English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a machine learning pipeline and model that uses the entireuncurated EHR for prediction of in-hospital mortality at arbitrary timeintervals, using all available chart, lab and output events, without the needfor pre-processing or feature engineering. Data for more than 45,000 AmericanICU patients from the MIMIC-III database were used to develop an ICU mortalityprediction model. All chart, lab and output events were treated by the model inthe same manner inspired by Natural Language Processing (NLP). Patient eventswere discretized by percentile and mapped to learnt embeddings before beingpassed to a Recurrent Neural Network (RNN) to provide early prediction ofin-patient mortality risk. We compared mortality predictions with theSimplified Acute Physiology Score II (SAPS II) and the Oxford Acute Severity ofIllness Score (OASIS). Data were split into an independent test set (10%) and aten-fold cross-validation was carried out during training to avoid overfitting.13,233 distinct variables with heterogeneous data types were included withoutmanual selection or pre-processing. Recordings in the first few hours of apatient's stay were found to be strongly predictive of mortality, outperformingmodels using SAPS II and OASIS scores within just 2 hours and achieving a stateof the art Area Under the Receiver Operating Characteristic (AUROC) value of0.80 (95% CI 0.79-0.80) at 12 hours vs 0.70 and 0.66 for SAPS II and OASIS at24 hours respectively. Our model achieves a very strong performance of AUROC0.86 (95% CI 0.85-0.86) for in-patient mortality prediction after 48 hours onthe MIMIC-III dataset. Predictive performance increases over the first 48 hoursof the ICU stay, but suffers from diminishing returns, providing rationale fortime-limited trials of critical care and suggesting that the timing of decisionmaking can be optimised and individualised.

Dynamic survival prediction in intensive care units from heterogeneous time series without the need for variable selection or pre-processing