scholarly journals The pneumonia severity index: assessment and comparison to popular machine learning classifiers

2021 ◽  
Author(s):  
Dawei Wang ◽  
Deanna R. Willis ◽  
Yuehwern Yih
Keyword(s):  
Machine Learning ◽  
Random Forest ◽  
Pneumonia Severity Index ◽  
Severity Index ◽  
Community Acquired Pneumonia ◽  
Prediction Performance ◽  
Pneumonia Severity ◽  
Machine Learning Classifiers ◽  
Learning Classifiers ◽  
Accuracy Performance

AbstractPneumonia is the top communicable cause of death worldwide. Accurate prognostication of patient severity with Community Acquired Pneumonia (CAP) allows better patient care and hospital management. The Pneumonia Severity Index (PSI) was developed in 1997 as a tool to guide clinical practice by stratifying the severity of patients with CAP. While the PSI has been evaluated against other clinical stratification tools, it has not been evaluated against multiple classic machine learning classifiers in various metrics over large sample size. In this paper, we evaluated and compared the prediction performance of nine classic machine learning classifiers with PSI over 34720 adult (age 18+) patient records collected from 749 hospitals from 2009 to 2018 in the United States on Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) and Average Precision (Precision-Recall AUC). Machine learning classifiers, such as Random Forest, provided a significant improvement (∼29% in PR AUC and ∼5% in ROC AUC) compared to PSI and required only 7 input values (compared to 20 parameters used in PSI). There were also statistically significant differences (p<0.05) between Random Forest and PSI among various races/ethnicities. Because of its ease of use, PSI remains a very strong clinical decision tool, but machine learning classifiers can provide better prediction accuracy performance. Comparing prediction performance across multiple metrics such as PR AUC, instead of ROC AUC alone can provide additional insight.Key MessagesThis work compared the prognostication accuracy performance of patient severity with Community Acquired Pneumonia (CAP) between Pneumonia Severity Index (PSI) and nine machine learning classifiers and found machine learning classifiers provided a significant improvement.

Soluble form of suppression of tumorigenicity-2 predicts clinical stability of inpatients with community-acquired pneumonia

2021 ◽  
pp. 153537022110271
Author(s):  
Yifeng Zeng ◽  
Mingshan Xue ◽  
Teng Zhang ◽  
Shixue Sun ◽  
Runpei Lin ◽  
...  
Keyword(s):  
Regression Model ◽  
Predictive Value ◽  
Cox Regression ◽  
Pneumonia Severity Index ◽  
Severity Index ◽  
Community Acquired Pneumonia ◽  
Risk Classification ◽  
Soluble Form ◽  
Pneumonia Severity ◽  
Clinical Stability

The soluble form of the suppression of tumorigenicity-2 (sST2) is a biomarker for risk classification and prognosis of heart failure, and its production and secretion in the alveolar epithelium are significantly correlated with the inflammation-inducing in pulmonary diseases. However, the predictive value of sST2 in pulmonary disease had not been widely studied. This study investigated the potential value in prognosis and risk classification of sST2 in patients with community-acquired pneumonia. Clinical data of ninety-three CAP inpatients were retrieved and their sST2 and other clinical indices were studied. Cox regression models were constructed to probe the sST2’s predictive value for patients’ restoring clinical stability and its additive effect on pneumonia severity index and CURB-65 scores. Patients who did not reach clinical stability within the defined time (30 days from hospitalization) have had significantly higher levels of sST2 at admission ( P <  0.05). In univariate and multivariate Cox regression analysis, a high sST2 level (≥72.8 ng/mL) was an independent reverse predictor of clinical stability ( P < 0.05). The Cox regression model combined with sST2 and CURB-65 (AUC: 0.96) provided a more accurate risk classification than CURB-65 (AUC:0.89) alone (NRI: 1.18, IDI: 0.16, P < 0.05). The Cox regression model combined with sST2 and pneumonia severity index (AUC: 0.96) also provided a more accurate risk classification than pneumonia severity index (AUC:0.93) alone (NRI: 0.06; IDI: 0.06, P < 0.05). sST2 at admission can be used as an independent early prognostic indicator for CAP patients. Moreover, it can improve the predictive power of CURB-65 and pneumonia severity index score.

scholarly journals Oropharyngeal squamous cell carcinoma: radiomic machine-learning classifiers from multiparametric MR images for determination of HPV infection status

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Chong Hyun Suh ◽  
Kyung Hwa Lee ◽  
Young Jun Choi ◽  
Sae Rom Chung ◽  
Jung Hwan Baek ◽  
...  
Keyword(s):  
Machine Learning ◽  
Squamous Cell Carcinoma ◽  
Logistic Regression ◽  
Random Forest ◽  
Cell Carcinoma ◽  
Squamous Cell ◽  
Oropharyngeal Squamous Cell Carcinoma ◽  
Machine Learning Classifiers ◽  
Learning Classifiers ◽  
Hpv Status

Abstract We investigated the ability of machine-learning classifiers on radiomics from pre-treatment multiparametric magnetic resonance imaging (MRI) to accurately predict human papillomavirus (HPV) status in patients with oropharyngeal squamous cell carcinoma (OPSCC). This retrospective study collected data of 60 patients (48 HPV-positive and 12 HPV-negative) with newly diagnosed histopathologically proved OPSCC, who underwent head and neck MRIs consisting of axial T1WI, T2WI, CE-T1WI, and apparent diffusion coefficient (ADC) maps from diffusion-weighted imaging (DWI). The median age was 59 years (the range being 35 to 85 years), and 83.3% of patients were male. The imaging data were randomised into a training set (32 HPV-positive and 8 HPV-negative OPSCC) and a test set (16 HPV-positive and 4 HPV-negative OPSCC) in each fold. 1618 quantitative features were extracted from manually delineated regions-of-interest of primary tumour and one definite lymph node in each sequence. After feature selection by using the least absolute shrinkage and selection operator (LASSO), three different machine-learning classifiers (logistic regression, random forest, and XG boost) were trained and compared in the setting of various combinations between four sequences. The highest diagnostic accuracies were achieved when using all sequences, and the difference was significant only when the combination did not include the ADC map. Using all sequences, logistic regression and the random forest classifier yielded higher accuracy compared with the that of the XG boost classifier, with mean area under curve (AUC) values of 0.77, 0.76, and 0.71, respectively. The machine-learning classifier of non-invasive and quantitative radiomics signature could guide the classification of the HPV status.

scholarly journals Characterizing and Developing Strategies for the Treatment of Community-Acquired Pneumonia at a Community Hospital

2002 ◽  
Vol 9 (4) ◽  
pp. 247-252 ◽  
Author(s):  
Mark C Fok ◽  
Zahra Kanji ◽  
Rajesh Mainra ◽  
Michael Boldt
Keyword(s):  
Peer Group ◽  
Pneumonia Severity Index ◽  
Retrospective Chart Review ◽  
Severity Index ◽  
Early Discharge ◽  
Primary Diagnosis ◽  
Community Acquired Pneumonia ◽  
Pneumonia Severity ◽  
Pneumonia Severity Index Score ◽  
Step Down

BACKGROUND: Patients admitted to Lions Gate Hospital, North Vancouver, British Columbia, with a primary diagnosis of community-acquired pneumonia (CAP) have a mean length of stay (LOS) of 9.1 days compared with 7.9 days for peer group hospitals. This difference of 1.2 days results in an annual potential savings of 406 bed days and warranted an investigation into the management of CAP.OBJECTIVE: To characterize and provide recommendations for the management of CAP.METHODS: A retrospective chart review of patients admitted with a primary diagnosis of CAP between May 1, 2000 and August 31, 2000.RESULTS: Fifty-one patients were included in the study, with a mean LOS of 9.9 days and a median LOS of five days. Based on pneumonia severity index scores calculated for each patient, eight patients (16%) were admitted inappropriately. Initial empirical antibiotic choices were consistent with the Canadian CAP guidelines in 27 patients (53%), with inconsistencies arising mainly because cephalosporin or azithromycin monotherapy regimens were prescribed. Step-down from intravenous to oral antibiotics occurred in approximately 20 patients (39%). An additional 12 patients (24%) could have undergone step-down, and step-down was not applicable in 19 patients (37%). The potential annual cost avoidance from implementing admission criteria based on a pneumonia severity index score, applying step-down criteria and promoting early discharge criteria was estimated to be $220,000.CONCLUSIONS: Considerable variability exists in the treatment of CAP. A CAP preprinted order sheet was developed to address the issues identified in the present study and provide consistency in the management of CAP at Lions Gate Hospital.

CURB-65 PNEUMONIA SEVERITY INDEX AND APACHE II SCORES TO ASSESS MORTALITY RISK IN COMMUNITY-ACQUIRED PNEUMONIA PATIENTS FROM PROWESS

CHEST Journal ◽  
2007 ◽  
Vol 132 (4) ◽  
pp. 559A ◽  
Author(s):  
Guy Richards ◽  
Howard Levy ◽  
Pierre-Francois Laterre ◽  
Charles Feldman ◽  
Becky M. Bates ◽  
...  
Keyword(s):  
Mortality Risk ◽  
Pneumonia Severity Index ◽  
Severity Index ◽  
Community Acquired Pneumonia ◽  
Apache Ii ◽  
Pneumonia Severity
Sign in / Sign up

Export Citation Format

Share Document