The Path to Continuously Learning Health Care

Health care in the United States has experienced an explosion in biomedical knowledge, dramatic innovations in therapies and surgical procedures, and expanded capacity to manage conditions that previously were debilitating or fatal—and ever more exciting H clinical capabilities are on the horizon. Yet, paradoxically, health care is falling short on basic dimensions of quality, outcomes, cost, and equity. Actions that could improve the health care system’s performance—developing knowledge, organizing and translating new information into medical evidence, applying the new evidence to patient care—are marred by significant shortcomings and inefficiencies that result in missed opportunities, waste, and harm to patients.

The human and economic impacts are great. An estimated 75,000 deaths could have been averted in 2005 alone if every state had delivered care on par with the best performing state. Current waste—an estimated $750 billion in unnecessary health spending in 2009—diverts valuable and limited resources from productive use.

It is important to note that individual physicians, nurses, technicians, pharmacists, and others involved in patient care work diligently to provide high-quality, compassionate care to their patients. The problem is not that they are not working hard enough. Rather, it is that the health care system does not adequately support them in their work. The system lags in adjusting to new discoveries, disseminating data in real time, organizing and coordinating the enormous volume of research and recommendations, and providing incentives for choosing the smartest route to health, not just the newest, shiniest—and often most expensive—tool. These broader issues prevent clinicians from providing the best care to their patients and limit their ability to continuously learn and improve.

The shortcomings are especially apparent when considering how other industries routinely operate compared with many aspects of health care. Builders rely on blueprints to coordinate the work of carpenters, electricians, and plumbers. Banks offer customers financial records that are updated in real time. Automobile manufacturers produce thousands of vehicles that are standardized at their core, while tailored at the margins. Although health care may face unique challenges in accommodating many competing priorities and human factors, the health care system could learn from these other industries how to better meet specific needs, expand choices, and shave costs.

The bottom line is that the nation, and its citizens, would be better served by a more nimble health care system that is consistently reliable and that constantly, systematically, and seamlessly improves. In short, the nation needs health care that learns by avoiding past mistakes and adopting newfound successes.

A vision—and a pathway

The Institute of Medicine has provided a roadmap for reaching this goal. It is detailed in Best Care at Lower Cost: The Path to Continuously Learning Health Care in America, a report released in September 2012. The good news is that opportunities for improving health care exist that were not available just a decade ago. Vast computational power is increasingly affordable and widely available, and connectivity allows information to be accessed in real time virtually anywhere by professionals and patients, permitting unprecedented diffusion of information cheaply, quickly, and on demand. Human and organizational capabilities offer expanded ways to improve the reliability and efficiency of health care. And health care organizations and providers increasingly recognize that effective care must be delivered by collaborative teams of clinicians, each member playing a vital role.

Yet simply acknowledging such opportunities does not necessarily result in putting them to good use. Indeed, building a learning health care system within current clinical environments requires overcoming substantial challenges. Clinicians routinely report moderate or high levels of stress, feel there is not enough time to meet their patients’ needs, and find their work environments chaotic. They struggle to deliver care while confronting inefficient workflows, administrative burdens, and uncoordinated systems, preventing them from focusing on additional tasks and initiatives, even those that have important goals for improving care.

Given such real-world impediments, crafting and implementing initiatives that focus merely on incremental improvements and add to a clinician’s daily workload are unlikely to succeed in fundamentally improving health care. Significant change can occur only if the environment, context, and systems in which health care professionals practice are reconfigured to support learning and improvement.

Realizing these objectives will require efforts in four main areas: generating and using real-time knowledge to improve outcomes; engaging patients, families, and communities; achieving high-value care; and creating a new culture of care.

Advancing real-time knowledge

Although unprecedented levels of information are available, clinicians and patients often lack practical access to guidance that is relevant, timely, and useful for the circumstances at hand. For example, of the clinical guidelines for the nine most common chronic conditions, fewer than half address the issues of patients who experience two or more of the conditions at the same time, even though 75 million patients fit this category. Bridging gaps in how knowledge is gathered and used will require applying computing capabilities and analytic approaches to develop real-time insights from routine patient care and then using new technological tools to disseminate the emerging knowledge.

One key step will be to strengthen the digital infrastructure of the health care system to better capture data on clinical care and patient outcomes, on the care delivery process, and on the costs of care. Data should be digitally collected, compiled, and protected as reliable and accessible resources for managing care, assessing results, improving processes, strengthening public health, and generating new knowledge.

Large quantities of clinical data are now generated every day in the regular process of care, but most of the information remains locked inside paper records that are difficult to access, transfer, and query. Digital systems have the potential to turn each of those bothersome traits on its head. Care must be taken, however, to integrate the new electronic methods seamlessly into providers’ daily workflow so as not to disrupt the clinical routine.

To complement the development of better digital systems, efforts are needed to promote expanded access to data and expanded data sharing. The idea is that the capacity for learning experiences increases exponentially when a system can draw knowledge from multiple sources. In one promising example, called distributed data networks, each participating organization stores its information locally, often in a common format. When a researcher seeks to answer a specific research question, all of the organizations in the network execute identical computer programs that analyze the data, create a summary from each site, and share those summaries with the entire network. In other efforts to expand data collection and access, insurance companies and other payer groups, health care delivery organizations, and companies that make medical products should be encouraged to contribute data to ongoing and new research efforts. And patients can play an important role by fully participating in self-reporting systems designed to gather data on patient outcomes, and by using new communication tools, such as personal portals, to better manage and record their own care.

Beyond technical matters, various legal and regulatory restrictions can be barriers to real-time learning and improvement. The privacy and security rules under the Health Insurance Portability and Accountability Act (HIPPA) pose particular challenges. In several surveys, researchers have reported that the rules increase the time and cost of research, impede collaboration among researchers, and make it difficult to recruit volunteers for studies. Protecting patient privacy is, of course, the basic starting point. But the current rules, with their inconsistent interpretation, offer a relatively limited security advantage to patients while impeding health research and the improvement of care. HHS is currently reviewing HIPPA rules, along with the policies of various institutional review boards that oversee research at many locations, with respect to actual or perceived regulatory impediments to the use of clinical data.

As more and better data become available, the obvious job will be to identify and adopt improved approaches for delivering accurate information to clinicians and patients in a timely manner. This will require making decision support tools and knowledge management systems routine features of health care delivery. Accelerating their use requires developing tools that deliver reliable, current clinical knowledge, in a clear and understandable format, to providers at the point of care, in addition to incentives that encourage the use of these tools. This also requires a shift in health professional education to teach skills for engaging in lifelong learning on how best to deliver safe care in an interdisciplinary environment. Furthermore, there are still multiple poorly understood barriers to dissemination and use of scientific evidence at the point of care. Addressing these barriers will require additional research and the development of practical tools that can improve the usefulness and accessibility of such data for clinicians and patients.

Empowering patients

An effective, efficient, and continuously learning system requires patients who are actively engaged in their own care. Clinicians supply information and advice based on their scientific expertise in treatment and their best assessment of potential outcomes, while patients, their families, and other caregivers bring personal knowledge on the suitability—or lack thereof—of different treatments for the patient’s circumstances and preferences. Both perspectives are needed to select the right care. Of course, providing what has come to be called “patient-centered” care does not mean that providers simply agree to every patient request. Rather, it entails meaningful awareness, discussion, and engagement among patient, family, and the care team on the evidence, risks and benefits, options, and decisions in play.

The structure, incentives, and culture of the current health care system, however, are poorly aligned to engage patients and respond to their needs—and patients are often insufficiently involved in their care decisions. Even when encouraged to play a role in decisions about their care, they often lack understandable, reliable information—from evidence on the efficacy and risks of different treatment options to information on the quality of different health care providers and organizations—that is customized to their needs, preferences, and health goals.

Patient-centered care takes on increasing importance in light of research that links such care to better health outcomes, lower costs, and customers—the patients themselves— who are happier with their experience, among other benefits. With these rewards in mind, health care providers and organizations will need to draw on a full toolkit of actions.

Providers should begin by placing a higher premium on involving patients in their own health care to the extent that patients choose, encouraging them and their families to be active participants. From this base, clinicians should employ high-quality, reliable tools and skills that are customized to a patient’s situation to aid in shared decision making. New technologies offer opportunities for clinicians to engage patients by meeting them where they are, rather than in traditional clinical settings. Further efforts may include providing new online sources of information and assisting patients in managing their own health—options that highlight the need for health professionals to assume new roles in partnering with patients.

Several actions can increase patient centeredness more broadly. First, there is a need for new tools that can assist individuals in managing their health and health care. Furthermore, public and private payers can promote and measure patient-centered care through payment models, contracting policies, and public reporting programs. There are also gaps in our ability to measure patient-centered care, which will require the development of a reliable set of measures of patient-centeredness for consistent use across the health care system. These measures can be used both to incentivize patient centered care and to assist organizations as they measure their improvement.

Fostering high-value care

Health care payment policies strongly influence how care is delivered, whether new scientific knowledge and insights about best care are diffused broadly, and whether improvement initiatives succeed. The prevailing approach to paying for health care, based predominantly on paying set fees for individual services and products, encourages wasteful and ineffective care. New models of paying for care and organizing care delivery are emerging to improve quality and value. Although evidence is conflicting on which models work best and under what circumstances, it is clear that a learning health care system would incorporate incentives aligned to encourage continuous improvement, identify and reduce waste, and reward high-value care.

The system would also be transparent. It would systematically monitor the safety, quality, processes, costs, and outcomes of care and make the information available for clinicians, patients, and families to use in making informed choices. This type of information on health care options, quality, price, and outcomes can then spur conversations among individuals and health care providers to promote informed decision making.

Multiple strategies exist for increasing the value of health care. Health care delivery organizations can use systems engineering tools and process improvement methods to eliminate inefficiencies, remove unnecessary burdens on clinicians and staff, enhance patient experience, and improve patient health outcomes. Furthermore, these organizations can reward continuous learning and improvement through internal practice incentives. For their part, public and private payers can adopt outcome- and value-oriented payment models and contracting policies that would support high-quality, team-based care focused on the needs and goals of patients and families. With an eye toward ongoing improvement, payment models, contracting policies, and benefit designs need to continuously refined to better reward high-value care that improves health outcomes.

Creating a new culture

Although each step along the path to a learning health care system is important, none by itself is sufficient. Rather, the host of needed changes must be interwoven to achieve a common goal: health care organizations that are devoted at their very core to optimizing care delivery practices, continually improving the value achieved by care and streamlining processes to provide the best patient health outcomes. Reaching this point will require broad participation by patients, families, clinicians, care leaders, and those who support their work. Health care delivery organizations, however, will play an especially important role. Because of their size and care capacities, they can set an example for improvement across the health care system by using new practice methods, setting standards, and sharing resources and information with smaller facilities and individual care providers.

Although details may vary among organizations, some key concepts will remain constant. A learning health care organization harnesses its internal wisdom—staff expertise, patient feedback, financial data, and other knowledge— to improve its operation. It also engages continuous feedback loops monitoring internal practices, assessing what can be improved, testing and adjusting it response to data, and implementing its findings across the organization.

Simply put, an organization that promotes continuous learning and improvement is one that makes the right thing easy to do. Its environment simplifies procedures and workflows so that providers can operate at peak performance to care for patients, and embraces support tools, such as checklists, that make providers’ jobs easier. This not only improves care delivery and patient outcomes; it also reduces stress on front-line care providers, improves job satisfaction, and prevents staff burnout.

Many organizations still struggle to implement such transformational system changes. They face both external obstacles, such as financial incentives that emphasize quantity of service over quality, and internal challenges to achieving constant improvement. To evolve successfully, health care organizations must develop a culture that supports improvement efforts, by adopting systematic problem-solving techniques, building operational models that encourage and reward sustained quality, and becoming transparent on costs and outcomes.

Leadership will be vital, as an organization’s leadership and governance set the tone for the entire system. The visibility of leaders at the highest level makes them uniquely positioned to define the organization’s quality goals, communicate these goals and gain acceptance from the staff, and make learning a priority. Leaders also have the ability to align activities to ensure that staff members have the necessary resources, time, and energy to accomplish the organization’s goals. By defining and visibly emphasizing a vision that rewards continuous learning and improvement, leadership encourages an organization’s disparate elements to work together toward a common end.

To complement leadership at the top, a continuously learning organization also requires leadership on the part of the managers and front-line workers who translate an expressed vision into practice. Middle managers play a crucial role in on-the-ground, day-to-day management of the units that, collectively, make up an organization. Unit leaders therefore must often challenge the prevailing mental models—deep-seated assumptions and ways of thinking about problems—and refocus attention on the barriers to learning and improvement. To this end, middle managers must be able to set priorities for improvement efforts, establish and implement continuous learning cycles, and foster a culture of respect among staff that empowers them to undertake continuous learning and improve patient care.

To promote continuous learning, health care organizations also need to adopt dedicated learning processes— mechanisms that help in constantly capturing knowledge and using the lessons to implement improvements. Achieving this type of systems-based problem solving requires an organizational culture that incentivizes experimentation among staff. While success is the goal, the system should recognize failure as key to the learning process and not penalize employees if their experiments are unsuccessful. Systems that continuously learn also need to be adept at transferring the knowledge they gain throughout the organization. Although each of these factors is important, it is the organization’s operational model—the way it aligns goals, resources, and incentives—that makes learning actionable. In this way, an organization’s operating model can promote continuous learning, help control variability and waste that do not contribute to quality care, and recoup savings to invest in improving care processes and patient health, and make improvement sustainable.

Pioneering health care organizations that successfully become continuously learning operations—fully or even partially— should also take the lead in diffusing the lessons learned more broadly. In this way, they not only can stand as beacons of opportunity, but also can provide the type of granular, hard-won information that can encourage and speed similar transformations across the entire health care system.

The entrenched challenges of the U.S. health care system demand a transformed approach. Left unchanged, health care will continue to underperform; cause unnecessary harm to patients; and strain national, state, community, and family budgets. The actions required to reverse current trends will be notable, substantial, sometime disruptive—and absolutely necessary.

The challenges are clear. But the imperatives are also clear, the changes are possible, and there are at least signs of success. Moving ahead, following the path to a continuously learning health care system, offers the prospect for best care at lower cost for individuals and society.