Literature DB >> 32065628

A customizable deep learning model for nosocomial risk prediction from critical care notes with indirect supervision.

Travis R Goodwin1, Dina Demner-Fushman1.   

Abstract

OBJECTIVE: Reliable longitudinal risk prediction for hospitalized patients is needed to provide quality care. Our goal is to develop a generalizable model capable of leveraging clinical notes to predict healthcare-associated diseases 24-96 hours in advance.
METHODS: We developed a reCurrent Additive Network for Temporal RIsk Prediction (CANTRIP) to predict the risk of hospital acquired (occurring ≥ 48 hours after admission) acute kidney injury, pressure injury, or anemia ≥ 24 hours before it is implicated by the patient's chart, labs, or notes. We rely on the MIMIC III critical care database and extract distinct positive and negative cohorts for each disease. We retrospectively determine the date-of-event using structured and unstructured criteria and use it as a form of indirect supervision to train and evaluate CANTRIP to predict disease risk using clinical notes.
RESULTS: Our experiments indicate that CANTRIP, operating on text alone, obtains 74%-87% area under the curve and 77%-85% Specificity. Baseline shallow models showed lower performance on all metrics, while bidirectional long short-term memory obtained the highest Sensitivity at the cost of significantly lower Specificity and Precision. DISCUSSION: Proper model architecture allows clinical text to be successfully harnessed to predict nosocomial disease, outperforming shallow models and obtaining similar performance to disease-specific models reported in the literature.
CONCLUSION: Clinical text on its own can provide a competitive alternative to traditional structured features (eg, lab values, vital signs). CANTRIP is able to generalize across nosocomial diseases without disease-specific feature extraction and is available at https://github.com/h4ste/cantrip. Published by Oxford University Press on behalf of the American Medical Informatics Association 2020. This work is written by US Government employees and is in the public domain in the US.

Entities:  

Keywords:  artificial intelligence; clinical; decision support systems; deep learning; machine learning; medical informatics; natural language processing

Year:  2020        PMID: 32065628      PMCID: PMC7075529          DOI: 10.1093/jamia/ocaa004

Source DB:  PubMed          Journal:  J Am Med Inform Assoc        ISSN: 1067-5027            Impact factor:   4.497


  43 in total

1.  PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals.

Authors:  A L Goldberger; L A Amaral; L Glass; J M Hausdorff; P C Ivanov; R G Mark; J E Mietus; G B Moody; C K Peng; H E Stanley
Journal:  Circulation       Date:  2000-06-13       Impact factor: 29.690

2.  KDIGO clinical practice guidelines for acute kidney injury.

Authors:  Arif Khwaja
Journal:  Nephron Clin Pract       Date:  2012-08-07

3.  Highly conservative phlebotomy in adult intensive care--a prospective randomized controlled trial.

Authors:  C R Harber; K J Sosnowski; R M Hegde
Journal:  Anaesth Intensive Care       Date:  2006-08       Impact factor: 1.669

Review 4.  Pressure ulcers in intensive care patients: a review of risks and prevention.

Authors:  B Paul J A Keller; Jan Wille; Bert van Ramshorst; Christian van der Werken
Journal:  Intensive Care Med       Date:  2002-09-07       Impact factor: 17.440

5.  Do blood tests cause anemia in hospitalized patients? The effect of diagnostic phlebotomy on hemoglobin and hematocrit levels.

Authors:  Paaladinesh Thavendiranathan; Akshay Bagai; Albert Ebidia; Allan S Detsky; Niteesh K Choudhry
Journal:  J Gen Intern Med       Date:  2005-06       Impact factor: 5.128

6.  Cost of Acute Kidney Injury in Hospitalized Patients.

Authors:  Samuel A Silver; Jin Long; Yuanchao Zheng; Glenn M Chertow
Journal:  J Hosp Med       Date:  2017-02       Impact factor: 2.960

7.  Hospital-acquired anemia: prevalence, outcomes, and healthcare implications.

Authors:  Colleen G Koch; Liang Li; Zhiyuan Sun; Eric D Hixson; Anne Tang; Shannon C Phillips; Eugene H Blackstone; J Michael Henderson
Journal:  J Hosp Med       Date:  2013-07-19       Impact factor: 2.960

8.  Anemia, transfusion, and phlebotomy practices in critically ill patients with prolonged ICU length of stay: a cohort study.

Authors:  Clarence Chant; Gail Wilson; Jan O Friedrich
Journal:  Crit Care       Date:  2006       Impact factor: 9.097

9.  Prediction of Acute Kidney Injury With a Machine Learning Algorithm Using Electronic Health Record Data.

Authors:  Hamid Mohamadlou; Anna Lynn-Palevsky; Christopher Barton; Uli Chettipally; Lisa Shieh; Jacob Calvert; Nicholas R Saber; Ritankar Das
Journal:  Can J Kidney Health Dis       Date:  2018-06-08

10.  A Probabilistic Reasoning Method for Predicting the Progression of Clinical Findings from Electronic Medical Records.

Authors:  Travis Goodwin; Sanda M Harabagiu
Journal:  AMIA Jt Summits Transl Sci Proc       Date:  2015-03-25
View more
  2 in total

1.  Examining the Concordance in the Documented Pressure Injury Site, Stage, and Count in Medical Information Mart for Intensive Care-III.

Authors:  Wenhui Zhang; Mani Sotoodeh; Joyce C Ho; Roy L Simpson; Vicki S Hertzberg
Journal:  Appl Clin Inform       Date:  2021-09-29       Impact factor: 2.762

2.  Current status and trends in researches based on public intensive care databases: A scientometric investigation.

Authors:  Min Li; Shuzhang Du
Journal:  Front Public Health       Date:  2022-09-15
  2 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.