Improving the healthcare of Americans with multiple chronic conditions is one of the top aims of the U.S. national healthcare research agenda. One of the driving forces behind this objective is the disproportionately high concentration of healthcare expenditures that are attributed to this population. The 2009 Medical Expenditure Panel Survey (MEPS) found that the sickest 10 percent of patients account for 65 percent of all health expenses for the U.S. population (from July 2012 meeting of the National Advisory Council for Healthcare Research and Quality, which focused on trends in healthcare costs and the concentration of medical expenditures). Further, the MEPS identified a number of chronic conditions that most influence high levels of expenditures, including heart disease, cancer, mental disorders, chronic obstructive pulmonary disease (COPD) and diabetes. Potentially preventable readmissions are among the disproportionately high concentration of healthcare expenditures attributed to the sickest 10 percent of patients. In their June 2007 report to Congress, the MedPAC (Medical Payment Advisory Commission) estimated that 75 percent of Medicare readmissions are potentially preventable. Potentially preventable hospitalizations have been linked to discontinuity of primary care. In a 1984 randomized trial, elderly male Veterans who saw a different physician at each visit had a statistically significantly higher rate of emergent hospital admissions and longer average length of stays than those who saw the same primary care doctor at every visit. Since the 1980's, the problem of high-risk patients with frequent hospital admissions has persisted despite dramatic changes in the practice environment. One important change is the redesign of primary care. The Veterans Health Administration (VHA) Primary Care Program Office is implementing a Patient Aligned Care Teams (PACT) model at all primary care sites, in which continuous care is delivered by interdisciplinary teams who serve as the first point of contact for a broad range of fully-integrated health services and community resources. The VA's PACT model and other Patient-Centered Medical Home (PCMH) models are based on earlier innovations such as the Chronic Care Model, the Idealized Design of Clinical Office Practices, and Clinical Microsystems. They seek to implement primary care that is continuous, comprehensive, efficient, patient-driven and team-based., The PCMH models are not primarily aimed at patients who are at high risk of hospital admission, although most include reducing hospitalizations and emergency room use among their secondary aims. In the VA PACT, for example, hospital care and specialized services are provided outside of the PACT, while coordinators focus on smoothing hand-offs between care settings including those involving VA and non-VA providers. A recent Agency for Healthcare Research and Quality (AHRQ) systematic review of 19 comparative studies found that implementation of these models had no effect on hospital admissions (RR 0.96; 95% CI 0.84-1.10), but it reduced emergency room visits (0.81, 0.67-0.98) and modestly improved staff and patient experiences. Disease management and care coordination programs have also sought to improve the quality and delivery of care to patients with high-cost chronic illnesses. The U.S. Congressional Budget Office (CBO) issued a brief report (January 2012) on the effects of 34 Medicare demonstration projects on disease management and care coordination programs, and cited the following approaches as helpful in reducing hospital admissions: (1) use of team-based care, especially those with larger teams that include pharmacists; and (2) the smoothing of transitions between a primary care provider and a specialist. However, a recent AHRQ review of nurse-led case management programs for adults with medical illnesses and complex care needs found no effect on rates of hospitalization and a variable effect on emergency room use. The nurse-led interventions encompassed patient self-management education, health status monitoring, and coordination of care, typified by the Medicare Coordinated Care Demonstration (MCCD). From the viewpoint of caring for frequently hospitalized patients, the rise of hospitalism also poses a challenge to continuity of care. David Meltzer, MD, PhD, Associate Professor of Medicine and Director of the Hospitalist Program at the University of Chicago Medicine, has studied the changing medical workforce in the United States and found that the trend toward increased medical specialization has had the unintended consequence of increased fragmentation of primary care. One of the key challenges for the primary care system in meeting the complex needs of high-risk patients with multiple chronic conditions is adapting to the increased demands of collaborating with a larger number of associated healthcare providers within a decentralized, fee-for-service healthcare delivery system which does not pay for or facilitate communication or coordination among providers. Some would say that we have reached a point where we have to decide if we should continue to invest in programs with an increased need for coordination or in different models that reduce the need for coordination in the first place. Applying team theory literature, Dr. Meltzer has said “if you're spending all of your time coordinating, you should change the product.” For example, Dr. Meltzer has proposed a ‘comprehensive care physician model,’ to reduce the need for coordination between the primary care and hospitalist settings. In this model, a physician with expertise in both inpatient medicine and primary care leads an interdisciplinary team that carries a panel of approximately 200 frequently hospitalized patients, who they will treat both inside and outside of the hospital. The Centers for Medicare & Medicaid Services (CMS) have funded a demonstration of this program at the University of Chicago. Others focus on reducing the need for coordination and handoffs between primary care and specialty care. From this viewpoint, admissions from primary care could be reduced through accelerated consults or direct access to intense short stay and complex diagnostic unit services. The persistence of high inpatient utilizers despite the spread of case management and PCMH models has led to interest in ‘intensive primary care’ models. For example, a new round of CMS-funded demonstration projects focus on intensive models for patients at high risk of using inpatient resources. In their Research Brief published by the National Institute for Health Care Reform, Yee et al. proposed a taxonomy for categorizing ‘High-Intensity Primary Care’ programs, also sometimes called ‘Ambulatory Intensive Care Units,’ which use PCMH-based approaches to managing the sickest, highest-cost patients. In a ‘freestanding’ model, ongoing care is fully transferred from patients' regular primary care physicians to a dedicated clinic that exclusively or chiefly provides high-intensity primary care to a select group of patients. In contrast, in a ‘practice-based’ model, patients continue to receive care from their regular primary care physician, but are offered additional, high-intensity services, often managed by a care coordinator. In a hybrid model, care is temporarily shifted from the patients' regular primary care physicians to a dedicated clinic, and returned once their conditions stabilize. These approaches, which involve physician-led, team-based discussion and coordination that serve as the source of primary care, were excluded from the two AHRQ reviews mentioned earlier. As part of the PACT model initiative, the Health Delivery Committee is proposing to develop a primary care intensivist model that deploys well-trained interdisciplinary teams that identify and proactively manage Veterans at highest risk for hospital admission and death. The goal of the model is to reduce emergency department and urgent care utilization, hospitalizations and mortality among complex, high-risk patients. For healthcare system decision-makers, evidence is only one of many different factors taken into account. If a health system waits until there is traditional hard and fast evidence on the effectiveness of a new healthcare model, they would always be 10 years out of date. However, decision-makers do need to consider the findings of the best available research and the strength and applicability of that evidence. This report was produced in response to the Health Delivery Committee's request for an evidence brief to assist with their evaluation of the effectiveness of existing intensive primary care programs involving multimodal interventions delivered by interdisciplinary teams. An evidence brief differs from a full systematic review in that the scope of work is more narrowly defined and the traditional review methods are streamlined in order to synthesize evidence within a shortened timeframe. An evidence brief cannot capture the actual day-to-day program operations of evolving programs. While decision-makers can benefit from knowing about the best available research and the strength and applicability of that evidence, an evidence brief cannot encompass the full range of policy options or novel programs, many of which have not been evaluated in formal studies.