Explore UAB

A Video is Worth 1,000 Operative Notes
Presented by Justin B. Dimick, MD
University of Michigan
Watch the recording here.

On March 1, 2022, Dr. Justin Dimick joined us from the University of Michigan Medicine to share his experience with using a learning health system (LHS) to improve surgeon technique and surgical outcomes. Dr. Dimick and his team recruited a cohort of bariatric surgeons to participate in a pilot study that involved video recording surgical procedures and having other surgeons review the footage and provide feedback in a “coaching” manner. These videos became invaluable data that revealed disparities in how surgeons perceived their own work during surgery, and how their surgical work was perceived by their colleagues. In addition to quality improvement at the University of Michigan Medicine, the archived data (i.e., video recordings) has informed surgical policy formation throughout the state. Dr. Dimick’s work demonstrates the potential of using video recordings as data to contribute to the ongoing learning and improvement cycles of an LHS, and provides evidence for the adoption of video recordings for quality improvement and research more broadly.

UAB's Accountable Care Teams
Moderated by David McCollum, MD
Panelists: Rockell Burks, RN; Kellie Flood, MD; Shikha Mehta, MD; Katherine Meese, PhD; Matthew Painter, PhD; Amy Stone, MSHA, LSSGB
Watch the recording here. 

The March 8^th LHS Month session featured a panel discussion by one of UAB Medicine’s Accountable Care Teams (ACT). The presentation began with the introduction of the underlying concept of UAB’s ACT: “How do we keep people working in healthcare and our organizations?” The answer is “accountable care,” which refers to a structured proactive approach to improve the way we collaborate to deliver quality care. ACTs brings together all members of a microsystem, that is, all interprofessional team members who work together, to problem solve. The sessions’ panelists are members of the “Transplant ACT” and define themselves as a learning community trained in discipline-specific problem solving for care quality and efficiency. The Transplant ACT aims to improve outcomes for patients, care teams, and the organization as a whole by following a model of reviewing data, utilizing disciplined problem solving, and disseminating solutions for the identified problem. As the team puts it, an ACT helps front-line personnel redesign their system by asking them to do differently rather than do more. In conclusion, ACT should be considered across services UAB Medicine, as research shows statistically significant improvement in , resilience, and an increased sense of belonging among healthcare workers. 

Using Automated Surveillance Reporting to Prevent Acute Kidney Injury: Implications for Learning Health Systems
Presented by Jeremiah Brown, PhD
The Dartmouth Institute
Watch the recording here. 

Dr. Jeremiah Brown joined us from the Dartmouth Geisel School of Medicine on March 15, 2022, to share his experience with utilizing an LHS to prevent acute kidney injury (AKI) among patients. Dr. Brown and his team demonstrated vast opportunities for improvement by using an automated surveillance and team-based coaching intervention to reduce occurrence of AKI. The automated surveillance reporting tool employed access to electronic health record and registry data from a corporate data warehouse and The National Repository for Cath Lab Data, respectively. Using this data, they developed a dashboard to compare overall risk-adjusted site-level performance to all other CART sites, allow for risk-adjusted site-level statistical analyses, and enable access to local sites’ patient-level identifiable data to support quality improvement. Dr. Brown and his team revealed an innovative implementation tool that provides readily available data for future needs within the healthcare system and beyond, while also utilizing this data to submit grant applications to sustain this work and continue to improve outcomes in, and reduction of, AKI.

The University of Florida Learning Health System and One Florida
Presented by Betsy Shenkman, PhD
The University of Florida
Watch the recording here. 

In the fourth presentation of our LHS month on March 22, 2022, Dr. Betsy Shenkman joined us from the University of Florida (UF) to discuss UF’s LHS and OneFlorida. OneFlorida is a partnership with the UF clinical and translational LHS initiative and dissemination and implemental science core.¹ Within this partnership, UF clinicians, the research community, UF health administration, and a quality improvement team come together to form a LHS review process. This process offers pilot funding and guides teams through project development and recommendations. Both real-time data and key stakeholders contribute to the success of this initiative. Their common data model includes procedures, lab results, death data, patients in clinical trials, and medication administration. Further, stakeholder input includes review from department/clinical teams, researchers, UF Health quality teams, informatics teams, patients/citizen scientists, and community members. Dr. Shenkman envisions the future of OneFlorida through 5 main pillars:

1. Establishing the FL, GA, and AL partnership as the vanguard for LHS initiatives, implementation science, and scholarship enhancement.
2. Expanding utilization of natural language processing to focus on social determinants of health.
3. Expanding engagement with partner sites and clinicians and leveraging county extension office alliances.
4. Adding genomic, pathology, imaging, and device-based information to the data library.
5. Identifying opportunities to enhance the impact of OneFlorida, particularly with diverse populations for LHS and learning health communities.

 The University of Alabama at Birmingham is now a member of OneFlorida. For more information on our involvement visit here.

___________________

1. Hogan, W. R., Shenkman, E. A., Robinson, T., Carasquillo, O., Robinson, P. S., Essner, R. Z., Bian, J., Lipori, G., Harle, C., Magoc, T., Manini, L., Mendoza, T., White, S., Loiacono, A., Hall, J., & Nelson, D. (2022). The OneFlorida Data Trust: a centralized, translational research data infrastructure of statewide scope. Journal of the American Medical Informatics Association : JAMIA, 29(4), 686–693. https://doi.org/10.1093/jamia/ocab221

Collaborative Learning Health Systems: Science and Practice
Presented by Michael Seid, PhD
Cincinnati Children’s Hospital
Watch the recording here. 

On March 29, 2022, Dr. Michael Seid joined us from Cincinnati Children’s Hospital to share a range of approaches for improving health outcomes in a healthcare system. A clinical psychologist by training, Dr. Seid argues that the healthcare system is a human system and while there is a lot of data and technology involved, fundamentally, a healthcare system is people working together. Dr. Seid envisions the science behind changing healthcare as a social and behavioral science. He discusses the “Actor Oriented Architecture” as the architecture of collaboration, suggesting that individuals who have a shared purpose, the capabilities and values to self-organize, a way to create and share resources, and shared processes that enable peer-to-peer collaboration, can effectively collaborate. By understanding that people want to cooperate and make a difference, we can begin to redesign systems to enable collaboration. Dr. Seid provides examples of leveraging the insight and participation of individuals who are on the opposite side of the healthcare system (i.e., patients who have, or have had, a medical diagnosis we are looking to improve), and how learning from the patient can take science beyond what was previously imagined in terms of clinical outcomes, academic literature, , policy change, so on and so forth. Dr. Seid concluded his talk by suggesting the need to change the system (not the people) and to make it easier to coproduce. There is a need to lead complex system change and to “be more of a gardener than a foreman,” meaning, for leaders to plant and water the seeds, making it possible for others to produce change. Though still in its infancy, Dr. Seid provided a thought-provoking shift in perspective when looking to redesign healthcare into learning health systems.

Articles from our speakers:

Additional resources:

• AHRQ: Supporting Learning Health Systems
  
• The Learning Healthcare Project: What is a Learning Health System
• National Academy of Medicine; The Learning Health System Series

• Defining learning health systems as a health system that acquires the ability to continuously and routinely improve health through discovery and implementation- once a health system acquires this capability, it then becomes a learning health system.
• End points to achieve while developing a learning health system:
• Having…
  • Characteristics, events and context captured as data to learn from- “Data to Knowledge”
  • Trusted knowledge that is rapidly available to support decisions (Knowledge to Performance)
  • Improvement that is continuous and enduring through an ongoing cyclic activity.
  • An infrastructure that enables this all to happen routinely and with the economy of scale.
  • All of the above is part of the health system’s culture.
• The three characteristics that make the LHS approach to health improvements different from other approaches:
  • Embraced uncertainty requiring discovery
  • Multi-stakeholder learning communities enabling continuity
  • Infrastructure: shared services enabling scalability
• The learning cycle should start at 5 o’clock, not 12 o’clock!
  • Starting at 12: the community has been told what to do or has made an “evidence-free” decision- no!
  • Starting at 5: the community collaboratively discovers what to do – yes!

• There is growing recognition about broader issues that impact health and we see this in some of the changes in the way healthcare is delivered- changing market and regulatory environment:
  • Transitioning from fragmented care to more coordinated care
  • Reimbursement focused on bundles or population management
  • Reimbursement tied to quality metrics and population health
  • There is now an increased focus on patient-centered medical homes and accountable care organizations.

• Why we need to focus on translational research and what this means.
  • Translation of new data into the clinic and health decision making.

• Learning and Innovation Health Community
  • Adding “Innovation” in the title because Dr. Rothman believes that a lot of the research that we do isn’t just about taking existing evidence and figuring out how to best deliver, but it’s also about how we create innovative approaches to care and then test those innovative approaches to care before we scale them out.
  • Adopting a learning health system approach can lead to rapid performance of research studies and then rapid implementation of those findings!

• This pressure injury case study demonstrated the impact of improvement tools and interventions on a specific improvement opportunity in an academic medical center and how this exemplifies health system learning.

• A custom-built application was integrated into UAB’s electronic health records which allowed users (nurses in this case) to document HAPI (hospital acquired pressure injuries) data on a daily basis. The pressure injury incidence data was available at both the unit and the organization level.

• This innovative technology provided structured reports of summary data, as well as individual patient details.

• Improvements in the health system have been implemented based on actionable data!
  • This dissemination at UAB resulted in 3 publications and 3 national presentations within a 3-year time span.

• Michael J. Mugavero, MD, MHSc: COVID-19 Core Overview
• Carrie Howell, PhD: Social determinants of health, insulin resistance and poor COVID-19 outcomes: A tale of two datasets
• Yulia Khodneva, MD, PhD: Clinical Outcomes of Covid-19 in a Racially Diverse Sample: Role of Renin-Angiotensin-Aldosterone System Inhibitors
• Katherine Meese, PhD: Rapid Data Collection and Dissemination with a new CORE team during COVID-19

• Dr. Mugavero presents on the formation of the CORE workgroups & development of the gateway process:
  • Concept proposal --> scientific gateway (based on expertise) --> working with IRB for either NHS or Exempt IRB --> Data Gateway --> Extracted Dataset --> Transformed Study-Specific Dataset

• The inevitable issue of extracting data from the EHR: raw data that typically involves errors and then the cleaned data that involves us manually going through and uploading datasheets correctly.

• The scientific protentional in leveraging existing electronic medical records to create a cohort study.

• Dr. Meese reflects on her rapid data collection and dissemination study during the COVID-19 pandemic by emphasizing the importance of collecting data from all stakeholders, including healthcare employees, to help inform prioritize

• Attendee participation indicated the following as some of the most important aspects in operationalizing a learning health system at UAB:
  • Access to usable data
  • Time and funding
  • Buy in from all stakeholders
  • Integrated academic practice partnerships
  • Creating a data group that connects the research data side to the clinical data side

• Attendees ranked the importance of each LHS theme as relative to the others as follows:
  1. Right data, right time, improved performance
  2. Realigning resources to support LHS priorities
  3. Creating a structure for aligned and informed prioritization
  4. Enhancing and integrating stakeholder engagement

• Why LHS? Why now?
  • We now are able to digitalize health care by use of electronic health records.
  • Transparency and public reports.
  • LHS drives value-based payment models.
  • Greater patient and stakeholder engagement.

• Lessons learned by successes and failures in LHS experiences
  1. “Face validity does not translate into efficacy/effectiveness. Rigorous evaluation should be pursued when possible.”
  2. “Learning health system opportunities exist beyond clinical care delivery. There are many opportunities for natural experiments.”
  3. “Randomization, when possible, should be utilized.”
  4. “Value of being embedded” (ex. How his opioid prescribing after surgery study came to be: Investigator was embedded in surgery and having discussions with surgical colleagues to identify problems.”
     • This study would have never come to surface if the colleague wasn’t embedded in surgery and at these leadership meetings where these questions and concerns come up.
  5. Fidelity to evidence-based interventions is important to achieve the desired outcomes.
  6. Patient and stakeholder engagement is important. Mixed methods are necessary for learning health system research.
  7. Engineering-based approaches can significantly enhance the operations and outcomes of a learning health system.
  8. Building and leveraging a continuous data monitoring system can identify opportunities for improvement.
  9. Evaluation of digital tools can be conducted as part of the implementation. User-centered design should be incorporated as part of the implementation.
  10. Multi-disciplinary teams can significantly enhance the impact in learning health systems.