Photo Credit: mi-viri
The following is a summary of “Harnessing Artificial Intelligence for Predictive Modelling of ICU Admissions in Patients with Acute Exacerbation of COPD: A Prospective Study,” published in the October 2024 issue of Pulmonology by Priyatha et al.
Artificial intelligence (AI) can predict ICU admission for patients presenting with acute exacerbation of chronic obstructive pulmonary disease (COPD).
Researchers conducted a retrospective study exploring the potential of AI to predict the need for ICU admission in patients with acute exacerbation of COPD.
They enrolled patients with acute exacerbation of COPD over 1 year and collected baseline demographic data, clinical parameters, laboratory results, and radiographic findings upon admission. An AI-driven predictive model was developed using machine learning algorithms, incorporating variables identified through feature selection techniques. Trained the model on a subset of the dataset and validated it on a separate cohort. The performance metrics were calculated, including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve (AUC), to assess the predictive accuracy of the model.
The results showed 500 patients with acute exacerbation of COPD, of which 150 required ICU admission. The AI-driven predictive model showed promising performance in identifying patients at risk of ICU admission, with an AUC of 0.85 (95% CI: 0.80–0.90) in the validation cohort. At a predefined threshold, the model exhibited a sensitivity of 0.80, specificity of 0.78, PPV of 0.65, and NPV of 0.89. Key predictors of ICU admission included the severity of dyspnea, arterial blood gas parameters, comorbidities, and prior exacerbation history.
They concluded that AI-driven predictive modeling can accurately identify patients with COPD at high risk of ICU admission, enabling early intervention and improved outcomes.
Source: journal.chestnet.org/article/S0012-3692(24)00938-3/fulltext
