Novel pneumonia score based on a machine learning model for predicting mortality in pneumonia patients on admission to the intensive care unit

BACKGROUND

Scores for predicting the long-term mortality of severe pneumonia are lacking. The purpose of this study is to use machine learning methods to develop new pneumonia scores to predict the 1-year mortality and hospital mortality of pneumonia patients on admission to the intensive care unit (ICU).

METHODS

The study population was screened from the MIMIC-IV and eICU databases. The main outcomes evaluated were 1-year mortality and hospital mortality in the MIMIC-IV database and hospital mortality in the eICU database. From the full data set, we separated patients diagnosed with community-acquired pneumonia (CAP) and ventilator-associated pneumonia (VAP) for subgroup analysis. We used common shallow machine learning algorithms, including logistic regression, decision tree, random forest, multilayer perceptron and XGBoost.

RESULTS

The full data set of the MIMIC-IV database contained 4697 patients, while that of the eICU database contained 13760 patients. We defined a new pneumonia score, the "Integrated CCI-APS", using a multivariate logistic regression model including six variables: metastatic solid tumor, Charlson Comorbidity Index, readmission, congestive heart failure, age, and Acute Physiology Score III. The area under the curve (AUC) and accuracy of the integrated CCI-APS were assessed in three data sets (full, CAP, and VAP) using both the test set derived from the MIMIC-IV database and the external validation set derived from the eICU database. The AUC value ranges in predicting 1-year and hospital mortality were 0.784-0.797 and 0.691-0.780, respectively, and the corresponding accuracy ranges were 0.723-0.725 and 0.641-0.718, respectively.

CONCLUSIONS

The main contribution of this study was a benchmark for using machine learning models to build pneumonia scores. Based on the idea of integrated learning, we propose a new integrated CCI-APS score for severe pneumonia. In the prediction of 1-year mortality and hospital mortality, our new pneumonia score outperformed the existing score.

The management of pneumonia in internal medicine

Pneumonia generates a high workload for internal medicine departments. Management of this disease is challenging, because patients are usually elderly and have multiple comorbid conditions. Furthermore, the interpretation and adherence to guidelines are far from clear in this setting. We report the opinion of 43 internists especially interested in infectious diseases that were questioned at the 2011 XXXII National Conference of Spanish Society of Internal Medicine about the main issues involved in the management of pneumonia in the internal medicine departments, namely, classification, admission criteria, microbiological workup, therapeutic management, discharge policy, and prevention of future episodes. Participants were asked to choose between 2 options for each statement by 4 investigators. Consensus could not be reached in many cases. The most controversial issues concerned recognition and management of healthcare-associated pneumonia (HCAP). Most participants were aware of the differences in terms of underlying diseases, etiological distribution, and outcome of HCAP compared with community-acquired pneumonia, but only a minority agreed to manage HCAP as hospital-acquired pneumonia, as suggested by some guidelines. A clinical patient-to-patient approach proved to be the option preferred by internists in the management of HCAP.

Timing of correct parenteral antibiotic initiation and outcomes from severe bacterial community-acquired pneumonia in children

BACKGROUND

The impact of timing of appropriate antibiotic initiation for critically ill children with severe bacterial community-acquired pneumonia (CAP) is unknown. We hypothesized that longer time to initiation of correct parenteral antibiotic would be associated with longer durations of mechanical ventilation, intensive care unit length of stay, and hospital length of stay.

METHODS

We retrospectively reviewed medical records of children admitted to Nationwide Children's Hospital between January 2004 and December 2006 with bacterial CAP treated with mechanical ventilation, excluding those with documented viral infection. Time to correct antibiotic was defined as time from presentation to any emergency department to the initiation of a parenteral antibiotic to which cultured pathogens were susceptible.

RESULTS

In all, 45 patients, median age 17 months, were identified. Median time to correct antibiotic was 10.3 hours, with 71% of patients receiving correct empiric therapy. After adjusting for severity of illness, longer time to correct antibiotic was independently associated with longer hospital stay (P = 0.007). For the 23 patients in the cohort for whom pneumonia was the primary diagnosis, longer time to correct antibiotic was independently associated with longer durations of mechanical ventilation (P = 0.01), intensive care unit stay (P = 0.001), and hospital stay (P = 0.006). Delays in antibiotic administration as short as 2 to 4 hours were associated with adverse outcomes in this group.

CONCLUSIONS

In our critically ill children with severe bacterial CAP, longer delays in receipt of appropriate empiric antibiotics were independently associated with adverse outcomes.