Grand Rounds Archive

Grand Rounds October 16, 2026: Catalyzing Change: How First-in-Patient Trials Can Fix US Clinical Trials (Adrian F. Hernandez, MD, MHS)

Posted on by Sophia Ramirez

Speaker

Adrian F. Hernandez, MD, MHS
Vice Dean and Executive Director
Duke Clinical Research Institute
Duke University School of Medicine

Keywords

Early-Phase Research; Clinical Trials; Competition; Operation TrialBlazer; First-in-Patient Trials

Key Points

  • U.S. trial efficiency is declining. The country’s share of industry-sponsored early-phase research fell 9% from 2017 to 2025, and a 2025 analysis showed that cardiovascular clinical trials in the U.S. had the most sites but enrolled significantly fewer patients than those in other parts of the world. These trends suggest underlying legal, regulatory, and cost-related barriers, highlighting the need for improved clinical trial infrastructure.
  • The Reagan Udall Report and Operation TrialBlazer offer hope for reforms. Academic centers need predictable activation without duplicate reviews; potential safe harbor or pooled insurance; sufficient study volume for dedicated teams and infrastructure; and performance metrics with recognition and rewards.
  • Under the proposed model, serial handoffs and repeated site-by-site setup would be replaced by parallel workstreams with 1 coordinated readiness check. Successful reforms could extend to later-phase trials.

Discussion Themes

Past initiatives have cautioned that multi-component reforms risk failing if all key institutional pieces and stakeholders are not fully aligned simultaneously.

Voluntary academic cooperation may not succeed against centralized international efforts in other countries. One attendee argued for a national policy backed by government authority, central leadership, and financial incentives such as tax credit.

Transparent tracking tools and process reporting could drive friendly institutional competition and continuous quality improvement across academic medical centers.

Grand Rounds October 9, 2026: An End-to-End Agentic AI Platform to Automate Clinical Trials: A Proof-of-Concept and Evaluation Study (Christoph Hornik, MD, PhD, MPH; Moulik Shah)

Posted on by Sophia Ramirez

Speakers

Christoph Hornik, MD, PhD, MPH
Duke Clinical Research Institute

Moulik Shah
Founder and CEO
Maxis AI

Keywords

Agentic AI; Clinical Research; Clinical Trials; Chatbot; AI Training

Key Points

  • Project Loom is a proof-of-concept study aimed at evaluating agentic artificial intelligence (AI) workflows end-to-end. The research team utilized pre-built agents with specific roles: writer, reviewer, builder, recruiter, and reporter. The agents were assembled on one platform to execute the trial. Human oversight took place at each step.
  • The agents were successful in 4 out of 5 operational domains and completed the work in 182 working hours. The presenters outlined several changes that they are making in response to failures in eligibility screening, chatbot hand-offs, and others.
  • The presenters concluded that end-to-end agentic AI execution is feasible and fast. This technology is meant to empower researchers and expedite their work, not to replace them. Training and testing of AI agents and workflows will be critical.

Discussion Themes

Document drafting, case report form generation, and backend data extraction are considered ready for live use today under human review. Direct participant-facing interactions and patient safety monitoring are viewed as the highest-risk areas requiring further refinement and testing.

The platform operates within institutional firewalls using strict role-based access control, bounded scopes of execution, and an observability layer that tracks every computation and handoff.

Resolving the EHR identification and screening issues is technically feasible by profiling unstructured clinical text alongside structured data and extending formal handoff contracts to third-party tools.

The term "Project" signifies that the system is an evolving, iterative initiative rather than a static commercial product.

Grand Rounds October 2, 2026: Self-Directed Versus Clinician-Delivered Cognitive Behavioral Therapy for Chronic Pain: Results From the COPES ExTRA Pragmatic Trial (Alicia Heapy, PhD; Diana Higgins, PhD)

Posted on by Sophia Ramirez

Speakers

Alicia Heapy, PhD
Associate Professor
Yale School of Medicine
Department of Psychiatry
Co-Director VA Pain/Opioid Consortium of Research
VA Connecticut Healthcare System

Diana Higgins, PhD
Clinical Psychologist
MBHI-Neurology
VA Connecticut Healthcare System

Keywords

Chronic Pain; Cognitive Behavioral Therapy; Veterans Administration; Superiority Trial; Barriers

Key Points

  • Cognitive behavioral therapy for chronic pain (CBT-CP) is an underutilized first-line treatment that improves pain and functional outcomes. Barriers to use remain, from patient barriers like the need to travel and mobility issues to system barriers like clinician scarcity and the prevalence of pain.
  • COPES ExTRA was a pragmatic, hybrid type 1 superiority trial that sought to assess the effectiveness of self-directed CBT-CP relative to clinician-delivered CBT-CP among adults with chronic musculoskeletal pain. The research team sought to evaluate the strengths and weaknesses of each approach and inform future clinical use of CBT-CP in the Veterans Health Administration.
  • They found that, while both arms saw clinically significant improvements, patients who received self-directed CBT-CP showed superior improvements on pain interference, pain intensity, sleep, and depressive symptoms durable to 12 months. Convenient, scalable self-directed CBT-CP may be an effective alternative to clinician-directed therapy.
  • Self-directed participants waited fewer days to start treatment and attended more treatment sessions. Factors that contributed to the success of self-directed CBT-CP included fidelity or consistency; the therapeutic elements of daily diary/feedback; walking; and faster time to treatment.

Discussion Themes

Pragmatic methods yielded a sample representative of patients receiving care in the Veterans Health Administration.

Adapting the model outside integrated health systems faces major billing and administrative hurdles, particularly regarding CPT coding and insurance reimbursement for asynchronous coaching.

Standardized effect sizes reached 0.43–0.44 for the primary outcome and 0.3–0.4 across secondary measures, comparing favorably to or exceeding effect sizes typically achieved in non-pharmacological chronic pain trials.

Recruiting and onboarding clinical sites required extensive effort; some clinicians were wary of automated electronic health record alerts generating overwhelming surges in patient referrals or impacting workload metrics.

Grand Rounds September 25, 2026: Research During Active War in Ukraine: Experiences From the PACT-ARROW Study (The Prevention and Treatment Clinical Trials for Antimicrobial Resistance Research to Improve Outcomes of Traumatic Wounds in Ukraine) (Corey Bills, MD, MPH, FACEP)

Posted on by Sophia Ramirez

Speaker

Corey Bills, MD, MPH, FACEP
Department of Emergency Medicine
Associate Professor | Emergency Medicine
Director | Global Emergency Care Initiative
Director | CU-WHO Collaborating Centre for Acute Care
Fellowship Director | Global Emergency Medicine and Public Health
Co-Principal Investigator | ATLAS Research Program
Colorado School of Public Health
Senior Investigator | Center for Global Health

Keywords

Conflict; War; Wounds; Wound Care; Antibiotics; Antimicrobial Resistance; Master Protocol; Ukraine

Key Points

  • Armed conflict has been shown to contribute to the development of antimicrobial resistance, which poses a major threat to human health. In the war in Ukraine, prolonged casualty-care in facilities with high circulation of resistant organisms, weak stewardship, and limited infection control increase the risk of infection.
  • There is little trial evidence for wound care in active conflict. Factors like data completeness, clinical priorities, logistics, layered approvals, an unstable environment, and attacks on healthcare make it difficult to conduct research. Without trials, treatment must rely on extrapolation and caution.
  • A master protocol facilitates research conducted during an armed conflict by pre-positioning ethics approvals, infrastructure, and standard outcomes. The PACT-ARROW master protocol applies a platform model to acute wound care in Ukraine, providing a framework for testing the safety, efficacy, and effectiveness of wound care interventions.
  • The research team began with an observational cohort, collecting longitudinal data from 905 adults with acute traumatic wounds: wound biospecimens, standardized photographs, clinical care data, and clinical outcomes. This showed which data were feasible before any intervention was tested. They found that many isolated bacteria had at least 1 concerning resistance gene to known commonly used antibiotics.

Discussion Themes

Dr. Bills shared lessons learned from implementation in an active war zone. He recommended keeping the data burden small with a minimum necessary dataset and limited follow-up; solving logistics locally; and working with partners for the sake of regulatory speed.

The trial developed from a confluence of U.S. Department of Defense funding opportunities, global acute care research expertise, and established pre-war trial partnerships with Ukrainian clinician-researchers.

The team partnered with Ukraine's Ministry of Health and Ministry of Defense to select 5 civilian hospital sites along active evacuation routes, where the median time from frontline injury to hospital study enrollment is approximately 6 days.

Grand Rounds September 18, 2026: The 4th Trimester Initiative: A Randomized Controlled Trial of Support for Childbearing Medical Trainees (Erika Rangel, MD)

Posted on by terren.green@duke.edu

Speaker

Erika Rangel, MD
Associate Professor of Surgery
Department of Surgery
Harvard Medical School
Medical Director for Clinician Wellbeing
Mass General Brigham

Keywords

Childbearing medical trainees, Burnout, Parental support, Pregnancy support

Key Points

  • Previous studies have shown that surgical residents experience difficulties during medical training and pregnancy and early parenthood. Modified schedules were rare, less than 6 weeks parental leave was common, pumping was difficult, and mentorship regarding balancing family and career was lacking. Many questioned the sustainability of a surgical career affecting an already existing workforce shortage.
  • The aim of this study was to improve the work environment and culture for the vulnerable group of childbearing medical trainees while preserving the rigor and quality of the training program.
  • Study outcomes were chosen in 3 domains that reflect positive return on investment for the workplace. Domains included well-being (burnout), workforce outcomes, and functioning at work and home. Burnout was used as the primary outcome; burnout costs healthcare organizations $4.6 billion a year and has a human cost to the trainee as well.
  • The pragmatic design of the study lessened the burden on participating institutions by including interventions that could be delivered directly to the trainee. The parental support package included a wearable breast pump, a smart bassinet, faculty mentorship, and 24/7 virtual prenatal care via an app.
  • Across 7 institutions, 156 pregnant residents and fellows were enrolled and randomized to either usual support or the parental support package group. All participants were given a fitbit to track sleep and all were asked to track medical errors. The research question was deliberately phrased as ‘what happens when this parental support package was offered at an institution.’
  • The parental support package was used variably by participants. Institutional environments also varied. These variable factors were analyzed separately. The support package appeared to buffer the rise in burnout for the intervention group. Interpersonal disengagement was significantly improved and improved sleep decreased medical error.

Discussion Themes

Institutions understand that burnout is an issue for childbearing medical trainees, and response to the study was enthusiastic. Originally only 3 institutions were targeted, but 7 were ultimately included in the study.

Medical trainees were also enthusiastic to participate with a 92% retention rate. Word of mouth and social media were instrumental in recruitment.

 

Read more about the The 4th Trimester Initiative.

 

Grand Rounds September 11, 2026: Embedded Trials in a Learning Health System: A Hospital Policy of Tranexamic Acid to Reduce Transfusion in Major Noncardiac Surgery (Ryan Zarychanski, MD, MSc, FRCPC, FCAHS)

Posted on by Sophia Ramirez

Speaker

Ryan Zarychanski, MD, MSc, FRCPC, FCAHS
Professor of Medicine (Hematology and Critical Care), Pharmacy, and Community Health Sciences
Tier 1 Canada Research Chair in Clinical Trials
Max Rady College of Medicine | University of Manitoba
Hematologist and Senior Scientist
CancerCare Manitoba and Research Institute
Scientific Director
George and Fay Yee Centre for Healthcare Innovation & Aurora Clinical Research

Keywords

Infusion; Surgery; Tranexamic Acid; Noncardiac Surgery

Key Points

  • Tranexamic acid (TXA) is an inexpensive, widely available resource with an established safety profile. Evidence shows that the use of TXA during cardiac surgeries reduces the rate of blood transfusion – an expensive procedure involving a limited resource.
  • The TRACTION trial sought to assess the use of TXA in non-cardiac surgery through a multicenter randomized controlled trial with 2 co-primary outcomes: the proportion of patients who received a blood transfusion and the rate of venous thromboembolism (VTE) within 90 days of the surgery.
  • Patients undergoing surgeries with over 5% risk of transfusion were eligible. Ultimately, 8,400 patients were enrolled across 10 sites in Canada. The study team utilized a cluster-randomized embedded design; each site was randomized to either placebo or TXA every 4 weeks.
  • TXA reduced the rate of transfusion during non-cardiac surgeries, with 7.4% of patients in the intervention arm versus 9.8% in the control arm. Thrombotic event rates were equivalent between the 2 groups. Novel methods like the cluster-randomized embedded design, risk-adapted consent, and electronic data capture increased efficiency and substantially reduced the cost of the trial.

Discussion Themes

Years of preparatory studies, statistical modeling, and navigating complex provincial and governmental data-sharing agreements were required before trial launch.

A decision not to include pregnant women in the trial was driven by hospital leadership's risk aversion and liability concerns rather than regulatory or IRB refusals.

A strict 1% non-inferiority margin for thrombotic safety was established through direct consultation with patient groups and clinicians.

Grand Rounds August 28, 2026: Evaluating Clinical Agentic AI (Haider Warraich, MD)

Posted on by Sophia Ramirez

Speaker

Haider Warraich, MD
Program Manager
ARPA-H (Advanced Research Projects Agency for Health)

Keywords

AI; Outpatient Care; Agentic AI; Heart Failure; Clinical Care

Key Points

  • The current state of outpatient care is sporadic and reactive. Low-value care leads to excess costs and harms, and clinical resources are wasted on administrative tasks. Clinical agentic AI could provide patients with round-the-clock access to specialty care and unlock the full potential of the Electronic Health Record and wearables.
  • ADVOCATE is envisioned as a patient-facing agentic AI system that can provide 24/7 clinical care as an extension of the care team. It would include administrative and wellness functions, like care navigation and diet and exercise recommendations; advanced decision support in regard to treatment and diagnosis; and high-risk functions, like prescription management. The ARPA-H team hopes it will transform advanced cardiovascular disease management.
  • Their current focuses are the development of an agent for patients with heart failure; the development of an independent supervisory agent to ensure continued agent accuracy and patient safety, particularly after clinical deployment; and recruitment of health systems for the design, development, and deployment of the technologies.
  • The clinical evaluation strategy for ADVOCATE will involve 3 phases: Human Factors Evaluation and simulated virtual patient testing; shadow-mode evaluation in clinically deployed settings; and a stepped-wedge multicenter cardiovascular outcomes trial.

Discussion Themes

While underlying foundation AI models possess broad medical knowledge, their autonomous actions must be strictly constrained to their labeled, intended use to maintain safety.

Heart failure was selected – as opposed to upstream prevention – in part because it is characterized by a highly algorithmic care path, a high event rate that generates clinical study data quickly, and strong financial incentives.

While the program initially targeted non-inferiority compared to a cardiologist, the metric was updated to require a superiority analysis against usual care.

Grand Rounds August 21, 2026: Off-Label Intelligence: Books, Language Models, and AI as a Medical Device (Eric Karl Oermann, MD)

Posted on by Sophia Ramirez

Speaker

Eric Karl Oermann, MD
Associate Professor of Neurosurgery, Radiology, and Data Science
Director, OLAB – Health AI Research Lab of NYU Langone Health
Associate Program Director, Neurosurgery Residency program
Investigator, Neuroscience Institute
NYU Langone Medical Center
New York University

Keywords

Artificial Intelligence; Large Language Models; AI in Medicine; Clinical Benefit

Key Points

  • While clinicians often treat large language models (LLMs) like textbooks, books are peer-reviewed, static, and universal. In contrast, LLMs are stochastic and probabilistic, meaning different users will see different responses. AI should be treated as a clinical tool.
  • There is a gap between the amount of work and publishing being done in this field and its impact on clinical practice. Since 2020, roughly 5,000 papers have been published on the topic of LLMs in medicine. Of those, only 21 generated legitimate medical evidence.
  • Process benefit does not imply clinical benefit. The success of medical AI lies in how the tool is paired with a clinical workflow. Dr. Oermann demonstrated this with examples from 4 real-world clinical trials. Future trials should test both models and workflows; how clinicians utilize a model is as important as the model they use.

Discussion Themes

Traditional randomized controlled trials can take years to plan and execute, during which time an AI model may go through several generations of updates. To match the speed of technology, Dr. Oermann advocated for adaptive trial designs and rapid learning health systems where institutions can run silent deployments, AB test AI tools in-house, and observe how they affect local care delivery.

Health systems are ultimately responsible for patient outcomes, placing the accountability for monitoring and validating AI performance over time onto the individual institutions rather than relying solely on third-party vendors.

Grand Rounds August 14, 2026: Designing for Impact: Introducing Translational Design Studies (Christopher J. Lindsell, PhD; Lloyd Michener, MD; Justin B. Moore, PhD, MS; Boris Volkov, PhD; Reuben Adatorwovor, PhD)

Posted on by Sophia Ramirez

Speakers

Christopher J. Lindsell, PhD
Vice Dean for Data Science and AI
Duke University School of Medicine

Lloyd Michener, MD
Professor & Chair Emeritus
Department of Family Medicine & Community Health
Duke University School of Medicine
Adjunct Professor, Public Health Leadership
UNC Gillings School of Global Public Health

Justin B. Moore, PhD, MS
Professor and Chair (interim)
Department of Implementation Science
Director for Dissemination, Implementation, and Continuous Quality Improvement
Clinical and Translational Science Institute
Wake Forest School of Medicine

Boris Volkov, PhD
Evaluation Director and Assistant Professor
University of Minnesota Clinical and Translational Science Institute & Institute for Health Informatics

Reuben Adatorwovor, PhD
Assistant Professor of Biostatistics
Institute of Informatics, Data Science and Biostatistics
Department of Emergency Medicine
Washington University in St. Louis

Keywords

Study Design; Translation; Translational Science; Failure; Engagement; Biostatistics; Informatics; Implementation; Evaluation

Key Points

  • Translational failure often begins before the trial. The labels we use – eg, recruitment failure, stakeholder resistance, and poor generalizability – indicate where failure becomes visible, not where it originates. However, trial design remains largely unevidenced. If a decision determines whether downstream work can succeed, the presenters argued, the basis for that decision belongs in the scientific record.
  • A translational design study is a systematic inquiry conducted before, or alongside, planned translational work. It seeks to reduce a named design uncertainty. The category is defined by intent, not by method; qualitative, quantitative, and mixed methods all qualify. The presenters introduced 4 lenses on design uncertainty: community engagement; evaluation; biostatistics and informatics; and implementation science. These are to be used in conjunction, not held as separate disciplines.
  • Attendees were encouraged to submit their design-phase work for publication, especially if their findings forced them to change course. If it includes a clear design question, an explicit assumption, reproducible methods, and results that others can apply, adapt, and build on, it qualifies.

Discussion Themes

While industry focuses on scalability, the presenters argued that the academic community should lead in design innovation.

ACTIV-6 is a real-world example. In that large decentralized trial, evidence-based protocol changes served as a model for transferable design learning.

Grand Rounds August 7, 2026: Adaptive AI for CV Adjudication: Adjudication Across Different Definitions in the ODYSSEY OUTCOMES and EUCLID Trials (Sreek Vemulapalli, MD; Ricardo Henao, PhD)

Posted on by Sophia Ramirez

Speakers

Sreek Vemulapalli, MD
Associate Professor of Medicine with Tenure
Director, CV Imaging, DUHS
Duke Clinical Research Institute
Duke-Margolis Institute for Health Policy

Ricardo Henao, PhD
Associate Professor
Department of Biostatistics & Bioinformatics
Department of Electrical and Computer Engineering
Department of Surgery
Associate Director of Clinical Trials AI, Duke Clinical Research Institute (DCRI)
Duke University

Keywords

Adjudication; AI; Clinical Event; Clinical Event Committees; Scalability

Key Points

  • The promise of clinical endpoint or event adjudication is standardization and the improvement of endpoint credibility. However, a human-centered review model utilizing Clinical Endpoint Committees (CECs) represents a major scalability constraint due to cost pressure, operational drag, and between-group variability.
  • AI could reduce the burden on CECs by separating confident events from cases requiring human review. Previous trials testing this concept have derived separate models for each event type and have no demonstrated ability to adapt to new definitions or endpoints across trials. The research team sought to develop and train an AI algorithm to perform clinical endpoint adjudication that can extend to new endpoint and trials. They then compared that algorithm’s performance to Chat GPT 4.0 and hybrid (AI + human) adjudication.
  • The research team found that a single AI algorithm trained on a single trial can, after brief adaptation, identify the same events with a new definition and identify new events. This was validated in the EUCLID trial, when the algorithm alone or as a hybrid strategy reproduced the EUCLID primary endpoint and effect size. While further prospective studies are needed, the potential impact of AI-enabled automated adjudication of outcomes on developing evidence is substantial.

Discussion Themes

The role of clinical adjudicators could shift toward auditing the AI, similar to the evolution of EKG reading, where experts focus on the 10% of cases where the algorithm is most likely to be wrong or uncertain.

The primary barriers to implementing AI systems that comb through electronic health records for trial data are governance, data security, and access rights, rather than a lack of technological capability.

While demonstrated here in cardiovascular trials, this methodology is theoretically applicable to other complex domains. To gain regulatory acceptance, the presenters suggested that prospective validation across multiple clinical research groups and a wide variety of endpoints will be necessary.