Speakers

Mark Sendak, MD, MPP
Population Health & Data Science lead
Duke Institute for Health Innovation (DIHI)

Suresh Balu, MD, MBA
Director, Duke Institute for Health Innovation (DIHI)
Associate Dean, Innovation and Partnership
Duke School of Medicine

Slides

Keywords

AI, ML, health equity

Key Points

• The Duke Institute for Health Innovation’s (DIHI) mission is to catalyze transformative innovation in health and healthcare through high-impact research, leadership development and workforce training and the cultivation of a community of entrepreneurship.

• DIHI approaches this work through four pillars of innovation: implementation and health delivery science, health technology innovation, leadership and workforce development, and best practices development and dissemination.

• In 2021, DIHI started a Health AI Partnership to empower healthcare professionals to use AI effectively, safely, and equitably through community-informed up-to-date standards. We have continued to see the digital divide, where a small number of teams have expertise. Through this partnership, we are trying to build those skills in low resource settings and build a community of practice.

• The Health AI Partnership started with 7 organization partners and has expanded to about 20 organizations. The main deliverables for the first phase of the project were developing standard key decision points for the AI product lifecycle and developing the Health Equity Across the AI Lifecycle (HEAAL) Framework.

• There are 8 key decision points in the AI product lifecycle: identify and prioritize a problem; evaluate AI as a viable component of the solution; develop measures of outcomes and success of the AI product; design a new optimal workflow to facilitate integration; evaluate pre-integration safety and effectiveness of the AI product; execute change management, workflow integration, and scaling strategy; monitor and maintain the AI product; and update or decommission the AI product.

• The first key decision point is procurement, which begins with identifying a problem and ends with allocation of resources to either build or buy an AI product or solution.

• Key decision point 1 is procurement. Frontline staff have to have buy-in to submit a proposal so organizational resources can be used to sustain the innovation. Align frontline staff and organizational leaders – create alignment throughout project selection.

• The second key decision point is development and adaptation, which is either building or adapting an external solution for internal clinical use. The first step is to develop measures of success, the second step is designing the workflow, and the third step is evaluating pre-integration safety and effectiveness before a solution is put into clinical use.

• The next key decision point is clinical integration. DIHI built a modular infrastructure to support many projects via a flexible data pipeline technology infrastructure that started with one model and now includes dozens of models. The final key decision point is lifecycle management, which includes monitoring and maintaining the AI product and updating or decommissioning the product.

• The Health AI Partnership held a workshop that examined how to assess the potential future impact of a new AI solution on health inequities. The discussion resulted in five assessment domains to be evaluated across the span of the AI adoption process: accountability, fairness, fitness for purpose, reliability and validity, and transparency.

Learn more

Health AI Partnership

Duke Institute for Health Innovation

Discussion Themes

-Can you talk about how DIHI/Health AI Partnership go about with the challenges in medical AI adoption outside of academic medical centers? There is a massive inequity and teams like DIHI are very rare. The next phase of work is building out a practice network in building AI capabilities. It will be a massive undertaking. We have a small philanthropic gift to get started. There is a need for support of infrastructure to bring AI partnerships to low resource settings.

–How do you get reliable information and data for care that is received outside of Duke? How do you mitigate bias because of incomplete data? For the CKD example, we combined data from Duke and claims data.

 

Tags

#pctGR, @Collaboratory1

Speaker

Pamela S. Douglas, MD, MACC, FASE, FAHA
Ursula Geller Professor of Research in Cardiovascular Diseases
Duke Clinical Research Institute – Duke University
Past President, American College of Cardiology
Past President, American Society of Echocardiography

Slides

Keywords

Chest Pain, Cardiovascular Imaging, Coronary Artery Disease

Key Points

• Imaging has transformed cardiology and many other fields. In 2024, despite several large randomized controlled trials (RCTs) comparing different evaluation approaches there is no universal consensus on initial imaging strategies, who to test and how; there are ongoing concerns about over imaging lowest risk patients but there is no consensus on testing deferral pathways; and new imaging technologies may offer value but are untested.

• Similar to most types of trials, there are pros and cons to both pragmatic and explanatory designs in imaging trials. Feasibility and generalizability affect design choices including inclusion/exclusion criteria, flexibility of imaging intervention being tested, guidance/control of subsequent care after imaging, and endpoints and outcomes.

• People with angina-like symptoms are often not patients with a disease. Most do not have obstructive coronary artery disease (CAD), but a few are very high risk. There is a potential for over testing with significant false positive rate and potential for missed diagnosis with stress imaging.

• RCT design will vary depending on which CAD manifestation(s) are reflected in the information provided by the imaging test being studied, which affects the treatment target(s) being evaluated in a therapeutic trial.

• There are many imaging findings that are important for optimal care. There is no single CAD phenotype that can be targeted diagnostically or therapeutically. Imaging strategies must be multidimensional or account for this heterogeneity.

• There are considerations for the flexibility of the intervention. When we evaluate imaging strategies for chest pain, is “usual testing” the appropriate comparator? Coronary CT angiography (CTA) may be the preferred test. Researchers also need to ask what is the optimal CTA intervention and should downstream care be required?

• What are appropriate endpoints for imaging RCTs and what events are we trying to avoid? Angina may not effectively discriminate between strategies. Intermediate endpoints such as diagnostic and therapeutic thinking are useful with impact on treatment being a major determinant of long-term value. Process of care/efficiency measures are important. Costs are rarely a significant factor in comparing different testing approaches.

 

Discussion Themes

-Why aren’t payors more willing to fund these trials? I have not seen a payor fund a trial in this space. There has only been a hand full of trials. They are trying to follow the guidelines for symptoms and disease. The tests are so common in aggregate it is a big expense.

–What is the sensitivity of portable AI ECHO devices? They are pretty good. We think about it as a screening tool to see if a patient needs an intervention.

 

Tags

#pctGR, @Collaboratory1

Speaker

Erin K. McCreary, PharmD, BCIDP
Director of Infectious Diseases Improvement and Clinical Research Innovation, UPMC
Clinical Assistant Professor of Medicine, University of Pittsburgh
President-Elect, Society of Infectious Diseases Pharmacists (SIDP)

Slides

Keywords

COVID-19, therapeutics, health equity, pragmatic clinical trial

Key Points

• In April 2020, UPMC established a COVID Therapeutics Committee to determine a process for allocating experimental COVID-19 therapies. UPMC did not allow patients to receive experimental COVID-19 therapies outside of the context of a clinical trial and used the REMAP-CAP platform, a global pragmatic adaptive trial platform, in all clinic sites.  Any patient admitted to the hospital with COVID-19 completed a COVID intake form in their admission record, and if they opted to learn about experimental therapies, a FaceTime call was set up with a research coordinator.

• The COVID Therapeutics Committee worked with the state health department to develop a policy for fair allocation of scarce medications to treat COVID-19. The guiding principles of the policy were developed to safeguard the public’s health by allocating scarce treatments to maximize community benefit; to lessen the impact of social inequities, to ensure that no patient is refused access to treatment based on age, disability, religion, race, ethnicity, national origin, immigration status, gender, perceived quality of life, sexual orientation, or gender identify; and to ensure that all patients receive individualized assessments by clinicians based on the best available medical evidence.

• The first drug that used the new process was Remdesivir in May 2020. UPMC set clear criteria based on the published data at the time for who could and could not receive Remdesivir. Setting clear inclusion/exclusion criteria based on highest level clinical evidence is crucial.

• UPMC established weighted factors for the Remdesivir lottery that were established by the ethics and clinical committee. Criteria included essential workers (25% increased odds of drug allocation), patients with an area of deprivation index score 80-100 (25% increased odds), end-of-life prognosis, defined as greater than 50% chance of death within one year from underlying conditions (50% decreased odds). The criteria were updated in June 2020 when it was determined that Remdesivir was safe for pregnant people, increasing odds because 2 lives could be saved.

• The COVID Therapeutics Committee created communication materials to be transparent and to get buy-in. The committee worked with communications specialists who partnered with palliative care providers and others to draft FAQs and other communication guidance for clinician teams who were having conversations about drug allocations with patients and families.

• There was a daily process for the Remdesivir lottery. Every patient with COVID-19 was automatically considered. Patients were screened by local teams at each site. If a patient was part of the daily allocation, the committee informed the local team by email and the drug was shipped from a central location to the local site.

• The first monoclonal antibody (mABs) treatment for treatment of COVID-19 was approved in November 2020. UPMC built upon the success of the Remdesivir lottery and in December 2020, UPMC began infusing monoclonal antibodies across 45 UPMC locations. Clinicians wrote a consult for mABs and the pharmacy delivered whichever mAB that was available.

• When Evusheld was approved in December 2021 a similar weighted lottery process was put into place. All eligible patients were gathered from data warehouse and manual EHR entry. The therapeutics and ethics committees grouped immunocompromised patients into 3 risk categories, with a lottery for those patients in group 1 (highest risk). Patients from disadvantaged communities had increased odds of receiving treatment. The lottery was repeated weekly with notification of allocation.

• Through this work, UPMC learned that allocation does not actually mean receiving the drug. The lottery structurally improved allocation. Association of lottery with underrepresented racial and ethnic groups other than Black was not evaluated due to small numbers. Factors associated with not receiving drug need further explored

 

Learn More

Read more in JAMA Network.

Discussion Themes

-The allocation didn’t match the likelihood that an individual would receive the drug. Any hypothesis for why this occurred? We have a few hypotheses. We went through every iteration of how do you contact patients. Only about 40% of our patients are enrolled in My UPMC. We know there is a digital gap in disadvantaged patients and elderly patients. Many patients did not answer the phone. There were also patients who wanted to talk to their doctor, and they needed time to pull in their provider.

 

Tags

#pctGR, @Collaboratory1

Speaker

Roxana Mehran, MD, FACC, FAHA, MSCAI, FESC
Mount Sinai Endowed Professor of Cardiovascular Clinical Research and Outcomes
Professor of Medicine (Cardiology), and Population Health Science and Policy
Director, Interventional Cardiovascular Research and Clinical Trials
Director, Women’s Heart and Vascular Center at Mount Sinai Heart
Icahn School of Medicine at Mount Sinai 

Slides

Keywords

Cardiovascular Health, Interventional Cardiovascular Research, Diversity, Health Disparities

Key Points

• There are important gender disparities in cardiovascular health. Globally cardiovascular disease (CVD) prevalence had decreased between 1990 and 2010, but it has slightly increased since 2010.

• CVD is the leading cause of mortality in women. The median survival time after a first myocardial infarction (MI) for adults 45 years and older is 8.2 years for males and 5.5 years for females at age 45 years or older. Of those who have had a first MI, the percentage with a recurrent MI or fatal CHD within 5 years is 17% of males and 21% of females age 45 years or older. Hospital admissions with acute coronary syndrome in women younger than 55 years increased 21% in 1995-1999 and 31% in 2010-2014.

• There are also ethnic and racial disparities. By 2045, more than 50% of the population in the U.S. is expected to be other than non-Hispanic white. It is important to note when you look at the ARIC surveillance data, there is consistently higher risk in black females and males across all ages.

• Yet enrollment of ethnic minorities in NIH clinical trials and for trials studying approved devices and drugs remains low.

• Diversity in clinical trials is important for generalizability of results, to provide equal opportunities, practice precision medicine, tailor practical guidelines, improve public health outcomes, detect potential differences in safety and efficacy, and to address health disparities.

• There are several strategies that will help increase diversity in CVD trials. First, increasing diversity among trial participants must be a top priority in order to address health disparities and allow for optimal diagnosis and management of CVD in all.

• Increasing diversity in trial leadership is one of the most important strategies to increase diversity among RCT participants.

• Further efforts are urgently needed to increase diversity in the cardiology workforce, which will improve clinical trial diversity and cardiovascular health for all.

• Approaches from the whole scientific community to tackle the inequality in workforce, trial leadership and trial participants have to be developed.

 

Discussion Themes

-What have been the challenges in developing and identifying people interested in working in clinical trials? Many high school and college students are moving away from the sciences. We have to change how we are practicing. The biggest barrier for CV medicine is the lack of women leaders in CV. Trainees don’t see a way forward. Work-life balance is a very important issue and will be for the next generation of women and men.

-Have you looked at the differential recruitment approaches between provider led and EHR-led approaches? Can we use blinded HER-led recruitment efforts to identify participants? This is a really important point because we know that providers have bias, and it is the providers who do not approach women for enrollment.

–What about women who are excluded because they are pregnant or planning to be pregnant? It is an excellent question. With the current ability to become pregnant this can affect women up to the age of 50. These are really excluding many women on the basis of this exclusion. We need to think about removing some of this and evaluate.

Tags

#pctGR, @Collaboratory1

Speakers

Ruth Engelberg, PhD
Research Professor of Medicine
Division of Pulmonary, Critical Care and Sleep Medicine
University of Washington

Erin Kross, MD
Associate Professor of Medicine,
Division of Pulmonary, Critical Care and Sleep Medicine
University of Washington

Robert Lee, MD, MS
Assistant Professor of Medicine
Division of Pulmonary, Critical Care and Sleep Medicine
University of Washington

Slides

Keywords

Jumpstart, Advanced Planning, Goals of Care

Key Points

• Researchers know that goals of care discussions between patients and clinicians are associated with important patient and family outcomes. And yet goals of care discussions and their documentation remain a shortcoming in many health systems.

• The Jumpstart intervention is a communication-priming intervention. It has been studied in prior contexts in a PCORI trial and a pilot inpatient trial where the intervention increased goals of care discussions from 8% to 21%.

• For the Jumpstart trial, a number of refinements were made to make it more pragmatic, including the creation of Jumpstart using EHR data rather than patient or family member surveys, delivering the intervention to clinicians only, automated population of Jumpstart guide fields, and automated Jumpstart delivery to clinicians by email.

• The research question Jumpstart set out to answer is can a patient-specific, clinician-facing communication priming intervention with discussion prompts effectively promote goals of care discussions between clinicians and hospitalized older adults with serious illness?

• Jumpstart is a pragmatic randomized trial of Jumpstart compared to usual care. It utilized a waiver of consent; all eligible patients were randomized, and randomization was stratified by hospital and history of dementia at 3 hospitals in University of Washington system.

• Patients who were eligible had been hospitalized at least 12 hours but no more than 96 hours, age 55 and older with at least one Dartmouth Atlas chronic condition, or they were age 80 or older. The Jumpstart Guide was delivered to members of the primary hospital team on day of randomization via a secure email with a reminder message via pager.

• The primary outcome was the proportion of patients with EHR-documented goals of care discussion within 30 days of randomization. Goals of care discussions were defined as discussions about overarching goals for medical care but going beyond “just code status” (e.g. DNR/DNI). The goals of care discussions were identified by a natural language processing (NLP) called BERT and screened human abstraction.

• The study team trained and validated the NLP model by adjudicating 2,500 notes, and compared human decisions with NLP decisions to come up with an assessment with how NLP preformed. In order to classify a patient with goals of care conversation, the study used the NLP to screen the records and pull out excerpts of EHR that had high probability to contain a goals of care conversation and then conducted a human review. EHR passages were adjudicated in pseudo-random order, blinded to patient ID.

• Jumpstart also obtained secondary outcomes from the EHR, including ICU admissions, ED visits, palliative care consultation, ICU and hospital-free days, death and hospital readmission within 7 days of discharge

• Among hospitalized older adults with serious illness, Jumpstart found that a pragmatic clinician-facing communication-priming intervention significantly improved documentation of goals of care discussions in the electronic health record, with a greater effect size in racially or ethnically minoritized patients.

• Jumpstart provides evidence that a low-touch intervention can nudge clinicians to change behavior. Overall prevalence of goals of care discussions is low suggestions opportunity for improvement. Jumpstart may be useful in enhancing equity in serious illness communication among racially or ethnically minoritized patients.

Learn more

Read more in JAMA and a JAMA editorial.

Discussion Themes

-Within individual physicians, did you see an increase in goals of care discussions? We have not looked within clinician practices for documentation. In some previous trials we have randomized at the clinician level. What’s hard is that patients have a number of different physicians. When we sent out the jumpstart we sent it to the entire team, from interns to attendings. We don’t have much data on the overall effect on individual physicians behavior.

-Were you to do this again would you use BERT again or go in a different direction? BERT was state of the art in 2018 and out preformed all other NLPs at the time. The same family of deep learning models is still considered state of the art. The key thing with these models is that they have gotten bigger to accommodate the language. We could possibly take out the training phase.

–What is the concern about prompt fatigue in these settings? It is something we are thinking about a lot in the outpatient trial. Clinicians could receive prompts on an almost daily basis. We are limiting it so clinicians can only receive prompts x times day.

Tags

#pctGR, @Collaboratory1

Speaker

Harlan M. Krumholz, MD, SM
Harold H. Hines, Jr. Professor of Medicine
Department of Internal Medicine
Section of Cardiovascular Medicine
Yale University School of Medicine
Director, Yale-New Haven Hospital Center for Outcomes Research and Evaluation

Slides

Keywords

Decentralized Trial, Digital Trial, Yale PaxLC Trial, Long COVID

Key Points

• The PaxLC Trial is a decentralized Phase 2, 1:1 randomized double-blind superiority placebo-controlled study on non-hospitalized high symptomatic adult participants with Long Covid to determine the efficacy, safety, and tolerability of 15 days of Nirmatrelvir/Ritonavir compared with a Placebo/Ritonavir.

• This talk describes the PaxLC Trial’s strategies to implement a digital, decentralized, and democratized approaches to multidisciplinary research that address the deficiencies of traditional research studies, which can tend to be hierarchical, siloed, slow, expensive, and inconvenient.

• A key question for the study was how do we put the interest of the participants first and foremost and produce knowledge rapidly? Research can improve by simultaneously leveraging advances in technology and culture. Studies must be convenient, meaningful, respectful, efficient, rapid, and fair.

• The research optimization requires partnership with participants, including involvement in study design and workflow, data collection and analysis, and data and results sharing. Participants should have access to investigators and the results of studies.

• PaxLC brings together many innovations including online screening, digital medical record review, e-consent, home-delivery of medications, local clinical blood draws, home-based biospecimen collection, online diaries and surveys, digital medical record outcomes, and participant-centricity, and return of results. During the presentation, members from across the research team shared how PaxLC implemented all of these innovations through the course of the trial.

Discussion Themes

-What have you experienced on the scalability of the approaches you have taken, such as regulations and IRB? Is this knowledge generalizable to other similar trials? Using the local Yale IRB was an asset. We had to have the right people in the room to initiate communication with collaborators and their IRB representatives. The Yale IRB and Trusted Medical helped facilitate the other states and anything we need to take in account for recruitment. We will have a repository with Trusted Medical for future studies. We had to operate the trial with an institution that would allow zero risk. We had to come up with solutions.

-Can you comment on the decision to make the PROMIS scale the primary outcome? Does it capture most what matters to patients? We spent a lot of time discussing what would be the best primary outcome. Pfizer provided feedback. We wanted to measure a lot of stuff. We were familiar with the PROMIS scale and  felt PROMIS would capture whether people were feeling better. We wanted to say if this is working, people’s general health should improve.

–How do you organize recruitment to ensure you have the diversity you want and how do you explain to people you are turning away? It’s a multipronged strategy. Be welcoming and be authentically committed to health equity and be worthy of the trust. One of the main reasons we turn people away is because they are taking a medication that does not work with Paxlovid. A lot of participants have worked with their physician to stop a drug so they can participate.

Tags

#pctGR, @Collaboratory1