“When I come here, they know me,” says Mary Febus, a severe chronic asthma patient, of the Family Physicians Group in Kissimmee, Florida. For years, Febus was a victim of inadequate, poorly coordinated, and unnecessarily expensive care for her uncontrolled asthma and other complex health conditions. More than a dozen hospitalizations per year for many days at a time took her away from her young son, who was himself a patient with complex needs.

All this ended when Febus came upon the Family Physicians Group, one of a handful of high-performing primary care practices identified by America’s Most Valuable Care, a research study funded by the Peterson Center on Healthcare and conducted at Stanford University’s Clinical Excellence Research Center (CERC). The study identified US primary care practices scoring in the top quartile on quality (based on the Healthcare Effectiveness Data and Information Set measures) and the lowest quartile on cost, after risk adjustment. Fewer than 5 percent of the roughly 15,000 sites assessed by the CERC team met these criteria. Eleven practices identified as exemplary are described in the research study.

That Mary Febus feels her practice knows her is a simple statement that masks the complexity of the personnel, clinical, operating, and information management systems required to provide effective, efficient, and personalized care. That the physicians and staff at this practice are engaged and find joy in their work is further indication that theirs is an innovative model worthy of replication.

The CERC study is a great example of health services research anchored in the scientific method to identify and validate a health care delivery innovation. Such solid science gives us confidence that we know that the intervention works in the precise settings in which it was tested. Researchers, practitioners, policy makers, and executives value scientific research for decision making. However, there exists a complementary need for rapid-cycle research focused more on speed and learning how to implement an intervention across different settings.

The Peterson Center on Healthcare is testing how to replicate the features of the 11 exemplary practices to improve health care practice across the country on four dimensions: clinical and functional outcomes, the experience of care, clinician engagement, and total cost of care. After a year of hands-on facilitation with practice teams to establish clear care team roles, reorganize workflows, and use data to measure change and integrate patients into the redesign, we are now replicating the model in three practices, in New York, Missouri, and Minnesota. We will ask ourselves: Does the replication result in significant enough improvement to warrant further investment to spread the model more broadly across the country? Do the benefits outweigh the costs by a significant margin? Can we reduce the costs of replication as we scale the effort while we learn how to generate even greater benefits?

With these questions, we would like to start a dialog on diffusion of practice transformation within the broader health care community. As we dig deeper into the replication, we will report what we have learned, reflect on ongoing challenges, and highlight what we still have yet to learn.

Meanwhile, the market has not waited for us. New primary care organizations, built from scratch, have independently begun to emulate many of the best practices observed in that one clinic in Florida and in the other exemplars. Organizations such as Chen-Med, CareMore, and One Medical are proliferating, led by entrepreneurs and infused by private investors who see opportunity in profiting from value-based reimbursement.

Their success as defined by relative growth confirms what we all intuit: There is demand for high-quality care at lower costs. Yet, even with all this market activity, Mary Febus’ experience is still rare, because the vast majority of the health care sector has not been able to meet that demand by effectively translating health services research into systemwide changes in delivery.

Health services researchers and many of us Health Affairs readers live on the “supply side,” where we identify and validate what works in health care policy and delivery. Health services research methods guide us with rigor toward a high level of confidence in a model or intervention. In this world, we value sample sizes, controlled variables, and rigorous statistical analyses. The bar is high, and rightly so, for certitude in measurement. The tradeoff for that certitude is time and perhaps a lack of focus on execution.

Those driven by investment models think in terms of time and cost of execution and the returns they can expect. If value is not delivered, they are out of business. They see opportunity driven by value-based payment; population demographics; and a growing demand from consumers, employers, insurers, and government to get better outcomes for their spending. In this world, implementation speed, rapid customer feedback, network effects, predictable returns on investment, and brand engagement signal success.

It is in this gap between health services research and market-driven service design where we believe the core of a rapidly self-improving health care system can be built. We believe that our collective ability as a health care sector to understand the relationship between these two worlds and to apply learning from each will determine how quickly we can evolve to higher-value health care.

Our imperative at the Peterson Center on Healthcare is to improve a delivery system that underperforms for patients, burns out providers, and is financially unsustainable. Our starting point was to identify primary care exemplars and validate the features that drove high performance—work conducted by CERC using traditional health services research methods. Our next task is to develop a sustainable strategy for implementing and replicating the exemplary features. This process has also revealed to us a new set of non-traditional health care sector participants from private equity investors to technology start-ups and service designers, who are working to solve these challenges using a market-driven approach.

Our endeavor has generated many questions:

  • What are the key adaptable interventions that will improve outcomes and the experience of care from all stakeholder perspectives while lowering costs?
  • How might we complement our evidence-driven orientation and learn from start-ups that are experimenting in rapid cycles with customer-centered service design, technology, and marketing?
  • How can we design a scaling strategy that combines research, onsite facilitation, self-learning networks, and software development to raise the benefits and lower the cost of adaptation and replication?
  • What can we learn from practice transformation that can inform payment and benefit reforms to incentivize providers who are being driven by public and private payers?

Atul Gawande describes the need for the health care sector to “professionalize spread and scale.” In a 2012 article, he used the Cheesecake Factory as an example of a company that reengineered a high-end restaurant experience into a consistently high-quality and affordable option for millions. To scale such an experience requires know-how in every aspect of service delivery, including supply chain management, experience design, personnel management, marketing, and flawless execution at the point of service.

Translating delivery research into implementation know-how and excellence in execution does occur in the US health care system, but it is not a pervasive capability. To achieve Gawande’s goal, US health care needs a design or what we at the Peterson Center call a “scale architecture.” Ultimately, we see this as a dynamic network of networks that combines the capabilities of a “learning health system,” which identifies what is worth replicating, and a scalable implementation system, which creates insight into how to adapt and replicate innovations for different environments.

There are many technical challenges to creating such an ecosystem. Peer, personal, and organizational networks already exist but do not effectively connect practitioners, administrators, and executives with content, tools, and data in real time. A dynamic network requires technical standards that enable organizations to “plug in,” innovate, collaborate, and contribute. It requires access to data to understand whether improvement is occurring.

Yet, even with this technical capability, we acknowledge that the adaptive challenges such as deeply held institutional norms, long-standing relationships, and engrained processes must also be addressed. Building a culture of improvement requires the engagement of patients, their families, providers, and staff in their own care delivery redesign. It requires a focus on process and facilitation as well as content. The Peterson Center on Healthcare’s experience in primary care has taught us that this starts with helping care givers and support staff with small tests of change at the point of care and showing value to their patients and colleagues as well as themselves.

Our common starting point is the unifying idea that everyone’s experience of the US health care system can and should be improved, for patients and their families first, but also for providers and the millions who support them. From there, we can begin harnessing a productive tension between “true science” and practical application that, while vexing, is at the core of a scale architecture for accelerating adoption of innovations in health care delivery.

These concepts are what we at the Peterson Center hope will be explored in this section of the Health Affairs Blog. Over the next year, we hope these pages will host a place for sharing knowledge and creating a dialog among research-driven and market-driven actors. That dialog can advance a common understanding of what constitutes actionable evidence and the capabilities we need as a community to build a network of networks equipped to translate validated research into many different practice environments.