Early machine learning prediction of hospitalized patients at low risk of respiratory deterioration or mortality in community-acquired pneumonia: Derivation and validation of a multivariable model

Strategies for preventing and reducing the impact of acute respiratory failure from pneumonia

INTRODUCTION

Pneumonia remains a leading cause of morbidity and mortality, particularly in critically ill patients with acute respiratory failure (ARF). This review discusses prevention strategies for pneumonia-induced ARF, categorized into primary, secondary, and tertiary prevention.

AREAS COVERED

A literature search was conducted through PubMed covering the years 2000-2024, using the keywords 'acute respiratory failure,' pneumonia prevention," 'risk stratification,' and 'preventive strategies.' Primary prevention focuses on reducing pneumonia risk through vaccination, smoking cessation, and comorbidity management. Secondary prevention involves early detection, risk assessment using clinical tools like the Pneumonia Severity Index (PSI) biomarkers, such as procalcitonin and C-reactive protein, appropriate antibiotic use, and emerging machine learning tools for real-time stratification. Tertiary prevention focuses on optimizing care with noninvasive respiratory support, lung-protective ventilation strategies, and ventilator bundles for intubated patients. Emerging therapies, including targeted use of corticosteroids and other immunomodulatory agents, are also discussed as promising adjuncts to current standards of care.

EXPERT OPINION

While these prevention strategies show potential, continued research is necessary to refine these interventions, explore newer therapies and evaluate long-term outcomes. Implementation of these strategies aims to reduce the impact of pneumonia-induced ARF on healthcare systems and improve patient survival and quality of care.

Prediction of mortality risk in patients with severe community-acquired pneumonia in the intensive care unit using machine learning

The aim of this study was to develop and validate a machine learning-based mortality risk prediction model for patients with severe community-acquired pneumonia (SCAP) in the intensive care unit (ICU). We collected data from two centers as the development and external validation cohorts. Variables were screened using the Recursive Feature Elimination method. Five machine learning algorithms were used to build predictive models. Models were evaluated through nested cross-validation to select the best one. The model was interpreted using Shapley Additive Explanations. We selected the optimal model to generate the web calculator. A total of 23 predictive features were selected. The Light Gradient Boosting Machine (LightGBM) model had an area under the receiver operating characteristic curve (AUC) of 0.842 (95% CI: 0.757-0.927), with an external 5-fold cross-validation average AUC of 0.842 ± 0.038, which was superior to the other models. External validation results also demonstrated good performance by the LightGBM model with an AUC of 0.856 (95% CI: 0.792-0.921). Based on this, we generated a web calculator by combining five high importance predictive factors. The LightGBM model was confirmed to be efficient and stable in predicting the mortality risk of patients with SCAP admitted to the ICU. The web calculator based on the LightGBM model can provide clinicians with a prognostic evaluation tool.

Forecasting hospital discharges for respiratory conditions in Costa Rica using climate and pollution data

Respiratory diseases represent one of the most significant economic burdens on healthcare systems worldwide. The variation in the increasing number of cases depends greatly on climatic seasonal effects, socioeconomic factors, and pollution. Therefore, understanding these variations and obtaining precise forecasts allows health authorities to make correct decisions regarding the allocation of limited economic and human resources. We aimed to model and forecast weekly hospitalizations due to respiratory conditions in seven regional hospitals in Costa Rica using four statistical learning techniques (Random Forest, XGboost, Facebook's Prophet forecasting model, and an ensemble method combining the above methods), along with 22 climate change indices and aerosol optical depth as an indicator of pollution. Models were trained using data from 2000 to 2018 and were evaluated using data from 2019 as testing data. During the training period, we set up 2-year sliding windows and a 1-year assessment period, along with the grid search method to optimize hyperparameters for each model. The best model for each region was selected using testing data, based on predictive precision and to prevent overfitting. Prediction intervals were then computed using conformal inference. The relative importance of all climatic variables was computed for the best model, and similar patterns in some of the seven regions were observed based on the selected model. Finally, reliable predictions were obtained for each of the seven regional hospitals.

The Neutrophil/Lymphocyte Ratio and Outcomes in Hospitalized Patients with Community-Acquired Pneumonia: A Retrospective Cohort Study

We aimed to assess the prognostic role of the neutrophil/lymphocyte ratio (NLR) in community-acquired pneumonia (CAP) via a single-center retrospective cohort of hospitalized adult patients from 1/2009 to 12/2019. Patients were dichotomized into lower NLR (≤12) and higher NLR (>12). The primary outcome was mortality. ICU admission and hospital- and ICU-free days were secondary outcomes. The pneumonia severity index (PSI) and the NLR's ability to predict outcomes was also tested. An NLR ≤12 was observed in 2513 (62.2%) patients and >12 in 1526 (37.8%). After adjusting for PSI, the NLR was not associated with hospital mortality (odds ratio [OR] 1.115; 95% confidence interval [CI] 0.774, 1.606; p = 0.559), but it was associated with a higher risk of ICU admission (OR 1.405; 95% CI 1.216, 1.624; p < 0.001). The PSI demonstrated acceptable discrimination for mortality (area under the receiver operating characteristic curve [AUC] 0.78; 95% CI 0.75, 0.82) which was not improved by adding the NLR (AUC 0.78; 95% CI 0.75, 0.82, p = 0.4476). The PSI's performance in predicting ICU admission was also acceptable (AUC 0.75; 95% CI 0.74, 0.77) and improved by including the NLR (AUC 0.76, 95% CI 0.74, 0.77, p = 0.008), although with limited clinical significance. The NLR was not superior to the PSI for predicting mortality in hospitalized CAP patients.