Pulse Brain · Growing Health Evidence Index
Tier 4 — Narrative / commentaryPeer-reviewed

Use of deep learning to develop continuous-risk models for adverse event prediction from electronic health records

Nenad Tomašev, Natalie Harris, Sebastien Baur, Anne Mottram, Xavier Glorot, Jack W. Rae, Michał Zieliński, Harry Askham, André Saraiva, Valerio Magliulo, Clemens Meyer, Suman Ravuri, Ivan Protsyuk, Alistair Connell, Cían Hughes, Alan Karthikesalingam, Julien Cornebise, Hugh Montgomery, Geraint Rees, Chris Laing, Clifton R. Baker, Thomas F. Osborne, Ruth Reeves, Demis Hassabis, Dominic King, Mustafa Suleyman, Trevor Back, Christopher Nielson, Martin Seneviratne, Joseph R. Ledsam, Shakir Mohamed

Nature Protocols · 2021

Read source ↗ All evidence

Summary

This Nature Protocols paper describes a methodological framework for applying deep learning techniques to electronic health records to develop continuous-risk models for adverse event prediction. Published by researchers at DeepMind and partner institutions, the work outlines protocol-based approaches for translating machine learning into clinical decision support. As suggested by the title and journal type, the paper documents procedures and best practices rather than reporting primary clinical outcomes.

UK applicability

The methodology could be directly applicable to UK NHS electronic health record systems and clinical informatics infrastructure, though implementation would require alignment with UK data governance, GDPR compliance, and existing NHS digital standards. The protocol-based approach may facilitate adoption across UK healthcare trusts seeking to implement AI-assisted clinical risk prediction.

Key measures

Model performance metrics for adverse event prediction; clinical risk stratification; deep learning architecture and validation approaches applied to EHR data

Outcomes reported

The study demonstrates development and validation of continuous-risk prediction models using deep learning applied to electronic health record data to forecast adverse clinical events. The work is presented as a methodological protocol for implementing such models in clinical settings.

Theme
Nutrition & health
Subject
Other / interdisciplinary
Study type
Guideline
Study design
Methodology/Protocol paper
Source type
Peer-reviewed study
Status
Published
Geography
United Kingdom
System type
Human clinical
DOI
10.1038/s41596-021-00513-5
Catalogue ID
BFmor3g48f-x9vjdn

Topic tags

Pulse AI · ask about this record

Dig deeper with Pulse AI.

Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.