A5079293522- Machine Learning in Healthcare
- Sepsis Diagnosis and Treatment
- Bayesian Methods and Mixture Models
- Statistical Methods and Inference
- Explainable Artificial Intelligence (XAI)
- Hemodynamic Monitoring and Therapy
- COVID-19 epidemiological studies
- Clinical Reasoning and Diagnostic Skills
- Statistical Methods and Bayesian Inference
- Artificial Intelligence in Healthcare and Education
- Respiratory Support and Mechanisms
- Data-Driven Disease Surveillance
- Metabolomics and Mass Spectrometry Studies
- Data Stream Mining Techniques
- Time Series Analysis and Forecasting
- COVID-19 Digital Contact Tracing
- Healthcare Technology and Patient Monitoring
- Reinforcement Learning in Robotics
- Frailty in Older Adults
- Gaussian Processes and Bayesian Inference
- American Sports and Literature
- Advanced Clustering Algorithms Research
- Generative Adversarial Networks and Image Synthesis
- Ethics in Clinical Research
- Health, Environment, Cognitive Aging
Apple (United States)
2021-2022
Harvard University
2019-2021
Apple (United Kingdom)
2020-2021
Duke University
2015-2020
Harvard University Press
2019-2020
Dartmouth Hospital
2012-2013
Dartmouth College
2012-2013
Risk sharing arrangements between hospitals and payers together with penalties imposed by the Centers for Medicare Medicaid (CMS) are driving an interest in decreasing early readmissions. There a number of published risk models predicting 30day readmissions particular patient populations, however they often exhibit poor predictive performance would be unsuitable use clinical setting. In this work we describe compare several models, some which have never been applied to task outperform...
Machine learning technologies are increasingly developed for use in healthcare. While research communities have focused on creating state-of-the-art models, there has been less focus real world implementation and the associated challenges to fairness, transparency, accountability that come from actual, situated use. Serious questions remain underexamined regarding how ethically build interpret explain model output, recognize account biases, minimize disruptions professional expertise work...
Successful integrations of machine learning into routine clinical care are exceedingly rare, and barriers to its adoption poorly characterized in the literature.This study aims report a quality improvement effort integrate deep sepsis detection management platform, Sepsis Watch, care.In 2016, multidisciplinary team consisting statisticians, data scientists, engineers, clinicians was assembled by leadership an academic health system radically improve treatment sepsis. This follows framework...
<h3>Importance</h3> The ability to accurately predict in-hospital mortality for patients at the time of admission could improve clinical and operational decision-making outcomes. Few machine learning models that have been developed death are both broadly applicable all adult across a health system readily implementable. Similarly, few implemented, none evaluated prospectively externally validated. <h3>Objectives</h3> To validate model predicts hospital design using commonly available...
Determine if deep learning detects sepsis earlier and more accurately than other models. To evaluate model performance using implementation-oriented metrics that simulate clinical practice.We trained internally temporally validated a (multi-output Gaussian process recurrent neural network [MGP-RNN]) to detect encounters from adult hospitalized patients at large tertiary academic center. Sepsis was defined as the presence of 2 or systemic inflammatory response syndrome (SIRS) criteria, blood...
We present a scalable end-to-end classifier that uses streaming physiological and medication data to accurately predict the onset of sepsis, life-threatening complication from infections has high mortality morbidity. Our proposed framework models multivariate trajectories continuous-valued time series using multitask Gaussian processes, seamlessly accounting for uncertainty, frequent missingness, irregular sampling rates typically associated with real clinical data. The process is directly...
Accurate measurement of daily infection incidence is crucial to epidemic response. However, delays in symptom onset, testing, and reporting obscure the dynamics transmission, necessitating methods remove effects stochastic from observed data. Existing estimators can be sensitive model misspecification censored observations; many analysts have instead used that exhibit strong bias. We develop an estimator with a regularization scheme cope delays, which we term robust deconvolution estimator....
Restricting in-person interactions is an important technique for limiting the spread of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). Although early research found strong associations between cell phone mobility and infection during initial outbreaks in United States, it unclear whether this relationship persists across locations time. We propose interpretable statistical model to identify spatiotemporal variation association rates. Using 1 year US county-level data, we that...
Sepsis is a poorly understood and potentially life-threatening complication that can occur as result of infection. Early detection treatment improves patient outcomes, such it poses an important challenge in medicine. In this work, we develop flexible classifier leverages streaming lab results, vitals, medications to predict sepsis before occurs. We model clinical time series with multi-output Gaussian processes, maintaining uncertainty about the physiological state while also imputing...
We present two elegant solutions for modeling continuous-time dynamics, in a novel model-based reinforcement learning (RL) framework semi-Markov decision processes (SMDPs), using neural ordinary differential equations (ODEs). Our models accurately characterize dynamics and enable us to develop high-performing policies small amount of data. also approach optimizing time schedules reduce interaction rates with the environment while maintaining near-optimal performance, which is not possible...
Machine learning technologies are increasingly developed for use in healthcare. While research communities have focused on creating state-of-the-art models, there has been less focus real world implementation and the associated challenges to accuracy, fairness, accountability, transparency that come from actual, situated use. Serious questions remain under examined regarding how ethically build interpret explain model output, recognize account biases, minimize disruptions professional...
OBJECTIVE: Specific factors affecting generalizability of clinical prediction models are poorly understood. Our main objective was to investigate how measurement indicator variables affect external validity in for predicting onset vasopressor therapy. DESIGN: We fit logistic regressions on retrospective cohorts predict using two classes variables: seemingly (vital signs and laboratory measurements) more subjective denoting recency measurements. SETTING: Three from tertiary-care academic...
Off-policy evaluation in reinforcement learning offers the chance of using observational data to improve future outcomes domains such as healthcare and education, but safe deployment high stakes settings requires ways assessing its validity. Traditional measures confidence intervals may be insufficient due noise, limited confounding. In this paper we develop a method that could serve hybrid human-AI system, enable human experts analyze validity policy estimates. This is accomplished by...
Abstract Accurate measurement of daily infection incidence is crucial to epidemic response. However, delays in symptom onset, testing, and reporting obscure the dynamics transmission, necessitating methods remove effects stochastic from observed data. Existing estimators can be sensitive model misspecification censored observations; many analysts have instead used that exhibit strong bias or do not account for delays. We develop an estimator with a regularization scheme cope these sources...