A recurrent neural network approach to predicting hemoglobin trajectories in patients with End-Stage Renal Disease

作者：

Highlights：

• A recurrent neural network approach is used to predict hemoglobin values.

• Future ESA and iron dosing is considered simultaneously.

• The hemoglobin predictions are patient specific.

• The predicted 3 month hemoglobin trajectory can forecast abrupt trend changes.

• The prediction error is small enough that the predictions are clinically relevant.

摘要

•A recurrent neural network approach is used to predict hemoglobin values.•Future ESA and iron dosing is considered simultaneously.•The hemoglobin predictions are patient specific.•The predicted 3 month hemoglobin trajectory can forecast abrupt trend changes.•The prediction error is small enough that the predictions are clinically relevant.

论文关键词：Hemoglobin prediction,Recurrent neural network,ESA dosing,Iron dosing,Anemia

论文评审过程：Received 15 August 2019, Revised 30 December 2019, Accepted 17 February 2020, Available online 19 February 2020, Version of Record 6 April 2020.

论文官网地址：https://doi.org/10.1016/j.artmed.2020.101823