EUVIMEDEUVIMED entdecken
Crossref

Machine learning prediction model and influencing factors for low back pain in middle-aged and older adults based on CHARLS cohort data

Jie Cheng, Jietao Li, Kai Liu

European Spine Journal · 06.11.2026

DOI: 10.1007/s00586-026-10467-5 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Spine Journal, online erschienen am 06.11.2026. Verfasst von Jie Cheng und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Orthopedics, The Second Affiliated Hospital of Harbin Medical University, Harbin, China. Die Arbeit nennt 39 Literaturangaben. Gefördert von: The mechanism of m6A-modified hsa_circ_0007292 regulating SATB2 through competitive binding of miR-508-3p to promote ossification of the posterior longitudinal….

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

BACKGROUND: Low back pain (LBP) is a common condition that leads to functional impairment and reduced quality of life in the modern population. Given the complexity of its etiology, this study aims to systematically identify precise risk factors at the population level and establish predictive models to provide a scientific basis for the early prevention of low back pain. METHODS: Using data from the China Health and Retirement Longitudinal Study (CHARLS) spanning 2011 to 2020, a total of 93,022 middle-aged and older adults were included in the analysis. Feature selection was performed using L1 regularization (LASSO) and recursive feature elimination (RFE). Eight machine learning models including logistic regression, random forest, and LightGBM were constructed and their predictive performances were compared. Model interpretability was assessed using SHAP values. RESULTS: LASSO and RFE jointly identified 20 core risk features. Among the eight machine learning models, LightGBM demonstrated optimal predictive performance, achieving an AUC of 0.898. SHAP analysis revealed that the co-occurrence of pain in multiple sites, such as "leg pain-knee pain-back pain," constituted the strongest predictor cluster for LBP. Poor self-rated health and depressive symptoms were also identified as significant risk factors. CONCLUSION: By integrating population-level big data, this study developed a highly accurate and easily interpretable predictive model for low back pain. The findings suggest that comprehensive prevention and treatment strategies for low back pain should focus on the coordinated management of pain in the "back-leg-knee" region, while also taking psychosocial factors into account.

Schlagwörter

Stichwörter der Autor:innen

CHARLSIntervertebral disc degenerationLow back painMachine learningRisk prediction model

Autor:innen und Institutionen

  • Department of Orthopedics, The Second Affiliated Hospital of Harbin Medical University, Harbin, China.
  • State Key Laboratory of Woody Oil Resources Utilization, Northeast Forestry University, Harbin, China.
  • Department of Orthopedics, The Second Affiliated Hospital of Harbin Medical University, Harbin, China. liukai_g3@126.com.

Förderung

  • The mechanism of m6A-modified hsa_circ_0007292 regulating SATB2 through competitive binding of miR-508-3p to promote ossification of the posterior longitudinal…

Zitieren

APA

Cheng, J., Li, J., & Liu, K. (2026). Machine learning prediction model and influencing factors for low back pain in middle-aged and older adults based on CHARLS cohort data. European Spine Journal. https://doi.org/10.1007/s00586-026-10467-5

Vancouver

Cheng J, Li J, Liu K. Machine learning prediction model and influencing factors for low back pain in middle-aged and older adults based on CHARLS cohort data. European Spine Journal. 2026. doi:10.1007/s00586-026-10467-5

BibTeX
@article{Cheng2026,
  author = {Cheng, Jie and Li, Jietao and Liu, Kai},
  title = {Machine learning prediction model and influencing factors for low back pain in middle-aged and older adults based on CHARLS cohort data},
  journal = {European Spine Journal},
  year = {2026},
  doi = {10.1007/s00586-026-10467-5}
}
Kostenlos starten →