No abstract

<b><i>Introduction:</i></b> Future digital health research hinges on methodologies to conduct remote clinical assessments and in-home monitoring. The Collaborative Aging Research Using Technology (CART) initiative was introduced to establish a digital technology research platform that could widely assess activity in the homes of diverse cohorts of older adults and detect meaningful change longitudinally. This paper reports on the built end-to-end design of the CART platform, its functionality, and the resulting research capabilities. <b><i>Methods:</i></b> CART platform development followed a principled design process aiming for scalability, use case flexibility, longevity, and data privacy protection while allowing sharability. The platform, comprising ambient technology, wearables, and other sensors, was deployed in participants’ homes to provide continuous, long-term (months to years), and ecologically valid data. Data gathered from CART homes were sent securely to a research server for analysis and future data sharing. <b><i>Results:</i></b> The CART system was created, iteratively tested, and deployed to 232 homes representing four diverse cohorts (African American, Latinx, low-income, and predominantly rural-residing veterans) of older adults (<i>n</i> = 301) across the USA. Multiple measurements of wellness such as cognition (e.g., mean daily computer use time = 160–169 min), physical mobility (e.g., mean daily transitions between rooms = 96–155), sleep (e.g., mean nightly sleep duration = 6.3–7.4 h), and level of social engagement (e.g., reports of overnight visitors = 15–45%) were collected across cohorts. <b><i>Conclusion:</i></b> The CART initiative resulted in a minimally obtrusive digital health-enabled system that met the design principles while allowing for data capture over extended periods and can be widely used by the research community. The ability to monitor and manage health digitally within the homes of older adults is an important alternative to in-person assessments in many research contexts. Further advances will come with wider, shared use of the CART system in additional settings, within different disease contexts, and by diverse research teams.

Implantable-cardioverter defibrillators (ICD) detect and terminate life-threatening ventricular tachyarrhythmia with electric shocks after they occur. This puts patients at risk if they are driving or in a situation where they can fall. ICD’s shocks are also very painful and affect a patient’s quality of life. It would be ideal if ICDs can accurately predict the occurrence of ventricular tachyarrhythmia and then issue a warning or provide preventive therapy. Our study explores the use of ICD data to automatically predict ventricular arrhythmia using heart rate variability (HRV). A 5 minute and a 10 second warning system are both developed and compared. The participants for this study consist of 788 patients who were enrolled in the ICD arm of the Sudden Cardiac Death–Heart Failure Trial (SCD-HeFT). Two groups of patient rhythms, regular heart rhythms and pre-ventricular-tachyarrhythmic rhythms, are analyzed and different HRV features are extracted. Machine learning algorithms, including random forests (RF) and support vector machines (SVM), are trained on these features to classify the two groups of rhythms in a subset of the data comprising the training set. These algorithms are then used to classify rhythms in a separate test set. This performance is quantified by the area under the curve (AUC) of the ROC curve. Both RF and SVM methods achieve a mean AUC of 0.81 for 5-minute prediction and mean AUC of 0.87–0.88 for 10-second prediction; an AUC over 0.8 typically warrants further clinical investigation. Our work shows that moderate classification accuracy can be achieved to predict ventricular tachyarrhythmia with machine learning algorithms using HRV features from ICD data. These results provide a realistic view of the practical challenges facing implementation of machine learning algorithms to predict ventricular tachyarrhythmia using HRV data, motivating continued research on improved algorithms and additional features with higher predictive power.

This work attempts to reduce the number of false alarms generated by bedside monitors in the intensive care unit (ICU), as a majority of current alarms are false. In this study, we applied methods that can be categorized into three stages: signal processing, feature extraction, and optimized machine learning. At the stage of signal processing, we ensured that the heartbeats were properly annotated. During feature extraction, besides extracting features that are relevant to the arrhythmic alarms, we also extracted a set of signal quality indices (SQIs), which we used to distinguish noise/artifact from normal physiological signals. When applying a machine learning algorithm (Random Forest), we performed feature selection in order to reduce the complexity of the models and improve the efficiency of the algorithm. The dataset used is from Reducing False Arrhythmia Alarms in the ICU: the PhysioNet/Computing in Cardiology Challenge 2015. Using the performance metric “score” from the Challenge, we achieved a score of 83.08 in the real-time category on the hidden test set, which is the highest in all published work.

Background In the SCD-HeFT a significant fraction of the congestive heart failure (CHF) patients ultimately did not die suddenly from arrhythmic causes. CHF patients will benefit from better tools to identify if ICD therapy is needed. Objective To identify predictor variables from baseline SCD-HeFT patients’ RR intervals that correlate to arrhythmic sudden cardiac death (SCD) and mortality and to design an ICD therapy screening test. Methods Ten predictor variables were extracted from pre-randomization Holter data from 475 patients enrolled in the SCD-HeFT ICD arm using novel and traditional heart rate variability methods. All variables were correlated to SCD using Mann Whitney-Wilcoxon test and receiver operating characteristic analysis. ICD therapy screening tests were designed by minimizing the cost of false classifications. Survival analysis, including log-rank test and Cox models, was also performed. Results α1 and α2 from detrended fluctuation analysis, the ratio of low to high frequency power, the number of PVCs per hour and heart rate turbulence slope are all statistically significant for predicting the occurrences of SCD (p<0.001) and survival (log-rank p<0.01). The most powerful multivariate predictor tool using the Cox Proportional Hazards was α2 with a hazard ratio of 0.0465 (95% CI: 0.00528 – 0.409, p<0.01). Conclusion Predictor variables from RR intervals correlate to the occurrences of SCD and distinguish survival among SCD-HeFT ICD patients. We believe SCD prediction models should incorporate Holter based RR interval analysis to refine ICD patient selection especially in removing patients who are unlikely to benefit from ICD therapy.