Enhanced prediction of hemoglobin concentration in a very large cohort of hemodialysis patients by means of deep recurrent neural networks
作者：
Highlights：
• A model for hemoglobin prediction based on recurrent neural networks is proposed.
• It can model whole patient histories, providing predictions at every time step.
• Trained on 110,000 patients from 12 different countries taking various medications.
• The results show an improvement in accuracy over state-of-the-art anemia prediction.
• It brings the benefits of algorithmic anemia control to a large cohort of patients.
摘要
•A model for hemoglobin prediction based on recurrent neural networks is proposed.•It can model whole patient histories, providing predictions at every time step.•Trained on 110,000 patients from 12 different countries taking various medications.•The results show an improvement in accuracy over state-of-the-art anemia prediction.•It brings the benefits of algorithmic anemia control to a large cohort of patients.
论文关键词：Anemia,Chronic kidney disease,Erythropoetin stimulating agents,Deep learning,Recurrent neural networks
论文评审过程：Received 23 January 2019, Revised 30 May 2020, Accepted 1 June 2020, Available online 22 June 2020, Version of Record 22 June 2020.
论文官网地址：https://doi.org/10.1016/j.artmed.2020.101898
