Center for Value-Based Health Care & Sciences

Overview

Our Mission

Through a joint commitment to research and clinical care, we aim to develop and implement health-care delivery models that improve quality of care, patient experience, patient-generated outcomes, and health-related quality of life across care delivery settings. We achieve our mission in three ways:

1. Value-based Care (Real-Life Implementation)
   We continuously gather evidence from learning health systems and apply that evidence to clinical care through practice innovation that can have a real impact on patient-reported outcomes. We accomplish this through collaboration and engagement with our patients, leveraging the Mass General Brigham Patient-Reported Outcomes Program and the Mass General Brigham Virtual Care Program.
2. Valid Information (Economic & Policy Analysis)
   We collect accurate measurements of Value (patient-reported outcomes or system’s quality/costs) and uncover opportunities for innovation with state-of-the art population health science approaches and attention to the impact of economic and policy changes on the quality of data and care.
3. Education & Dissemination
   We support and instruct investigators interested in Population Health Sciences, such as Clinical Epidemiology, Quality Assessment, Causal Inference, Comparative Effectiveness and Cost-effectiveness.

Research Projects

We translate lessons from clinical discovery, performance evaluation, and QI efforts into clinical care innovation through collaboration and engagement with our patients, caregivers, and patient-centered community centers. These systems utilize principles translatable to any clinical specialty to enhance health care delivery by implementing patient-reported outcome measures, identifying variations in outcomes and access to health interventions across demographic factors, and targeting health inequities.

Acute Ischemic Stroke

Anticonvulsants are commonly prescribed to older adults after stroke to prevent seizures, but these drugs may not be necessary or the harms could outweigh the benefits. There is limited rigorous evidence guiding when and how physicians should prescribe this type of treatment. This project will use the real-world experience of millions of Medicare beneficiaries to generate this evidence and inform both policy and future clinical decisions.

Comparative Effectiveness of EEG-guided treatment in acute brain injury

Anti-seizure medications are commonly prescribed to suppress seizures and seizure-like epileptiform electrical activity of the brain detected by electroencephalography (EEG) in patients with acute brain injury. While suppressing epileptiform electrical activity of the brain may minimize its harmful effects, anti-seizure medications have significant adverse effects and there is limited to no evidence on the risks vs. benefits of treatment. This project utilizes observational data from a multi-center EEG dataset of 46,000 patients to determine whether treatment of epileptiform electrical activity with anti-seizure medications improves mortality and functional outcomes, and to provide high quality evidence to guide clinical decisions in patients with acute brain injuries.

Traumatic Brain Injury

This study evaluates the effectiveness and safety of antiseizure medication (ASM) prophylaxis after a traumatic brain injury (TBI).

Luna AI

Timely communication between patients and clinicians is essential for safe and effective care; however, electronic patient portals now generate far more messages than providers can manage, resulting in significant delays and high rates of clinician burnout. This project will test a secure, human-in-the-loop artificial intelligence (AI) assistant, embedded directly within the electronic health record, that classifies urgency, summarizes clinical context, and drafts editable replies while maintaining complete control of communication for clinicians. Through improvements in response time, reductions in after-hours work, high-quality assurance, and accurate triage across diverse patient groups, this research will generate the first rigorous evidence that an EHR-integrated AI tool can enhance patient safety, protect clinician well-being, and provide a scalable model for health systems nationwide.

A Quality Improvement Program to Optimize Triage for Acute Care EEG monitoring

Electroencephalography (EEG) is utilized in the critical care setting for diagnosis and treatment of seizures, prognostication, monitoring depth of sedation, and ischemia monitoring. National studies have demonstrated increased routine and continuous EEG utilization in the critical care setting, and that EEG use may be associated with improved outcomes in patients with seizures or those that are comatose. Yet, there are significant resource limitations hindering access to EEG particularly in smaller community healthcare settings. This project addresses a critical need to develop efficient EEG triage systems that can accurately identify patients for urgent EEG monitoring, increase access to EEG, and optimize patient outcomes relative to resource utilization.

Traumatic Brain Injury

This study evaluates the effectiveness and safety of antiseizure medication (ASM) prophylaxis after a traumatic brain injury (TBI) in adults aged ≥65 years, a population with the highest TBI-related hospitalization and mortality rates. Although guidelines recommend ASM prophylaxis for up to seven days after moderate-to-severe TBI, real-world practice we’ve observed a prolonged use despite limited evidence of benefit and substantial risk of adverse events, including falls, cognitive decline, recurrent TBI, and mortality. This is especially concerning among vulnerable older adults who have multiple comorbidities including Alzheimer’s disease and related dementias. ASM use has increased dramatically among Medicare beneficiaries, yet existing evidence is largely extrapolated from younger populations and clinical trials are impractical in this setting. Leveraging linked Medicare claims, electronic health records, and a TBI registry, this study will use advanced causal inference and machine learning methods to improve TBI severity classification, estimate the comparative effectiveness and harms of differing ASM durations, and assess treatment effect heterogeneity by age, dementia status, and sociodemographic factors. The overarching hypothesis is that, in older adults, prolonged ASM prophylaxis confers greater harm than benefit.

Research Team

Lidia Maria Moura, MD, MPH, PhD
Co-Director

Associate Professor of Neurology, Harvard Medical School
Associate Professor of Associate Professor of Epidemiology, Harvard T.H Chan School of Public Health
Director of Population Health, Neurology Department, Massachusetts General Hospital
Director, Center for Value-Based Population Health Care & Sciences, Massachusetts General Hospital
Director, NeuroValue Lab
Course co-Director, Advanced Epidemiology of Aging, Harvard T.H Chan School of Public Health
Co-director, Center for Healthcare Intelligence, Harvard Business School
Chair, Quality Informatics Subcommittee, American Academy of Neurology
Email: Lidia.moura@mgh.harvard.edu

Sahar Zafar, MD, MBBS
Co-Director

Associate Professor of Neurology, Harvard Medical School
Vice Chair for Quality, MGB Department of Neurology
Assistant Chief Medical Officer, Mass General Brigham
Director, MGH Center for Value-Based Population Health Care & Sciences
Executive Committee Brain Data Science Collaborators Consortium: bdsp.io
sfzafar@mgh.harvard.edu

Maria A. Donahue, MD
Lab Manager

Julianne Brooks, MPH
Data Manager

Aya ElHassan
Clinical Research Coordinator II

Madhav Sankaranarayanan
Clinical Research Coordinator

Aidan McDonald Wojciechowski
Clinical Research Coordinator

Sophia Muqaddas
Clinical Research Coordinator

Mamoon Habib
Data Scientist

Marta Bento Fernandes, PhD
Data Scientist

Alexandra Tautan
Postdoc

Jhansi Anjana Rayapureddy
Data Analyst

Publications

Lidia’s publications

Sahar’s publications