Novita Dwi Safitri, Salwa Zahrotul Jannah

Survival Analysis of Kidney Failure Patients Using Nelson-Aalen Cumulative Hazard Function Estimation and Log-Rank Test

Introduction

Survival analysis of kidney failure patients using nelson-aalen cumulative hazard function estimation and log-rank test. Analyze kidney failure patient survival using Nelson-Aalen cumulative hazard and Log-Rank test. Explores death risk by severity, sex, and age, finding no significant statistical difference.

Abstract

Kidney failure is a critical condition that requires intensive monitoring of the patient's survival. This study aims to analyze the survival of kidney failure patients using a survival analysis approach with Nelson-Aalen cumulative hazard function estimation and the Log-Rank Test. The data used consisted of 106 patients with kidney failure who underwent hospitalization for 35 days, with characteristic variables including disease severity (chronic or acute), sex, and age group (≤ 50 years or > 50 years). The results of Nelson-Aalen's estimation showed that patients with acute conditions, male and > age 50 years had a visually higher risk of death. However, the results of the Log-Rank test showed that there was no significant difference in survival function between the three categories (p-value > 0.05). These findings suggest that although there are visual indications of a difference in risk, statistically, these clinical characteristics have not had a significant effect on patient survival in the observation period. This study suggests the need for further studies with data and a broader duration of observations to gain a deeper understanding.

Review

This study tackles the vital issue of survival in kidney failure patients, employing standard survival analysis techniques, specifically Nelson-Aalen cumulative hazard function estimation and the Log-Rank test. The research involved a cohort of 106 hospitalized patients observed over a 35-day period, investigating the impact of disease severity, sex, and age on survival. Initial insights from the Nelson-Aalen estimation suggested visually higher mortality risks for patients with acute conditions, males, and those over 50 years of age, setting up an expectation for potential group differences. However, the primary finding, which requires critical attention, is the subsequent lack of statistical significance from the Log-Rank test across all analyzed characteristics (p-value > 0.05). While the abstract appropriately notes that "statistically, these clinical characteristics have not had a significant effect," the initial "visual indications" of difference are considerably weakened by this result. This discrepancy strongly points towards potential limitations in the study's power, primarily stemming from the relatively small sample size (106 patients) and, more critically, the very short observation period of only 35 days. For a chronic and complex condition like kidney failure, such a brief window might not capture a sufficient number of survival events (deaths) or the full trajectory of the disease, thus hindering the ability to detect statistically significant differences even if true effects exist over a longer term. The authors' own recommendation for future studies with broader data and extended observation durations is highly commendable and absolutely necessary. To build upon this preliminary work, future investigations should consider substantially increasing both the sample size and the follow-up period to provide adequate power for detecting meaningful differences. Additionally, incorporating a richer set of covariates, such as specific underlying causes of kidney failure, common comorbidities, and the type or intensity of medical interventions, would greatly enhance the clinical relevance. Employing more sophisticated multivariate survival models, like Cox proportional hazards regression, would also be beneficial to adjust for confounding factors and provide a more comprehensive understanding of the independent predictors of survival in this critical patient population.

Full Text

Comments

You need to be logged in to post a comment.