Division of Nephrology, University of British Columbia, Vancouver, British Columbia, Canada.
Division of Nephrology, University of British Columbia, Vancouver, British Columbia, Canada.
Am J Kidney Dis. 2016 May;67(5):779-86. doi: 10.1053/j.ajkd.2015.11.007. Epub 2015 Dec 23.
Predicting outcomes to guide clinical care, decision making, and resource allocation is a challenging undertaking in chronic kidney disease (CKD). Many prediction models have been developed, but few have been appropriately externally validated and even fewer have been assessed to be usable in the clinical setting. This contributes to the currently infrequent use of existing prediction models. Patients with CKD are a particularly heterogeneous group with significant biological variability, making the development of useful prediction models even more challenging. This article explores the different challenges in the development, validation, and application of prediction models in CKD. We explore the notion that newer biomarkers offer potential for enhancing existing and future prediction models and that modern technology is an opportunity to make prediction models more accessible and less cumbersome to use in clinical practice. Despite the challenges associated with their development and implementation, clinical prediction models have the potential to be a powerful tool for clinicians, researchers, and policy makers alike.
预测结局以指导临床护理、决策制定和资源分配是慢性肾脏病（CKD）面临的一项挑战。已经开发了许多预测模型，但很少有经过适当的外部验证，更少的模型被评估为可在临床环境中使用。这导致了现有预测模型的使用频率目前仍然较低。CKD 患者是一个特别具有异质性的群体，存在显著的生物学变异性，这使得开发有用的预测模型更加具有挑战性。本文探讨了在 CKD 中开发、验证和应用预测模型所面临的不同挑战。我们探讨了这样一种观点，即新型生物标志物有可能增强现有和未来的预测模型，而现代技术则为使预测模型在临床实践中更易于使用和更精简提供了机会。尽管在开发和实施方面存在挑战，但临床预测模型有可能成为临床医生、研究人员和决策者的有力工具。
