, investigators will execute a 24-month quasi-experimental pre-post study across two emergency departments to evaluate targeted HepC-EnD screening in comparison with standard nontargeted screening programs. Formatted as a type 2 hybrid implementation-effectiveness study, this phase will assess the tool’s usability, acceptability, feasibility, and adoption across participating healthcare teams.
The primary clinical outcome evaluated throughout the trial is the rate of new hepatitis C virus or HIV diagnoses. Secondary evaluation metrics include clinical alert rates, hepatitis C and HIV testing rates, linkage to care rates, and cost-effectiveness. Registered under ClinicalTrials.gov identifier NCT07406490, the study protocol notes that if successful, HepC-EnD could offer a scalable framework for integrating machine learning-driven targeted screening into broader clinical practice while reducing the resource burden of universal nontargeted testing.
The structured implementation plan explicitly addresses operational challenges that previously limited routine viral screening programs within emergency department settings. Guided by the Exploration, Preparation, Implementation, Sustainment framework across each project phase, the study protocol outlines steps to evaluate implementation processes alongside diagnostic performance. In the initial development phase, risk prediction algorithms are developed and validated using electronic health record data from OneFL+, incorporating sociodemographic predictors, clinical factors, and social determinants of health. This predictive modeling strategy is designed to identify patients at increased risk of infection prior to alert generation, separating retrospective algorithm development from live clinical deployment.
To support practical adoption and usability, the protocol incorporates the clinical decision support Five Rights framework during tool development. Rather than issuing universal prompts, the resulting tool notifies emergency department healthcare professionals when an individual’s predicted risk score surpasses a prespecified threshold. The research team plans to refine this alert threshold and interface iteratively through semi-structured clinician interviews and stakeholder meetings. These collaborative sessions focus directly on optimizing usability and workflow integration, ensuring the clinical decision support system functions effectively within busy emergency workflows.
The prospective evaluation compares this targeted strategy against nontargeted hepatitis C virus screening across two emergency departments over a 24-month period. Because nontargeted emergency department screening programs are resource-intensive and may lack long-term sustainability, this type 2 hybrid implementation-effectiveness study evaluates whether targeted screening improves screening efficiency and diagnostic yield while reducing operational burden. Investigators will assess tool usability, acceptability, feasibility, and adoption across participating healthcare settings. Primary evaluation centers on new hepatitis C virus or HIV diagnoses, with secondary assessments tracking alert rates, testing rates, linkage to care rates, and cost-effectiveness. If successful, HepC-EnD may provide a scalable and sustainable framework for integrating machine learning-based targeted screening into routine clinical practice. Because this publication establishes a study protocol rather than trial results, definitive conclusions regarding screening performance await full trial completion.
Subject of Research: Implementation protocol for a machine learning clinical decision support tool for emergency department hepatitis C screening
Article Title: Assessing performance of the Hepatitis C Emergency Department (HepC-EnD) screening tool: implementation protocol
Article References: Hong, J.-W. J., Stultz, A. A., Salloum, R. G., Nguyen, K. A., Gage, C. N., Lo-Ciganic, W.-H., Petrauskis, M., Wu, Y., Cook, R. L., Nelson, D. R., & Park, H. (2026). Assessing performance of the Hepatitis C Emergency Department (HepC-EnD) screening tool: implementation protocol. Implementation Science Communications. https://doi.org/10.1186/s43058-026-01159-8
Image Credits: AI Generated
DOI: 10.1186/s43058-026-01159-8
Keywords: hepatitis C virus, emergency department screening, HepC-EnD, clinical decision support, machine learning, OneFL+, electronic health records, implementation science, EPIS framework, Five Rights, social determinants of health, HIV diagnosis, linkage to care, cost-effectiveness, direct-acting antiviral therapies, targeted screening, NCT07406490
Cite Scienmag News
Kristina Jarvis. (November 17, 2026). Machine Learning Protocol Aims to Target Emergency Department Hepatitis C Screening. Scienmag. https://scienmag.com/machine-learning-protocol-aims-to-target-emergency-department-hepatitis-c-screening/
Kristina Jarvis. "Machine Learning Protocol Aims to Target Emergency Department Hepatitis C Screening." Scienmag, 17 November 2026, https://scienmag.com/machine-learning-protocol-aims-to-target-emergency-department-hepatitis-c-screening/. Accessed 17 November 2026.
Kristina Jarvis. "Machine Learning Protocol Aims to Target Emergency Department Hepatitis C Screening." Scienmag. November 17, 2026. https://scienmag.com/machine-learning-protocol-aims-to-target-emergency-department-hepatitis-c-screening/








