Artificial intelligence in sepsis early prediction and diagnosis using unstructured data in healthcare

Kim Huat Goh; Le Wang; Adrian Yong Kwang Yeow; Hermione Poh; Ke Li; Joannas Jie Lin Yeow; Gamaliel Yu Heng Tan

doi:10.1038/s41467-021-20910-4

Author(s)

Kim Huat Goh
Adrian Yong Kwang Yeow
Hermione Poh
Ke Li
Joannas Jie Lin Yeow
Gamaliel Yu Heng Tan

Detail(s)

Original language	English
Article number	711
Journal / Publication	Nature Communications
Volume	12
Online published	29 Jan 2021
Publication status	Published - 2021
Externally published	Yes

Link(s)

DOI	DOI https://doi.org/10.1038/s41467-021-20910-4 Final Published version
	Check@CityULib
Attachment(s)	Documents 79131247.pdf Final Published version, 1.04 MB, PDF Publisher's Copyright Statement This full text is made available under CC-BY 4.0. https://creativecommons.org/licenses/by/4.0/
Link to Scopus	https://www.scopus.com/record/display.uri?eid=2-s2.0-85100110345&origin=recordpage
Permanent Link	https://scholars.cityu.edu.hk/en/publications/publication(72809d63-da57-4062-9f6c-b4f7f397f71d).html

Abstract

Sepsis is a leading cause of death in hospitals. Early prediction and diagnosis of sepsis, which is critical in reducing mortality, is challenging as many of its signs and symptoms are similar to other less critical conditions. We develop an artificial intelligence algorithm, SERA algorithm, which uses both structured data and unstructured clinical notes to predict and diagnose sepsis. We test this algorithm with independent, clinical notes and achieve high predictive accuracy 12 hours before the onset of sepsis (AUC 0.94, sensitivity 0.87 and specificity 0.87). We compare the SERA algorithm against physician predictions and show the algorithm’s potential to increase the early detection of sepsis by up to 32% and reduce false positives by up to 17%. Mining unstructured clinical notes is shown to improve the algorithm’s accuracy compared to using only clinical measures for early warning 12 to 48 hours before the onset of sepsis.