A Bipolar Learning Health Network: An innovation whose time has come
Editorial
The system of care for people with bipolar disorder “…is perfectly designed to get the results it gets.” (Attribution disputed: Paul Batalden or Arthur Jones). Those results are not good enough. We propose to improve results for people with bipolar disorder by disrupting the current care system with a Learning Health Network. First, we describe the system's current status and why it needs to be disrupted and improved. Then we describe a Learning Health Network model from the Anderson Center for Health Systems Excellence at Cincinnati Children's Hospital. Finally, we describe our plan to develop the Bipolar Learning Health Network. Few clinicians have insight with respect to bipolar disorder outcomes or the quality of care within their healthcare systems. To highlight this gap, we challenge clinicians to answer the following questions within their system, clinic, or own practices: How many bipolar patients are:
Getting reliable diagnostic assessments, measurement-based care, and guideline-informed care?
Being provided effective treatment or evidence-based psychotherapy?
Receiving adequate treatment for comorbid medical conditions?
Experiencing sustained remission and thriving?
Persistently ill? Rapid cycling? Experiencing treatment-resistant bipolar depression? Have psychotic symptoms?
Going to the emergency department? Hospitalized at any given time? Re-hospitalized within a year?
Having suicidal thoughts or behaviors? Dying from suicide?
Responding to unstudied combinations of treatments?
Participating in research studies?
CURRENT STATUS
Multiple challenges exist in the field of bipolar disorder. Illness etiology and pathophysiology are mostly unknown, so the reasons for heterogeneity in illness trajectories are poorly understood. Advances from research to practice take too long, often decades, and too few patients participate in research. When they do, patients and their families do not have enough opportunities to help guide research. Discovering new treatments primarily relies on serendipity without good etiologic models; treatment advances occur slowly. Consequently, new drug targets and treatments that yield better results are lacking. More than half of individuals needing care cannot access it. When they do, too few treating clinicians use systematic assessments or guideline-informed care or measurement-based care, exposing patients to wide variations in practice. We lack methods to personalize care, partly because we lack objective clinically relevant biomarkers and do not have the data to know what is happening in our systems or patients. Finally, as with the rest of healthcare, significant inequities exist in diagnosing and treating bipolar disorder.1 With all these challenges taken together, it is no surprise that poor outcomes for people with bipolar disorder, including premature mortality, have not improved in more than 20 years.2 We must do better.
WHAT IS A LEARNING HEALTH NETWORK?
The Learning Health Network model was designed and developed for the American Board of Pediatrics by Cincinnati Children's James M. Anderson Center for Health Systems Excellence (https://www. cincinnatichildrens.org/research/divisions/j/anderson-center/learn ing-networks). Since 2007, The Anderson Center has helped develop 15 learning health networks, including for chronic diseases, children's hospital safety, and community health. Learning Health Networks have achieved dramatically improved health results for their target population, for example, a> 60% decrease in mortality for high-risk children with hypoplastic left heart syndrome3 and a 72.5% decrease in adverse event drug rates in children's hospitals (https:// www.solutionsforpatientsafety.org/our-results). These networks have reduced the gap between research findings and the clinic by years. This successful Anderson Center Learning Health Network model has never been used for serious mental illness; the Bipolar Learning Health Network will be the first. We start with bipolar disorder as an initial step to quickly develop learning health network core processes and systems that could eventually be applied to other psychiatric diagnoses.
The Learning Health Network is an actor-oriented network organizational model in which participants (actors) actively collaborate to identify critical issues, test and validate methods to improve those issues systematically and reproducibly, develop best practices and evidence and then disseminate the resulting knowledge and know-how continuously.4 The Learning Health Network sets a collective aspirational goal that guides all activities.5 Learning Health Networks are characterized by:
• An unrelenting focus on exceptional, equitable outcomes;
• An engaged community of people with a culture of trust, respect, openness, humility, generosity, curiosity, and willingness to contribute and develop best practices and evidence together;
• A shared infrastructure of technology, governance policies, processes, and incentives so that participants can find what's needed when it's needed and where it's needed to make decisions that directly focus on improving health;
• A rapid learning system with members who can continuously improve their healthcare systems, reduce inequities, and use data to eliminate the boundaries between clinical care, quality improvement, and research.
Clinicians and patients meet within a supportive and collaborative learning community. Clinicians in the network access data to routinely assess which treatments are working best to help their patients get healthy and to discover, share, and learn about innovations from colleagues; clinicians work together with a community of stakeholders that includes patients and their families and advocacy groups, who also contribute ideas and solutions to help their loved ones cope, and to determine the research and innovation priorities of the network. Researchers are also part of this community and collaborate to design, conduct, and spread research results at unprecedented speed. All can teach, and all can learn.
The Learning Health Network develops analytic tools, training and support to improve outcomes continuously. First, the network relies on dashboards that all sites can access (with data never to be used to compete), updated at least monthly, to assess whether improvements in outcomes and care processes align with shared goals. Dashboards allow clinical teams to “drill down” to learn from variations across organizations and sites, individual physicians, and patients to determine what enables them to perform better (e.g., are they diagnosing reliably?), have better processes (e.g., is their system of patient-reported outcomes working?), and have better results for their patients. This information identifies areas that should be targeted for future quality improvement, innovation, or research activities. These dashboards also drive annual strategic planning to identify the most critical improvement and research priorities. Equity for marginalized and minoritized populations is crucial across all analyses, as inequity is a leading indicator of poor quality. These data can identify disparities as a target for improvement. The “community wisdom” element adds crucial context to data points and empowers stakeholder groups (e.g., patients and caregivers) often marginalized in favor of payer and health system priorities in traditional care settings.
Second, the network trains sites on changing systems to advance more reliable and integrated care by adapting evidence and innovations. The Learning Health Network uses experimental methods such as A/B testing, factorial experiments, and conjoint analysis to identify, test, and implement innovative, efficient interventions that improve sites' performance.
Finally, participants learn and integrate new insights into the network. They incorporate results from newly generated evidence into network best practices documents and guidelines, learning events, and coaching. This occurs via monthly all-network calls, emails, blogs, biannual conferences, tips, tools, case studies, and live and on-demand resources continuously available to all network sites and stakeholders, leveraging the communications infrastructure platform. The network disseminates information externally on websites and through social media to attract additional sites interested in joining. This continuous analysis and feedback process iteratively improves the network. It enhances network leaders' knowledge and expertise in optimizing the network's performance from the network system level to individual patients.
The learning health network can provide solutions to problems that are invisible to clinicians. For example, an adolescent in the Improve Care Now inflammatory bowel disease network designed a hoodie sweatshirt with a zipper up the sleeve to make it more comfortable to get infusions (https://www.improvecarenow.org/stayi ng_warm_and_comfortable_during_infusions). Clinicians can find support for aspects of clinical practice and research that fall outside formal clinical training (e.g., fundraising, stakeholder empowerment) or using innovative research designs such as n-of-1 trials, cohort multiple designs, sequential multiple assignment randomized trial (SMART) designs, pragmatic virtual clinical trials, and cluster randomization. Designing the learning health network correctly helps bridge a research and care landscape that makes it easier to manage complexity and shares the intellectual, financial, and administrative burden to maximize the benefit for all.
HOW WILL WE BUILD THE BIPOLAR LEARNING HEALTH NETWORK?
Our Learning Health Network will start by working with those health systems willing to take the risk of an uncertain journey to improve the health of people living with bipolar disorder. These health systems will work with clinicians, patients and their families, payers and purchasers, advocacy groups, and non-profit organizations. We plan to initially recruit 10–20 healthcare systems that will identify pilot sites, each caring for at least 750–1500 children or adults, along with a diverse and inclusive cohort of patients, families, and other key stakeholders to co-design, test, build, and ultimately scale more reliable and impactful approaches to bipolar care. Preliminary criteria for organizations to participate include:
• Commitment to improve outcomes for people with bipolar disorder;
• Commitment to equitable improvement and to eventual impact at scale in their organization. We are looking for senior organizational leaders' strategic commitment to transforming the mental health care system for people living with bipolar disorder across their entire organization's patient population;
• Commitment to share data and measures transparently across the network;
• Commitment to working as a collective system with mental health, primary care, emergency medicine and hospital care, and any other part of the health system that touches people with bipolar disorder;
• Willingness to participate in design activities and contribute ideas and innovations;
• Agreement to become a member of the network and to share responsibility for the ongoing operational infrastructure and finances of the network;
• Agreement to support a team including a physician/mental health leader and administrative support to lead and provide day-to-day leadership of activities to change local systems;
• An identified champion who can spend 4–6 hours a month on improvement activities and attend several meetings per year
CONCLUSIONS
Our healthcare system for people with bipolar disorder and their families is broken. We intend to find solutions by disrupting the flawed systems of care and research using a field-tested, proven method to get better results. We welcome any healthcare system ready to improve bipolar care and outcomes to join us in this journey.
ACKNOWLEDGMENTS
Supported, in part, by the Dauten Family Foundation and Kent and Liz Dauten and the Thomas P. Hackett, MD Endowed Chair in Psychiatry at Massachusetts General Hospital (AAN). Thanks to Emina Berbic for administrative support.
DATA AVAILABILITY STATEMENT
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Andrew A. Nierenberg1,2 Peter Margolis3 Stephen Strakowski4,5 Madhukar Trivedi6 Lakshmi N. Yatham7 Bipolar Disorder Learning Health Network*
1 Dauten Family Center for Bipolar Treatment Innovation, Massachusetts General Hospital, Boston, Massachusetts, USA 2 Harvard Medical School, Boston, Massachusetts, USA 3 James M. Anderson Center for Health Systems Excellence at Cincinnati Children's Hospital Medical Center, Cincinnati, USA 4 Indiana University School of Medicine, Indianapolis, Indiana, USA 5 Dell Medical School, The University of Texas at Austin, Austin, Texas, USA 6 Division of Mood Disorders and the Center for Depression Research and Clinical Care, University of Texas Southwestern Medical Center, Dallas, Texas, USA 7 Department of Psychiatry, Institute of Mental Health, University of British Columbia, Vancouver, British Columbia, Canada
Correspondence Andrew A. Nierenberg, Massachusetts General Hospital, Dauten Family Center for Bipolar Treatment Innovation, Suite 580, 50 Staniford Street, Boston, MA, USA. Email: anierenberg@mgh.harvard.edu
ORCID Andrew A. Nierenberg https://orcid.org/0000-0003-2897-0458 Stephen Strakowski https://orcid.org/0000-0003-2837-6280
References
Akinhanmi MO, Biernacka JM, Strakowski SM, et al. Racial disparities in bipolar disorder treatment and research: a call to action. Bipolar Disord 2018; 20:506-514. 2018/03/13. 10.1111/bdi.12638
He H, Hu C, Ren Z, Bai L, Gao F, Lyu J. Trends in the incidence and DALYs of bipolar disorder at global, regional, and national levels: Results from the global burden of Disease Study 2017. J Psychiatr Res. 2020;125:96-105. 2020/04/07. doi:10.1016/j.jpsychires.2020.03.015
Brown TN, Brown DW, Tweddell JS, et al. Digoxin associated with greater transplant-free survival in high- vs low-risk interstage patients. Ann Thorac Surg. 2022;114:1453-1459. 2021/10/24. doi:10.1016/j.athoracsur.2021.08.082
Fjeldstad O, Snow C, Miles R, et al. The architecture of collaboration. Strateg Manag J. 2012;33:734-750.
Ackoff R, Magdison J, Addison H. Idealized Design. Creating an Organization's Future. Prentice Hall; 2006.
APPENDIX A
Bipolar Disorder Learning Health Network Hilary P. Blumberg—Center for Neurocognition and Behavior, Center for Neurocomputation and Machine Intelligence, Center for Neurodevelopment and Plasticity Department of Psychiatry, Yale School of Medicine, New Haven, CT. Melissa DelBello— Mood Disorders Center, Department of Psychiatry, University of Cincinnati College of Medicine, Cincinnati, OH. Ken Duckworth— National Alliance on Mental Illness and Beth Israel Deaconess Medical Center, Boston, MA. Tristan Gorrindo—Optum Behavioral Care, Eden Prairie, MN. Dan Iosifescu—Department of Psychiatry, NYU School of Medicine, New York, NY and Nathan Kline Institute, Orangeburg, NY. Jonathan Jackson—CARE Research Center, Department of Neurology, Massachusetts General Hospital, Boston, MA and Harvard Medical School, Boston, MA. Masoud Kamali—Dauten Family Center for Bipolar Treatment Innovation, Massachusetts General Hospital, Boston, MA and Harvard Medical School, Boston, MA. Douglas Katz—Eliot Community Behavioral Health Centers, Dauten Family Center for Bipolar Treatment Innovation, Massachusetts General Hospital, and Harvard Medical School Boston, MA. Anil Malhotra—The Feinstein Institutes for Medical Research, Zucker Hillside Hospital, Zucker School of Medicine at Hofstra/Northwell, Glen Oaks, NY. Charles R. Marmar— Center for Precision Medicine in Alcohol Use Disorder and PTSD, Department of Psychiatry, NYU Grossman School of Medicine, New York, NY. Melvin McInnis—Heinz C. Prechter Bipolar Research Program, University of Michigan Medical School, Ann Arbor, MI. David J. Miklowitz—Max Gray Child and Adolescent Mood Disorders Program, Division of Child and Adolescent Psychiatry, UCLA Semel Institute, Los Angeles, CA. Mark Rapaport—The Huntsman Mental Health Institute at the University of Utah, Salt Lake City, UT. Perry Renshaw—The Huntsman Mental Health Institute, University of Utah, Salt Lake City, UT. Erika F.H. Saunders—Penn State College of Medicine, Hershey, PA and Penn State Health Milton S. Hershey Medical Center, Hershey, PA. Manpreet K. Singh—Stanford Pediatric Mood Disorders Research Program, Department of Psychiatry and Behavioral Science—Child and Adolescent Psychiatry and Child Development, Stanford University, Stanford, CA. Michael Sorter—Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, OH. Holly A. Swartz— University of Pittsburgh, Center For Advanced Psychotherapy, Pittsburgh, PA. Louisa G. Sylvia—Dauten Family Center for Bipolar Treatment Innovation, Massachusetts General Hospital, Boston, MA and Harvard Medical School, Boston, MA. Christina Temes—Dauten Family Center for Bipolar Treatment Innovation, Massachusetts General Hospital, Boston, MA and Harvard Medical School, Boston, MA. Mauricio Tohen—Department of Psychiatry and Behavioral Sciences, University of New Mexico School of Medicine, Albuquerque, NM. Katherine L. Wisner—Northwestern University Feinberg School of Medicine, Asher Center for the Study and Treatment of Depressive Disorders, Chicago, IL. Janet Wozniak— Clinical and Research Program in Pediatric Psychopharmacology and Adult ADHD, Massachusetts General Hospital, Boston, MA and Department of Psychiatry, Harvard Medical School, Boston, MA. Deborah Yurgelun-Todd—The Huntsman Mental Health Institute, University of Utah, Salt Lake City, UT.