Aulia Azira Putri, Dian Desti Ananda

Survival Analysis of Pneumonia Patients Using Kaplan-Meier and Cox Proportional Hazard Regression

Introduction

Survival analysis of pneumonia patients using kaplan-meier and cox proportional hazard regression. Analyze pneumonia patient survival in ICU using Kaplan-Meier and Cox regression. Uncover survival trends for elderly and complicated cases, emphasizing disease severity in risk assessment.

Abstract

Pneumonia is one of the leading infectious diseases in the world, especially in the elderly and patients with complications. This study aims to analyze the survival of pneumonia patients treated in the ICU room of Hospital X by 2024 using the Kaplan-Meier method and Cox Proportional Hazard regression. The data used consisted of 37 patients, with independent variables in the form of age and type of pneumonia (mildness, sepsis, and other complications), and dependent variables in the form of patient survival time. The Kaplan-Meier analysis showed that patients with >60 years of age and complicated types of pneumonia tended to have lower survival rates. However, the results of the Log-Rank test showed that there was no statistically significant difference between age groups and types of pneumonia on survival (p > 0.05). The results of the Cox regression also showed that no variables had a significant effect on the patient's risk of death, although types of pneumonia with other complications showed a significant tendency at the level of 10% (p = 0.0699). This study demonstrates the importance of considering disease severity in the evaluation of the risk of death of pneumonia patients, as well as the need for further research with larger sample counts and additional clinical variables.

Review

This study addresses a highly relevant and critical area of public health: survival outcomes for pneumonia patients, particularly in the vulnerable ICU setting. The authors appropriately employ established survival analysis methods, Kaplan-Meier and Cox Proportional Hazard regression, which are standard tools for evaluating time-to-event data in medical research. The initial findings suggesting lower survival rates for older patients and those with complicated pneumonia types align with existing clinical understanding, underscoring the importance of patient stratification based on disease severity, even if statistical significance was not achieved in all analyses. However, the most significant limitation of this study is the remarkably small sample size of 37 patients. This severely restricts the statistical power to detect true effects and leads directly to the non-significant findings for both the Log-Rank test (p > 0.05 for age and type of pneumonia) and the Cox regression, where no variables reached conventional statistical significance. While the authors note a "significant tendency" for complicated pneumonia at the 10% level (p = 0.0699), this falls short of widely accepted thresholds for statistical inference in medical literature and likely reflects the underpowered nature of the study rather than a strong effect. The broad categorization of "type of pneumonia" (mildness, sepsis, and other complications) could also benefit from more granular and clinically defined classifications to capture the heterogeneity of the disease. For future research, it is imperative to conduct studies with substantially larger sample sizes, ideally across multiple centers, to enhance generalizability and statistical power. Incorporating a wider array of relevant clinical variables, such as comorbidities (e.g., Charlson Comorbidity Index), severity scores (e.g., SOFA, APACHE II, CURB-65), specific pathogen identification, initial treatment regimens, and ventilator status, would provide a more comprehensive understanding of factors influencing survival. While this study serves as a valuable preliminary exploration, its current findings should be interpreted with extreme caution and are best considered hypothesis-generating, warranting robust validation in adequately powered and more detailed investigations.

Full Text

Comments

You need to be logged in to post a comment.