About this Research Topic
Big Data and Artificial Intelligence are at the forefront of contemporary innovation in healthcare, particularly among the critically ill. Often, this is framed in the context of personalized care. Major advances in biospecimen analysis such as "omics" platforms can unlock the biochemical pathophysiology of many syndromes and diseases that afflict the critically ill population. However, these technologies are not immediately available to bedside clinicians, and it is not clear how they will have a tangible impact on treatment or outcomes. Clinicians need AI tools rooted in the physiologic response to diagnostic testing and therapy.
The electronic medical record (EMR) is a widely used, readily-available alternative data source for many machine learning-based models of syndromes like sepsis and acute respiratory distress syndrome. However, model designers relying on data from the EMR may not incorporate features representing physiologic responses to a diagnostic procedure or a therapeutic intervention. Physiologic changes to intentional iatrogenic stressors are a guiding principle in contemporary critical care, and represent an underutilized resource of Big Data. The goals of this themed article collection are:
(1) to highlight the specific ways in which dynamic physiologic challenges to hospitalized patients can be captured using established principles such as functional hemodynamic monitoring, and newer technologies such as videomicroscopy of microcirculatory flow;
(2) to demonstrate how physiologic challenges can be paired with powerful machine learning algorithms to create a useful data resource for decision support.
This Research Topic will provide an exhaustive review of the data sources that capture elements of dynamic physiology. These include simple diagnostic assessments such as capillary refill or murmurs heard through a stethoscope, more advanced diagnostic tools such as point-of-care ultrasound, and bedside assessments of fluid responsiveness such as pulse pressure or stroke volume variation. These tools and techniques can all diagnose clinical conditions of critical illness and organ perfusion and can test the response to therapeutic interventions over time. Authors addressing these topics should provide a summary of the state of each tool or technique in the context of AI-based clinical decision support, discuss specific challenges to the use of each, and suggestions on how those can be overcome. Topics of particular interest are those proposing the use of AI in the following contexts: (1) to diagnose or predict a disease; or (2) to facilitate noninvasive estimates of more invasive dynamic metrics. This Research Topic also aims to provide a summary of the challenges common to all modalities of dynamic physiological assessment that limit their use in AI designed for critical care, suggestions on how they can be overcome, and suggest future directions.
Topic editor Andre L. Holder declares that he received funding and speaker fees from Baxter International and consulting fees from Philips for clinical input. All Topic Editors declare no competing interests with regard to the Research Topic subject.
The electronic medical record (EMR) is a widely used, readily-available alternative data source for many machine learning-based models of syndromes like sepsis and acute respiratory distress syndrome. However, model designers relying on data from the EMR may not incorporate features representing physiologic responses to a diagnostic procedure or a therapeutic intervention. Physiologic changes to intentional iatrogenic stressors are a guiding principle in contemporary critical care, and represent an underutilized resource of Big Data. The goals of this themed article collection are:
(1) to highlight the specific ways in which dynamic physiologic challenges to hospitalized patients can be captured using established principles such as functional hemodynamic monitoring, and newer technologies such as videomicroscopy of microcirculatory flow;
(2) to demonstrate how physiologic challenges can be paired with powerful machine learning algorithms to create a useful data resource for decision support.
This Research Topic will provide an exhaustive review of the data sources that capture elements of dynamic physiology. These include simple diagnostic assessments such as capillary refill or murmurs heard through a stethoscope, more advanced diagnostic tools such as point-of-care ultrasound, and bedside assessments of fluid responsiveness such as pulse pressure or stroke volume variation. These tools and techniques can all diagnose clinical conditions of critical illness and organ perfusion and can test the response to therapeutic interventions over time. Authors addressing these topics should provide a summary of the state of each tool or technique in the context of AI-based clinical decision support, discuss specific challenges to the use of each, and suggestions on how those can be overcome. Topics of particular interest are those proposing the use of AI in the following contexts: (1) to diagnose or predict a disease; or (2) to facilitate noninvasive estimates of more invasive dynamic metrics. This Research Topic also aims to provide a summary of the challenges common to all modalities of dynamic physiological assessment that limit their use in AI designed for critical care, suggestions on how they can be overcome, and suggest future directions.
Topic editor Andre L. Holder declares that he received funding and speaker fees from Baxter International and consulting fees from Philips for clinical input. All Topic Editors declare no competing interests with regard to the Research Topic subject.
Keywords: Artificial Intelligence, Big Data, physiology, heart-lung interaction, clinical phenotyping, precision medicine, critical care
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
