Healthcare Ethical AI Lab · OHSU

Mohammad
Adibuzzaman

Building the frameworks, governance, and infrastructure to translate AI responsibly into healthcare — and to ask who benefits when it works.

Assistant Professor, Dept. of Medicine · OHSU
Joint Appointment, Biomedical Engineering · OHSU (2026)
Director, OCTRI Informatics
PI, HEAL Lab
View Publications Get in Touch Google Scholar
Mohammad Adibuzzaman
✉ adibuzza@ohsu.edu 🏥 OHSU Profile in LinkedIn ⌥ GitHub S Google Scholar Ω ORCID 𝕏 @adibzaman

About

Unchaining data
to improve health

I am an Assistant Professor in the Department of Medicine at Oregon Health & Science University, with a Joint Appointment in the Department of Biomedical Engineering (from 2026). I lead the Healthcare Ethical AI (HEAL) Lab and serve as Director of Informatics at OCTRI — the Oregon Clinical and Translational Research Institute.

A question keeps me up at night: Can we build AI systems that genuinely improve healthcare, navigate real institutions, survive deployment, and serve patients fairly — all at the same time? My lab exists to answer that question through frameworks, governance infrastructure, and translational science.

My background is in computational science (PhD, Marquette University), with deep experience in causal inference, machine learning, and clinical data infrastructure. Before OHSU, I served as Assistant Director of Data and Computing at the Regenstrief Center for Healthcare Engineering at Purdue, where I built research infrastructure from the ground up and established partnerships with MIT's Laboratory for Computational Physiology.

My current work spans AI validation and assurance, translational gaps in clinical AI adoption, healthcare AI governance and policy, equity and the distribution of AI-enabled value, and building the computational infrastructure for ethical AI at scale.

Research Areas

Six lenses on
responsible AI in healthcare

Our work lives at intersections others avoid — where healthcare systems meet technology policy, where technical rigor meets ethical obligation, where publishing a model is not enough.

⚖
Ethical AI Infrastructure
Extending the Belmont principles for the age of algorithmic medicine. Building executable governance, machine-readable consent, and equity enforcement directly into AI data systems — not as afterthoughts.
Key frameworks: EMBRACE-X, Belmont-as-Infrastructure
🔄
AI Translation & Institutional Readiness
Most AI systems developed in academia never reach patients. We study the translational gaps, institutional barriers, and infrastructure needed for trustworthy deployment in real health systems.
Key frameworks: PATH (Principles of AI Translation in Healthcare)
🏛
Healthcare AI Governance & Policy
Who is accountable when clinical AI fails? We engage with governance at institutional, system, and policy levels — including the EU AI Act, FDA regulation, and evolving regulatory science frameworks.
Focus areas: EU AI Act, FDA, HIPAA, learning health systems
⇌
Causal AI & Clinical Decision-Making
Structural causal models, counterfactual reasoning, and causal AI for treatment policy. Answering what an intervention actually causes — not just what correlates with an outcome.
Methods: SCM, backdoor adjustment, reinforcement learning
🔬
AI Validation & Benchmarking
Developing scalable assurance frameworks for evaluating healthcare AI across institutions. Federated evaluation approaches, validation networks, and lifecycle monitoring — the SAFE-AI initiative.
Initiative: SAFE-AI (multi-site validation network)
◎
Equity & Distribution of AI Value
When AI improves patient outcomes, who benefits? We analyze incentive structures that shape how value from AI-enabled care is distributed — and develop methods to make disparities visible.
Framework: "Who Wins in AI-Enabled Healthcare?"
Research Timeline

How the work has evolved

From computational physiology to ethical AI governance — an interactive map of the research. Click any card to read more.

2026
Belmont-as-Infrastructure: Computationally Extending Ethics into AI Data Ecosystems
AI Ethics
Specific Aims — Active Grant Proposal
Proposes a transformative expansion of Belmont from study ethics to data ecosystem ethics. Redefines Respect for Persons as machine-readable agency within AI pipelines; Beneficence as continuous model validation; Justice as measurable representational equity enforced through dataset composition tracking.
2026
CARE-AI: Community-based Assistive Platform for Resilient Engagement using AI
Infrastructure
Grant Proposal — FHIR-based patient-directed platform
A patient-led digital health platform operationalizing FHIR standards and AI at the point of patient engagement. Integrates EHR data, wearables, and patient-generated health data for chronic disease self-management, with pilot studies in pancreatic cancer (COMPASS) and sickle cell disease (SickleSense).
2026
Who Wins in AI-Enabled Healthcare? Aligning Incentives and Policies
Policy
Perspective — Under Review
Advances a guiding principle: AI-enabled healthcare must be structured so durable gains are anchored in demonstrable improvements in patient health. Examines four foundational domains — financial incentives, digital/built environment stewardship, ethical/legal frameworks, and societal readiness — as prerequisites for alignment. Co-authored with Adrian Zai (UMass Chan) and Elizabeth Johnson (Montana State).
2025–26
Principles for AI Translation in Healthcare — CTSA Progress
Translation
NCATS CTSA UL1 — OCTRI
Six-month interim report documenting the PATH (Principles for AI Translation in Healthcare) framework development. Establishes generalizable principles for responsible AI translation across CTSA consortium institutions, with focus on governance, institutional readiness evaluation, and equity in deployment.
2025
Leveraging Causal AI for Robust Treatment Policy Formulation: Ensemble Causal Tree Approach
Causal AI
Conference Paper · ML for Healthcare
Ensemble causal tree approach with expert insights for critical care decision-making, combining structural causal models with reinforcement learning for treatment policy formulation in ICU settings.
2025
SAFE-AI: Scalable Assurance Framework for AI Evaluation in Healthcare
Infrastructure
Multi-site Validation Initiative — UG3/UH3
Multi-institutional initiative developing federated evaluation frameworks for clinical AI systems. Addresses the gap between how AI models are tested in research settings and what real-world deployment actually requires, including lifecycle monitoring and post-deployment performance tracking.
2024
Identification of Predictive Patient Characteristics for COVID-19 In-Hospital Mortality
Clinical AI
PLOS Digital Health, Vol. 3(4)
Machine learning study identifying patient factors predictive of severe COVID-19 outcomes, with focus on inflammatory mediator roles in acute respiratory distress syndrome. Collaborative work with Bartek Rajwa, Md. Mobasshir Arshed Naved, and team at Purdue/OHSU.
PLOS Digital Health · April 2024
2024
AI-READI Consortium: Responsible AI Data Infrastructure for Health
AI Ethics
Nature Metabolism — Commentary
Consortium commentary on responsible AI data infrastructure, addressing data governance, equity in AI training datasets, and the infrastructure requirements for trustworthy AI in health research contexts.
Nature Metabolism · 2024
2023
Structural Causal Model with Expert-Augmented Knowledge to Estimate Oxygen Therapy Effect on ICU Mortality
Causal AI
Artificial Intelligence in Medicine
Integrates domain expert knowledge into structural causal models for estimating treatment effects in critical care — demonstrating how expert priors can improve causal identification in observational ICU data. Collaboration with Shravan Kethireddy (Cleveland Clinic) and Md Osman Gani.
AI in Medicine · 2023
2022
Identifying and Analyzing Sepsis States: A Retrospective Study in ICUs
Clinical AI
PLOS Digital Health
Machine learning-driven identification of distinct sepsis states from ICU data, with collaboration across Purdue and OHSU teams including Chih-Hao Fang, Vikram Ravindra, and Shankar Subramaniam.
PLOS Digital Health · Nov 2022
2022
Acute Kidney Disease in Patients with Cirrhosis and Acute Kidney Injury
Clinical AI
Journal of Hepatology
Multi-site study demonstrating that acute kidney disease is common and associated with poor outcomes in cirrhosis patients presenting with AKI. Published in one of hepatology's highest-impact journals.
Journal of Hepatology · Vol. 77(1) · 2022
2020
A Causally Formulated Hazard Ratio Estimation through Backdoor Adjustment on Structural Causal Model
Causal AI
Machine Learning for Healthcare Conference
Demonstrates how backdoor adjustment in structural causal models can be used to estimate hazard ratios from observational data — providing a causally valid alternative to traditional survival analysis in settings with confounding.
ML4H 2020 · PMLR
2020
Collaborative Cloud Computing Framework for Health Data with Open Source Technologies
Infrastructure
ACM BCB 2020
Presents a cloud computing framework for large-scale health data analysis built on open-source technologies, enabling HIPAA-compliant collaborative research across institutional boundaries.
ACM Conference on Bioinformatics, Computational Biology, and Health Informatics · 2020
2019
Methods for Quantifying Efficacy–Effectiveness Gap of Randomized Controlled Trials: Examples in ARDS
Causal AI
Critical Care Medicine
Formalizes the efficacy-effectiveness gap problem in RCTs using structural causal frameworks — demonstrating how results from controlled trials fail to translate to heterogeneous real-world ICU populations, with ARDS as the primary case study.
Critical Care Medicine · 2019
2018
Classification of Short Single-Lead ECGs for Atrial Fibrillation Detection using Piecewise Linear Spline and XGBoost
Clinical AI
Physiological Measurement
Novel approach combining piecewise linear spline feature extraction with gradient boosting for AF detection from short single-lead ECG recordings — with applications in wearable cardiac monitoring.
Physiological Measurement · 2018
2017
Big Data in Healthcare — The Promises, Challenges, and Opportunities from a Research Perspective
Infrastructure
AMIA Annual Symposium 2017
Foundational paper examining the infrastructure, governance, and methodological challenges of big data research in healthcare. Used the Cerner Health Facts 69M patient cohort as a case study in the promise and complexity of real-world clinical data.
AMIA Annual Symposium · 2017
2016
Closing the Data Loop: An Integrated Open Access Analysis Platform for the MIMIC Database
Infrastructure
Computing in Cardiology 2016
Describes the collaborative analysis platform developed in partnership with MIT's Laboratory for Computational Physiology for open-access clinical research on the MIMIC ICU database — enabling reproducible computational physiology research.
Computing in Cardiology · 2016
2015
Assessment of Pain Using Facial Pictures Taken with a Smartphone
Clinical AI
IEEE COMPSAC 2015
Mobile health application for objective pain assessment using facial expression analysis — early work in integrating computational sensing into clinical practice.
IEEE COMPSAC · 2015
2014
Mixing Rate of Arterial Blood Pressure Waveform Markov Chain for Hemorrhage Detection
Causal AI
IEEE EMBC 2014 — PhD Research
PhD research applying Markov chain mixing rates (second largest eigenvalue) to detect hemorrhage from arterial blood pressure waveforms — the computational physiology foundation that underpins later causal and time-series work in ICU settings.
IEEE EMBC · 2014
2012
e-ESAS: Mobile-Based Symptom Monitoring for Breast Cancer Patients in Bangladesh
Global Health
CHI 2012 — Best Paper Nomination
Pioneering mHealth application for symptom monitoring in resource-limited settings — evaluating the Edmonton Symptom Assessment Scale (ESAS) through mobile devices for breast cancer patients in rural Bangladesh. Best Paper Nomination at CHI 2012.
ACM CHI · 2012 · Best Paper Nomination
Publications

Selected works

Who Wins in AI-Enabled Healthcare? Aligning Incentives and Policies to Deliver on AI's Promise
2026
Perspective — Under Review | With Adrian Zai (UMass Chan), Elizabeth Johnson (Montana State)
Policy AI Governance
Leveraging Causal AI for Robust Treatment Policy Formulation: An Ensemble Causal Tree Approach with Expert Insights in Critical Care
2025
Conference Paper · ML for Healthcare 2025
Causal AI Critical Care
Timing Matters: A Machine Learning–Driven Comparison of Community and Hospital-Onset Sepsis
2026
BMC Medical Informatics and Decision Making | OHSU — Akram Khan group
Clinical AI → View
Large Language Model Architectures in Health Care: Comparative Study of BERT vs. GPT for Clinical Applications
2024
Journal of Medical Internet Research (JMIR) | With Florian Leiser, Richard Guse
Clinical AI LLM
Identification of Predictive Patient Characteristics for Assessing the Probability of COVID-19 In-Hospital Mortality
2024
PLOS Digital Health, Vol. 3(4) | Rajwa, Naved, Adibuzzaman et al.
Clinical AI → View
Structural Causal Model with Expert Augmented Knowledge to Estimate the Effect of Oxygen Therapy on Mortality in the ICU
2023
Artificial Intelligence in Medicine · Gani, Kethireddy, Adib, Hasan, Griffin, Adibuzzaman
Causal AI Critical Care
Identifying and Analyzing Sepsis States: A Retrospective Study on Patients with Sepsis in ICUs
2022
PLOS Digital Health, Vol. 1(11) · Fang, Ravindra, Akhter, Adibuzzaman et al.
Clinical AI
Acute Kidney Disease is Common and Associated with Poor Outcomes in Patients with Cirrhosis and Acute Kidney Injury
2022
Journal of Hepatology, Vol. 77(1) · Patidar, Naved, Grama, Adibuzzaman et al.
Clinical AI
A Causally Formulated Hazard Ratio Estimation through Backdoor Adjustment on Structural Causal Model
2020
Machine Learning for Healthcare Conference (ML4H) · PMLR · Adib, Griffin, Ahamed, Adibuzzaman
Causal AI → arXiv
Collaborative Cloud Computing Framework for Health Data with Open Source Technologies
2020
ACM BCB 2020
Infrastructure → arXiv
Fast-Track Long Term Continuous Heart Monitoring in a Stroke Clinic: A Feasibility Study
2020
Frontiers in Neurology · 2020
Clinical AI → PMC
Methods for Quantifying Efficacy–Effectiveness Gap of Randomized Controlled Trials: Examples in ARDS
2019
Critical Care Medicine · 2019
Causal AI Critical Care
Classification of Short Single-Lead ECGs for Atrial Fibrillation Detection using Piecewise Linear Spline and XGBoost
2018
Physiological Measurement · Chen, Wang, Jung, Abedi, Zand, Bikak, Adibuzzaman
Clinical AI
Big Data in Healthcare — The Promises, Challenges, and Opportunities from a Research Perspective: A Case Study with a Model Database
2017
AMIA Annual Symposium 2017 · Adibuzzaman, DeLaurentis, Hill, Benneyworth
Infrastructure → PMC
Closing the Data Loop: An Integrated Open Access Analysis Platform for the MIMIC Database
2016
Computing in Cardiology · 2016
Infrastructure
Symptom Levels in Care-Seeking Bangladeshi and Nepalese Adults with Advanced Cancer
2017
Journal of Global Oncology · 2017
Global Health
Findings of e-ESAS: A Mobile Based Symptom Monitoring System for Breast Cancer Patients in Rural Bangladesh
2012
ACM CHI 2012 · Best Paper Nomination
mHealth Best Paper Nomination
View All 50+ Publications on Google Scholar →
Grants & Proposals

Active research funding
& proposals

Translational, infrastructure, and ethics-focused work funded through NIH, NCATS, and institutional sources.

Active
NCATS / NIH · UL1 CTSA
Principles for AI Translation in Healthcare (PATH) — OCTRI CTSA Program
Developing generalizable frameworks for responsible AI translation across the CTSA consortium. Addresses institutional readiness, governance structures, and equity in AI deployment across diverse health systems.
RolePI / Director
InstitutionOCTRI, OHSU
Active
NIH · UG3/UH3
SAFE-AI: Scalable Assurance Framework for AI Evaluation in Healthcare
Multi-site initiative developing federated evaluation frameworks for clinical AI systems. Closes the gap between controlled research evaluation and real-world deployment, with lifecycle monitoring and performance tracking across institutions.
RolePrincipal Investigator
PhaseUG3 → UH3
Submitted
NIH · R01 Proposal
Belmont-as-Infrastructure: Computationally Extending Ethics into AI-Driven Data Ecosystems
High-risk, high-reward proposal to embed Belmont principles as executable constraints within AI data pipelines. Treats participant agency as a structural variable affecting representativeness, equity, and algorithmic performance.
RolePrincipal Investigator
InnovationExecutable ethics infrastructure
Submitted
NIH Proposal
CARE-AI: Community-based Assistive Platform for Resilient Engagement using AI
Patient-led FHIR-based digital health platform for chronic disease self-management, with AI-powered personalization. Pilot studies in pancreatic cancer (COMPASS) and sickle cell disease (SickleSense), with bi-directional data exchange to clinical systems.
RolePrincipal Investigator
PilotsCOMPASS, SickleSense
Prior Funding
Purdue Data Science Initiative
Explainable AI for Clinical Understanding and Translation in Healthcare
$300K data science initiative grant from Purdue University for explainable AI research in clinical contexts, including causal inference methods and interpretable machine learning for ICU decision support.
Amount$300,000
InstitutionPurdue / RCHE
Prior Funding
Cloud Computing Initiative
HIPAA-Compliant Cloud Environment for EHR and Physiological Data Research
$100K grant to develop cloud computing infrastructure for large-scale clinical data analysis — enabling access to Beth Israel Deaconess MIMIC data and 69M-patient Cerner Health Facts cohort.
Amount$100,000
InstitutionPurdue / RCHE
Academic Positions

Career trajectory

Assistant Professor, Department of Medicine
Oregon Health & Science University (OHSU)
Division of Informatics, Clinical Epidemiology & Translational Data Science · Portland, Oregon
2022 – Present
Joint Appointment, Department of Biomedical Engineering
Oregon Health & Science University (OHSU)
Portland, Oregon
2026 – Present
Director, Informatics Program
Oregon Clinical and Translational Research Institute (OCTRI)
NCATS CTSA Program · Oregon Health & Science University
2022 – Present
Assistant Director, Data and Computing
Regenstrief Center for Healthcare Engineering (RCHE)
Purdue University · West Lafayette, Indiana
2020 – 2022
Research Scientist → Senior Research Scientist
Regenstrief Center for Healthcare Engineering (RCHE)
Purdue University · West Lafayette, Indiana
2015 – 2020
Visiting Researcher / FDA Collaboration
U.S. Food and Drug Administration (FDA)
Under Dr. David Strauss · Mathematical models with clinical data
2013 – 2015
PhD, Computational Sciences
Marquette University
Milwaukee, Wisconsin · Advisors: Dr. Stephen Merrill, Dr. Sheikh Iqbal Ahamed
2010 – 2015
Junior Research Fellow, HCI Lab
National University of Singapore (NUS)
With Professor Shengdong Zhao
2009 – 2010
B.S. Computer Science & Engineering
Bangladesh University of Engineering and Technology (BUET)
Dhaka, Bangladesh
2004 – 2009
Service & Leadership

National & institutional
engagement

Governance, peer review, editorial work, and professional societies — connecting the research to the broader field.

National Leadership
Co-Chair, Principles of AI Translation in Healthcare Working Group
NCATS/CTSA Consortium
Leading and co-designing a national working group to dismantle translational barriers to responsible AI adoption across CTSA hubs. Develops shared frameworks for model evaluation and dissemination of best practices.
2025–present
Voting Member, CTSA Biostatistics, Informatics & Data Science (BIDS) Executive Committee
NCATS/CTSA
2025–present
Member, CTSA Rigor & Reproducibility Collaborative Workshop Steering Committee
NCATS/CTSA
2025–present
Voting Member, CTSA Informatics Executive Committee (IEC)
NCATS/CTSA
2021–2024
Member, Vulcan Accelerator, FHIR HL7
HL7 International
2021–present
NSF and NIH Reviewer
Smart and Connected Health Program
2023
Member, AMIA Intensive Care Unit Working Group
American Medical Informatics Association
2015–2018
Institutional Leadership
Co-Chair, OHSU AI Strategy Working Group
OHSU
Co-leading enterprise strategy for responsible AI adoption, evaluation, and governance across clinical and research settings. Recommendations are actively shaping major AI strategy at OHSU.
2025–present
Council Member, Center for AI-Enabled Learning Health Sciences (CAILHS)
OHSU
2025–present
Institutional Review Board (IRB) Member
OHSU
2025–present
Member, Information Security & Privacy Steering Committee
OHSU
2025–present
Member, Informatics Governance Group (IGG)
OHSU
2021–present
Member, Research Data Governance Committee (RDGC)
OHSU
2021–present
Organizer, Health Innovation Collaboration Laboratory
OHSU
2024
Editorial & Peer Review
AMIA Annual Symposium
Reviewer · 2012–present
JAMIA · JCTS · AI in Medicine · NEJM AI
Ad hoc Reviewer · 2015–present
COMPSAC
Reviewer · 2018–present
Professional Memberships
AMIA · 2016–present ACM · 2012–2019 IEEE · 2014–2018
Contact

Let's connect

Interested in collaborating on AI governance, clinical translation, or ethical AI infrastructure? I'm always happy to discuss research ideas, grant opportunities, or speaking invitations.

Email
adibuzza@ohsu.edu
Location
OCTRI · OHSU, Portland, OR
Twitter / X
@adibzaman
LinkedIn
mohammad-adibuzzaman
GitHub
github.com/adibzaman
ORCID
0000-0002-7984-071X