EUVIMEDEUVIMED entdecken
Multizentrische StudieValidierungsstudieQuerschnittstudieOpen Access · CC BY-NC 3.0

Development and External Validation of an Explainable Machine Learning Model to Identify Positive Dysphagia Screening Results in People with Dementia: A Multicenter Cross-Sectional Study

Jie Du iD, Huipin Zhang, Xiaomeng Wen, Xianwei Guo, Juanjuan Hu, Zhunzhun Liu iD, Xiaobao Li iD, Claire Shuiqing Zhang iD, Yun Ye iD

Clinical Interventions in Aging · 2026

DOI: 10.2147/cia.s625367 ↗

Kurz gefasst

Multizentrische Studie in der Zeitschrift Clinical Interventions in Aging, 2026. Verfasst von Jie Du und 8 weiteren Autor:innen. Beteiligte Einrichtung: Geriatric Psychiatry, Affiliated Mental Health Center of Jiangnan University, Wuxi, Jiangsu, People's Republic of China. Hauptthemen laut MEDLINE: Deglutition Disorders, Dementia, Machine Learning, Mass Screening. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 3.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

To develop and externally validate an interpretable machine learning model for identifying the likelihood of a positive Standardized Swallowing Assessment (SSA) result in people with dementia, particularly in settings where instrumental swallowing assessments are not readily available.

Methods

Candidate variables included demographic, physical, oral-health, and mental health variables routinely available in clinical practice. Least absolute shrinkage and selection operator (LASSO) regression was used for variable selection. Four machine learning models were developed and compared: logistic regression, support vector machines, random forest, and AdaBoost. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, calibration, and decision curve analysis. SHAP analysis was applied to enhance model interpretability. A web-based likelihood assessment tool was developed to estimate the individual probability of a concurrent positive SSA result.

Results

All four models showed good discrimination. AdaBoost achieved an AUC of 0.918 in internal cross-validation and AUCs of 0.900 and 0.889 in the domain and temporal external validation cohorts, respectively, with sensitivities ranging from 0.879 to 0.901. Since sensitivity was prioritized for case finding, AdaBoost was selected for model interpretation and online-tool development. SHAP analysis ranked body mass index as the leading contributor to the AdaBoost output, followed by number of teeth, eating ability, dietary type, and Clinical Dementia Rating score.

Conclusion

The model showed good performance for identifying people likely to have a concurrent positive SSA screening result. The web-based tool may support case finding and referral for further swallowing assessment, but it should not be used as a standalone diagnostic or prognostic instrument. Prospective and geographically diverse validation is required before routine clinical implementation. Why was this study done? Swallowing problems are common in people with dementia. Specialist tests are often unavailable, especially in nursing homes. We aimed to develop a simple tool to flag those who may have a positive Standardized Swallowing Assessment screen result. What did the researchers do and find? We analysed data from 620 people with dementia from a hospital and a nursing home. We compared four machine learning models. The selected model used five simple characteristics: body mass index, number of teeth, eating ability, diet type, and dementia severity. It performed consistently in the development and validation cohorts. We created a free online tool that estimates the likelihood of a positive bedside result. What do the results mean? The tool helps nurses and clinicians identify who may need further swallowing assessment. It does not diagnose dysphagia. It cannot detect silent aspiration (when food or liquid enters the airway without a cough). It should never replace professional clinical judgment.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deglutition DisordersDementiaMachine LearningMass ScreeningHumansCross-Sectional StudiesFemaleMaleAgedLogistic ModelsAged, 80 and overRandom ForestBoosting Machine Learning AlgorithmsROC CurvePredictive Learning ModelsSupport Vector MachineArea Under CurveSensitivity and Specificity

Stichwörter der Autor:innen

SHAP analysisdementiadysphagiamachine learningscreening model

Autor:innen und Institutionen

  • Geriatric Psychiatry, Affiliated Mental Health Center of Jiangnan University, Wuxi, Jiangsu, People's Republic of China.
  • Department of Nursing, The First People's Hospital of Changzhou, The Third Affiliated Hospital of Soochow University, Changzhou, Jiangsu, People's Republic of China.
  • Center for Evidence-Based Medicine Research, The First People's Hospital of Changzhou, The Third Affiliated Hospital of Soochow University, Changzhou, Jiangsu, People's Republic of China.
  • School of Nursing, Wuxi Taihu University, Wuxi, Jiangsu, People's Republic of China.
  • China-Australia International Research Centre for Chinese Medicine, RMIT University, Bundoora, VIC 3083, Australia.

ORCID-Profile: Jie Du, Zhunzhun Liu, Xiaobao Li, Claire Shuiqing Zhang, Yun Ye

Zitieren

APA

Du, J., Zhang, H., Wen, X., Guo, X., Hu, J., Liu, Z., Li, X., Zhang, C. S., & Ye, Y. (2026). Development and External Validation of an Explainable Machine Learning Model to Identify Positive Dysphagia Screening Results in People with Dementia: A Multicenter Cross-Sectional Study. Clinical Interventions in Aging. https://doi.org/10.2147/cia.s625367

Vancouver

Du J, Zhang H, Wen X, Guo X, Hu J, Liu Z, et al. Development and External Validation of an Explainable Machine Learning Model to Identify Positive Dysphagia Screening Results in People with Dementia: A Multicenter Cross-Sectional Study. Clinical Interventions in Aging. 2026. doi:10.2147/cia.s625367

BibTeX
@article{Du2026,
  author = {Du, Jie and Zhang, Huipin and Wen, Xiaomeng and Guo, Xianwei and Hu, Juanjuan and Liu, Zhunzhun and Li, Xiaobao and Zhang, Claire Shuiqing and Ye, Yun},
  title = {Development and External Validation of an Explainable Machine Learning Model to Identify Positive Dysphagia Screening Results in People with Dementia: A Multicenter Cross-Sectional Study},
  journal = {Clinical Interventions in Aging},
  year = {2026},
  doi = {10.2147/cia.s625367}
}
Kostenlos starten →