Inside BD²’s Integrated Network: How the Learning Health Network is Improving Clinical Care and Outcomes

A Learning Health Network is a relatively new, innovative concept that brings healthcare delivery together with the research ecosystem to develop and evaluate care improvement in real-world clinical settings. Networks, large and small, that have been organized as Learning Health Networks build the capacity of everyone who participates – from community members to researchers to clinicians – to work together to accelerate research and improve health.

BD² has taken the novel approach of a Learning Health Network a step further, intentionally embedding a traditional Longitudinal Cohort Study. A deepened connection between research and care means the network can prospectively capture and analyze cutting-edge research data across neuroimaging, cognitive, clinical, and biological domains. With more connected data and collaborative teams, the Integrated Network facilitates the rapid adaptation of new evidence and accelerates our ability to generate new knowledge in bipolar disorder. The outcome of this means that, for example, a  Longitudinal Cohort Study participant at our Michigan site can benefit from the aggregate data and knowledge of all BD²’s Integrated Network sites.

Clinicians at each Integrated Network site are engaging people with bipolar disorder who have joined the study as participants in coordination with their care teams, including a physician leader, RN, coordinator, other staff, and family members. The clinicians and care staff collaborate with organizational leaders, the Integrated Network’s leadership team, the Scientific Steering Committee (SSC), and BD² partners. To make information as accessible as possible, researchers, clinicians, and study participants share data, outcomes, and clinical care strategies between institutions through one system. In doing so, BD² is hoping to more quickly impact the lives of people with bipolar disorder by driving continuous care and learning. And more broadly and over time, the Integrated Network will inform new care guidelines for all those living with bipolar disorder.

To learn more about the Learning Health Network, we spoke to BD² Integrated Network Scientific Steering Committee members Philip Wang, MD, DrPH, and Peter Zandi, PhD, about its potential impact on people with bipolar disorder.

How is BD²’s Learning Health Network unique? 

The BD² Learning Health Network is unique not just because of its size, but also the comprehensive scope of the scientific questions it will be able to answer. Few efforts to establish learning health systems have been able to address the breadth of current challenges faced by patients with mental disorders. BD² is an exception in that it will address a wide range of needs, including discovering novel biomarkers and mechanisms, as well as treatment and prevention targets, developing and testing clinical innovations, and rapidly and widely implementing effective innovations into practice.
– Philip Wang, MD, DrPH

Learning Health Networks are an exciting new approach to health care that is only now beginning to be considered in psychiatry, and BD² is leading the way in bringing this approach to advance the care of individuals with bipolar disorder. At the heart of a Learning Health Network is the idea of building an infrastructure and culture that enables us to learn in collaboration with our patients as we care for them and then use what we learn in real time to continuously improve outcomes. BD² has flipped the script and started by engaging a large and diverse cohort of individuals with bipolar disorder receiving care at leading medical centers that will closely follow the latest medical technologies over several years to learn what factors shape their journeys with the illness.

The deeply followed cohort will accelerate our understanding of the longitudinal course of bipolar disorder in a way that would not have been possible otherwise, and it will yield insights that can be rapidly disseminated to improve the outcomes of all patients.  In addition, it will cement the foundation for working together in a fully realized Learning Health Network that continues to drive progress in the care of bipolar disorder in the future.
– Peter Zandi, PhD

What does an embedded longitudinal cohort study mean for a Learning Health Network? 

Although longitudinally measuring participants over time can be methodologically challenging, it is essential for answering pressing scientific questions that are not possible to answer in traditional cross-sectional studies. For example, it is increasingly being recognized that the group-level differences from traditional cross-sectional studies cannot shed light on the dynamic nature of the mechanisms underlying psychopathology. Likewise, without longitudinally following participants with repeated deep phenotyping, it is not possible to identify modifiable risk factors that could be targeted with preventive interventions.
– Philip Wang, MD, DrPH

The embedded longitudinal cohort study is the linchpin to BD²’s approach to building a Learning Health Network for bipolar disorder. It will marshal an unprecedented amount of resources in an effort to learn in collaboration with individuals with bipolar disorder what factors determine the different trajectories of illness and what strategies are most effective in improving these trajectories. We simply do not know enough about what causes the illness and how best to manage it.

The embedded longitudinal cohort which is unmatched in scope, will provide a unique opportunity to address these shortcomings. By being embedded in a new Learning Health Network, we will be able to take what we learn from the cohort and rapidly translate it into improved outcomes for all patients in the network. Moreover, it will help establish the infrastructure and culture for working together that is needed to support future collaborative efforts in a growing and vibrant Learning Health Network on bipolar disorder.
– Peter Zandi, PhD

How has the development of the BD² Learning Health Network created new ways for numerous data to be transformed into actionable insights? Can you provide an example or scenario of how participants/researchers/clinicians might experience being involved in a Learning Health Network?

The strong commitment of BD² to open science will allow investigators both inside and outside the BD2 Learning Health Network to continually develop new actionable insights as well as clinical innovations. For example, by collecting a standard set of common clinical data elements or “anchors”, investigators will be able to study how adding novel phenotyping tools can improve the accuracy of biosignatures in the future. Investigators will also be able to link the growing sources of “real-world data” (e.g. from EHRs, administrative claims, national mortality databases, etc.) using privacy-preserving record linkages, to make new actionable discoveries — such as identifying new drug repurposing targets for bipolar disorder.
– Philip Wang, MD, DrPH

An exciting goal of the BD² effort is to prospectively collect a broad array of data on individuals with bipolar disorder — including genomic, molecular, imaging, digital, and electronic health record data. The various data types hold great promise, especially with the possibility of being analyzed with novel machine learning and artificial intelligence (AI) methods, for providing actionable insights that inform treatment decisions. The BD² effort will help determine which data is most clinically valuable and should be collected as part of the standard of care for all people with bipolar disorder. This process of collecting and using new data types as standard of care has already begun in the BD² Learning Health Network.

There is considerable evidence that routinely collecting patient-rated outcome measures as part of the standard of care to objectively track how patients are doing over time improves patient outcomes. Until very recently, this has not been done routinely in psychiatry. However, patients in the BD² Learning Health Network will complete a set of well-established patient-rated outcome measures as part of their standard of care, and clinicians will be able to review the results with the patients during their clinic visits to make more informed collaborative treatment decisions.

At the same time, these outcome measures will be collected in a database along with the other novel data types that are being collected on patients in BD². This database can then be analyzed by researchers to develop ever more sophisticated models with different data types that can better inform treatment decisions by more accurately predicting patient trajectories of illness or responses to different treatments. As the evidence base grows stronger, clinicians and patients can start to use these more sophisticated models in routine care to further improve outcomes. This process of continuously collecting data and learning from it to improve the standard of care is what the Learning Health Network is all about at its core.
– Peter Zandi, PhD

How will insights from the Learning Health Network impact people with bipolar disorder in the short and long term? 

I think in the short term, the BD² Learning Health Network will provide the field with a platform to identify and de-risk new treatment and prevention targets for bipolar illness. In the long term, especially with an ever-growing BD² data commons as well as advances in AI, we will be able to learn many more fundamental insights about the brain, mood, cognition, and behavior—it is exciting to dream about!
– Philip Wang, MD, DrPH

Treatment guidelines based on the latest evidence already exist for bipolar disorder, but these are not always followed and they are imperfect due to our limited understanding of the illness. The Learning Health Network provides mechanisms for disseminating best-evidence practices according to current guidelines, which will offer immediate benefit for outcomes in people with bipolar disorder across the network who may be receiving suboptimal care. At the same time, the Learning Health Network will continue to build the evidence base for improving the best-evidence practices and treatment guidelines which will benefit outcomes even more over the longer term.
– Peter Zandi, PhD