BACKGROUND: The detection of change in magnitude of directional coupling between two non-linear time series is a common subject of interest in the biomedical domain, including studies involving the respiratory chemoreflex system. Although transfer entropy is a useful tool in this avenue, no study to date has investigated how different transfer entropy estimation methods perform in typical biomedical applications featuring small sample size and presence of outliers. METHODS: With respect to detection of increased coupling strength, we compared three transfer entropy estimation techniques using both simulated time series and respiratory recordings from lambs. The following estimation methods were analyzed: fixed-binning with ranking, kernel density estimation (KDE), and the Darbellay-Vajda (D-V) adaptive partitioning algorithm extended to three dimensions. In the simulated experiment, sample size was varied from 50 to 200, while coupling strength was increased. In order to introduce outliers, the heavy-tailed Laplace distribution was utilized. In the lamb experiment, the objective was to detect increased respiratory-related chemosensitivity to O2 and CO2 induced by a drug, domperidone. Specifically, the separate influence of end-tidal PO2 and PCO2 on minute ventilation (V˙E) before and after administration of domperidone was analyzed. RESULTS: In the simulation, KDE detected increased coupling strength at the lowest SNR among the three methods. In the lamb experiment, D-V partitioning resulted in the statistically strongest increase in transfer entropy post-domperidone for PO2 → VE. In addition, D-V partitioning was the only method that could detect an increase in transfer entropy for PCO2 → VE, in agreement with experimental findings. CONCLUSIONS: Transfer entropy is capable of detecting directional coupling changes in non-linear biomedical time series analysis featuring a small number of observations and presence of outliers. The results of this study suggest that fixed-binning, even with ranking, is too primitive, and although there is no clear winner between KDE and D-V partitioning, the reader should note that KDE requires more computational time and extensive parameter selection than D-V partitioning. We hope this study provides a guideline for selection of an appropriate transfer entropy estimation method.

Clinical data management systems typically provide caregiver teams with useful information, derived from large, sometimes highly heterogeneous, data sources that are often changing dynamically. Over the last decade there has been a significant surge in interest in using these data sources, from simply reusing the standard clinical databases for event prediction or decision support, to including dynamic and patient-specific information into clinical monitoring and prediction problems. However, in most cases, commercial clinical databases have been designed to document clinical activity for reporting, liability, and billing reasons, rather than for developing new algorithms. With increasing excitement surrounding 'secondary use of medical records' and 'Big Data' analytics, it is important to understand the limitations of current databases and what needs to change in order to enter an era of 'precision medicine.' This review article covers many of the issues involved in the collection and preprocessing of critical care data. The three challenges in critical care are considered: compartmentalization, corruption, and complexity. A range of applications addressing these issues are covered, including the modernization of static acuity scoring; online patient tracking; personalized prediction and risk assessment; artifact detection; state estimation; and incorporation of multimodal data sources such as genomic and free text data.

With the booming of new technologies, biomedical science has transformed into digitalized, data intensive science. Massive amount of data need to be analyzed and interpreted, demand a complete pipeline to train next generation data scientists. To meet this need, the transinstitutional Big Data to Knowledge (BD2K) Initiative has been implemented since 2014, complementing other NIH institutional efforts. In this report, we give an overview the BD2K K01 mentored scientist career awards, which have demonstrated early success. We address the specific trainings needed in representative data science areas, in order to make the next generation of data scientists in biomedicine

OBJECTIVES: Sepsis is among the leading causes of morbidity, mortality, and cost overruns in critically ill patients. Early intervention with antibiotics improves survival in septic patients. However, no clinically validated system exists for real-time prediction of sepsis onset. We aimed to develop and validate an Artificial Intelligence Sepsis Expert algorithm for early prediction of sepsis. DESIGN: Observational cohort study. SETTING: Academic medical center from January 2013 to December 2015. PATIENTS: Over 31,000 admissions to the ICUs at two Emory University hospitals (development cohort), in addition to over 52,000 ICU patients from the publicly available Medical Information Mart for Intensive Care-III ICU database (validation cohort). Patients who met the Third International Consensus Definitions for Sepsis (Sepsis-3) prior to or within 4 hours of their ICU admission were excluded, resulting in roughly 27,000 and 42,000 patients within our development and validation cohorts, respectively.None. MEASUREMENTS AND MAIN RESULTS: High-resolution vital signs time series and electronic medical record data were extracted. A set of 65 features (variables) were calculated on hourly basis and passed to the Artificial Intelligence Sepsis Expert algorithm to predict onset of sepsis in the proceeding T hours (where T = 12, 8, 6, or 4). Artificial Intelligence Sepsis Expert was used to predict onset of sepsis in the proceeding T hours and to produce a list of the most significant contributing factors. For the 12-, 8-, 6-, and 4-hour ahead prediction of sepsis, Artificial Intelligence Sepsis Expert achieved area under the receiver operating characteristic in the range of 0.83-0.85. Performance of the Artificial Intelligence Sepsis Expert on the development and validation cohorts was indistinguishable. CONCLUSIONS: Using data available in the ICU in real-time, Artificial Intelligence Sepsis Expert can accurately predict the onset of sepsis in an ICU patient 4-12 hours prior to clinical recognition. A prospective study is necessary to determine the clinical utility of the proposed sepsis prediction model.

Medication dosing in a critical care environment is a complex task that involves close monitoring of relevant physiologic and laboratory biomarkers and corresponding sequential adjustment of the prescribed dose. Misdosing of medications with narrow therapeutic windows (such as intravenous [IV] heparin) can result in preventable adverse events, decrease quality of care and increase cost. Therefore, a robust recommendation system can help clinicians by providing individualized dosing suggestions or corrections to existing protocols. We present a clinician-in-the-loop framework for adjusting IV heparin dose using deep reinforcement learning (RL). Our main objectives were to learn a new IV heparin dosing policy based on the multi-dimensional features of patients, and evaluate the effectiveness of the learned policy in the presence of other confounding factors that may contribute to heparin-related side effects. The data used in the experiments included 2598 intensive care patients from the publicly available MIMIC database and 2310 patients from the Emory University clinical data warehouse. Experimental results suggested that the distance from RL policy had a statistically significant association with anticoagulant complications (p< 0.05), after adjusting for the effects of confounding factors.

Objective: Changes in heart rate (HR) and locomotor activity reflect changes in autonomic physiology, behavior, and mood. These systems may involve interrelated neural circuits that are altered in psychiatric illness, yet their interactions are poorly understood. We hypothesized interactions between HR and locomotor activity could be used to discriminate patients with schizophrenia from controls, and would be less able to discriminate non-psychiatric patients from controls. Approach: HR and locomotor activity were recorded via wearable patches in 16 patients with schizophrenia and 19 healthy controls. Measures of signal complexity and interactions were calculated over multiple time scales, including sample entropy, mutual information, and transfer entropy. A support vector machine was trained on these features to discriminate patients from controls. Additionally, time series were converted into a network with nodes comprised of HR and locomotor activity states, and edges representing state transitions. Graph properties were used as features. Leave-one-out cross validation was performed. To compare against non-psychiatric illness, the same approach was repeated in 41 patients with atrial fibrillation (AFib) and 53 controls. Main results: Network features enabled perfect discrimination of schizophrenia patients from controls with an areas under the receiver operating characteristic curve (AUC) of 1.00 for training and test data. Other bivariate measures of interaction achieved lower AUCs (train 0.98, test 0.96), and univariate measures of complexity achieved the lowest performance. Conversely, interaction features did not improve discrimination of AFib patients from controls beyond univariate approaches. Significance: Interactions between HR and locomotor activity enabled perfect discrimination of subjects with schizophrenia from controls, but these features were less performant in a non-psychiatric illness. This is the first quantitative evaluation of interactions between physiology and behavior in patients with psychiatric illness.

Objective: This work aims to validate a set of data processing methods for variability metrics, which hold promise as potential indicators for autonomic function, prediction of adverse cardiovascular outcomes, psychophysiological status, and general wellness. Although the investigation of heart rate variability (HRV) has been prevalent for several decades, the methods used for preprocessing, windowing, and choosing appropriate parameters lacks consensus among academic and clinical investigators. Moreover, many of the important steps are omitted from publications, preventing reproducibility. Approach: To address this, we have compiled a comprehensive and open-source modular toolbox for calculating HRV metrics and other related variability indices, on both raw cardiovascular time series and RR intervals. The software, known as the PhysioNet Cardiovascular Signal Toolbox, is implemented in the MATLAB programming language, with standard (open) input and output formats, and requires no external libraries. The functioning of our software is compared with other widely used and referenced HRV toolboxes to identify important differences. Main results: Our findings demonstrate how modest differences in the approach to HRV analysis can lead to divergent results, a factor that might have contributed to the lack of repeatability of studies and clinical applicability of HRV metrics. Significance: Existing HRV toolboxes do not include standardized preprocessing, signal quality indices (for noisy segment removal), and abnormal rhythm detection and are therefore likely to lead to significant errors in the presence of moderate to high noise or arrhythmias. We therefore describe the inclusion of validated tools to address these issues. We also make recommendations for default values and testing/reporting.

In this work, we propose a stacked switching vector-autoregressive (SVAR)-CNN architecture to model the changing dynamics in physiological time series for patient prognosis. The SVAR-layer extracts dynamical features (or modes) from the time-series, which are then fed into the CNN-layer to extract higher-level features representative of transition patterns among the dynamical modes. We evaluate our approach using 8-hours of minute-by-minute mean arterial blood pressure (BP) from over 450 patients in the MIMIC-II database. We modeled the time-series using a third-order SVAR process with 20 modes, resulting in first-level dynamical features of size 20×480 per patient. A fully connected CNN is then used to learn hierarchical features from these inputs, and to predict hospital mortality. The combined CNN/SVAR approach using BP time-series achieved a median and interquartile-range AUC of 0.74 [0.69, 0.75], significantly outperforming CNN-alone (0.54 [0.46, 0.59]), and SVAR-alone with logistic regression (0.69 [0.65, 0.72]). Our results indicate that including an SVAR layer improves the ability of CNNs to classify nonlinear and nonstationary time-series.

OBJECTIVES: To train a model to predict vasopressor use in ICU patients with sepsis and optimize external performance across hospital systems using domain adaptation, a transfer learning approach. DESIGN: Observational cohort study. SETTING: Two academic medical centers from January 2014 to June 2017. PATIENTS: Data were analyzed from 14,512 patients (9,423 at the development site and 5,089 at the validation site) who were admitted to an ICU and met Center for Medicare and Medicaid Services definition of severe sepsis either before or during the ICU stay. Patients were excluded if they never developed sepsis, if the ICU length of stay was less than 8 hours or more than 20 days or if they developed shock up to the first 4 hours of ICU admission. MEASUREMENTS AND MAIN RESULTS: Forty retrospectively collected features from the electronic medical records of adult ICU patients at the development site (four hospitals) were used as inputs for a neural network Weibull-Cox survival model to derive a prediction tool for future need of vasopressors. Domain adaptation updated parameters to optimize model performance in the validation site (two hospitals), a different healthcare system over 2,000 miles away. The cohorts at both sites were randomly split into training and testing sets (80% and 20%, respectively). When applied to the test set in the development site, the model predicted vasopressor use 4-24 hours in advance with an area under the receiver operator characteristic curve, specificity, and positive predictive value ranging from 0.80 to 0.81, 56.2% to 61.8%, and 5.6% to 12.1%, respectively. Domain adaptation improved performance of the model to predict vasopressor use within 4 hours at the validation site (area under the receiver operator characteristic curve 0.81 [CI, 0.80-0.81] from 0.77 [CI, 0.76-0.77], p < 0.01; specificity 59.7% [CI, 58.9-62.5%] from 49.9% [CI, 49.5-50.7%], p < 0.01; positive predictive value 8.9% [CI, 8.5-9.4%] from 7.3 [7.1-7.4%], p < 0.01). CONCLUSIONS: Domain adaptation improved performance of a model predicting sepsis-associated vasopressor use during external validation.

Objectives: To identify circumstances in which repeated measures of organ failure would improve mortality prediction in ICU patients. Design: Retrospective cohort study, with external validation in a deidentified ICU database. Setting: Eleven ICUs in three university hospitals within an academic healthcare system in 2014. Patients: Adults (18 yr old or older) who satisfied the following criteria: 1) two of four systemic inflammatory response syndrome criteria plus an ordered blood culture, all within 24 hours of hospital admission; and 2) ICU admission for at least 2 calendar days, within 72 hours of emergency department presentation. Intervention: None Measurements and Main Results: Data were collected until death, ICU discharge, or the seventh ICU day, whichever came first. The highest Sequential Organ Failure Assessment score from the ICU admission day (ICU day 1) was included in a multivariable model controlling for other covariates. The worst Sequential Organ Failure Assessment scores from the first 7 days after ICU admission were incrementally added and retained if they obtained statistical significance (p < 0.05). The cohort was divided into seven subcohorts to facilitate statistical comparison using the integrated discriminatory index. Of the 1,290 derivation cohort patients, 83 patients (6.4%) died in the ICU, compared with 949 of the 8,441 patients (11.2%) in the validation cohort. Incremental addition of Sequential Organ Failure Assessment data up to ICU day 5 improved the integrated discriminatory index in the validation cohort. Adding ICU day 6 or 7 Sequential Organ Failure Assessment data did not further improve model performance. Conclusions: Serial organ failure data improve prediction of ICU mortality, but a point exists after which further data no longer improve ICU mortality prediction of early sepsis.