Automated Identification of Cardiopulmonary Disease Cases for Preoperative Risk Stratification Using Machine Learning: A Retrospective Analysis

Clinical Snapshot

Key Findings

• P Value: Statistical significance via McNemar's test (specific p-values not provided)

• Effect Size: F1 score 0.71 for best performing SVM model

• Primary Outcome: Classification of cardiopulmonary-only vs mixed/other cases

• Nnt Or Sensitivity: Precision 0.94, False Positive Rate 0.6% (1/172), ~100× efficiency gain over manual review

• Confidence Interval: Bootstrap confidence intervals calculated but specific values not reported

Clinical Application

Requires implementation of NLP infrastructure and ML models, near-instantaneous processing reported Would need adaptation to Australian clinical documentation standards, TGA medical device regulations for clinical decision support, and alignment with ANZCA perioperative guidelines rather than US-based 2024 Mult Society guidelines Patients undergoing preoperative evaluation for noncardiac surgery with documented cardiopulmonary conditions

Abstract

BACKGROUND: Preoperative chart review is time-consuming and prone to errors, particularly for cardiopulmonary conditions that impact anesthetic planning. We developed a guideline-aligned "clinical insight bot" that mines free-text documentation to surface perioperative cardiovascular risk signals relevant to the 2024 Mult Society perioperative guideline for noncardiac surgery. METHODS: We analyzed 1000 de-identified medical cases from the PhysioNet MIMIC database. Medical terminology was extracted using regex-based NLP and categorized into 13 clinical specialties. Text features were encoded using TF-IDF vectorization and 1536-dimensional semantic embeddings stored in a PostgreSQL vector database (pgvector). Four machine learning models-Logistic Regression, Random Forest, Support Vector Machine (SVM), and Naive Bayes-were trained with stratified fivefold cross-validation to classify cases as "cardiopulmonary-only" versus "mixed/other." Performance was evaluated using accuracy, precision, recall, and F1 score, with statistical comparison via McNemar's test and bootstrap confidence intervals. RESULTS: In a held-out test set of 200 notes (28 positive; 172 negatives; ~14% prevalence), a linear support vector machine achieved the best overall balance (F1 ≈ 0.71), with high precision (positive predictive value 0.94) and very low false positive rate (FPR) (1/172 ≈ 0.6%). False negatives were the dominant residual error class. The pipeline processed documents near-instantaneously and, when scaled to 1000 notes, replaced on the order of tens of clinician review hours (≈100× efficiency gain) while maintaining performance across common preoperative document types. CONCLUSIONS: A lightweight, guideline-aligned insight bot can transform unstructured preoperative notes into concise, stepwise prompts that flag cardiovascular risk signals before the day of surgery. High precision with a very low FPR supports safe integration with anesthesiology workflows by minimizing paging noise, whereas time savings create operational and financial value. Future work should emphasize multicenter validation, structured data fusion (including labs, imaging, and vitals) to improve sensitivity, and prospective evaluation of downstream clinical and operational outcomes.

References

1. 1.Aggarwal, I., Rhee, C., Chura, M., Bora, V., & Reddy, D. M. (2026). Automated identification of cardiopulmonary disease cases for preoperative risk stratification using machine learning: A retrospective analysis. A&A Practice. https://doi.org/10.1213/XAA.0000000000002183