Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP.
R, L., Q, W., X, Z., & F, Z. (2026). Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP.. International journal of chronic obstructive pulmonary disease. https://doi.org/10.2147/COPD.S623174
R L, Q W, X Z, F Z. Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP.. International journal of chronic obstructive pulmonary disease. 2026; doi: 10.2147/COPD.S623174
R L, Q W, X Z, et al. Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP.[J]. International journal of chronic obstructive pulmonary disease. 2026. DOI: 10.2147/COPD.S623174.
@article{r2026,
author = {Li R and Wang Q and Zhang X and Zhou F},
title = {Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP.},
journal = {International journal of chronic obstructive pulmonary disease},
year = {2026},
doi = {10.2147/COPD.S623174},
note = {PMID: 42453371},
}
TY - JOUR AU - Li R AU - Wang Q AU - Zhang X AU - Zhou F TI - Multi‑feature Prediction Model for Coronary Heart Disease Comorbidity in Middle‑aged and Older Adults with COPD Based on Machine Learning and SHAP. T2 - International journal of chronic obstructive pulmonary disease PY - 2026 DO - 10.2147/COPD.S623174 AN - PMID:42453371 ER -
BACKGROUND: Chronic obstructive pulmonary disease (COPD) frequently coexists with coronary heart disease (CHD), markedly worsening prognosis in middle-aged and older patients. Early identification of CHD comorbidity in this population remains clinically imperative. METHODS: This single-center, cross-sectional study included COPD patients aged 45 years or older admitted between 2020 and 2025. Missing data were imputed using random forest, and least absolute shrinkage and selection operator regression was applied for feature selection. Nine machine learning models were constructed and evaluated by the area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis. SHapley Additive exPlanations and restricted cubic splines (RCS) were employed for model interpretation and dose-response exploration. RESULTS: Of 17,862 eligible patients, 7,211 (40.37%) had coexisting CHD. Sixteen predictors were identified. The XGBoost model demonstrated moderate predictive performance (training AUC 0.871, 95% CI: 0.864-0.877; validation AUC 0.743, 95% CI: 0.730-0.756), significantly outperforming all other models in the training set and showing comparable performance to GBDT in the validation set. Age, hypertension, total cholesterol (TC), chronic gastritis, and uric acid (UA) were the top five predictors. RCS identified various dose-response patterns, including nonlinear associations for pulse rate, diastolic blood pressure, TC, and platelet count, and linear positive associations for prothrombin time and UA. CONCLUSION: The XGBoost model showed moderate discriminative ability for predicting CHD comorbidity in middle-aged and older COPD patients. However, further external validation is required before clinical application, and the findings should be interpreted with caution given the single-center, cross-sectional design.