Abstract
This cross-sectional study of 2,250 hospital workers attempted to identify potential physiological risk structures for night workers by utilizing the nonlinear manifold learning framework. Unlike conventional mean-centered linear analysis, this study applied the Potential of Heat-diffusion for Affinity-based Transition Embedding (PHATE) for topological visualization and Neural Additive Models (NAM) to predict interpretable risk. The ‘cholesterol homeostasis profile’ was defined as the main latent indicator. Analysis showed that the systemic metabolic load (PHATE-1) was higher after age correction, despite the night-shift group being younger than the non-night shift group (p < 0.001). This axis showed a positive correlation with body mass index (r = 0.71) and triglyceride (r = 0.56). The NAM model captured the risk slope with an average AUC of 0.864 (range 0.832–0.901) and confirmed a nonlinear risk transition when it exceeded the triglyceride 2 mmol/L threshold. The predicted risk of night workers in the highest risk cluster (cluster 1) was 0.71, and the metabolic load and cholesterol homeostasis profile risk were increased. Ultimately, the PHATE-NAM framework moves beyond traditional group averages. It provides a purely data-driven way to pinpoint individual physiological vulnerabilities, giving hospitals the exact evidence needed to design safer, more sustainable shift-work environments.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to the institution’s Electronic Medical Record (EMR) data protection policy regarding participant privacy. However, the raw data supporting the conclusions of this article are available from the corresponding author (JeongBeom Lee, leejb@sch.ac.kr) upon reasonable request and subject to Institutional Review Board (IRB) approval. To fully comply with the journal’s Code Availability policy and support research reproducibility, the official Python implementation of the analytical pipeline (including PHATE, LightGBM, NAM, and VaDE models) has been made publicly available. The code repository can be accessed at: https://github.com/Crescendo-OEM/DL-Physiological-Risk-Stratification. To ensure permanent accessibility, the version of the code used in this study is also archived at: https://doi.org/10.5281/zenodo.18872957.
Abbreviations
- BMI:
-
body mass index
- LDL/HDL:
-
Low-/high-density lipoprotein
- G-GTP:
-
Gamma-glutamyl transpeptidase
- ALT/AST:
-
Alanine/aspartate aminotransferase
- ROC:
-
Receiver operating characteristic
- AUC:
-
Area under the curve
References
Pan, A., Schernhammer, E. S., Sun, Q. & Hu, F. B. Rotating night shift work and risk of type 2 diabetes: two prospective cohort studies in women. PLoS Med. 8, e1001141. https://doi.org/10.1371/journal.pmed.1001141 (2011).
Wang, F. et al. Meta-analysis on night shift work and risk of metabolic syndrome. Obes. Rev. 15, 709–720. https://doi.org/10.1111/obr.12194 (2014).
Morris, C. J., Purvis, T. E., Hu, K. & Scheer, F. A. J. L. Circadian misalignment increases cardiovascular disease risk factors in humans. Proc. Natl. Acad. Sci. USA 113, E1402–E1411 (2016). https://doi.org/10.1073/pnas.1516953113
Leproult, R., Holmbäck, U. & Van Cauter, E. Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. Diabetes 63, 1860–1869. https://doi.org/10.2337/db13-1546 (2014).
Kecklund, G. & Axelsson, J. Health consequences of shift work and insufficient sleep. BMJ 355, i5210. https://doi.org/10.1136/bmj.i5210 (2016).
International Agency for Research on Cancer. Night Shift Work (IARC Monographs on the Identification of Carcinogenic Hazards to Humans, Vol. 124). IARC, (2020). https://publications.iarc.fr/593
International Labour Organization. Night Work Convention, 1990 (No. 171). ILO, (1990). https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:C171
Caruso, C. C. et al. NIOSH Training for Nurses on Shift Work and Long Work Hours (DHHS (NIOSH) Publication No. 2015 – 115, National Institute for Occupational Safety and Health, 2015). https://www.cdc.gov/niosh/docs/2015-115/
Vetter, C. et al. Night shift work, genetic risk, and type 2 diabetes in the UK Biobank. Diabetes Care. 41, 762–769. https://doi.org/10.2337/dc17-1933 (2018).
Puttonen, S., Härmä, M. & Hublin, C. Shift work and cardiovascular disease: pathways from circadian stress to morbidity. Scand. J. Work Environ. Health. 36, 96–108. https://doi.org/10.5271/sjweh.2894 (2010).
Sooriyaarachchi, P., Jayawardena, R., Pavey, T. & King, N. A. Shift work and the risk for metabolic syndrome among healthcare workers: a systematic review and meta-analysis. Obes. Rev. 23, e13489. https://doi.org/10.1111/obr.13489 (2022).
Biglari, H. et al. Relationship between occupational stress and cardiovascular diseases risk factors in drivers. Int. J. Occup. Med. Environ. Health. 29, 895–901. https://doi.org/10.13075/ijomeh.1896.00125 (2016).
Khaleghi, S. et al. Association between blood pressure and oral temperature rate with sleepiness changes among clinical night workers. Iran. J. Public. Health. 49, 2232–2234. https://doi.org/10.18502/ijph.v49i11.4747 (2020).
Saberinia, A. et al. Investigation of relationship between occupational stress and cardiovascular risk factors among nurses. Iran. J. Public. Health. 49, 1954–1958. https://doi.org/10.18502/ijph.v49i10.4699 (2020).
Foster, R. G. Sleep, circadian rhythms and health. Interface Focus. 10, 20190098. https://doi.org/10.1098/rsfs.2019.0098 (2020).
Shen, Y. et al. Personalized physician-assisted sleep advice for shift workers: algorithm development and validation study. JMIR Form. Res. 9, e65000. https://doi.org/10.2196/65000 (2025).
Moon, K. R. et al. Visualizing structure and transitions in high-dimensional biological data. Nat. Biotechnol. 37, 1482–1492. https://doi.org/10.1038/s41587-019-0336-3 (2019).
Agarwal, R. et al. Neural additive models: interpretable machine learning with neural nets. Adv. Neural Inf. Process. Syst. 34, 4699–4711. https://doi.org/10.48550/arXiv.2004.13912 (2021).
Jiang, S. et al. Enhancing vitiligo stage diagnosis through a reliable multimodal model with uncertainty calibration. Appl. Intell. 55, 1002. https://doi.org/10.1007/s10489-025-06839-x (2025).
Wang, Y., Kung, L., Wang, W. Y. C. & Cegielski, C. G. An integrated big data analytics-enabled transformation model: application to healthcare. Inf. Manag. 55, 64–79. https://doi.org/10.1016/j.im.2017.04.001 (2018).
Rajkomar, A., Dean, J. & Kohane, I. Machine learning in medicine. N Engl. J. Med. 380, 1347–1358. https://doi.org/10.1056/NEJMra1814259 (2019).
Sadeghi, Z. et al. A review of explainable artificial intelligence in healthcare. Comput. Electr. Eng. 118, 109370. https://doi.org/10.1016/j.compeleceng.2024.109370 (2024).
Morris, C. J. et al. Endogenous circadian system and circadian misalignment impact glucose tolerance via separate mechanisms in humans. Proc. Natl. Acad. Sci. USA 112, E2225–E2234 (2015). https://doi.org/10.1073/pnas.1418955112
Chellappa, S. L., Vujovic, N., Williams, J. S. & Scheer, F. A. J. L. Impact of circadian disruption on cardiovascular function and disease. Trends Endocrinol. Metab. 30, 767–779. https://doi.org/10.1016/j.tem.2019.07.008 (2019).
Mavaddat, N. et al. Prediction of breast cancer risk based on profiling with common genetic variants. J. Natl. Cancer Inst. 107, djv036. https://doi.org/10.1093/jnci/djv036 (2015).
Rose, G. Sick individuals and sick populations. Int. J. Epidemiol. 14, 32–38. https://doi.org/10.1093/ije/14.1.32 (1985).
Knutsson, A. Methodological issues in shift work research. Occup. Med. (Lond). 54, 193–198. https://doi.org/10.1093/occmed/kqh039 (2004).
McEwen, B. S. Protective and damaging effects of stress mediators. N Engl. J. Med. 338, 171–179. https://doi.org/10.1056/NEJM199801153380307 (1998).
Vyas, M. V. et al. Shift work and vascular events: systematic review and meta-analysis. BMJ 345, e4800 (2012). https://doi.org/10.1136/bmj.e4800
Reutrakul, S. & Knutson, K. L. Consequences of circadian disruption on cardiometabolic health. Sleep. Med. Clin. 10, 455–468. https://doi.org/10.1016/j.jsmc.2015.07.005 (2015).
Czeisler, C. A. Perspective: casting light on sleep deficiency. Nature 497, 13. https://doi.org/10.1038/497S13a (2013).
Buxton, O. M. et al. Adverse metabolic consequences in humans of prolonged sleep restriction combined with circadian disruption. Sci. Transl Med. 4, 129ra43. https://doi.org/10.1126/scitranslmed.3003200 (2012).
Wong, I. S., McLeod, C. B. & Demers, P. A. Shift work trends and risk of work injury among Canadian workers. Scand. J. Work Environ. Health. 37, 54–61. https://doi.org/10.5271/sjweh.3124 (2011).
Huang, H., Liu, Z., Xie, J. & Xu, C. Association between night shift work and NAFLD: a prospective analysis of 281,280 UK Biobank participants. BMC Public. Health. 23, 1282. https://doi.org/10.1186/s12889-023-16204-7 (2023).
De Bacquer, D. et al. Rotating shift work and the metabolic syndrome: a prospective study. Int. J. Epidemiol. 38, 848–854. https://doi.org/10.1093/ije/dyn360 (2009).
Partch, C. L., Green, C. B. & Takahashi, J. S. Molecular architecture of the mammalian circadian clock. Trends Cell. Biol. 24, 90–99. https://doi.org/10.1016/j.tcb.2013.07.002 (2014).
Koruca, H. İ., Emek, M. S. & Gulmez, E. Development of a personalized staff-scheduling method with work-life balance perspective: case of a hospital. Ann. Oper. Res. https://doi.org/10.1007/s10479-023-05244-2 (2023).
Whittaker, D. S. et al. Circadian modulation by time-restricted feeding rescues brain pathology and improves memory in mouse models of Alzheimer’s disease. Cell. Metab. 35, 1704–1721e6. https://doi.org/10.1016/j.cmet.2023.07.014 (2023).
Adamovich, Y. et al. Oxygen and carbon dioxide rhythms are circadian clock controlled and differentially directed by behavioral signals. Cell. Metab. 29, 1092–1103e3. https://doi.org/10.1016/j.cmet.2019.01.007 (2019).
Tian, H., Zhao, X., Zhang, Y. & Xia, Z. Research progress of circadian rhythm in cardiovascular disease: a bibliometric study. Heliyon 10, e28738. https://doi.org/10.1016/j.heliyon.2024.e28738 (2024).
Gachon, F. et al. Potential bidirectional communication between the liver and the central circadian clock in MASLD. NPJ Metab. Health Dis. 3, 15. https://doi.org/10.1038/s44324-025-00058-1 (2025).
Shi, S. Q. et al. Circadian disruption leads to insulin resistance and obesity. Curr. Biol. 23, 372–381. https://doi.org/10.1016/j.cub.2013.01.048 (2013).
Chellappa, S. L., Morris, C. J. & Scheer, F. A. J. L. Daily circadian misalignment impairs human cognitive performance task-dependently. Sci. Rep. 8, 3041. https://doi.org/10.1038/s41598-018-20707-4 (2018).
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 285, 2486–2497. https://doi.org/10.1001/jama.285.19.2486 (2001).
Zhang, L., Sun, D. M., Li, C. B. & Tao, M. F. Influencing factors for sleep quality among shift-working nurses. Asian Nurs. Res. 10, 277–282. https://doi.org/10.1016/j.anr.2016.09.002 (2016).
Eastman, C. I. How to reduce circadian misalignment in rotating shift workers. ChronoPhysiol Ther. 6, 41–46. https://doi.org/10.2147/CPT.S111424 (2016).
Chellappa, S. L. et al. Daytime eating during simulated night work mitigates changes in cardiovascular risk factors. Nat. Commun. 16, 57846. https://doi.org/10.1038/s41467-025-57846-y (2025).
Molzof, H. E. et al. Nightshift work and nighttime eating are associated with higher insulin and leptin levels in hospital nurses. Front. Endocrinol. 13, 876752. https://doi.org/10.3389/fendo.2022.876752 (2022).
Ezerins, M. E. et al. Advancing safety analytics: a diagnostic framework for assessing system readiness within occupational safety and health. Saf. Sci. 146, 105569. https://doi.org/10.1016/j.ssci.2021.105569 (2022).
Baron, K. G. & Reid, K. J. Circadian misalignment and health. Int. Rev. Psychiatry. 26, 139–154. https://doi.org/10.3109/09540261.2014.911149 (2014).
Salehi, A. & Khedmati, M. Identifying at-risk patients for congenital heart disease using integrated predictive models and fuzzy clustering analysis: A cross-sectional study. Heliyon 10, e39609. https://doi.org/10.1016/j.heliyon.2024.e39609 (2024).
Epstein, M., Arakelian, E., Tucker, P. & Dahlgren, A. Managing sustainable working hours within participatory scheduling for nurses and assistant nurses: a qualitative interview study. J. Nurs. Manag. 2023 (8096034). https://doi.org/10.1155/2023/8096034 (2023).
Lee, I. et al. Metabolic risk stratification of night shift workers in a large retail workplace through clustering and SHAP interpretation. Front. Public. Health. 13, 1704046. https://doi.org/10.3389/fpubh.2025.1704046 (2026).
Chemkomnerd, N. et al. A scenario-driven simulation approach to sustainable hospital resource management: aging society, pandemic preparedness and referral enhancement. BMC Health Serv. Res. 25, 982. https://doi.org/10.1186/s12913-025-13221-7 (2025).
European Agency for Safety and Health at Work (EU-OSHA). Current and emerging issues in the healthcare sector, including home and community care (Publications Office of the European Union, 2014). https://doi.org/10.2802/72773
Acknowledgements
We extend our thanks to the subjects whose participation made this study possible.
Funding
This research was supported by the Soonchunhyang University Research Fund. The funder had no role in study design, data collection, analysis, decision to publish, or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
Conceptualization: EC.J., JB.L., and IH.L.; Methodology: IH.L., JB.L., SH.H., and EC.J.; Software: IH.L.; Validation: JB.L. and EC.J.; Formal analysis: IH.L. and JB.L.; Investigation: IH.L., SH.H., JH.L., and HY.L.; Resources: SC.K., YS.M., and EC.J.; Data Curation: IH.L. and SH.H.; Visualization: IH.L.; Supervision: JB.L. and EC.J.; Writing – original draft: IH.L.; Writing – review & editing: IH.L., JB.L., and EC.J.
Corresponding authors
Ethics declarations
Competing interest
The authors declare no competing interests.
Consent for publication
Not applicable.
Ethical approval and consent to participate
All participants provided written informed consent, and the study protocol was reviewed and approved by the Soonchunhyang University Cheonan Hospital Institutional Review Board (IRB No. SCHCA-2025-10-010).
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Lee, I., Hong, S., Lee, J. et al. Deep learning-based physiological risk stratification in night-shift hospital workers. Sci Rep (2026). https://doi.org/10.1038/s41598-026-43982-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-43982-y