Development and Implementation of an AI System for Generating Clinical Urine Drug Test Sign-Outs
JAMA Network Open · 23.06.2026 · 9 (6) · e2619816
DOI: 10.1001/jamanetworkopen.2026.19816 ↗Kurz gefasst
Fachartikel in der Zeitschrift JAMA Network Open, online erschienen am 23.06.2026. Verfasst von Nathan Laha und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Laboratory Medicine & Pathology, University of Washington School of Medicine, Seattle. Hauptthemen laut MEDLINE: Artificial Intelligence, Natural Language Processing, Substance Abuse Detection, Urinalysis. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
Importance Modern natural language processing tools have potential to improve clinical workflows, but few have been successfully deployed in practice.
Objective
To describe the development, deployment, and evaluation of an artificial intelligence (AI) language tool for generating preliminary sign-outs to support a urine drug testing service. Design, Setting, and Participants In this prognostic study, large language models (LLMs) were used to extract substance use patterns from clinical urine drug test interpretations at a single medical center between January 1, 2014, and February 29, 2024. An AI model using these data was trained to predict substance use from qualitative and quantitative urine testing results. Predicted substance use patterns were used to create preliminary clinical sign-out statements, which were then integrated into an existing clinical workflow. Main Outcomes and Measures Predeployment and postdeployment user studies were performed to evaluate model performance and user experience within the workflow. Statistical differences between event rates were calculated using χ 2 tests, and between means using t tests. Differences between human and LLM labelers were calculated using the McNemar test.
Results
A total of 83 553 urine tests from 26 459 patients (12 413 male [46.9%]; mean [SD] age, 47.5 [16.7] years) were analyzed. LLM-based extraction of substance-use patterns was 99.9% accurate (13 509 of 13 520 tests), outperforming human labeling. Substance use prediction was similarly accurate, with area under the receiver operating curve greater than 0.99 for 23 of 26 substances. Workflow integration of the AI tool reduced clinical sign-out times by 28.5 seconds per case (23% efficiency gain), and by 65 seconds per case (51% efficiency gain) when integrated alongside a second, non-AI workflow improvement. Conclusions and Relevance In this prognostic study, AI-based interpretation of urine drug testing was fast and accurate, providing notable efficiency gains to the clinical service. These findings suggest that natural language processing tool integration can provide substantial clinical benefit, without compromising quality of care.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Laboratory Medicine & Pathology, University of Washington School of Medicine, Seattle.
Zitieren
APA
Laha, N., Keebaugh, M., Liao, H. C., Amankwaa, B., Adesoye, O., Pablo, A., Phipps, W. S., Hoofnagle, A. N., Baird, G. S., Mathias, P. C., & Foy, B. H. (2026). Development and Implementation of an AI System for Generating Clinical Urine Drug Test Sign-Outs. JAMA Network Open, 9(6), e2619816. https://doi.org/10.1001/jamanetworkopen.2026.19816
Vancouver
Laha N, Keebaugh M, Liao HC, Amankwaa B, Adesoye O, Pablo A, et al. Development and Implementation of an AI System for Generating Clinical Urine Drug Test Sign-Outs. JAMA Network Open. 2026;9(6):e2619816. doi:10.1001/jamanetworkopen.2026.19816
BibTeX
@article{Laha2026,
author = {Laha, Nathan and Keebaugh, Michael and Liao, Hsuan-Chieh and Amankwaa, Bright and Adesoye, Olumuyiwa and Pablo, Abed and Phipps, William S and Hoofnagle, Andrew N and Baird, Geoffrey S and Mathias, Patrick C and Foy, Brody H},
title = {Development and Implementation of an AI System for Generating Clinical Urine Drug Test Sign-Outs},
journal = {JAMA Network Open},
year = {2026},
volume = {9},
number = {6},
pages = {e2619816},
doi = {10.1001/jamanetworkopen.2026.19816}
}
EUVIMED