文献速递慢性肾脏病

Klinrisk机器学习模型在大型代表性美国人群中预测CKD进展的验证

Validation of the Klinrisk Machine Learning Model for CKD Progression in a Large Representative US Population.
PubMed 文献 · J Am Soc Nephrol · 2025-08-06

本研究在美国商业保险、Medicare和Medicaid参保成人中验证Klinrisk机器学习模型对2年CKD进展的预测能力。

要点

  • Klinrisk模型预测2年CKD进展的AUC为0.80–0.87。
  • 与KDIGO热图分期标准相比,Klinrisk模型预测能力更优。
  • 模型在商业保险、Medicare和Medicaid参保成人中均表现准确。

研究设计与样本

  • 研究建立三个队列,分别纳入2007年1月1日至2020年12月31日期间参加商业保险、Medicare和Medicaid的参保成人,要求至少1次血清肌酐检测、eGFR在15至180 ml/min/1.73 m2之间,且19种其他实验室指标中至少7种可用。
  • 每个保险类型下评估两个主要亚队列:所有具有至少7种实验室指标的患者,以及其中具备尿检结果的患者。
  • 疾病进展定义为eGFR持续下降40%或肾衰竭的复合终点。
  • 采用受试者工作特征曲线下面积(AUC)、Brier评分和校准图评估区分度、准确性和校准度。

主要发现

  • 商业保险队列中,Klinrisk模型2年AUC范围为0.83(95%CI 0.82至0.83)至0.87(95%CI 0.86至0.87),最大Brier评分为0.005(95%CI 0.0005至0.005)。
  • Medicare患者中,AUC范围为0.80(95%CI 0.79至0.80)至0.81(95%CI 0.80至0.82),最大Brier评分为0.026(95%CI 0.025至0.027)。
  • Medicaid患者中,AUC范围为0.82(95%CI 0.82至0.82)至0.84(95%CI 0.82至0.86),最大Brier评分为0.014(95%CI 0.012至0.015)。
  • 与标准诊疗即基于KDIGO热图的分期系统相比,Klinrisk模型预测能力更优。

关键数值与推荐

  • 模型预测2年CKD进展的AUC在商业保险队列为0.83–0.87,Medicare队列为0.80–0.81,Medicaid队列为0.82–0.84。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

KEY POINTS: The Klinrisk model predicted risk of 2-year CKD progression with area under the receiver operating characteristic curves of 0.80–0.87. When compared with standard of care, the Kidney Disease Improving Global Outcomes (KDIGO) heatmap-based staging system, the Klinrisk model achieved superior predictive capacity. The Klinrisk machine learning model predicted CKD progression in US commercial, Medicare, and Medicaid insured adults. BACKGROUND: Early identification of high-risk CKD can facilitate optimal medical management and improve outcomes. We aimed to validate the Klinrisk machine learning model for prediction of CKD progression in large US commercial, Medicare, and Medicaid populations. METHODS: We developed three cohorts, consisting of insured adults enrolled in (1) commercial, (2) Medicare, and (3) Medicaid plans between January 1, 2007, and December 31, 2020, with ≥1 serum creatinine test, an eGFR between 15 and 180 ml/min per 1.73 m2, and ≥7 of the 19 other laboratory analytes available. Two primary subcohorts were evaluated within each insurer: (1) all patients with ≥7 laboratory analytes and (2) patients in (1) with available urinalysis results. Disease progression was defined as the composite outcome of a sustained 40% decline in eGFR or kidney failure. Discrimination, accuracy, and calibration were assessed using the area under the receiver operating characteristic curve (AUC), Brier scores, and calibration plots. RESULTS: In the commercial cohort, the Klinrisk model achieved AUCs ranging from 0.83 (95% confidence interval, 0.82 to 0.83) to 0.87 (0.86 to 0.87) and a maximum Brier score of 0.005 (0.0005 to 0.005) at 2 years. In Medicare patients, AUCs ranged from 0.80 (0.79 to 0.80) to 0.81 (0.80 to 0.82), with a maximum Brier score of 0.026 (0.025 to 0.027). In Medicaid patients, we found AUCs ranging from 0.82 (0.82 to 0.82) to 0.84 (0.82 to 0.86) and a maximum Brier score of 0.014 (0.012 to 0.015). CONCLUSIONS: The Klinrisk machine learning model was accurate in predicting CKD progression in 4.8 million US adults across commercial, Medicare, and Medicaid populations.

来源:J Am Soc Nephrol,查看原文。版权归原作者所有。

引用

  1. Tangri N, Ferguson TW, Teng CC, et al. Validation of the Klinrisk Machine Learning Model for CKD Progression in a Large Representative US Population. J Am Soc Nephrol. 2025. PMID: 40768294. doi:10.1681/ASN.0000000817

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。