Filipe R Lucini1,2, Karla D Krewulak1, Kirsten M Fiest1,3,4, Sean M Bagshaw5,6, Danny J Zuege1,6, Joon Lee2,3,7, Henry T Stelfox1,3. 1. Department of Critical Care Medicine, Cumming School of Medicine, University of Calgary and Alberta Health Services, Calgary, Alberta, Canada. 2. Data Intelligence for Health Lab, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada. 3. Department of Community Health Sciences and O'Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada. 4. Department of Psychiatry & Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada. 5. Department of Critical Care Medicine, Faculty of Medicine and Dentistry, University of Alberta, and Alberta Health Services, Edmonton, Alberta, Canada. 6. Critical Care Strategic Clinical Network, Alberta Health Services, Alberta, Canada. 7. Department of Cardiac Sciences, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada.
Abstract
OBJECTIVE: To apply natural language processing (NLP) techniques to identify individual events and modes of communication between healthcare professionals and families of critically ill patients from electronic medical records (EMR). MATERIALS AND METHODS: Retrospective cohort study of 280 randomly selected adult patients admitted to 1 of 15 intensive care units (ICU) in Alberta, Canada from June 19, 2012 to June 11, 2018. Individual events and modes of communication were independently abstracted using NLP and manual chart review (reference standard). Preprocessing techniques and 2 NLP approaches (rule-based and machine learning) were evaluated using sensitivity, specificity, and area under the receiver operating characteristic curves (AUROC). RESULTS: Over 2700 combinations of NLP methods and hyperparameters were evaluated for each mode of communication using a holdout subset. The rule-based approach had the highest AUROC in 65 datasets compared to the machine learning approach in 21 datasets. Both approaches had similar performance in 17 datasets. The rule-based AUROC for the grouped categories of patient documented to have family or friends (0.972, 95% CI 0.934-1.000), visit by family/friend (0.882 95% CI 0.820-0.943) and phone call with family/friend (0.975, 95% CI: 0.952-0.998) were high. DISCUSSION: We report an automated method to quantify communication between healthcare professionals and family members of adult patients from free-text EMRs. A rule-based NLP approach had better overall operating characteristics than a machine learning approach. CONCLUSION: NLP can automatically and accurately measure frequency and mode of documented family visitation and communication from unstructured free-text EMRs, to support patient- and family-centered care initiatives.
OBJECTIVE: To apply natural language processing (NLP) techniques to identify individual events and modes of communication between healthcare professionals and families of critically illpatients from electronic medical records (EMR). MATERIALS AND METHODS: Retrospective cohort study of 280 randomly selected adult patients admitted to 1 of 15 intensive care units (ICU) in Alberta, Canada from June 19, 2012 to June 11, 2018. Individual events and modes of communication were independently abstracted using NLP and manual chart review (reference standard). Preprocessing techniques and 2 NLP approaches (rule-based and machine learning) were evaluated using sensitivity, specificity, and area under the receiver operating characteristic curves (AUROC). RESULTS: Over 2700 combinations of NLP methods and hyperparameters were evaluated for each mode of communication using a holdout subset. The rule-based approach had the highest AUROC in 65 datasets compared to the machine learning approach in 21 datasets. Both approaches had similar performance in 17 datasets. The rule-based AUROC for the grouped categories of patient documented to have family or friends (0.972, 95% CI 0.934-1.000), visit by family/friend (0.882 95% CI 0.820-0.943) and phone call with family/friend (0.975, 95% CI: 0.952-0.998) were high. DISCUSSION: We report an automated method to quantify communication between healthcare professionals and family members of adult patients from free-text EMRs. A rule-based NLP approach had better overall operating characteristics than a machine learning approach. CONCLUSION: NLP can automatically and accurately measure frequency and mode of documented family visitation and communication from unstructured free-text EMRs, to support patient- and family-centered care initiatives.
Authors: Judy E Davidson; Rebecca A Aslakson; Ann C Long; Kathleen A Puntillo; Erin K Kross; Joanna Hart; Christopher E Cox; Hannah Wunsch; Mary A Wickline; Mark E Nunnally; Giora Netzer; Nancy Kentish-Barnes; Charles L Sprung; Christiane S Hartog; Maureen Coombs; Rik T Gerritsen; Ramona O Hopkins; Linda S Franck; Yoanna Skrobik; Alexander A Kon; Elizabeth A Scruth; Maurene A Harvey; Mithya Lewis-Newby; Douglas B White; Sandra M Swoboda; Colin R Cooke; Mitchell M Levy; Elie Azoulay; J Randall Curtis Journal: Crit Care Med Date: 2017-01 Impact factor: 7.598
Authors: Kirsten M Fiest; Karla D Krewulak; E Wesley Ely; Judy E Davidson; Zahinoor Ismail; Bonnie G Sept; Henry T Stelfox Journal: Crit Care Med Date: 2020-07 Impact factor: 7.598
Authors: Samiha Mohsen; Stephana J Moss; Filipe Lucini; Karla D Krewulak; Henry T Stelfox; Daniel J Niven; Khara M Sauro; Kirsten M Fiest Journal: Crit Care Med Date: 2022-08-26 Impact factor: 9.296
Authors: Tamryn F Gray; Anne Kwok; Khuyen M Do; Sandra Zeng; Edward T Moseley; Yasser M Dbeis; Renato Umeton; James A Tulsky; Areej El-Jawahri; Charlotta Lindvall Journal: JMIR Med Inform Date: 2022-06-15