Nakhonphanom Clinical Risk Score (NP score) for Predicting Mortality of Hospitalized Community-Acquired Pneumonia Patients of Nakhonphanom Hospital

Abstract

Objective: To develop a clinical risk score (Nakhonphanom score; NP score) to help in prediction death for hospitalized community-acquired pneumonia.
Methods: A case-control data analysis was conducted at Nakhonphanom hospital, Thailand. Retrospective data collection from 1^st January 2019 – 31^st December 2019, of 1,006 hospitalized community-acquired pneumonia, case were 355 death cases and control were 651 survival cases. The best predictors were selected by multivariable logistic regression and transformed into clinical predictive scores.
Results: The best combination of predictors included ECOG (Eastern Cooperative Oncology Group) score 3 and over, White Blood Cell less than 4,000 cells/mm³, serum lactate over than 4 mmol/L, multi-lobar or bilateral pneumonia, and shock. The scores predicted death correctly with an AuROC of 80.13% (95CI; 77.30, 82.96). The likelihood ratio of positive for death was 0.52 (95%CI; 0.43, 0.62) in the low-risk category, 2.47 (95%CI; 2.04, 2.99) in the moderate-risk and 9.96 (95%CI; 2.93, 33.77) in the high-risk.
Conclusion: NP score is a simple clinical risk score prediction model that may help physician predict death in hospitalized community-acquired pneumonia. Further score external validation and clinical treatment guideline can significantly improve community-acquired pneumonia outcomes.