A clinically applicable approach to continuous prediction of future acute kidney injury

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

doi:10.1038/s41586-019-1390-1

Cited by 707 publications

(585 citation statements)

References 34 publications

Supporting

Mentioning

570

Contrasting

Unclassified

Order By: Relevance

“…Second, we expect related models to be extended to patient admission decisions as well as continuous hospital monitoring. [40][41][42] The qCSI does not separate patients without any nasal cannula requirement from those with even a minimal oxygen requirement. We expect that future models for safe discharge of COVID-19 patients will more strongly weigh even low oxygen requirements as local practice patterns may likely necessitate admission of any patient on exogenous oxygen.…”

Section: Discussionmentioning

confidence: 99%

Development and validation of the COVID-19 severity index (CSI): a prognostic tool for early respiratory decompensation

Haimovich

Ravindra

Stoytchev

et al. 2020

Preprint

View full text Add to dashboard Cite

Objective: The goal of this study was to create a predictive model of early hospital respiratory decompensation among patients with COVID-19.Design: Observational, retrospective cohort study.Setting: Nine-hospital health system within the Northeastern United States.Populations: Adult patients (≥ 18 years) admitted from the emergency department who tested positive for SARS-CoV-2 (COVID-19) up to 24 hours after initial presentation. Patients meeting criteria for respiratory critical illness within 4 hours of arrival were excluded.Main outcome and performance measures: We used a composite endpoint of critical illness as defined by oxygen requirement (greater than 10 L/min by low-flow device, high-flow device, non-invasive, or invasive ventilation) or death within the first 24 hours of hospitalization. We developed models predicting our composite endpoint using patient demographic and clinical data available within the first four hours of arrival. Eight hospitals (n = 932) were used for model development and one hospital (n = 240) was held out for external validation. Area under receiver operating characteristic (AU-ROC), precision-recall curves (AU-PRC), and calibration metrics were used to compare predictive models to three illness scoring systems: Elixhauser comorbidity index, qSOFA, and CURB-65.

show abstract

Section: Discussionmentioning

confidence: 99%

Development and validation of the COVID-19 severity index (CSI): a prognostic tool for early respiratory decompensation

Haimovich

Ravindra

Stoytchev

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…(3) We will consider more important variable categories (e.g., interventions, transfusions, diagnostics) (Koyner et al, 2018;Tomašev et al, 2019) and replicate the identified sub-phenotypes on more data sets. relationship of white blood cells with acute kidney injury and mortality in critically ill patients.…”

Section: Discussionmentioning

confidence: 99%

Identifying sub-phenotypes of acute kidney injury using structured and unstructured electronic health record data with memory networks

Chou

Zhang

et al. 2020

Journal of Biomedical Informatics

View full text Add to dashboard Cite

Acute Kidney Injury (AKI) is a common clinical syndrome characterized by the rapid loss of kidney excretory function, which aggravates the clinical severity of other diseases in a large number of hospitalized patients. Accurate early prediction of AKI can enable in-time interventions and treatments. However, AKI is highly heterogeneous, thus identification of AKI sub-phenotypes can lead to an improved understanding of the disease pathophysiology and development of more targeted clinical interventions. This study used a memory network-based deep learning approach to discover AKI sub-phenotypes using structured and unstructured electronic health record (EHR) data of patients before AKI diagnosis. We leveraged a real world critical care EHR corpus including 37,486 ICU stays. Our approach identified three distinct sub-phenotypes: sub-phenotype I is with an average age of 63.03±17.25 years, and is characterized by mild loss of kidney excretory function (Serum Creatinine (SCr) 1.55 ± 0.34 mg/dL, estimated Glomerular Filtration Rate Test (eGFR) 107.65±54.98 mL/min/1.73m 2 ). These patients are more likely to develop stage I AKI. Sub-phenotype II is with average age 66.81±10.43 years, and was characterized by severe loss of kidney excretory function (SCr 1.96 ± 0.49 mg/dL, eGFR 82.19 ± 55.92 mL/min/1.73m 2 ).

show abstract

“…This approach will be useful for categorizing differences, including normal from diseased. A relevant recent example has been the development of “deep learning” approaches for continuous prediction of incident disease, including AKI . Nevertheless, the limitations of a statistical approach described in the previous section still apply to the “big data” omics approach employing “artificial intelligence” methods.…”

Section: How Do Systems Biology and “Omics” Fit In?mentioning

confidence: 99%

Renal oxygenation: From data to insight

et al. 2020

View full text Add to dashboard Cite

Computational models have made a major contribution to the field of physiology. As the complexity of our understanding of biological systems expands, the need for computational methods only increases. But collaboration between experimental physiologists and computational modellers (ie theoretical physiologists) is not easy. One of the major challenges is to break down the barriers created by differences in vocabulary and approach between the two disciplines. In this review, we have two major aims. Firstly, we wish to contribute to the effort to break down these barriers and so encourage more interdisciplinary collaboration. So, we begin with a “primer” on the ways in which computational models can help us understand physiology and pathophysiology. Second, we aim to provide an update of recent efforts in one specific area of physiology, renal oxygenation. This work is shedding new light on the causes and consequences of renal hypoxia. But as importantly, computational modelling is providing direction for experimental physiologists working in the field of renal oxygenation by: (a) generating new hypotheses that can be tested in experimental studies, (b) allowing experiments that are technically unfeasible to be simulated in silico, or variables that cannot be measured experimentally to be estimated, and (c) providing a means by which the quality of experimental data can be assessed. Critically, based on our experience, we strongly believe that experimental and theoretical physiology should not be seen as separate exercises. Rather, they should be integrated to permit an iterative process between modelling and experimentation.

show abstract

A clinically applicable approach to continuous prediction of future acute kidney injury

Cited by 707 publications

References 34 publications

Development and validation of the COVID-19 severity index (CSI): a prognostic tool for early respiratory decompensation

Development and validation of the COVID-19 severity index (CSI): a prognostic tool for early respiratory decompensation

Identifying sub-phenotypes of acute kidney injury using structured and unstructured electronic health record data with memory networks

Renal oxygenation: From data to insight

Contact Info

Product

Resources

About