| Literature DB >> 33953393 |
Christopher Nielson1,2, Martin G Seneviratne3, Joseph R Ledsam4,5,6, Shakir Mohamed7, Nenad Tomašev8, Natalie Harris9, Sebastien Baur9, Anne Mottram7, Xavier Glorot7, Jack W Rae7,10, Michal Zielinski7, Harry Askham7, Andre Saraiva7, Valerio Magliulo9, Clemens Meyer7, Suman Ravuri7, Ivan Protsyuk9, Alistair Connell9, Cían O Hughes9, Alan Karthikesalingam9, Julien Cornebise7,11, Hugh Montgomery12, Geraint Rees13, Chris Laing14, Clifton R Baker1, Thomas F Osborne15,16, Ruth Reeves1, Demis Hassabis7, Dominic King9, Mustafa Suleyman7, Trevor Back7.
Abstract
Early prediction of patient outcomes is important for targeting preventive care. This protocol describes a practical workflow for developing deep-learning risk models that can predict various clinical and operational outcomes from structured electronic health record (EHR) data. The protocol comprises five main stages: formal problem definition, data pre-processing, architecture selection, calibration and uncertainty, and generalizability evaluation. We have applied the workflow to four endpoints (acute kidney injury, mortality, length of stay and 30-day hospital readmission). The workflow can enable continuous (e.g., triggered every 6 h) and static (e.g., triggered at 24 h after admission) predictions. We also provide an open-source codebase that illustrates some key principles in EHR modeling. This protocol can be used by interdisciplinary teams with programming and clinical expertise to build deep-learning prediction models with alternate data sources and prediction tasks.Entities:
Mesh：
Year: 2021 PMID： 33953393 DOI： 10.1038/s41596-021-00513-5
Source DB: PubMed Journal: Nat Protoc ISSN： 1750-2799 Impact factor: 13.491