Interview

The Architecture of Creating New Knowledge

Genevieve Melton-Meaux, MD, PhD, Center for Learning Health System Sciences

What are learning health system sciences?

There’s really two parts to this question: What is a learning health system (LHS) and from what sciences does an LHS draw? A Learning Health System (LHS) is a virtuous cycle of data, knowledge and practice driven by culture, incentives and leadership. The Institute of Medicine (National Academy of Medicine) describes an LHS as:

A system in which science, informatics, incentives and culture are aligned for continuous improvement and innovation, with best practices seamlessly embedded in the care process, patients and families as active participants in all elements, and new knowledge captured as an integral by-product of the care experience.

The Journal of Learning Health Systems explains it as a constant spiral of learning where data flows into knowledge, knowledge flows into practice, and practice flows into data, repeating again and again, improving and benefiting patients and clinicians along the way.

To answer the second question, the strands of science or research this spiral of learning draws from are multidisciplinary. For example, the programs and units in the Center for Learning Health System Sciences (Center) support patient engagement, implementation science, evidence-based care, evaluation with rapid learnings, infrastructure to engage practices in research, and the use of digital solutions and artificial intelligence in clinical care. Our Center creates unique blends of all these sciences, bringing them together to drive our work. (https://med.umn.edu/clhss).

What are some of the benefits they present?

Bottom line: the patients we serve can positively and more quickly benefit from advances in research. Right now, the average translation gap for new science to reach patients is 17 years. The work we are doing as an LHS dramatically reduces that gap. We want to get to a point where routinely evidence informs health care delivery and health care delivery informs evidence.

What are the biggest challenges of incorporating these ideas into the health care delivery system?

This question really gets at why the Center was created—to provide the resources and support providers need to address a known issue in care delivery with evidence and to study design and the deep connection between health care delivery and researchers to integrate and disseminate evidence and best practices. That said, establishing an LHS is a significant change for all stakeholders, requiring consistent engagement at all levels, promotion by leadership, new logistics in care and research operations, data infrastructure, etc. For health care providers and staff, there is limited bandwidth to engage in new non-clinical activities, especially in the pandemic environment. Another major challenge is finding and training researchers with the skills needed to do this interdisciplinary work, which is why AHRQ (Agency for Healthcare Research and Quality) and PCORI (Patient-Centered Outcomes Research Institute) created funding opportunities to develop the skills of the next generation of LHS researchers. The Minnesota Learning Health System Mentored Career Development Program (MN-LHS) mentors junior faculty interested in gaining these skills. 

What are some examples of how they can be incorporated now?

Most recently, MN-LHS Scholar Carolyn Bramante, MD, MPH, led the nation’s first study, a randomized controlled trial, on whether metformin, fluvoxamine and ivermectin, or their combinations could serve as possible treatments to prevent ER visits or hospitalization, as well as Long-COVID. The results of this clinical trial were published in the New England Journal of Medicine.

Our program for Digital Technology Innovation is currently evaluating tools for educating patients and clinicians about dermatology images, including a tool using artificial intelligence to narrow down the differential diagnosis when a patient presents with a skin condition. Together with M Health Fairview, researchers are also partnering with Q-rounds, a health care software, as a service (SaaS) company whose flagship product is an inpatient virtual rounding queue that hospitals use to provide rounding schedule transparency for the patients and care teams. By creating time transparency, everyone knows when to be present for rounds, resulting in a more efficient rounding process for providers by decreasing missed connections from family and care team members and an increase in patient satisfaction. This tool will be piloted at M Health Fairview Masonic Children’s Hospital NICU this fall.

Currently, the Rapid Prospective Evaluation program has six projects underway, ranging in variety from using decision aids to improve utilization of cardiac monitoring and reducing chemotherapy toxicity in older adults with cancer, to utilization of medication therapy management or reduction of opiate overuse/dependency and improved care through sepsis microlearning and expanding specialist telestroke care.

the average translation gap for new science to reach patients is 17 years.

Additionally, we are committed to data and technology democratization for greater good. One example is a collaborative project focused on best practices with traumatic brain injury management and appropriate anticoagulation. The best practice tools and decision support will use an interoperable approach with the FHIR (Fast Health Interoperability Resources) standard so that the tools can be deployed at multiple organizations and multiple vendor systems. 

What are the goals of the new Center?

The Center for Learning Health System Sciences (CLHSS) is a collaboration between the University of Minnesota Medical School and the School of Public Health to create value in health delivery systems by establishing an iterative cycle of leveraging existing evidence and gathering new knowledge, applying our work into direct population health action and developing the learning health system field through education.

More specifically, the Center has three goals:

• Build a research and education LHS program of distinction.
• Decrease the time for science to successfully make it into patient care and into direct population health action.
• Seamlessly integrate research, care delivery and continuous improvement to create new knowledge and directly improve care. 

What is the Healthcare Innovation Program for Implementation & Evaluation (HI-PIE)?

HI-PIE studies how to translate and use evidence-based practices, interventions and policies effectively in real world settings within health care practices. The program is directed by Timothy Beebe, PhD, interim dean of the School of Public Health and consists of two units: Evidence Synthesis and Rapid Prospective Evaluation (RapidEval).

Evidence Synthesis is a collaboration of CLHSS and the Minnesota Evidence-Based Practice Center (EPC). The unit is co-led by Mary Butler, PhD, MBA, who also co-directs EPC and Josh Rhein, MD, assistant professor in the Division of Infectious Diseases and International Health. Together with their integration lead, Bronwyn Southwell, MD, assistant professor of Anesthesia, the team evaluates topic areas where evidence is emerging or evidence gaps exist to inform and adapt clinical practice (https://med.umn.edu/clhss/hi-pie/evidence-synthesis).

RapidEval features the unique opportunity for providers with an idea for improving care to be supported in generating high quality new evidence on health care practices. The unit focuses on and fosters rapid, iterative learning that builds upon the natural innovation taking place 8 within the health care system. The activities of the RapidEval Unit are aimed at and designed to increase adoption of best practices. RapidEval is led by Michael Usher, MD, a hospitalist and “triagist” managing patient flow and capacity at the M Health Fairview Systems Operations Center (https://med.umn.edu/clhss/hi-pie/rapideval)

How does clinical AI differ from digital technology innovation?

While they are complementary, each plays a different role. Digital Technology Innovation (DTI) evaluates and advances knowledge around the rapidly growing area of digital health and digital transformation, broadly defined as the use of technology to improve health and wellness. Digital health technologies cover a gamut of solutions such as wearable devices, virtual reality, mobile health apps, patient-reported data, smartphone-connected devices, etc. DTI is led by Rubina Rizvi, MD, PhD, as the Integration Lead, and myself (https://med.umn.edu/clhss/dti).

Think of the Program for Clinical AI as informational methods that crunch the data collected, so that health care delivery can be improved. The 9 Clinical AI team investigates AI-enabled tools in real-world settings, including monitoring AI model performance for drift, equity and fitness for answering questions across settings and subpopulations. A number of these Clinical AI algorithms may also be deployed into Digital Technology. Christopher Tignanelli, MD, MS, is a practicing trauma surgeon and leads this program (https://med.umn.edu/clhss/clinical-ai).

A helpful analogy is taking a road trip: while we know the basics of how to drive and read road signs, we can make better decisions, ones that improve the trip, when we use a GPS that has instant traffic monitoring and reroutes us to avoid delays and road closures.

Please tell us about the Practice-Based Research Network (PBRN).

Practice Based Research Networks (PBRNs) are groups of primary care clinicians and practices working together to answer community-based health care questions and translate research findings into evidence-based practice (https://pbrn.ahrq.gov/). The strength of a PBRN lies in its focus on community-driven and equity-focused research carried out in settings where participants/patients have pre-existing relationships with providers. This allows for research to be accessible, relevant and meaningful to communities.

The Primary Care Service Line (PCSL) PBRN is a partnership between the Department of Family Medicine and Community Health (DFMCH), CLHSS and M Health Fairview. The DFMCH is using their past success with their own PBRN to expand to all of the PCSL. Research facilitators and research champion providers are embedded throughout M Health Fairview primary care clinics to support building practice-based research capacity, opportunities and scholarship. We are actively engaged right now with the PCSL in developing this PBRN, which is led by Jerica Berge, PhD, MPH, LMFT, CFLE (https://med.umn.edu/clhss/pbrn).

How are you interacting with other health learning systems around the country?

The collaboration between the University of Minnesota Medical School and School of Public Health is novel. The LHS philosophy is spreading throughout the country and continues the deep research-clinical care integrations, which increased throughout the COVID-19 pandemic. In promotion of creating awareness for LHS, we host a monthly seminar, CLHSS iMpact, where we feature national experts in population health, informatics, PBRN, etc., to present their work and share their experiences in advancing the work of learning health systems.

We also collaborate nationally on various research studies. Recently, Dr. Tignanelli and I received an AHRQ R18 grant award, Evaluation of the SCALED (SCaling AcceptabLE cDs) Approach for the Implementation of Interoperable CDS for Venous Thromboembolism Prevention. The proposed project will adapt a currently deployed clinical decision support system (CDS) to scale, evaluate and maintain an interoperable CDS of venous thromboembolism prevention guideline for adult patients with traumatic brain injury across our collaborative network (Regenstrief Institute/Indiana University, University of California-Davis, Geisinger Health, Johns Hopkins University and Mayo Clinic Arizona). The current climate of each health care system developing “home-grown” CDS for the exact same guidelines is not tenable. Building capabilities to rapidly translate patient-centered outcomes research to the bedside at scale and share interoperable CDS routinely with an updated knowledge base (living evidence synthesis) is necessary.

Please tell us about the Clinical Translational Science Institute and how it interacts with the Community Engagement to Advance Research and Community Health (CEARCH) team.

The Clinical and Translational Science Institute (CTSI) is enhancing the way research is conducted to make a meaningful impact on people’s lives by providing a comprehensive infrastructure of research services, training, grants, tools and more. One of the components of the CTSI is the Community Engagement to Advance Research and Community Health (CEARCH) team. CEARCH provides the architecture for University of Minnesota researchers and local organizations to collaborate so they can address health issues in ways that are truly relevant to the community. CLHSS collaborates with CEARCH primarily in development of the PCSL-PBRN.

The Center is celebrating its first anniversary, congratulations.

Thank you! It’s been quite a year, focused on the establishment of our operations, hiring a talented team, strengthening partnerships and building awareness of LHS. As you’ve heard, our programs have developed robust portfolios. Earlier this year, Evidence Synthesis was consulted in preparation for a Sepsis Summit at M Health Fairview and provided a review of the impact of the CMS Severe Sepsis and Septic Shock Management Bundle on patient outcomes. This is an excellent example of how we would like to be engaging on a regular basis. We are looking ahead now to moving from development and starting to see the fruits of that labor deploy more interventions into practice, to learn from implementation and use that data to build new knowledge and eventually proliferate the LHS cycle throughout Minnesota.

Genevieve Melton-Meaux, MD, PhD, is a professor of surgery, director of the Center for Learning Health System Sciences and core faculty in the Institute for Health Informatics at the University of Minnesota (U of M). She is a practicing colorectal surgeon. She serves as the U of M associate director for the Clinical NLP-IE Research Group and nationally as president of the American College of Medical Informatics and as a board member for the American Medical Informatics Association.