Xu Ke, Kang Hakmook
Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USA.
Kidney Res Clin Pract. 2025 Jun 10. doi: 10.23876/j.krcp.25.046.
Nephrology research plays an important role in advancing our understanding of kidney disease and improving patient outcomes. However, the complexity of nephrology data and the application of advanced statistical methods present significant challenges. This review highlights key statistical considerations in nephrology research, focusing on common errors such as violations of statistical assumptions, multicollinearity, missing data, overfitting, and the integration of machine learning tools. It emphasizes the importance of applying appropriate statistical approaches to ensure the reliability of study findings. Additionally, the review underscores the need for transparency and reproducibility in nephrology research, particularly the importance of open access to data, code, and study protocols. By utilizing tools like R, RStudio, Git, and GitHub, researchers can integrate their code, results, and data into a transparent workflow, enhancing the reproducibility of their research. This review also presents a practical checklist for promoting reproducible research practices, which can help improve the quality, transparency, and reliability of nephrology studies. This review aims to contribute to the advancement of nephrology research and, ultimately, to support the long-term goal of improving patient care and outcomes.
肾脏病学研究在增进我们对肾脏疾病的理解以及改善患者预后方面发挥着重要作用。然而，肾脏病学数据的复杂性以及先进统计方法的应用带来了重大挑战。本综述强调了肾脏病学研究中的关键统计考量因素，重点关注诸如违反统计假设、多重共线性、数据缺失、过度拟合以及机器学习工具的整合等常见错误。它强调了应用适当统计方法以确保研究结果可靠性的重要性。此外，该综述强调了肾脏病学研究中透明度和可重复性的必要性，特别是开放获取数据、代码和研究方案的重要性。通过使用R、RStudio、Git和GitHub等工具，研究人员可以将他们的代码、结果和数据整合到一个透明的工作流程中，提高其研究的可重复性。本综述还提出了一份促进可重复研究实践的实用清单，这有助于提高肾脏病学研究的质量、透明度和可靠性。本综述旨在推动肾脏病学研究的发展，并最终支持改善患者护理和预后的长期目标。
