Machine learning-based predictive model for postoperative delirium of elderly patients with coronary heart disease undergoing non-cardiac surgery: a retrospective cohort study - Scorecard - MDSpire

Machine learning-based predictive model for postoperative delirium of elderly patients with coronary heart disease undergoing non-cardiac surgery: a retrospective cohort study

  • By

  • Wenjie Kong

  • Jing Jiang

  • Yuanlong Wang

  • Jiayi Chen

  • Bin Wang

  • Kun Wang

  • Yizhi Liang

  • Jiahan Wang

  • Chuan Li

  • Yanan Lin

  • Hongyan Gong

  • Yongxin Liang

  • Yanlin Bi

  • Xu Lin

  • March 27, 2026

  • 0 min

Share

Clinical Scorecard: Predictive Model Utilizing Machine Learning for Postoperative Delirium in Elderly Coronary Heart Disease Patients Undergoing Non-Cardiac Surgery: A Retrospective Cohort Analysis

At a Glance

CategoryDetail
ConditionPostoperative delirium (POD) in elderly patients with coronary heart disease (CHD)
Key MechanismsAcute and fluctuating impairment of consciousness, attention, cognition, and perception influenced by frailty, cognitive status, and insomnia
Target PopulationElderly patients (≥65 years) with CHD undergoing non-cardiac surgery
Care SettingPerioperative care in hospital surgical settings

Key Highlights

  • Gradient boosting machine (GBM) model based on seven key features effectively predicts POD with AUC of 0.856.
  • Clinical Frailty Scale (CFS) grade, Mini-mental State Examination (MMSE) score, and Athens Insomnia Scale (AIS) score are significant predictors of POD.
  • An online calculator based on the GBM model is available for clinical use to estimate POD risk.

Guideline-Based Recommendations

Diagnosis

  • Assess preoperative cognitive function using MMSE; exclude patients with MMSE <23.
  • Evaluate frailty status using Clinical Frailty Scale (CFS).
  • Assess sleep quality using Athens Insomnia Scale (AIS).

Management

  • Use predictive modeling (GBM) to identify high-risk elderly CHD patients for POD.
  • Implement early warning and preventive strategies in patients with higher CFS grade, lower MMSE score, and higher AIS score.

Monitoring & Follow-up

  • Monitor elderly CHD patients postoperatively for signs of delirium, especially those identified as high risk by the model.
  • Utilize the online calculator to support ongoing risk assessment.

Risks

  • Higher CFS grade increases POD risk.
  • Lower MMSE score indicates greater vulnerability to POD.
  • Higher AIS score correlates with increased POD occurrence.

Patient & Prescribing Data

Elderly patients with coronary heart disease undergoing non-cardiac surgery

Machine learning model identifies patients at increased risk for POD to guide personalized perioperative care and preventive interventions.

Clinical Best Practices

  • Incorporate frailty, cognitive, and sleep assessments preoperatively in elderly CHD patients.
  • Apply machine learning-based risk prediction tools to stratify POD risk.
  • Use predictive insights to tailor perioperative management and reduce POD incidence.
  • Validate predictive models externally before widespread clinical implementation.

References

Original Source(s)

Machine learning-based predictive model for postoperative delirium of elderly patients with coronary heart disease undergoing non-cardiac surgery: a retrospective cohort study

  • by Wenjie Kong, Jing Jiang, Yuanlong Wang, Jiayi Chen, Bin Wang, Kun Wang, Yizhi Liang, Jiahan Wang, Chuan Li, Yanan Lin, Hongyan Gong, Yongxin Liang, Yanlin Bi, Xu Lin

  • March 27, 2026

Related Content

Development of a machine learning-based depression risk prediction model for middle-aged and elderly Chinese heart disease patients: Evidence from CHARLS data

AI Scribes Lag Clinicians on Note Quality

A VHA study across 11 vendors finds AI-generated primary care notes score lower than clinician-written notes, with the largest deficits in thoroughness, organization, and usefulness

Late-life difficult-to-treat depression and dementia subtypes: a naturalistic cohort study using electronic health records