With the proliferation of electronic healthcare data, data science and artificial intelligence have been used to gain new insights and develop clinical decision support (CDS) tools. However, this has been tempered by new analyses revealing not only limitations on generalizability and validity, but also discovery of underappreciated biases and health disparities - often with significant implications.
This Research Topic explores these avenues in the fields of pulmonary, critical care, and sleep medicine. Traditionally, studies have been conducted in laboratory settings. With new avenues opened, this topic will consider research from in silico trials through real-world implementation. Furthermore, this collection will also consider evaluations of implementation.
The Guest Editors of this Research Topic welcome original research articles and/or reviews on the topic of AI and data science in Pulmonary and Critical Care Medicine which aim to:
• Leverage AI and/or data science in the identification, interpretation, prediction, and management of physiologic and/or pathologic processes. This includes, but is not limited to pulmonary physiology, critical care diagnostics, and sleep therapeutics.
• Characterization of implemented AI and data science to evaluate validity (e.g., temporal, geographic)
• Identify and validate subphenotypes of physiology or pathology with clinically relevant context
• Identify and propose remedies for biases and health disparities discovered relevant to pulmonary and/or critical care medicine. This prioritizes disparities other than social determinants of health (e.g., impaired patient access), and includes - but is not limited to: insufficient calibration (e.g., racial disparities in pulse oximetry), insufficient sampling (e.g., racial bias in pulmonary function test normal ranges), and biased formulae (e.g., eGFR equations).
Specific sub-topics include, but are not limited to:
• Machine learning to predict clinically relevant endpoints (e.g. acute respiratory failure)
• Evaluation of implementation and quantitative effects (e.g. epic sepsis model)
• Data science to identify disparities in care (e.g. Analysis of discrepancies in pulse oximetry)
• Data subphenotypes (e.g. Sepsis endotypes)
Submissions with shared models and/or data will be preferred.
Topic Editor Yuh-Chin Huang receives financial support from Genentech, Inc and Windtree Therapeutics, Inc; Topic Editor An-Kwok Ian Wong holds equity and has management roles at Ataia Medical. The other Topic Editors declare no competing interests with regarding the Research Topic subject.
With the proliferation of electronic healthcare data, data science and artificial intelligence have been used to gain new insights and develop clinical decision support (CDS) tools. However, this has been tempered by new analyses revealing not only limitations on generalizability and validity, but also discovery of underappreciated biases and health disparities - often with significant implications.
This Research Topic explores these avenues in the fields of pulmonary, critical care, and sleep medicine. Traditionally, studies have been conducted in laboratory settings. With new avenues opened, this topic will consider research from in silico trials through real-world implementation. Furthermore, this collection will also consider evaluations of implementation.
The Guest Editors of this Research Topic welcome original research articles and/or reviews on the topic of AI and data science in Pulmonary and Critical Care Medicine which aim to:
• Leverage AI and/or data science in the identification, interpretation, prediction, and management of physiologic and/or pathologic processes. This includes, but is not limited to pulmonary physiology, critical care diagnostics, and sleep therapeutics.
• Characterization of implemented AI and data science to evaluate validity (e.g., temporal, geographic)
• Identify and validate subphenotypes of physiology or pathology with clinically relevant context
• Identify and propose remedies for biases and health disparities discovered relevant to pulmonary and/or critical care medicine. This prioritizes disparities other than social determinants of health (e.g., impaired patient access), and includes - but is not limited to: insufficient calibration (e.g., racial disparities in pulse oximetry), insufficient sampling (e.g., racial bias in pulmonary function test normal ranges), and biased formulae (e.g., eGFR equations).
Specific sub-topics include, but are not limited to:
• Machine learning to predict clinically relevant endpoints (e.g. acute respiratory failure)
• Evaluation of implementation and quantitative effects (e.g. epic sepsis model)
• Data science to identify disparities in care (e.g. Analysis of discrepancies in pulse oximetry)
• Data subphenotypes (e.g. Sepsis endotypes)
Submissions with shared models and/or data will be preferred.
Topic Editor Yuh-Chin Huang receives financial support from Genentech, Inc and Windtree Therapeutics, Inc; Topic Editor An-Kwok Ian Wong holds equity and has management roles at Ataia Medical. The other Topic Editors declare no competing interests with regarding the Research Topic subject.
