Grand Rounds Archive

Grand Rounds September 19, 2025: Hurdles for the Delivery of Clinical Trials: Insights From the REMAP-CAP Trial in Europe (Denise van Hout, MD, PhD)

Posted on by terren.green@duke.edu

Speaker

Denise van Hout, MD, PhD
Postdoctoral Researcher
Julius Center for Health Sciences and Primary Care
University Medical Center Utrecht, the Netherlands

Keywords

Adaptive platform trial, Regulatory efficiency, REMAP-CAP, Study design, Study startup.

Key Points

  • Randomised Embedded Multifactorial Adaptive Platform trial for Community-Acquired Pneumonia (REMAP-CAP) began in 2016 as a data driven analysis of ethical, administrative, and logistical and ethical (EARL) delays in clinical trials studying respiratory infections. The goal was embedded trials that are flexible, efficient, and agile to provide clinicians with high-quality evidence to make the best treatment decisions.
  • REMAP-CAP is a global multifactorial adaptive platform trial with a master protocol that can investigate multiple interventions in different treatment domains for a single disease.
  • Over 8000 patients in REMAP-CAP were randomized to 44 interventions in 16 different treatment domains between January 2019 and June 2023. Patients could be randomized to more than 1 domain resulting in 15656 randomizations. Enrollment increased during the COVID-19 pandemic.
  • Dr. van Hout believes that to improve clinical trials, we should treat the challenges as a scientific problem and solve them with the same rigor.
  • Regulatory requirements and informed consent regulations differed among sites causing confusion for the researchers about what documents should be submitted with the contract and protocol in each country. Drug labeling requirements in some countries also slowed protocol approval. EARL processes also slowed trial initiation and patient enrollment.
  • It was clear that overall enrollment in the UK outpaced the other 257 sites worldwide. The UK had a shorter period of time to a fully signed study contract and protocol approval compared with sites in other countries (5 days in the UK compared with 183 days in non-UK countries). This quicker time to signed contract was accomplished by either accepting the contract as-is or rejecting the contract – without negotiating small details. The UK was also 3 months faster than non-UK countries at enrolling the first patient after study approval (1 month vs 4 months, respectively) leading to more enrollment and more research questions answered.
  • In January of 2022 the EU centralized regulatory submission to a single portal (CTIS) to ease and speed the process of starting a new trial.

Discussion Themes

Adaptive platform trials were uncommon before the COVID-19 pandemic, but their value became clear during the pandemic. After the pandemic, REMAP-CAP focuses on different treatment domains for pneumonia. Maintaining the infrastructure for an adaptive platform trial is difficult if there is not a clear need such as there was during the COVID-19 pandemic.

Centralizing approval for trials under one government body could speed the approval process for studies. During times of high need, prioritizing one or two good trials over a lot of smaller trials can also help speed the process.

 

Learn more about REMAP-CAP at https://www.remapcap.eu/

Grand Rounds September 5, 2025: The Non-Learning Health System (Robert Califf, MD)

Posted on by Sophia Ramirez

Speaker

Robert Califf, MD
Instructor in Medicine
Duke University Medical Center
Former Commissioner of Food and Drugs

Keywords

Healthcare; Learning Health System; Evidence-Based Practices; Health Outcomes

Q&A

The following reflects key takeaways from a fireside chat with Dr. Robert Califf, in which he shared his perspective on the “non-learning” health system. For a comprehensive account of Dr. Califf’s insights, watch the recording.

What do you mean by the “non-learning” health system?

25 years ago, certain visionaries looked at the advancement of computing, electronic health records, and other digital data and noted that data could and should be used to improve healthcare delivery and, in turn, health outcomes.

But increasingly, the healthcare system in the United States is “learning” based on institutional financial outcomes as opposed to patient outcomes. That’s not to say it’s a zero-sum game—but efforts are being directed towards expensive technologies that offer marginal benefit (but deliver good economic returns) as opposed to primary care, prevention, and interventions that address basic risk factors.

How can we reshape those incentives?

If the goal is to optimize the longevity, well-being, and functionality of the American population, incentives within the healthcare system should be aligned with health outcomes.

Why has it been so difficult to integrate evidence-based practices into healthcare settings? And how can we begin to change that?

If we align health care systems’ incentives with health outcomes, they will figure out how to operationalize these practices. But if we assume the incentives will not be realigned in the near future, we will need to eke out areas of alignment with decision-makers, incremental improvements that are not so disruptive that they get squashed. And finally, we need to develop disruptive external systems to challenge health systems.

What tasks should this community focus on?

Keep working on pragmatic trials; show that interventions have practical applications. Keep developing the skills to communicate about your work to the public. And be prepared to put our system back together when it breaks.

Discussion Themes

Other discussion themes included the critical role of randomized trials and the potential role of AI in answering scientific questions; what the research community can learn from other industries; and anticipated changes to the healthcare system and research landscape.

Grand Rounds August 22, 2025: Avoiding the Fumble: Building on a Decade of Lessons from Pragmatic Clinical Trials (Emily O’Brien, PhD, FAHA)

Posted on by Sophia Ramirez

Speaker

Emily O’Brien, PhD, FAHA
Associate Professor
Duke Clinical Research Institute
Duke University School of Medicine
Department of Population Health Sciences

Keywords

Pragmatic Trials; Best Practices; PCORnet; Evidence-Based Practices

Key Points

  • Historically, the healthcare industry has been limited by an insufficient body of evidence driving everyday clinical decision-making. Roughly a decade ago, pragmatic clinical trials (PCTs) began to gain traction as a promising solution.
  • There are several advantages of PCTs. They can be embedded within healthcare systems without disrupting the clinical workflow; answer questions of major public health importance; streamline procedures and infrastructure by making use of existing data; and include diverse, representative study populations for highly generalizable results.
  • But a recent analysis of clinical research site challenges noted that protocol complexity, site workload, and patient burden have increased since 2015. Though the analysis was not specific to pragmatic trials, a fundamental shift in how researchers think about study design is required across the clinical trials space.
  • Additionally, evidence-based practices – even those that have been stress-tested in PCTs – are not always adopted by health systems. Trial success does not necessarily coincide with system priorities; different audiences, i.e. systems and funders, require different kinds of evidence; and 5- to 10-year studies are misaligned with systems’ 2- to 3-year decision horizons.
  • The NIH Pragmatic Trials Collaboratory philosophy holds that fumbles are part of the game; we can’t improve if we only share wins, and transparency and teamwork has helped this community iterate and improve. Accordingly, the PCORnet team developed “The Playbook,” inspired by the NIH Collaboratory’s Living Textbook, as a tool for sharing and refining the best approach to national-scale research.
  • The Playbook contains practical “drills” for avoiding common fumbles in recruitment, workflow, and outcome capture, and was created using a user-centered design process. They engaged PCORnet groups, partners, and members of the Playbook’s intended audience to inform and guide the content.
  • Modules 1 – 5 of the Playbook, launching this year, will provide an introduction to the network. They include sections on getting started with PCORnet, utilizing the network’s resources, dissemination and implementation expectations for PCORnet studies, and case studies.
  • In the long-term, the PCORnet team plans to actively review, maintain, and expand the Playbook. Additional modules are in process and targeted for release in 2026.

Discussion Themes

The success of the Playbook may depend on the willingness of investigators to share both their “best plays” and their mistakes. Dr. O’Brien was optimistic that research teams will buy into this philosophy and acknowledge it as an important piece of the evidence generation process.

The case studies that the team selected serve to illustrate A) that PCORnet trials are unique, innovative, and approaching challenges in a thoughtful, inspiring way and B) the many ways to engage with the network.

Grand Rounds August 15, 2025: Dexmedetomidine- or Clonidine-Based Sedation Compared With Propofol in Critically Ill Patients: The A2B Randomized Clinical Trial – Addressing a Complex Clinical Question With Novel Integrated Methodological Approaches (Tim Walsh, MD, FFICM; Chris Weir, PhD; Richard Parker, MsC)

Posted on by terren.green@duke.edu

Speakers

Professor Tim Walsh, MD, FFICM
Chair of Critical Care, Usher Institute, University of Edinburgh

Professor Chris Weir, PhD
Chair of medical statistics, Edinburgh Clinical Trials Unit, Usher Institute, University of Edinburgh

Richard Parker MSc, Research Fellow
Edinburgh Clinical Trials Unit, Usher Institute, University of Edinburgh

Keywords

Dexmedetomidine; Clonidine; Propofol; Sedation; Critical care; Mechanical ventilation.

Key Points

  • Critically ill, mechanically ventilated patients often require sedation for comfort. Avoiding deep sedation, agitation, and delirium is a priority in these patients. To accomplish this, current usual care for sedation internationally is the GABAergic agent propofol along with an opioid or other analgesia, but the alpha-2 agonists Dexmedetomidine or Clonidine are increasingly being used in the ICU.
  • Dexmedetomidine is licensed for ICU sedation and is believed to achieve a lighter level of sedation compared with propofol, but has a higher cost and, some studies have shown, a potentially higher rate of mortality in younger ICU patients. Clonidine, despite widespread use in the UK, is not licensed for ICU sedation and has a limited evidence base but is lower in cost.
  • This 3-arm, A2B pragmatic open-label effectiveness trial compared dexmedetomidine plus an opioid analgesic or clonidine plus an opioid analgesic with propofol plus an opioid analgesic in 1404 critically ill, mechanically ventilated patients (457 dexmedetomidine, 476 clonidine, 471 standard care [propofol]) in 41 ICUs across the UK from December 2018-December 2024.
  • The trial’s primary outcome was time to successful extubation post-randomization, and secondary outcomes included sedation quality, time to optimum sedation, delirium, ICU mortality, ICU length of stay, and drug related adverse events. Hypothesis for clonidine and dexmedetomidine were determined through 2 separate 3-stage pathways.
  • There were no significant differences found in either dexmedetomidine or clonidine compared with propofol (sHR 1.09, 95% CR 0.96 to 1.25, P =0.20; sHR 1.05, 95% CR 0.95 to 1.17, P=0.34; respectively) for the cumulative incidence of time to extubation. There were also no statistically significant differences in the percent of patients successfully extubated within 7 days, patient survival time to ICU mortality, outcome by age, SOFA score, or PRE-DELIRIC risk score. Unlike previous studies, no interaction was found between age and mortality in this study.
  • Rates of agitation were higher in the dexmedetomidine and clonidine groups, but rates of pain behavior, unnecessary deep sedation, and overall optimum sedation did not reach statistically significant differences compared with the propofol group. There were higher rates of severe bradycardia and cardiac arrythmia in the dexmedetomidine arm.
  • Interviews with clinical and research staff showed that multiple factors influence sedation practice within the trial context including context-specific sedation ‘culture’; clinician preference and equipoise; staff capacity, training, and capability; safety concerns; and engrained practices such as more deeply sedating patients overnight.

Discussion Themes

The latest European Society of Intensive Care Medicine guidelines, which don’t include the data from this trial, leave the decision to use dexmedetomidine or propofol up to clinician judgement. Trial data don’t support using clonidine widely, but if used, ICU staff should be confident with and have support for any adverse effects.

While delirium is associated with a range of outcomes like mortality and cost of care, no trial has shown that altering delirium translates into a benefit on those outcomes. A primary outcome of mortality would have needed a large number of participants, making the trial much more difficult to accomplish. For this reason, time to successful extubation was chosen as the primary outcome for this trial.

This A2B trial was not a blinded trial. Blinded drug trials have their place but are not really feasible pragmatic real-world trials. In addition, the cost of blinding this A2B trial would have been very high. Also, patients had strong views that they wanted the clinicians to know what drug they were delivering so they would be prepared to handle any adverse events.

Read more about the protocol and results of this A2B randomized clinical trial of sedation in clinically ill patients.

Grand Rounds August 8, 2025: Youth Nicotine Vaping Cessation: RCT of Varenicline Added to Remote Young Adult Lay Counselor Delivered Behavioral Cessation Support Vs. Texting Support (A. Eden Evins, MD, MPH)

Posted on by Sophia Ramirez

Speaker

A. Eden Evins, MD, MPH
Cox Family Professor of Psychiatry, Harvard Medical School
Founding Director, MGH Center for Addiction Medicine
Director for Faculty Development, Mass General Hospital Department of Psychiatry

Keywords

Vaping; Nicotine; Cessation; Behavioral Support; Texting

Key Points

  • Vaping, though technically a less harmful alternative to cigarette smoking, is the primary route to nicotine addiction in youth. Initially promoted to help with smoking cessation, vaping devices are increasingly marketed towards young people and have transformed teen nicotine and cannabis use.
  • Few treatments for vaping cessation have been tested. The research team hypothesized that the pharmacotherapy varenicline could help young adults abstain from nicotine vaping.
  • They conducted a randomized clinical trial with 2 aims: evaluate the efficacy of varenicline in addition to behavioral support for cessation of vaped nicotine and assess the safety and tolerability of varenicline. The target population was adolescents attempting vaping cessation.
  • According to the original design, participants would be randomized to 1 of 2 arms: varenicline + behavioral therapy and placebo + behavioral therapy. The possibility arose that the behavioral intervention would be particularly effective, they wouldn’t be able to differentiate between the arms, and the trial would fail. So they added a third arm: a referral to a widely-available messaging app supporting youth vaping cessation.
  • The research team made a few additional modifications to cut costs and enhance adherence. These included fully remote intervention and assessment; non-clinical personnel delivering the behavioral interventions; and video documentation of adherence, with compensation of $1 per video.
  • Over half of participants offered varenicline + counseling quit vaping and were abstinent for the last 4 weeks of treatment. The 4-week abstinence rate was 14% in the placebo + counseling group and 6% in the group that received texting support.
  • The study had a 97% completion rate using remote methods for intervention delivery and data collection. The research team received video evidence for 52% of varenicline doses and 42% of placebo doses. Behavioral counseling attendance was higher in the varenicline group (84%) than in the placebo group (66%).
  • Nausea, vivid dreams, and insomnia were more common in the group that received varenicline.

Discussion Themes

There’s interest in testing the effectiveness of varenicline without the addition of behavioral support. It’s an important question, Dr. Evins noted, given that behavioral support can be hard to come by.

Varenicline has been associated with exacerbation of psychiatric symptoms, including suicidal ideation. However, this has not been replicated in large clinical trials or epidemiologic trials. Explanations for the association include the effects of nicotine withdrawal symptom or the manifestation of comorbid psychiatric illnesses.

Grand Rounds August 1, 2025: Clinical Trial Notifications Triggered by Artificial Intelligence-Detected Cancer Progression (Kenneth L. Kehl, MD, MPH)

Posted on by Sophia Ramirez

Speaker

Kenneth L. Kehl, MD, MPH
Assistant Professor of Medicine and Physician
Dana-Farber Cancer Institute

Keywords

Artificial Intelligence; Cancer; Notification; Enrollment; Patient Identification

Key Points

  • Historically, less than 10% of adults with cancer enroll in clinical trials. At the same time, many trials struggle to reach their accrual goals. One possible contributor is that many trials of novel therapies for cancer have specific molecular criteria.
  • Dana Farber Cancer Institute (DFCI) developed MatchMiner, a computational matching tool, to connect patients to trials. However, identified patients often weren’t at a place in their treatment when information about trials was relevant. The research team was interested in whether they could train an artificial intelligence (AI) model to identify “trial-ready” patients.
  • The team conducted an implementation pilot, providing clinicians and research staff with weekly spreadsheets containing predictions of clinical trial “readiness” as identified by AI. The majority of identified patients were found to be ineligible upon RN review. Of those who were eligible, the majority opted not to move forward with the trial referral. At the end of the 9-month pilot, 6 AI-identified patients had been consented and enrolled in a therapeutic trial.
  • To assess the impact of AI-driven identification of trial-ready patients, the team launched OPTIONS (Optimizing Precision Trials with an artificial Intelligence driven Oncologist Notification System). The primary outcome of the trial was enrollment in any DFCI therapeutic clinical trial.
  • Patients with solid tumors were randomized into either a control group, in which they could be identified by the standard MatchMiner workflow, or 1 of 2 intervention groups. In the intervention arms, treating oncologists for genomically-matched patients with progressive disease and anticipated changes in treatment were contacted via email. In group 3, patients who met the readiness criteria were manually reviewed before the oncologists were contacted.
  • They found that, though the AI models successfully predicted which patients with active or progressive cancer may need treatment changes, sharing the trial information with oncologists did not increase trial enrollment.
  • This intervention addressed 1 barrier to trial participation. Other barriers may include eligibility criteria that goes beyond genomics and recent progression; and factors related to patient or oncologist preference, such as the motivation for participating, the complexity of the trial, and time toxicity.
  • Dr. Kehl concluded with a reminder that while AI can accelerate clinical cancer research by rapidly identifying clinical trial options for patients, impact requires integration. AI must be applied thoughtfully and continuously evaluated, and researchers should be aware of the pitfalls and shortcuts associated with the technology.

Discussion Themes

The DFCI team is currently working on MatchMiner-AI: an open-source tool that they hope will improve the accessibility of clinical trials for all patients by providing a list of relevant clinical trials. They’re running a pilot study focused on incorporating MatchMiner-AI with the historical tool.

It’s easier to train a model than it is to deploy it in a complicated healthcare context. Given that the tool performs as hoped, there are evidently implementation challenges that still need to be worked out.

The study team considered training the model on a more proximal task – i.e., “Predict whether this patient will enroll in a clinical trial.” However, they were concerned that this would introduce biases – a pertinent concern with AI models – based on which patients typically have the opportunity to enroll in clinical trials.

While there may be use cases in which providing the trial information directly to patients would be more efficient, this would need to be done carefully. Information about worsening cancer, for instance, is best contextualized in a conversation with an oncologist.

Grand Rounds July 18, 2025: State of Clinical Trials: An Analysis of ClinicalTrials.gov (Adrian F. Hernandez, MD, MHS; Rebecca D. Sullenger, MPH; Sara Bristol Calvert, PharmD; Karen Chiswell, PhD; Christopher J. Lindsell, PhD)

Posted on by Sophia Ramirez

Speakers:

Adrian F. Hernandez, MD, MHS
Executive Director
Duke Clinical Research Institute

Rebecca D. Sullenger, MPH
Duke University School of Medicine
MD Student | Class of 2026

Panelists:

Sara Bristol Calvert, PharmD
Director of Projects
Clinical Trials Transformation Initiative

Karen Chiswell, PhD
Statistical Scientist
Duke Clinical Research Institute

Christopher J. Lindsell, PhD
Director, Data Science and Biostatistics
Duke Clinical Research Institute

Keywords

Clinical Trials; Enrollment; Pragmatic Clinical Trials; Policy; Data Science

Key Points

  • A study of clinical trials from 2007 to 2010 found that the field was dominated by small trials and contained significant heterogeneity in methodological approaches, including reported use of randomization, blinding, and Data Monitoring Committees.
  • Clinical trials in the United States may be limited by legal, regulatory, and cost-related barriers. In a study of patient enrollment for cardiovascular clinical trials, the authors concluded that the U.S. had more trial sites than Eastern Europe or South America, but enrolled significantly fewer patients per site. These trends highlight the need for improved clinical trial infrastructure.
  • The presenters noted several promising trends in the field: growth in pragmatic clinical trials; high interest in clinical trial innovation from regulatory bodies and funding agencies; and the rapidly evolving capacity of clinical trials, particularly around accessibility.
  • The presenters provided an updated picture of the clinical trials landscape in the U.S., based on retrospective analyses of interventional clinical trials registered on ClinicalTrials.gov between 2018 and 2024.
  • They found that many trials remain small, lack a control group, and are incomplete after 5 years. Although small clinical trials without controls may be appropriate or necessary in specific contexts, such trials are also less likely to produce actionable data.
  • National policies prioritizing a more rapid, rigorous evidence generation system will likely be necessary to create a clinical trial ecosystem best equipped to advance public health.
  • In light of these insights, the team shared 5 potential policy approaches to improve the evidence-generation system in the U.S.:
    • Streamline trial start-up processes, institutional review board approvals, and contracting;
    • Enable scalable technologies to support greater participation;
    • Invest in modern clinical trial design strategies;
    • Require public reporting of key performance indicators and pay-for-performance results; and
    • Create stronger data sharing requirements and accountability rules.

Discussion Themes

Though the team utilized existing fields in ClinicalTrials.gov for their data, future research may utilize the key word search (i.e. adaptive platform trials) or natural language processing to investigate the state of clinical trials.

The value of small (<100 participants) trials was debated by the panelists. Though they do have a time and place, the high proportion of Phase III trials that enrolled less than 100 participants was surprising and concerning.

There are some limitations to ClinicalTrials.gov, namely in data entry. The more complex the trial, the more difficult it is to submit in a timely fashion. The system may post a barrier to embracing modern clinical trial design strategies.

Academia will also need to make policy changes to facilitate a healthier clinical trials ecosystem. The way career development and promotion pathways are structured, researchers are incentivized to lead small, potentially duplicative trials. Institutions need to reward, compensate, and value individual contributions to large-scale programs; i.e., the informative trial over the individually led trial.

Grand Rounds July 11, 2025: Novel Approaches to Recruiting Clinical Sites for Embedded Pragmatic Clinical Trials: Insights from the AIM-Back Trial (Trevor Lentz, PT, PhD and Tyler Cope, PT, DPT, ACT)

Posted on by terren.green@duke.edu

Speakers

Trevor Lentz, PT, PhD
Tyler Cope, PT, DPT, ACT
Duke Clinical Research Institute
Duke Department of Population Health Sciences
Durham Veterans Administration

Keywords

AIM-Back; Clinical site recruitment; Cluster randomized trial; Low back pain; Recruitment funnel

Key Points

  • Low back pain is an impactful condition that is more common in the veteran population. Typical low back pain care involves imaging and pharmacologic treatments that don’t always resolve pain issues and may lead to more invasive injection-based or surgical measures that often don’t result in better outcomes.
  • Research has shown that non-drug treatments (eg, cognitive behavioral therapy [CBT], yoga, physical therapy [PT]) are effective but not often used.
  • The AIM-Back trial (Improving Veteran Access to Integrated Management of Back Pain), an embedded pragmatic cluster randomized trial, sought to restructure care practices in Veteran’s Administration (VA) healthcare systems to promote and facilitate 2 clinical non-drug pathways that are supported by established guidelines as first-line treatment for low back pain.
  • Two care pathways were developed in coordination with VA clinicians, veterans, and care givers: (1) Sequenced Care Pathway – This pathway provided an initial onsite physical therapy evaluation and treatment session followed by weekly telehealth physical activity training for 6 weeks. The patient then saw the physical therapist again and was either discharged or provided with 6 weeks of training in psychologically-informed practices to help patients manage pain. (2) Pain Navigator Pathway – In this pathway, a local site clinician who was trained by the study team as a pain navigator discussed and facilitated alternative treatments for low back pain (eg, PT, yoga, CBT, massage). Patient follow up at both 6 and 12 weeks assessed progress and outcomes.
  • AIM-Back used a novel and intentional recruitment method, borrowing the concept of the business sales funnel, to generate as many site leads as possible. The recruitment process was systematic involving a 3 step framework: (1) Identify leads, (2) Approach leads, (3) Engage and select sites.
  • In step 1, leads were identified through Warm Market methods (sites known to the researchers), by Leveraging Data (evaluating lists of providers for potential fit), and through traditional Promotional Outreach efforts (advertising through networks and listservs). AIM-Back identified 184 leads from 53 VA healthcare systems.
  • Step 2 involved approaching leads through email messages. AIM-Back learned that promoting the trial in a way that helps clinicians solve their problems instead of asking clinicians to help with the research was more likely to yield the site. AIM-Back received responses from 23 VA healthcare systems.
  • In step 3, AIM-Back engaged personnel at all levels, from leadership to clinicians, to assess feasibility and buy-in at the site. AIM-Back selected 19 participant sites within 10 VA healthcare systems.
  • The Promotional Outreach strategy proved most effective with 9 (47.4%) of sites resulting from this strategy. The Leveraging Data strategy netted 6 (31.6%) sites, and 4 (21.1%) sites came from the Warm Market strategy. Site recruitment took approximately 3.6-3.8 months on average.
  • 17 sites enrolled 1817 Veterans with most sites (n=16) meeting or exceeding the minimum enrollment goal. When sites chose not to participate, they cited a reluctance to change their existing programs, a lack of clinicians or resources, or they were already participating in similar trials.

Discussion Themes

AIM-Back messaging evolved over the course of recruitment from a more traditional trial marketing email to an email that was more personal, short, and leveraged the standing of Duke University. This more personal approach to recruitment led to better relationships with sites during the trial.

Project management software can be helpful for tracking follow up with site leads and communication during the recruitment process.

One overall goal of AIM-Back was to set up a new clinical program that could continue after the end of the trial. Sites were given training materials for the centralized study components and support from AIM-Back was stepped down slowly. Sites that chose to continue the intervention trained a physical activity/whole health coach and a PT for the psychologically informed PT portion of the intervention.

Indicators of a potentially successful site included qualitative components that reflect a high level of engagement such as high interest and excitement in the study along with a sufficient patient population.

Read more about the AIM-Back trial design.

Grand Rounds June 27, 2025: Building Electronic Tools To Enhance and Reinforce CArdiovascular REcommendations for Heart Failure (BETTER CARE-HF) (Amrita Mukhopadhyay, MD, MS)

Posted on by Sophia Ramirez

Speaker

Amrita Mukhopadhyay, MD, MS
Eugene Braunwald, MD Assistant Professor of Cardiology
The Leon H. Charney Division of Cardiology Department of Medicine
Division of Healthcare Delivery Science Department of Population Health
NYU School of Medicine
NYU Langone Health

Keywords

Heart Failure; Electronic Health Record; Prescribing

Key Points

  • Heart failure is a major public health issue and a leading cause of hospitalization, affecting over 6 million Americans. Mineralocorticoid antagonists (MRA) are a potentially life-saving treatment but are under-prescribed in patients with heart failure with reduced ejection fraction (HFrEF). Closing this treatment gap could save over 20,000 lives in the U.S annually.
  • Electronic Health Record (EHR) tools could be a low-cost, scalable way to improve prescribing. However, there’s wide variability in EHR tool development and design. The optimal delivery and timing of EHR tools is unknown.
  • EHR tools fall into 2 categories: alerts and messages. Alerts apply to a single patient at a time and pop up during a clinical encounter; messages apply to multiple patients at once and are seen between encounters. The BETTER CARE-HF team designed both in accordance with Cognitive Load Theory and Nudge theory, applying the concepts of positioning, the split attention effect, default option, the transient information effect, and social influence.
  • They hypothesized that A) among patients with HFrEF who are evaluated by a cardiologist in the outpatient setting, an alert or a message will improve prescribing of MRA as compared to usual care, and B) the alert would be more effective than the message.
  • The researchers approached the pilot study as a “qualitative phase,” in which they would solicit feedback from participants and refine the intervention. They made several modifications to the EHR alerts and messages in response, and noted that guiding frameworks and pilot-testing were critical to designing an electronic intervention.
  • The pilot study was followed by a pragmatic trial that took place in over 60 practices in the NYU Langone Health System. Patients were cluster-randomized to an alert arm, message arm, or usual care. The primary outcome was new MRA prescription during the study period.
  • In the alert arm, nearly 30% of MRA-eligible patients were newly prescribed MRA – a highly statistically significant increase. The alerts were effective across all practice settings but were especially effective in high-volume settings.
  • In the message arm, 15.6% of MRA-eligible patients were newly prescribed MRA. Compared to 11.7% in the usual care arm, this was still a statistically significant increase, but was less effective than the alerts. Looked at another way, the number of MRA-eligible patients needed to result in one prescription was 25.6 in the message arm, compared to 5.6 in the alert arm.
  • An automated, EHR-embedded, tailored, and selective alert delivered at the time of the visit more than doubled prescribing of MRA as compared to usual care. Well-designed EHR tools could save lives.
  • Despite EHR tool effectiveness, busy physicians may still be hesitant. Too many tools can cause fatigue and burnout; concerns about workload and time costs can hinder uptake. Conversely, EHR tools that save time and reduce cognitive load may be more beneficial in busy practices. A post-trial survey indicated that cardiologist perceptions were generally favorable towards the BETTER CARE-HF tools, with some notable differences when asked about workflow.
  • The research team is conducting a multi-center trial to assess the effectiveness of the alert at other institutions, specifically across 3 high-volume health systems around the country. They are actively seeking other institutions to join the trial and encouraged attendees to reach out if interested.

Discussion Themes

The research team started by compiling EHR data on the current gap in care at NYU Langone. Having that real-time data helped the health system, and the physicians were a part of it, recognize that the intervention was necessary – despite their predisposition that they were delivering high-quality care.

This intervention was targeted to a specific population (cardiologists at NYU Langone) and a specific treatment (MRA) for a specific condition (HFrEF). In a different setting or if there was a different treatment involved, implementation may need to be adjusted.

Dr. Mukhopadhyay noted that folks who saw how the intervention worked were often surprised by how rarely the alert was triggered. She suspects that the selective nature of the intervention helped drive the intervention’s effectiveness by preventing burnout.

Working with a single IRB that understood the intention behind a learning health system helped standardize regulatory expectations across sites and facilitated onboarding.

Grand Rounds June 20, 2025: The BedMed Trials: Does the Timing of Blood Pressure Medication Matter? (Scott Garrison, MD, PhD, CCFP)

Posted on by Sophia Ramirez

Speaker

Scott Garrison MD, PhD, CCFP
Professor, University of Alberta, Department of Family Medicine
Director, Pragmatic Trials Collaborative

Keywords

Hypertension; Blood Pressure; Blood Pressure Medication; Medication Timing

Key Points

  • In 2010, the Monitorización Ambulatoria para Predicción de Eventos Cardiovasculares (MAPEC) trial found that hypertensive patients who took once-daily blood pressure (BP) medication at night, as opposed to in the morning, had a 61% reduction in major adverse cardiovascular events (MACE).
  • These results came with a credible rationale: BP is higher during the day than overnight, and overnight BP is a better predictor of cardiovascular events than daytime BP. Theoretically, patients taking BP medication at bedtime could preferentially lower overnight BP. But there were also good reasons to be skeptical of the results, and clinical guidelines remained unchanged.
  • To further investigate whether the timing of BP medication had an effect on MACE, Dr. Garrison and his team conducted 2 randomized controlled trials: BedMed and BedMed Frail. The former was conducted in a hypertensive primary care population; the latter in a hypertensive continuing care population. They conducted them separately, given the differing risks and benefits for the populations and the likely underrepresentation of frail or complex older patients in BedMed.
  • In the early stages, Dr. Garrison came across several unexpected challenges. There were restrictions around data access; regulations around billing for trial-related procedures in British Columbia; and the time it took to identify a data partner.
  • In both trials, the intervention group took a once-daily BP medication when getting ready for bed. In BedMed, the control took a once-daily BP medication upon waking up in the morning; in BedMed Frail, the control had no change in their existing routine (which typically meant taking their BP medication in the morning). Given the unique needs of the trial population, BedMed Frail utilized opt-out consenting.
  • The primary outcome for both trials was all-cause death or hospitalization/emergency department visit for stroke, acute coronary syndrome, or heart failure. In BedMed, they used an intent-to-treat analysis, with patients participating (via active or passive follow-up) in the study for a median of 4.6 years. In BedMed Frail, they used a modified intent-to-treat analysis, with patients participating for a median of 1.1 years due to high mortality in the study population.
  • The research team found that no additional cardiovascular benefit is conveyed from taking BP medication at bedtime. Conversely, their results concluded that these medications can be safely taken at bedtime, so patients should incorporate them into their routine whenever they are least likely to forget it.

Discussion Themes

Dr. Garrison noted that he was more confident in the negative result for the BedMed trial than for BedMed Frail, given that the adjusted hazard ratio of 0.88 and the unadjusted ratio of 0.93 in the latter. A 12% reduction in the outcome (which was largely driven by death) may still be meaningful to patients.

Designing a trial that was workflow-friendly for physicians was a top priority for the research team and was critical to obtaining buy-in for and executing this trial.

A major accomplishment of BedMed and BedMed Frail was developing a network of volunteer physicians and a data partner who would collaborate with the Pragmatic Trials Collaborative on future trials.