What Causes Postoperative Aspiration?

TL;DR

A machine learning model can effectively predict postoperative aspiration risk, enabling targeted preventative measures and important implications for improving postoperative care protocols and aspiration prevention strategies.

Abstract

Background: Aspiration, the inhalation of foreign material into the lungs, significantly impacts surgical patient morbidity and mortality. This study develops a machine learning (ML) model to predict postoperative aspiration, enabling timely preventative interventions. Methods: From the MIMIC-IV database of over 400,000 hospital admissions, we identified 826 surgical patients (mean age: 62, 55.7\% male) who experienced aspiration within seven days post-surgery, along with a matched non-aspiration cohort. Three ML models: XGBoost, Multilayer Perceptron, and Random Forest were trained using pre-surgical hospitalization data to predict postoperative aspiration. To investigate causation, we estimated Average Treatment Effects (ATE) using Augmented Inverse Probability Weighting. Results: Our ML model achieved an AUROC of 0.86 and 77.3\% sensitivity on a held-out test set. Maximum daily opioid dose, length of stay, and patient age emerged as the most important predictors. ATE analysis identified significant causative factors: opioids (0.25 +/- 0.06) and operative site (neck: 0.20 +/- 0.13, head: 0.19 +/- 0.13). Despite equal surgery rates across genders, men were 1.5 times more likely to aspirate and received 27\% higher maximum daily opioid dosages compared to women. Conclusion: ML models can effectively predict postoperative aspiration risk, enabling targeted preventative measures. Maximum daily opioid dosage and operative site significantly influence aspiration risk. The gender disparity in both opioid administration and aspiration rates warrants further investigation. These findings have important implications for improving postoperative care protocols and aspiration prevention strategies.

Figures (17)

• Figure 1: Flowchart of cohort selection. Our goal is to build a predictive model for postoperative aspiration using a cohort of positive and negative samples from MIMIC-IV. Of about 400K admissions, roughly 15% involved surgery. Admissions were labeled positive if the patient had no prior J69 ICD code, a chest X-ray confirming aspiration, and the X-ray occurred within 7 days of surgery. Patients with a history of aspiration were excluded to avoid an easier prediction task. To balance classes, negatives were downsampled to match the number of positives.
• Figure 2: Feature selection for a positive case. Extracting data from admission to the last surgery before aspiration (or last surgery for non-aspirating patients) to predict aspiration within 7 days post-operation.
• Figure 3: Correlation between different medications and postoperative aspiration. We see a higher correlation between opioids and saline flush with aspiration
• Figure 4: ATE of operative sites on postoperative aspiration, showing significant effects for neck, head, upper abdomen, and spine surgeries ($p < 0.05$)
• Figure 5: Difference in the distribution based on gender. Both MIMIC IV and patients who underwent surgeries were about 50% male-female. However, among the patients who aspirated post-operation, 61.3% of them were male.