New approaches to physiological informatics in neurocritical care

State of the Art of Machine Learning-Enabled Clinical Decision Support in Intensive Care Units: Literature Review

Background

Modern clinical care in intensive care units is full of rich data, and machine learning has great potential to support clinical decision-making. The development of intelligent machine learning–based clinical decision support systems is facing great opportunities and challenges. Clinical decision support systems may directly help clinicians accurately diagnose, predict outcomes, identify risk events, or decide treatments at the point of care.

Objective

We aimed to review the research and application of machine learning–enabled clinical decision support studies in intensive care units to help clinicians, researchers, developers, and policy makers better understand the advantages and limitations of machine learning–supported diagnosis, outcome prediction, risk event identification, and intensive care unit point-of-care recommendations.

Methods

We searched papers published in the PubMed database between January 1980 and October 2020. We defined selection criteria to identify papers that focused on machine learning–enabled clinical decision support studies in intensive care units and reviewed the following aspects: research topics, study cohorts, machine learning models, analysis variables, and evaluation metrics.

Results

A total of 643 papers were collected, and using our selection criteria, 97 studies were found. Studies were categorized into 4 topics—monitoring, detection, and diagnosis (13/97, 13.4%), early identification of clinical events (32/97, 33.0%), outcome prediction and prognosis assessment (46/97, 47.6%), and treatment decision (6/97, 6.2%). Of the 97 papers, 82 (84.5%) studies used data from adult patients, 9 (9.3%) studies used data from pediatric patients, and 6 (6.2%) studies used data from neonates. We found that 65 (67.0%) studies used data from a single center, and 32 (33.0%) studies used a multicenter data set; 88 (90.7%) studies used supervised learning, 3 (3.1%) studies used unsupervised learning, and 6 (6.2%) studies used reinforcement learning. Clinical variable categories, starting with the most frequently used, were demographic (n=74), laboratory values (n=59), vital signs (n=55), scores (n=48), ventilation parameters (n=43), comorbidities (n=27), medications (n=18), outcome (n=14), fluid balance (n=13), nonmedicine therapy (n=10), symptoms (n=7), and medical history (n=4). The most frequently adopted evaluation metrics for clinical data modeling studies included area under the receiver operating characteristic curve (n=61), sensitivity (n=51), specificity (n=41), accuracy (n=29), and positive predictive value (n=23).

Conclusions

Early identification of clinical and outcome prediction and prognosis assessment contributed to approximately 80% of studies included in this review. Using new algorithms to solve intensive care unit clinical problems by developing reinforcement learning, active learning, and time-series analysis methods for clinical decision support will be greater development prospects in the future.

Hierarchical Cluster Analysis Identifies Distinct Physiological States After Acute Brain Injury

BACKGROUND

Analysis of intracranial multimodality monitoring data is challenging, and quantitative methods may help identify unique physiological signatures that inform therapeutic strategies and outcome prediction. The aim of this study was to test the hypothesis that data-driven approaches can identify distinct physiological states from intracranial multimodality monitoring data.

METHODS

This was a single-center retrospective observational study of patients with either severe traumatic brain injury or high-grade subarachnoid hemorrhage who underwent invasive multimodality neuromonitoring. We used hierarchical cluster analysis to group hourly values for heart rate, mean arterial pressure, intracranial pressure, brain tissue oxygen, and cerebral microdialysis across all included patients into distinct groups. Average values for measured physiological variables were compared across the identified clusters, and physiological profiles from identified clusters were mapped onto physiological states known to occur after acute brain injury. The distribution of clusters was compared between patients with favorable outcome (discharged to home or acute rehab) and unfavorable outcome (in-hospital death or discharged to chronic nursing facility).

RESULTS

A total of 1704 observations from 20 patients were included. Even though the difference in mean values for measured variables between patients with favorable and unfavorable outcome were small, we identified four distinct clusters within our data: (1) events with low brain tissue oxygen and high lactate-to-pyruvate ratio-values (consistent with cerebral ischemia), (2) events with higher intracranial pressure values without evidence for ischemia (3) events which appeared to be physiologically "normal," and (4) events with high cerebral lactate without brain hypoxia (consistent with cerebral hyperglycolysis). Patients with a favorable outcome had a greater proportion of cluster 3 (normal) events, whereas patients with an unfavorable outcome had a greater proportion of cluster 1 (ischemia) and cluster 4 (hyperglycolysis) events (p < 0.0001, Fisher-Freeman-Halton test).

CONCLUSIONS

A data-driven approach can identify distinct groupings from invasive multimodality neuromonitoring data that may have implications for therapeutic strategies and outcome predictions. These groupings could be used as classifiers to train machine learning models that can aid in the treatment of patients with acute brain injury. Further work is needed to replicate the findings of this exploratory study in larger data sets.

Challenges and Opportunities in Multimodal Monitoring and Data Analytics in Traumatic Brain Injury

PURPOSE OF REVIEW

Increasingly sophisticated systems for monitoring the brain have led to an increase in the use of multimodality monitoring (MMM) to detect secondary brain injuries before irreversible damage occurs after brain trauma. This review examines the challenges and opportunities associated with MMM in this population.

RECENT FINDINGS

Locally and internationally, the use of MMM varies. Practical challenges include difficulties with data acquisition, curation, and harmonization with other data sources limiting collaboration. However, efforts toward integration of MMM data, advancements in data science, and the availability of cloud-based infrastructures are now affording the opportunity for MMM to advance the care of patients with brain trauma. MMM provides data to guide the precision management of patients with traumatic brain injury in real time. While challenges exist, there are exciting opportunities for MMM to live up to this promise and to drive new insights into the physiology of the brain and beyond.

Neurocritical Care: Bench to Bedside (Eds. Claude Hemphill, Michael James) Integrating and Using Big Data in Neurocritical Care

The critical care environment drives huge volumes of data, and clinicians are tasked with quickly processing this data and responding to it urgently. The neurocritical care environment increasingly involves EEG, multimodal intracranial monitoring, and complex imaging which preclude comprehensive human synthesis, and requires new concepts to integrate data into clinical care. By definition, Big Data is data that cannot be handled using traditional infrastructures and is characterized by the volume, variety, velocity, and variability of the data being produced. Big Data in the neurocritical care unit requires rethinking of data storage infrastructures and the development of tools and analytics to drive advancements in the field. Preprocessing, feature extraction, statistical inference, and analytic tools are required in order to achieve the primary goals of Big Data for clinical use: description, prediction, and prescription. Barriers to its use at bedside include a lack of infrastructure development within the healthcare industry, lack of standardization of data inputs, and ultimately existential and scientific concerns about the outputs that result from the use of tools such as artificial intelligence. However, as implied by the fundamental theorem of biomedical informatics, physicians remain central to the development and utility of Big Data to improve patient care.

Linking Big Data and Prediction Strategies: Tools, Pitfalls, and Lessons Learned

OBJECTIVES

Modern critical care amasses unprecedented amounts of clinical data-so called "big data"-on a minute-by-minute basis. Innovative processing of these data has the potential to revolutionize clinical prognostics and decision support in the care of the critically ill but also forces clinicians to depend on new and complex tools of which they may have limited understanding and over which they have little control. This concise review aims to provide bedside clinicians with ways to think about common methods being used to extract information from clinical big datasets and to judge the quality and utility of that information.

DATA SOURCES

We searched the free-access search engines PubMed and Google Scholar using the MeSH terms "big data", "prediction", and "intensive care" with iterations of a range of additional potentially associated factors, along with published bibliographies, to find papers suggesting illustration of key points in the structuring and analysis of clinical "big data," with special focus on outcomes prediction and major clinical concerns in critical care.

STUDY SELECTION

Three reviewers independently screened preliminary citation lists.

DATA EXTRACTION

Summary data were tabulated for review.

DATA SYNTHESIS

To date, most relevant big data research has focused on development of and attempts to validate patient outcome scoring systems and has yet to fully make use of the potential for automation and novel uses of continuous data streams such as those available from clinical care monitoring devices.

CONCLUSIONS

Realizing the potential for big data to improve critical care patient outcomes will require unprecedented team building across disparate competencies. It will also require clinicians to develop statistical awareness and thinking as yet another critical judgment skill they bring to their patients' bedsides and to the array of evidence presented to them about their patients over the course of care.

Multimorbidity states associated with higher mortality rates in organ dysfunction and sepsis: a data-driven analysis in critical care

BACKGROUND

Sepsis remains a complex medical problem and a major challenge in healthcare. Diagnostics and outcome predictions are focused on physiological parameters with less consideration given to patients' medical background. Given the aging population, not only are diseases becoming increasingly prevalent but occur more frequently in combinations ("multimorbidity"). We hypothesized the existence of patient subgroups in critical care with distinct multimorbidity states. We further hypothesize that certain multimorbidity states associate with higher rates of organ failure, sepsis, and mortality co-occurring with these clinical problems.

METHODS

We analyzed 36,390 patients from the open source Medical Information Mart for Intensive Care III (MIMIC III) dataset. Morbidities were defined based on Elixhauser categories, a well-established scheme distinguishing 30 classes of chronic diseases. We used latent class analysis to identify distinct patient subgroups based on demographics, admission type, and morbidity compositions and compared the prevalence of organ dysfunction, sepsis, and inpatient mortality for each subgroup.

RESULTS

We identified six clinically distinct multimorbidity subgroups labeled based on their dominant Elixhauser disease classes. The "cardiopulmonary" and "cardiac" subgroups consisted of older patients with a high prevalence of cardiopulmonary conditions and constituted 6.1% and 26.4% of study cohort respectively. The "young" subgroup included 23.5% of the cohort composed of young and healthy patients. The "hepatic/addiction" subgroup, constituting 9.8% of the cohort, consisted of middle-aged patients (mean age of 52.25, 95% CI 51.85-52.65) with the high rates of depression (20.1%), alcohol abuse (47.75%), drug abuse (18.2%), and liver failure (67%). The "complicated diabetics" and "uncomplicated diabetics" subgroups constituted 9.4% and 24.8% of the study cohort respectively. The complicated diabetics subgroup demonstrated higher rates of end-organ complications (88.3% prevalence of renal failure). Rates of organ dysfunction and sepsis ranged 19.6-69% and 12.5-46.7% respectively in the six subgroups. Mortality co-occurring with organ dysfunction and sepsis ranges was 8.4-23.8% and 11.7-27.4% respectively. These adverse outcomes were most prevalent in the hepatic/addiction subgroup.

CONCLUSION

We identify distinct multimorbidity states that associate with relatively higher prevalence of organ dysfunction, sepsis, and co-occurring mortality. The findings promote the incorporation of multimorbidity in healthcare models and the shift away from the current single-disease paradigm in clinical practice, training, and trial design.

Vagal contributions to fetal heart rate variability: an omics approach

OBJECTIVE

Fetal heart rate variability (fHRV) is an important indicator of health and disease, yet its physiological origins, neural contributions, in particular, are not well understood. We aimed to develop novel experimental and data analytical approaches to identify fHRV measures reflecting the vagus nerve contributions to fHRV.

APPROACH

In near-term ovine fetuses, a comprehensive set of 46 fHRV measures was computed from fetal pre-cordial electrocardiogram recorded during surgery and 72 h later without (n  =  24) and with intra-surgical bilateral cervical vagotomy (n  =  15).

MAIN RESULTS

The fetal heart rate did not change due to vagotomy. We identify fHRV measures specific to the vagal modulation of fHRV: multiscale time irreversibility asymmetry index (AsymI), detrended fluctuation analysis (DFA) α ₁, Kullback-Leibler permutation entropy (KLPE) and scale-dependent Lyapunov exponent slope (SDLE α).

SIGNIFICANCE

We provide a systematic delineation of vagal contributions to fHRV across signal-analytical domains which should be relevant for the emerging field of bioelectronic medicine and the deciphering of the 'vagus code'. Our findings also have clinical significance for in utero monitoring of fetal health during surgery. Key points •Fetal surgery causes a complex pattern of changes in heart rate variability measures with an overall reduction of complexity or variability. •At 72 h after surgery, many of the HRV measures recover and this recovery is delayed by an intrasurgical cervical bilateral vagotomy. •We identify HRV pattern representing complete vagal withdrawal that can be understood as part of 'HRV code', rather than any single HRV measure. •We identify HRV biomarkers of recovery from fetal surgery and discuss the effect of anticholinergic medication on this recovery.

Data Dissemination: Shortening the Long Tail of Traumatic Brain Injury Dark Data

Translation of traumatic brain injury (TBI) research findings from bench to bedside involves aligning multi-species data across diverse data types including imaging and molecular biomarkers, histopathology, behavior, and functional outcomes. In this review we argue that TBI translation should be acknowledged for what it is: a problem of big data that can be addressed using modern data science approaches. We review the history of the term big data, tracing its origins in Internet technology as data that are "big" according to the "4Vs" of volume, velocity, variety, veracity and discuss how the term has transitioned into the mainstream of biomedical research. We argue that the problem of TBI translation fundamentally centers around data variety and that solutions to this problem can be found in modern machine learning and other cutting-edge analytical approaches. Throughout our discussion we highlight the need to pull data from diverse sources including unpublished data ("dark data") and "long-tail data" (small, specialty TBI datasets undergirding the published literature). We review a few early examples of published articles in both the pre-clinical and clinical TBI research literature to demonstrate how data reuse can drive new discoveries leading into translational therapies. Making TBI data resources more Findable, Accessible, Interoperable, and Reusable (FAIR) through better data stewardship has great potential to accelerate discovery and translation for the silent epidemic of TBI.

Unraveling the complexities of invasive multimodality neuromonitoring

Acute brain injuries are a major cause of death and disability worldwide. Survivors of life-threatening brain injury often face a lifetime of dependent care, and novel approaches that improve outcome are sorely needed. A delayed cascade of brain damage, termed secondary injury, occurs hours to days and even weeks after the initial insult. This delayed phase of injury provides a crucial window for therapeutic interventions that could limit brain damage and improve outcome. A major barrier in the ability to prevent and treat secondary injury is that physicians are often unable to target therapies to patients' unique cerebral physiological disruptions. Invasive neuromonitoring with multiple complementary physiological monitors can provide useful information to enable this tailored, precision approach to care. However, integrating the multiple streams of time-varying data is challenging and often not possible during routine bedside assessment. The authors review and discuss the principles and evidence underlying several widely used invasive neuromonitors. They also provide a framework for integrating data for clinical decision making and discuss future developments in informatics that may allow new treatment paradigms to be developed.

Multimodality Neuromonitoring in Adult Traumatic Brain Injury: A Narrative Review

Neuromonitoring plays an important role in the management of traumatic brain injury. Simultaneous assessment of cerebral hemodynamics, oxygenation, and metabolism allows an individualized approach to patient management in which therapeutic interventions intended to prevent or minimize secondary brain injury are guided by monitored changes in physiologic variables rather than generic thresholds. This narrative review describes various neuromonitoring techniques that can be used to guide the management of patients with traumatic brain injury and examines the latest evidence and expert consensus guidelines for neuromonitoring.

Severe traumatic brain injury: targeted management in the intensive care unit

Severe traumatic brain injury (TBI) is currently managed in the intensive care unit with a combined medical-surgical approach. Treatment aims to prevent additional brain damage and to optimise conditions for brain recovery. TBI is typically considered and treated as one pathological entity, although in fact it is a syndrome comprising a range of lesions that can require different therapies and physiological goals. Owing to advances in monitoring and imaging, there is now the potential to identify specific mechanisms of brain damage and to better target treatment to individuals or subsets of patients. Targeted treatment is especially relevant for elderly people-who now represent an increasing proportion of patients with TBI-as preinjury comorbidities and their therapies demand tailored management strategies. Progress in monitoring and in understanding pathophysiological mechanisms of TBI could change current management in the intensive care unit, enabling targeted interventions that could ultimately improve outcomes.

Multimodality Neuromonitoring

The monitoring of systemic and central nervous system physiology is central to the management of patients with neurologic disease in the perioperative and critical care settings. There exists a range of invasive and noninvasive and global and regional monitors of cerebral hemodynamics, oxygenation, metabolism, and electrophysiology that can be used to guide treatment decisions after acute brain injury. With mounting evidence that a single neuromonitor cannot comprehensively detect all instances of cerebral compromise, multimodal neuromonitoring allows an individualized approach to patient management based on monitored physiologic variables rather than a generic one-size-fits-all approach targeting predetermined and often empirical thresholds.

Multimodal neurologic monitoring

Neurocritical care has two main objectives. Initially, the emphasis is on treatment of patients with acute damage to the central nervous system whether through infection, trauma, or hemorrhagic or ischemic stroke. Thereafter, attention shifts to the identification of secondary processes that may lead to further brain injury, including fever, seizures, and ischemia, among others. Multimodal monitoring is the concept of using various tools and data integration to understand brain physiology and guide therapeutic interventions to prevent secondary brain injury. This chapter will review the use of electroencephalography, intracranial pressure monitoring, brain tissue oxygenation, cerebral microdialysis and neurochemistry, near-infrared spectroscopy, and transcranial Doppler sonography as they relate to neuromonitoring in the critically ill. The concepts and design of each monitor, in addition to the patient population that may most benefit from each modality, will be discussed, along with the various tools that can be used together to guide individualized patient treatment options. Major clinical trials, observational studies, and their effect on clinical outcomes will be reviewed. The future of multimodal monitoring in the field of bioinformatics, clinical research, and device development will conclude the chapter.

Data collection and interpretation

Patient monitoring is routinely performed in all patients who receive neurocritical care. The combined use of monitors, including the neurologic examination, laboratory analysis, imaging studies, and physiological parameters, is common in a platform called multi-modality monitoring (MMM). However, the full potential of MMM is only beginning to be realized since for the most part, decision making historically has focused on individual aspects of physiology in a largely threshold-based manner. The use of MMM now is being facilitated by the evolution of bio-informatics in critical care including developing techniques to acquire, store, retrieve, and display integrated data and new analytic techniques for optimal clinical decision making. In this review, we will discuss the crucial initial steps toward data and information management, which in this emerging era of data-intensive science is already shifting concepts of care for acute brain injury and has the potential to both reshape how we do research and enhance cost-effective clinical care.

Identification of Clinically Relevant Groups of Patients Through the Application of Cluster Analysis to a Complex Traumatic Brain Injury Data Set

In neurological intensive care units (NICUs) we are collecting an ever increasing quantity of data. These range from patient demographics and physiological monitoring to treatment strategies and outcomes. The BrainIT database is an example of this type of rich data source. It contains validated data on 264 patients who suffered traumatic brain injury (TBI) admitted to 22 NICUs in 11 European countries between March 2003 and July 2005 [1, 6].

Topological data analysis for discovery in preclinical spinal cord injury and traumatic brain injury

Data-driven discovery in complex neurological disorders has potential to extract meaningful syndromic knowledge from large, heterogeneous data sets to enhance potential for precision medicine. Here we describe the application of topological data analysis (TDA) for data-driven discovery in preclinical traumatic brain injury (TBI) and spinal cord injury (SCI) data sets mined from the Visualized Syndromic Information and Outcomes for Neurotrauma-SCI (VISION-SCI) repository. Through direct visualization of inter-related histopathological, functional and health outcomes, TDA detected novel patterns across the syndromic network, uncovering interactions between SCI and co-occurring TBI, as well as detrimental drug effects in unpublished multicentre preclinical drug trial data in SCI. TDA also revealed that perioperative hypertension predicted long-term recovery better than any tested drug after thoracic SCI in rats. TDA-based data-driven discovery has great potential application for decision-support for basic research and clinical problems such as outcome assessment, neurocritical care, treatment planning and rapid, precision-diagnosis.

Does heart rate variability reflect the systemic inflammatory response in a fetal sheep model of lipopolysaccharide-induced sepsis?

Fetal inflammatory response occurs during chorioamnionitis, a frequent and often subclinical inflammation associated with increased risk for brain injury and life-lasting neurologic deficits. No means of early detection exist. We hypothesized that systemic fetal inflammation without septic shock will be reflected in alterations of fetal heart rate (FHR) variability (fHRV) distinguishing baseline versus inflammatory response states. In chronically instrumented near-term fetal sheep (n = 24), we induced an inflammatory response with lipopolysaccharide (LPS) injected intravenously (n = 14). Ten additional fetuses served as controls. We measured fetal plasma inflammatory cytokine IL-6 at baseline, 1, 3, 6, 24 and 48 h. 44 fHRV measures were determined continuously every 5 min using continuous individualized multi-organ variability analysis (CIMVA). CIMVA creates an fHRV measures matrix across five signal-analytical domains, thus describing complementary properties of fHRV. Using principal component analysis (PCA), a widely used technique for dimensionality reduction, we derived and quantitatively compared the CIMVA fHRV PCA signatures of inflammatory response in LPS and control groups. In the LPS group, IL-6 peaked at 3 h. In parallel, PCA-derived fHRV composite measures revealed a significant difference between LPS and control group at different time points. For the LPS group, a sharp increase compared to baseline levels was observed between 3 h and 6 h, and then abating to baseline levels, thus tracking closely the IL-6 inflammatory profile. This pattern was not observed in the control group. We also show that a preselection of fHRV measures prior to the PCA can potentially increase the difference between LPS and control groups, as early as 1 h post LPS injection. We propose a fHRV composite measure that correlates well with levels of inflammation and tracks well its temporal profile. Our results highlight the potential role of HRV to study and monitor the inflammatory response non-invasively over time.

Windowed multitaper correlation analysis of multimodal brain monitoring parameters

Although multimodal monitoring sets the standard in daily practice of neurocritical care, problem-oriented analysis tools to interpret the huge amount of data are lacking. Recently a mathematical model was presented that simulates the cerebral perfusion and oxygen supply in case of a severe head trauma, predicting the appearance of distinct correlations between arterial blood pressure and intracranial pressure. In this study we present a set of mathematical tools that reliably detect the predicted correlations in data recorded at a neurocritical care unit. The time resolved correlations will be identified by a windowing technique combined with Fourier-based coherence calculations. The phasing of the data is detected by means of Hilbert phase difference within the above mentioned windows. A statistical testing method is introduced that allows tuning the parameters of the windowing method in such a way that a predefined accuracy is reached. With this method the data of fifteen patients were examined in which we found the predicted correlation in each patient. Additionally it could be shown that the occurrence of a distinct correlation parameter, called scp, represents a predictive value of high quality for the patients outcome.

Information technology in critical care: review of monitoring and data acquisition systems for patient care and research

There is a broad consensus that 21st century health care will require intensive use of information technology to acquire and analyze data and then manage and disseminate information extracted from the data. No area is more data intensive than the intensive care unit. While there have been major improvements in intensive care monitoring, the medical industry, for the most part, has not incorporated many of the advances in computer science, biomedical engineering, signal processing, and mathematics that many other industries have embraced. Acquiring, synchronizing, integrating, and analyzing patient data remain frustratingly difficult because of incompatibilities among monitoring equipment, proprietary limitations from industry, and the absence of standard data formatting. In this paper, we will review the history of computers in the intensive care unit along with commonly used monitoring and data acquisition systems, both those commercially available and those being developed for research purposes.

Employing heat maps to mine associations in structured routine care data

OBJECTIVE

Mining the electronic medical record (EMR) has the potential to deliver new medical knowledge about causal effects, which are hidden in statistical associations between different patient attributes. It is our goal to detect such causal mechanisms within current research projects which include e.g. the detection of determinants of imminent ICU readmission. An iterative statistical approach to examine each set of considered attribute pairs delivers potential answers but is difficult to interpret. Therefore, we aimed to improve the interpretation of the resulting matrices by the use of heat maps. We propose strategies to adapt heat maps for the search for associations and causal effects within routine EMR data.

METHODS

Heat maps visualize tabulated metric datasets as grid-like choropleth maps, and thus present measures of association between numerous attribute pairs clearly arranged. Basic assumptions about plausible exposures and outcomes are used to allocate distinct attribute sets to both matrix dimensions. The image then avoids certain redundant graphical elements and provides a clearer picture of the supposed associations. Specific color schemes have been chosen to incorporate preexisting information about similarities between attributes. The use of measures of association as a clustering input has been taken as a trigger to apply transformations which ensure that distance metrics always assume finite values and treat positive and negative associations in the same way. To evaluate the general capability of the approach, we conducted analyses of simulated datasets and assessed diagnostic and procedural codes in a large routine care dataset.

RESULTS

Simulation results demonstrate that the proposed clustering procedure rearranges attributes similar to simulated statistical associations. Thus, heat maps are an excellent tool to indicate whether associations concern the same attributes or different ones, and whether affected attribute sets conform to any preexisting relationship between attributes. The dendrograms help in deciding if contiguous sequences of attributes effectively correspond to homogeneous attribute associations. The exemplary analysis of a routine care dataset revealed patterns of associations that follow plausible medical constellations for several diseases and the associated medical procedures and activities. Cases with breast cancer (ICD C50), for example, appeared to be associated with radiation therapy (8-52). In cross check, approximately 60 percent of the attribute pairs in this dataset showed a strong negative association, which can be explained by diseases treated in a medical specialty which routinely does not perform the respective procedures in these cases. The corresponding diagram clearly reflects these relationships in the shape of coherent subareas.

CONCLUSION

We could demonstrate that heat maps of measures of association are effective for the visualization of patterns in routine care EMRs. The adjustable method for the assignment of attributes to image dimensions permits a balance between the display of ample information and a favorable level of graphical complexity. The scope of the search can be adapted by the use of pre-existing assumptions about plausible effects to select exposure and outcome attributes. Thus, the proposed method promises to simplify the detection of undiscovered causal effects within routine EMR data.

Physiological monitoring of the severe traumatic brain injury patient in the intensive care unit

Traumatic brain injury (TBI) is a major cause of morbidity and mortality worldwide. Despite encouraging animal research, pharmacological agents and neuroprotectants have disappointed in the clinical environment. Current TBI management therefore is directed towards identification, prevention, and treatment of secondary cerebral insults that are known to exacerbate outcome after injury. This strategy is based on a variety of monitoring techniques that include the neurological examination, imaging, laboratory analysis, and physiological monitoring of the brain and other organ systems used to guide therapeutic interventions. Recent clinical series suggest that TBI management informed by multimodality monitoring is associated with improved patient outcome, in part because care is provided in a patient-specific manner. In this review we discuss physiological monitoring of the brain after TBI and the emerging field of neurocritical care bioinformatics.

Nonconvulsive seizures after subarachnoid hemorrhage: Multimodal detection and outcomes

OBJECTIVE

Seizures have been implicated as a cause of secondary brain injury, but the systemic and cerebral physiologic effects of seizures after acute brain injury are poorly understood.

METHODS

We analyzed intracortical electroencephalographic (EEG) and multimodality physiological recordings in 48 comatose subarachnoid hemorrhage patients to better characterize the physiological response to seizures after acute brain injury.

RESULTS

Intracortical seizures were seen in 38% of patients, and 8% had surface seizures. Intracortical seizures were accompanied by elevated heart rate (p = 0.001), blood pressure (p < 0.001), and respiratory rate (p < 0.001). There were trends for rising cerebral perfusion pressure (p = 0.03) and intracranial pressure (p = 0.06) seen after seizure onset. Intracortical seizure-associated increases in global brain metabolism, partial brain tissue oxygenation, and regional cerebral blood flow (rCBF) did not reach significance, but a trend for a pronounced delayed rCBF rise was seen for surface seizures (p = 0.08). Functional outcome was very poor for patients with severe background attenuation without seizures and best for those without severe attenuation or seizures (77% vs 0% dead or severely disabled, respectively). Outcome was intermediate for those with seizures independent of the background EEG and worse for those with intracortical only seizures when compared to those with intracortical and scalp seizures (50% and 25% death or severe disability, respectively).

INTERPRETATION

We replicated in humans complex physiologic processes associated with seizures after acute brain injury previously described in laboratory experiments and illustrated differences such as the delayed increase in rCBF. These real world physiologic observations may permit more successful translation of laboratory research to the bedside.

Altering physiological networks using drugs: steps towards personalized physiology

Background

The rise of personalized medicine has reminded us that each patient must be treated as an individual. One factor in making treatment decisions is the physiological state of each patient, but definitions of relevant states and methods to visualize state-related physiologic changes are scarce. We constructed correlation networks from physiologic data to demonstrate changes associated with pressor use in the intensive care unit.

Methods

We collected 29 physiological variables at one-minute intervals from nineteen trauma patients in the intensive care unit of an academic hospital and grouped each minute of data as receiving or not receiving pressors. For each group we constructed Spearman correlation networks of pairs of physiologic variables. To visualize drug-associated changes we split the networks into three components: an unchanging network, a network of connections with changing correlation sign, and a network of connections only present in one group.

Results

Out of a possible 406 connections between the 29 physiological measures, 64, 39, and 48 were present in each of the three component networks. The static network confirms expected physiological relationships while the network of associations with changed correlation sign suggests putative changes due to the drugs. The network of associations present only with pressors suggests new relationships that could be worthy of study.

Conclusions

We demonstrated that visualizing physiological relationships using correlation networks provides insight into underlying physiologic states while also showing that many of these relationships change when the state is defined by the presence of drugs. This method applied to targeted experiments could change the way critical care patients are monitored and treated.

Clinical review: neuromonitoring - an update

Critically ill patients are frequently at risk of neurological dysfunction as a result of primary neurological conditions or secondary insults. Determining which aspects of brain function are affected and how best to manage the neurological dysfunction can often be difficult and is complicated by the limited information that can be gained from clinical examination in such patients and the effects of therapies, notably sedation, on neurological function. Methods to measure and monitor brain function have evolved considerably in recent years and now play an important role in the evaluation and management of patients with brain injury. Importantly, no single technique is ideal for all patients and different variables will need to be monitored in different patients; in many patients, a combination of monitoring techniques will be needed. Although clinical studies support the physiologic feasibility and biologic plausibility of management based on information from various monitors, data supporting this concept from randomized trials are still required.

Research and technology in neurocritical care

The daily practice of neurointensivists focuses on the recognition of subtle changes in the neurological examination, interactions between the brain and systemic derangements, and brain physiology. Common alterations such as fever, hyperglycemia, and hypotension have different consequences in patients with brain insults compared with patients of general medical illness. Various technologies have become available or are currently being developed. The session on "research and technology" of the first neurocritical care research conference held in Houston in September of 2009 was devoted to the discussion of the current status, and the research role of state-of-the art technologies in neurocritical patients including multi-modality neuromonitoring, biomarkers, neuroimaging, and "omics" research (proteomix, genomics, and metabolomics). We have summarized the topics discussed in this session. We have provided a brief overview of the current status of these technologies, and put forward recommendations for future research applications in the field of neurocritical care.

Informatics infrastructure for syndrome surveillance, decision support, reporting, and modeling of critical illness

OBJECTIVE

To develop and validate an informatics infrastructure for syndrome surveillance, decision support, reporting, and modeling of critical illness.

METHODS

Using open-schema data feeds imported from electronic medical records (EMRs), we developed a near-real-time relational database (Multidisciplinary Epidemiology and Translational Research in Intensive Care Data Mart). Imported data domains included physiologic monitoring, medication orders, laboratory and radiologic investigations, and physician and nursing notes. Open database connectivity supported the use of Boolean combinations of data that allowed authorized users to develop syndrome surveillance, decision support, and reporting (data "sniffers") routines. Random samples of database entries in each category were validated against corresponding independent manual reviews.

RESULTS

The Multidisciplinary Epidemiology and Translational Research in Intensive Care Data Mart accommodates, on average, 15,000 admissions to the intensive care unit (ICU) per year and 200,000 vital records per day. Agreement between database entries and manual EMR audits was high for sex, mortality, and use of mechanical ventilation (kappa, 1.0 for all) and for age and laboratory and monitored data (Bland-Altman mean difference +/- SD, 1(0) for all). Agreement was lower for interpreted or calculated variables, such as specific syndrome diagnoses (kappa, 0.5 for acute lung injury), duration of ICU stay (mean difference +/- SD, 0.43+/-0.2), or duration of mechanical ventilation (mean difference +/- SD, 0.2+/-0.9).

CONCLUSION

Extraction of essential ICU data from a hospital EMR into an open, integrative database facilitates process control, reporting, syndrome surveillance, decision support, and outcome research in the ICU.

Identification of complex metabolic states in critically injured patients using bioinformatic cluster analysis

Introduction

Advances in technology have made extensive monitoring of patient physiology the standard of care in intensive care units (ICUs). While many systems exist to compile these data, there has been no systematic multivariate analysis and categorization across patient physiological data. The sheer volume and complexity of these data make pattern recognition or identification of patient state difficult. Hierarchical cluster analysis allows visualization of high dimensional data and enables pattern recognition and identification of physiologic patient states. We hypothesized that processing of multivariate data using hierarchical clustering techniques would allow identification of otherwise hidden patient physiologic patterns that would be predictive of outcome.

Methods

Multivariate physiologic and ventilator data were collected continuously using a multimodal bioinformatics system in the surgical ICU at San Francisco General Hospital. These data were incorporated with non-continuous data and stored on a server in the ICU. A hierarchical clustering algorithm grouped each minute of data into 1 of 10 clusters. Clusters were correlated with outcome measures including incidence of infection, multiple organ failure (MOF), and mortality.

Results

We identified 10 clusters, which we defined as distinct patient states. While patients transitioned between states, they spent significant amounts of time in each. Clusters were enriched for our outcome measures: 2 of the 10 states were enriched for infection, 6 of 10 were enriched for MOF, and 3 of 10 were enriched for death. Further analysis of correlations between pairs of variables within each cluster reveals significant differences in physiology between clusters.

Conclusions

Here we show for the first time the feasibility of clustering physiological measurements to identify clinically relevant patient states after trauma. These results demonstrate that hierarchical clustering techniques can be useful for visualizing complex multivariate data and may provide new insights for the care of critically injured patients.

Infection in the intensive care unit alters physiological networks

BACKGROUND

Physicians use clinical and physiological data to treat patients every day, and it is essential for treating a patient appropriately. However, medical sources of clinical physiological data are only now starting to find use in bioinformatics research.

RESULTS

We collected 29 types of physiological and clinical data on a minute-by-minute basis from trauma patients in the intensive care unit along with whether they contracted an infection during their stay. Dividing the patients into two groups based on this criterion, we determined that the correlational network amongst pairs of physiological variables changes based on whether the patient contracted an infection.

CONCLUSION

Examining the variable pairs with the largest change in correlation across groups reveals potential changes in the way our treatments affect the patient's physiology and in how our bodies react to physiological insults. These findings highlight the usefulness of physiological informatics and suggest new relationships to study while also validating previously reported relationships.

Laboratory testing of the Pressio intracranial pressure monitor

OBJECTIVE

The Sophysa Pressio (Sophysa Ltd., Orsay, France) is a new intracranial pressure monitoring system. This study aimed to evaluate its accuracy and compare it with the popular Codman intracranial pressure transducer (Codman/Johnson & Johnson, Raynham, MA) in vitro.

METHODS

A computerized rig was used to test the Pressio and Codman transducers simultaneously. Properties that were tested included drift over 7 days, the effect of temperature on drift, frequency response, the accuracy of measurement of static and pulsatile pressures, and connectivity of the system.

RESULTS

Long-term (7 d) relative zero drift was less than 0.05 mmHg. The temperature drift was low (0.3 mmHg/207C). Absolute static accuracy was better than 0.5 mmHg over the range of 0 to 100 mmHg. Pulse waveform accuracy, relative to the Codman transducer, was better than 0.2 mmHg over the range of 1 to 20 mmHg. The frequency bandwidth of the Pressio transducer was 22 Hz. The Pressio monitor can transmit data directly to an external computer without the use of a pressure bridge amplifier.

CONCLUSION

The new Pressio transducer proved to be accurate for measuring static and dynamic pressure during in vitro evaluation.