A5060802442- Machine Learning in Healthcare
- COVID-19 Clinical Research Studies
- Artificial Intelligence in Healthcare and Education
- COVID-19 diagnosis using AI
- Acute Ischemic Stroke Management
- Statistical Methods and Inference
- Explainable Artificial Intelligence (XAI)
- Health Systems, Economic Evaluations, Quality of Life
- Sepsis Diagnosis and Treatment
- COVID-19 and healthcare impacts
- COVID-19 epidemiological studies
- Cerebrovascular and Carotid Artery Diseases
- Blood Pressure and Hypertension Studies
- Cardiovascular Function and Risk Factors
- Cardiovascular Health and Risk Factors
- Topic Modeling
- Diabetes Treatment and Management
- Insurance, Mortality, Demography, Risk Management
- demographic modeling and climate adaptation
- Bayesian Modeling and Causal Inference
- Cardiac, Anesthesia and Surgical Outcomes
- Statistical Methods and Bayesian Inference
- Cardiac Imaging and Diagnostics
- Hyperglycemia and glycemic control in critically ill and hospitalized patients
- Time Series Analysis and Forecasting
University of Oxford
2021-2024
Massachusetts Institute of Technology
2018-2023
Harvard University Press
2023
Science Oxford
2021-2022
Society of Thoracic Surgeons
2022
IIT@MIT
2020
Timely identification of COVID-19 patients at high risk mortality can significantly improve patient management and resource allocation within hospitals. This study seeks to develop validate a data-driven personalized calculator for hospitalized patients. De-identified data was obtained 3,927 positive from six independent centers, comprising 33 different Demographic, clinical, laboratory variables were collected hospital admission. The Mortality Risk (CMR) tool developed using the XGBoost...
Accurate, automated extraction of clinical stroke information from unstructured text has several important applications. ICD-9/10 codes can misclassify ischemic events and do not distinguish acuity or location. Expeditious, accurate data could provide considerable improvement in identifying large datasets, triaging critical reports, quality efforts. In this study, we developed report a comprehensive framework studying the performance simple complex stroke-specific Natural Language Processing...
The COVID-19 pandemic has created unprecedented challenges worldwide. Strained healthcare providers make difficult decisions on patient triage, treatment and care management a daily basis. Policy makers have imposed social distancing measures to slow the disease, at steep economic price. We design analytical tools support these combat pandemic. Specifically, we propose comprehensive data-driven approach understand clinical characteristics of COVID-19, predict its mortality, forecast...
Current stroke risk assessment tools presume the impact of factors is linear and cumulative. However, both novel their interplay influencing incidence are difficult to reveal using traditional additive models. The goal this study was improve upon established Revised Framingham Stroke Risk Score design an interactive Non-Linear Score. Leveraging machine learning algorithms, our work aimed at increasing accuracy event prediction uncovering new relationships in interpretable fashion. A...
Abstract Background Timely identification of COVID-19 patients at high risk mortality can significantly improve patient management and resource allocation within hospitals. This study seeks to develop validate a data-driven personalized calculator for hospitalized patients. Methods De-identified data was obtained 3,927 positive from six independent centers, comprising 33 different Demographic, clinical, laboratory variables were collected hospital admission. The Mortality Risk (CMR) tool...
Abstract Tree-based models are increasingly popular due to their ability identify complex relationships that beyond the scope of parametric models. Survival tree methods adapt these allow for analysis censored outcomes, which often appear in medical data. We present a new Optimal Trees algorithm leverages mixed-integer optimization (MIO) and local search techniques generate globally optimized survival demonstrate OST improves on accuracy existing methods, particularly large datasets.
State-of-the-art clustering algorithms use heuristics to partition the feature space and provide little insight into rationale for cluster membership, limiting their interpretability. In healthcare applications, latter poses a barrier adoption of these methods since medical researchers are required detailed explanations decisions in order gain patient trust limit liability. We present new unsupervised learning algorithm that leverages Mixed Integer Optimization techniques generate...
There is a growing amount of evidence that machine learning (ML) algorithms can be used to develop accurate clinical risk scores for wide range medical conditions. However, the degree which such affect decision-making not well understood. Our work attempts address this problem, investigating effect algorithmic predictions on human expert judgment. Leveraging an online survey providers and data from leading U.S. hospital, we ML algorithm compare its performance with experts in task predicting...
Current Society of Thoracic Surgeons (STS) risk models for predicting outcomes mitral valve surgery (MVS) assume a linear and cumulative impact variables. We evaluated postoperative MVS designed mortality morbidity calculators to supplement the STS score.Data from Adult Cardiac Surgery Database was used 2008 2017. The data included 383,550 procedures 89 Machine learning (ML) algorithms were employed train predict patients. Each model's discrimination calibration performance validated using...
The COVID-19 pandemic has created unprecedented challenges worldwide. Strained healthcare providers make difficult decisions on patient triage, treatment and care management a daily basis. Policy makers have imposed social distancing measures to slow the disease, at steep economic price. We design analytical tools support these combat pandemic. Specifically, we propose comprehensive data-driven approach understand clinical characteristics of COVID-19, predict its mortality, forecast...
The COVID-19 pandemic has prompted an international effort to develop and repurpose medications procedures effectively combat the disease. Several groups have focused on potential treatment utility of angiotensin-converting-enzyme inhibitors (ACEIs) angiotensin-receptor blockers (ARBs) for hypertensive patients, with inconclusive evidence thus far. We couple electronic medical record (EMR) registry data 3,643 patients from Spain, Italy, Germany, Ecuador, US a machine learning framework...
Abstract Introduction Inpatient hyperglycemia is an established independent risk factor among several patient cohorts for hospital readmission. This has not been studied after kidney transplantation. Nearly one‐third of patients who have undergone a transplant reportedly experience 30‐day Methods Data on first‐time solitary transplantations were retrieved between September 2015 and December 2018. Information was linked to the electronic health records determine diagnosis diabetes mellitus...
The COVID-19 pandemic has created unprecedented challenges worldwide. Strained healthcare providers make difficult decisions on patient triage, treatment and care management a daily basis. Policy makers have imposed social distancing measures to slow the disease, at steep economic price. We design analytical tools support these combat pandemic. Specifically, we propose comprehensive data-driven approach understand clinical characteristics of COVID-19, predict its mortality, forecast...
Malignant cerebral edema occurs when brain swelling displaces and compresses vital midline structures within the first week of a large middle artery stroke. Early interventions such as hyperosmolar therapy or surgical decompression may reverse secondary injury but must be administered judiciously. To optimize treatment reduce damage, clinicians need strategies to frequently quantitatively assess trajectory using updated, relevant information. However, existing risk assessment tools are...
Tree-based models are increasingly popular due to their ability identify complex relationships that beyond the scope of parametric models. Survival tree methods adapt these allow for analysis censored outcomes, which often appear in medical data. We present a new Optimal Trees algorithm leverages mixed-integer optimization (MIO) and local search techniques generate globally optimized survival demonstrate OST improves on accuracy existing methods, particularly large datasets.
Abstract Due to its prevalence and association with cardiovascular diseases premature death, hypertension is a major public health challenge. Proper prevention management measures are needed effectively reduce the pervasiveness of condition. Current clinical guidelines for provide physicians general suggestions first‐line pharmacologic treatment, but do not consider patient‐specific characteristics. In this study, longitudinal electronic record data utilized develop personalized predictions...
As machine learning algorithms start to get integrated into the decision-making process of companies and organizations, insurance products are being developed protect their owners from liability risk. Algorithmic differs human since it is based on a single model compared multiple heterogeneous decision-makers its performance known priori for given set data. Traditional actuarial tools do not take these properties consideration, primarily focusing distribution historical claims. We propose,...
Missing data is a common problem in real-world settings and particularly relevant healthcare applications where researchers use Electronic Health Records (EHR) results of observational studies to apply analytics methods. This issue becomes even more prominent for longitudinal sets, multiple instances the same individual correspond different observations time. Standard imputation methods do not take into account patient specific information incorporated multivariate panel data. We introduce...