Organizational and Nursing Management Factors Associated With Guideline‐Discordant Medical Radiation Protective Apparel Disinfection: A Multicenter Interpretable Machine Learning Study
- 1Department of Nursing Shanghai Guangming Hospital of Traditional Chinese Medicine Affiliated to Shanghai University of Traditional Chinese Medicine Shanghai China
- 2Clinical Research Center Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine Shanghai China shutcm.edu.cn
Abstract
Background: Effective management of Medical Radiation Protective Apparel (MRPA) is a global implementation challenge. In China, this is complicated by discordant guidance regarding MRPA disinfection, with some domestic specialty recommendations differing from manufacturer‐aligned material‐care guidance.
Aim: To identify organizational and nursing management factors associated with guideline‐discordant MRPA disinfection using an interpretable machine learning approach and to characterize their joint predictive patterns.
Participants and settings: A total of 123 public hospitals in two developed regions of eastern China (Shanghai and Zhejiang Province, China), including Tertiary A, Tertiary B, and Secondary A institutions.
Methods: In this cross‐sectional study, a self‐designed structured survey assessed MRPA management practices, with Critical Noncompliance (CNP)—a study‐defined endpoint referring to reported use of alcohol‐ or chlorine‐based disinfectants—as the primary outcome. This classification reflected discordance with manufacturer‐aligned material‐care recommendations rather than a formal regulatory designation. Data were analyzed using descriptive statistics and interpretable machine‐learning models, with random forest selected as the final model and SHAP used for model interpretation.
Results: Guideline‐discordant practices were common: alcohol‐ and chlorine‐based disinfection was reported by 63.4% and 65.9% of institutions, and 67.5% used manual or partly manual traceability. Nursing personnel held 81.7% of management roles. In the independent test set, the random forest model achieved an AUROC of 0.849 (95% CI, 0.558–0.922). Random forest and SHAP identified Tertiary Grade A status, critical‐area traceability, designated personnel, disinfection‐personnel training, housekeeping/environmental‐services execution, and third‐party laundering as prominent contributors to CNP prediction.
Conclusions: Guideline‐discordant MRPA disinfection was associated with multiple organizational and workflow characteristics rather than a single professional role or management feature. The random forest–SHAP patterns are hypothesis‐generating and identify organizational domains for prospective validation, particularly those involving guidance, training, traceability, and workflow design.
Implication for Nursing Management: Nursing roles should be considered within the broader organizational systems in which responsibility, training, traceability, and workflow are structured. Future implementation studies should evaluate whether coordinated alignment of these components with evidence‐based MRPA care improves guideline adherence, particularly in higher‐tier hospitals.
Related
- Knowledge attitudes and practices regarding toothbrush hygiene and disinfection among dental students in a multicenter cross sectional study
- Assessment of cleaning and disinfection of Intensive Care Unit beds: quality management with Lean.
- Automated on-site disinfection of urine collection bags: a sustainable waste management
- Multicenter evaluation of AI-based intraocular lens formulas and interpretable machine learning for cataract refractive prediction
- Evolution of Management of Postoperative Pain in Spain: A Nationwide Survey of Organizational Structure and Clinical Practice
BibTeX
@article{xin2026organizational,
title = {{Organizational and Nursing Management Factors Associated With Guideline‐Discordant Medical Radiation Protective Apparel Disinfection: A Multicenter Interpretable Machine Learning Study}},
author = {Xing Xin and Suwen Li and Chenxin Ye and Luyin Cheng and Yuting Wu and Gendi Lu},
journal = {Journal of Nursing Management},
year = {2026},
month = nov,
doi = {10.1155/jonm/2036458},
url = {https://doi.org/10.1155/jonm/2036458},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Organizational and Nursing Management Factors Associated With Guideline‐Discordant Medical Radiation Protective Apparel Disinfection: A Multicenter Interpretable Machine Learning Study
AU - Xing Xin
AU - Suwen Li
AU - Chenxin Ye
AU - Luyin Cheng
AU - Yuting Wu
AU - Gendi Lu
PY - 2026
DA - 2026/11/08
JO - Journal of Nursing Management
DO - 10.1155/jonm/2036458
UR - https://doi.org/10.1155/jonm/2036458
AB - Background: Effective management of Medical Radiation Protective Apparel (MRPA) is a global implementation challenge. In China, this is complicated by discordant guidance regarding MRPA disinfection, with some domestic specialty recommendations differing from manufacturer‐aligned material‐care guidance. Aim: To identify organizational and nursing management factors associated with guideline‐discordant MRPA disinfection using an interpretable machine learning approach and to characterize their joint predictive patterns. Participants and settings: A total of 123 public hospitals in two developed regions of eastern China (Shanghai and Zhejiang Province, China), including Tertiary A, Tertiary B, and Secondary A institutions. Methods: In this cross‐sectional study, a self‐designed structured survey assessed MRPA management practices, with Critical Noncompliance (CNP)—a study‐defined endpoint referring to reported use of alcohol‐ or chlorine‐based disinfectants—as the primary outcome. This classification reflected discordance with manufacturer‐aligned material‐care recommendations rather than a formal regulatory designation. Data were analyzed using descriptive statistics and interpretable machine‐learning models, with random forest selected as the final model and SHAP used for model interpretation. Results: Guideline‐discordant practices were common: alcohol‐ and chlorine‐based disinfection was reported by 63.4% and 65.9% of institutions, and 67.5% used manual or partly manual traceability. Nursing personnel held 81.7% of management roles. In the independent test set, the random forest model achieved an AUROC of 0.849 (95% CI, 0.558–0.922). Random forest and SHAP identified Tertiary Grade A status, critical‐area traceability, designated personnel, disinfection‐personnel training, housekeeping/environmental‐services execution, and third‐party laundering as prominent contributors to CNP prediction. Conclusions: Guideline‐discordant MRPA disinfection was associated with multiple organizational and workflow characteristics rather than a single professional role or management feature. The random forest–SHAP patterns are hypothesis‐generating and identify organizational domains for prospective validation, particularly those involving guidance, training, traceability, and workflow design. Implication for Nursing Management: Nursing roles should be considered within the broader organizational systems in which responsibility, training, traceability, and workflow are structured. Future implementation studies should evaluate whether coordinated alignment of these components with evidence‐based MRPA care improves guideline adherence, particularly in higher‐tier hospitals.
SN - 0966-0429
SN - 1365-2834
ER -