The value of health care — defined as the health outcomes achieved relative to the costs invested — is an increasing concern in the United States.1 Health care spending is expected to accelerate over the next 5 years, reaching 20% of the U.S. gross domestic product by 2025.2 Quality of care and health outcomes remain suboptimal.3 While environmental, social, and behavioral factors are the primary determinants of poor population health status and health outcomes in the U.S.,4 significant opportunities remain to improve the value of health care and the health care delivery system. Moreover, the health care system has an obligation to play a meaningful role in helping to overcome the effects of these external factors.
In this environment, the health care system is rightly moving away from the historical model of fee-for-service payment, which incentivizes the provision of larger quantities of health care services, and toward models based on value, through which hospitals and physicians receive higher or lower payments based the quality and cost of care. Nationally, public and private health care payers are implementing value- and risk-based payment models that include pay-for-performance, shared savings and losses, episode-based payment, and capitation, among others.5–8
As this transition occurs, it is important to recognize the diverse nature of today’s health care system, the changes that will be required to create the health system of the future, and the unique and pivotal roles that will be played by major academic health systems. Without modifications to current value-based payment models and the incentives that they introduce into health insurance markets, the ability of major academic centers to serve their communities and the nation will be strained. Nonetheless, academic medical centers welcome the opportunity to respond to appropriately structured changes to payment models that can spur improvements in the effectiveness of their multipart missions. At many institutions, improvements are already occurring in the operational and clinical efficiency of patient care services as well as in the medical education curriculum to imbue trainees with a greater understanding of population health and health systems science.
Historical Functions and Financial Characteristics of Academic Medical Centers
Academic health systems serve multiple important roles in addition to training most physicians, nurses, pharmacists, and other clinicians. As tertiary and quaternary referral centers, academic centers provide high volumes of advanced, technologically intensive, and sometimes experimental clinical services that are seldom available elsewhere. These centers lead advances in scientific research, driving innovation across one of the largest sectors of the U.S. economy and, secondarily, around the world.9 Most academic hospitals serve as vital community resources, working to address disparities in health care and health outcomes due to social determinants of health. Academic centers often represent critical elements of the first-responder system given their roles as trauma centers and leaders in disaster and pandemic infection preparedness.10
Major academic centers also function as large community-based health care systems that offer ambulatory primary and specialty care as well as basic hospital services to populations in their surrounding areas. Indeed, evidence suggests that teaching hospitals provide higher-quality care than non-teaching hospitals do — not only when caring for the sickest patients, but also when caring for patients with routine conditions. For example, from 2012 to 2014, Medicare beneficiaries had significantly lower 30-day mortality rates at major teaching hospitals than at non-teaching hospitals (8.3% vs. 9.5%, p < 0.001) after accounting for patient and hospital characteristics. Reductions in mortality were observed among patients with common medical diagnoses that are leading causes of hospitalization as well as among those undergoing major surgical procedures (Fig. 1).11 In fact, patients with low-severity medical illnesses or moderate-severity conditions requiring surgical treatment benefited from being admitted to major academic centers as much as or more than those with high-severity illnesses did.12 In addition to serving patients better, major teaching hospitals’ dual functions as community hospitals and regional referral centers enable trainees to become proficient in both basic and advanced types of care in an organizationally and economically efficient manner rather than in a diffuse, uncoordinated manner.
Given their diverse roles, major academic medical centers have unique financial structures, with high fixed costs as well as high costs per case. Tertiary and quaternary care require large investments to deliver routine care in ample case volumes and to ensure adequate standby capacity — i.e., keeping personnel, equipment, and facility infrastructure at the ready for patients with rare and catastrophic illnesses. While this profile makes academic centers more expensive than other settings, concentrating specialized clinical capabilities, education, and research at academic centers improves efficiency and limits total costs to the overall system because it avoids unnecessary duplication of scarce and expensive resources.
Research suggests that, together, standby capacity and case mix account for most of the increase in cost per case at academic centers relative to community hospitals, whereas medical education and research contribute modestly to the cost differences.13,14 In 2012 and 2013, for example, major academic hospitals were more than twice as likely to be in the top decile nationally in terms of caring for the highest-cost Medicare beneficiaries, who had average health care expenditures that were >10 times higher than those of other patients and who together accounted for 55% of all Medicare expenditures.15 In addition to differences in clinical case mix, academic centers treat many patients who have limited or no insurance, low incomes, and high social needs. Payments from Medicare and Medicaid often fall short of the average cost of providing care for these populations.16,17
In fee-for-service payment systems, the increased intensity of services and higher case mix at academic centers are partly accounted for through higher reimbursements per service. For example, more complicated procedures are assigned higher relative value units. Under Medicare policy, hospitalizations are reimbursed according to diagnosis-related group (DRG) weights, which are increased when complications and comorbidities are present.18 Because DRG weights do not fully account for the high case cost and high rates of unreimbursed costs at academic hospitals, Medicare makes indirect medical education (IME) payments to hospitals with accredited graduate medical education programs.19–22 Medicare calculates IME payments by adding an additional percentage per case based on the ratio of interns plus residents to hospital beds. IME payments can equal several million dollars per year.23 Other sources of government payments that help to account for the higher intensity of services at academic hospitals include Medicare outlier payments, which apply to cases involving extraordinarily high costs under the diagnosis-related group system, as well as Medicare payments to disproportionate share hospitals, which apply to hospitals that care for large percentages of low-income patients.18
Shifts to Value-Based Payment and Implications for Academic Medical Centers
In contrast to fee-for-service payment models, the value-based models being implemented today do not fully account for the fact that patients at academic hospitals are sicker, receive more complex services, and have higher social needs, leading to worse outcomes and increased costs per case without risk adjustment. Clinical factors as well as patient socioeconomic position, race/ethnicity, and community characteristics are known to influence health care utilization, care processes, and outcomes.24 Rewarding better outcomes and lower expenditures is bound to reduce payments to academic hospitals and physicians, unless differences in the populations treated and services provided are fully accounted for. This is true whether the payment model is pay-for-performance, shared savings and losses, episode-based payment, or capitation.8
Most Medicare value-based payment policies do not currently account for clinical, socioeconomic, or other differences between patients managed at teaching and non-teaching institutions. Moreover, Medicare value-based payment policies that lower payments to academic hospitals secondarily reduce their IME payments as well.20,25 A recent report by the National Academy of Medicine acknowledged that teaching hospitals have not fared well under value-based payment policies to date.24 This is true even for some academic medical centers that have achieved top-decile and higher risk-adjusted performance among all hospitals on quality measures, including mortality.
Value-Based Payment for Physicians
Under pay-for-performance, payment to physicians is increased or decreased according to the attainment of quality, patient satisfaction, and/or spending goals.26 Commercial payers have instituted many diverse pay-for-performance programs over the last decade or more.27 Most programs have rewarded better-quality care in primary care settings, particularly related to the management of chronic conditions.26 In 2017, the Centers for Medicare and Medicaid Services (CMS) began the gradual implementation of the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA), which requires physicians in fee-for-service settings to participate in value-based payment programs. About 500,000 physicians are expected to participate via the Merit-based Incentive Payment System (MIPS), a complex pay-for-performance program. Performance is based on quality of care, use of electronic health records, participation in improvement activities, and episode-based costs.28 The MIPS policy applies to all specialties and settings of care, but few quality measures exist for the highly intensive and specialized services at academic centers.29 It remains uncertain how physicians at academic centers will fare under MACRA and MIPS.
Value-Based Payment for Hospitals
Starting in 2012, Medicare implemented several mandatory programs that have penalized hospitals for lower relative performance on specified metrics: the Hospital Readmissions Reduction Program (HRRP), the Hospital Value-based Purchasing Program, and the Hospital Acquired Conditions Program. Under the original versions of these policies, the sizeable penalties were not adjusted to account for patient social characteristics or hospital teaching status. As a result, academic hospitals have been penalized more often than non-academic hospitals — even though the academic hospitals tend to provide higher-quality care.5,30–34 Recently, HRRP was modified to stratify eligible hospitals by the proportion of patients who are dually eligible for Medicare and Medicaid, to account for variation in patient socioeconomic status.35
Accountable care organizations (ACOs) are a value-based payment model used in fee-for-service settings through which hospitals, physicians, and other entities contract to provide coordinated high-quality care to a defined population of patients, reduce unnecessary spending, and share in any savings if health care expenditures decline. In risk-bearing ACOs, providers share in losses if expenditures rise.26,27 Participation in ACOs has grown rapidly since they were introduced in 2011. As of 2017, Medicare Shared Savings Program ACOs had 10.5 million enrollees, commercial ACOs had 19 million, and Medicaid ACOs had 4 million.38 Until recently, 82% of Medicare ACOs involved only shared savings, not shared losses.39 However, this is likely to change under MACRA, which allows physicians to participate in advanced alternative payment models, including risk-bearing ACOs, as an alternative to MIPS.28,40 The market may also be shifting toward risk-bearing ACOs in the private sector.41
In addition to the risk of shared losses if health care expenditures rise, the policies governing the Medicare ACO program create additional problems for academic medical centers. Because of the higher costs per case at academic centers, community-based physicians have incentives to direct routine patients to lower-cost institutions rather than to local academic centers, even though the latter have better treatment capabilities and achieve better outcomes. Yet, because of their enhanced treatment capabilities, academic centers continue to receive referrals of patients with complex needs, which is an example of “adverse selection.” The Medicare ACO program does not consider the effects of adverse selection when setting financial improvement goals. However, some commercial ACOs have implemented payment designs that better account for adverse selection. Nonetheless, many academic hospitals have chosen to participate in Medicare ACOs out of commitment to their missions and to serve the community.
Episode-based or “bundled” payment links multiple services that patients receive during an episode of care, such as hospital services, physician services, and postoperative care services. Each episode is assigned a fixed target price, which can be adjusted according to the attainment of quality and spending goals. Payment can be adjusted according to performance on quality measures (e.g., the rate of surgical site infection).42 This payment model gives hospitals an incentive to coordinate with other provider organizations, improve care transitions, and lower the cost per case.
The CMS Innovation Center has tested bundled payment models for several types of surgical episodes.43 One model, the mandatory Medicare Comprehensive Care for Joint Replacement Model, was scaled back from 67 to 34 major metropolitan areas in 2017 by federal administrators.44 Two other proposed episode-based payment programs for surgery were cancelled in 2017.45,46 However, under MACRA, episode-based payment is considered an Advanced Alternative Payment Model, offering physicians another alternative to MIPS, so episode-based payment will continue to play a role in the Medicare system.40,47 Medicare announced the first example of an Advanced Alternative Payment Model, the voluntary Bundled Payments for Care Improvement Advanced Model, in January 2018. This new policy applies to 32 types of outpatient and inpatient episodes, and uptake by both physician groups and hospitals has been robust.46,48 Private payers also are rapidly expanding episode-based payment programs.6,49 Academic centers often treat patients for whom surgery involves substantial risk as well as potential benefit; therefore, the hospitals’ assumption of responsibility for costs and outcomes in such situations involves the possibility of large, unpredictable financial losses.
Under capitation, payments for health care are based on a fixed amount per enrollee per month. In public managed care programs such as Medicare Advantage and Medicaid managed care programs, the public payer makes capitated payments to a private payer, typically a health maintenance organization (HMO). In turn, the private payer may make capitated, fee-for-service, or other types of payments to contracted providers.50–52 Capitation is primarily designed to manage spending, although contracts also can specify some incentives for quality of care.
For example, in Medicare Advantage, payers receive bonus payments when quality goals are attained.51,52 When teaching hospitals care for patients in Medicare Advantage plans, they continue to receive IME payments for those services.22 A substantial shift toward capitated payment has occurred in the Medicare and Medicaid systems in recent years. Nineteen million Medicare beneficiaries (33%) are now enrolled in Medicare Advantage plans, and 55 million Medicaid recipients are enrolled in diverse HMOs based in 47 states.50,51,53 Because of the nature of the patient populations that they serve, academic medical centers are highly vulnerable to the adequacy and quality of the acuity and quality measurement systems used in capitated payment models. Moreover, because patients with low incomes and high social needs are more likely to be enrolled in less-generous capitated payment programs such as Medicaid managed care plans, academic medical centers that treat large numbers of these patients may be at particular financial risk.
Potential Refinements to Value-Based Payment Policies
Value-based payment policies could be refined in several ways to avoid straining the finances of academic medical centers. First, Medicare policymakers could improve the risk-adjustment techniques used in value-based payment models to reduce the possibility of disproportionately penalizing hospitals that treat larger numbers of vulnerable patients. Although the Hospital Readmissions Reduction Program was recently modified to stratify hospitals according to the proportions of patients who are also eligible for Medicaid, the implications for academic centers have yet to be examined. Moreover, recent research has suggested that adding patients’ clinical and social characteristics to the policies’ original risk-adjustment procedures might be a more flexible and effective approach than stratification by Medicaid eligibility.54
Second, hospitalizations that count as outliers under the Medicare diagnosis-related group policy could be excluded from quality measures used in value-based payment programs in order to reduce the risk of penalizing teaching hospitals for taking on patients with complex conditions and an increased risk of poor outcomes. Third, Medicare could increase IME payments to hospitals to better account for adverse selection under value-based payment. Finally, major teaching hospitals and health care systems could be compared with their peers rather than with all hospitals and health care systems.
Responding to the Challenge of Value-Based Payment
Academic medical centers have an obligation — and most have a strong desire — to contribute to the national effort to make health care more affordable, providing the impetus to become as clinically and operationally efficient as possible. Moreover, with the rising cost of care and persistent questions about quality, academic centers have a responsibility and opportunity to lead by demonstrating what is possible in solving these problems.55 Innovation is part of the mission of academic medical centers. Just as they have historically been sources of advances in basic science and therapeutics, academic centers already have become sources of innovation in care delivery. Academic centers that are unable to improve performance on measures of quality and value will be at risk for reductions in the revenue that is needed to support clinical services and teaching programs.
Successes and Challenges Faced by Initiatives in Academic Health Systems
There are many examples of past successful efforts by academic health systems. An early innovation was Geisinger Health System’s ProvenCare program, established in 2006. Geisinger “guarantees” that care will adhere to evidence-based processes for surgical and medical conditions and bundles payment for an episode of care (including any complications), which it describes as a “warranty.” The program demonstrated adherence to evidence-based care processes, improvements in outcomes, and reductions in costs.56–58 Massachusetts General Hospital and Brigham and Women’s Hospital, through Partners HealthCare, began participating in the CMS Pioneer ACO demonstration project in 2011. In this endeavor, they engaged primary care practices, care managers for high-risk patients, and specialist providers to manage population health in a manner that concurrently educated trainee physicians about this emerging field.59 Virginia Mason Medical Center, among other efforts, established a dedicated outpatient clinic at which physiatrists and physical therapists assumed the initial care of patients presenting with low back pain, often seeing patients the same day. This program reduced unnecessary imaging, enabled appropriate management to begin immediately, and resulted in lower costs and fewer missed days of work.60 At the Cleveland Clinic, a clinical decision support intervention that was introduced in 2013 reduced inappropriate laboratory testing and demonstrated that certain types of alerts are more effective than others.61
Not all initiatives have succeeded, however. For example, the Oregon Health & Science University established — and 5 years later dismantled — Propel Health, a multi-institution partnership that was designed to enhance population health management but was unable to overcome multiple challenges.62
Several factors can create challenges for initiatives designed to improve quality and value in academic health systems. First, expert clinicians in multiple specialties need to reach consensus about what protocols and standards should be implemented. Such agreement can be difficult to achieve at teaching hospitals, where many patients have unusual and complex health problems and where different departments can have divergent philosophies or experience disproportionate financial effects from efforts to improve quality or lower cost. For example, primary care physicians may be encouraged to order fewer unnecessary tests and procedures, which may lower revenue for radiology and surgical departments. Even when agreement is reached on which protocols and standards should be implemented, however, independently minded academic physicians may not consistently align their practices with them.62
Second, attributing care to individual physicians, such as for accountability or pay-for-performance initiatives, can be hampered by the fact that supervising attendings change frequently on teaching services. Third, trainees lack clinical experience as well as training in efficiency principles, such as Lean, making it harder to ensure adherence to clinical quality measures and to trim unnecessary care and spending. Moreover, the influx of new trainees each academic year creates a continuous need for training in order to sustain value-based protocols and practices over the long term. Fourth, while some academic health systems are expanding to include both healthier and sicker populations, this expansion has advantages and disadvantages. On the one hand, such expansion can enable academic health systems to better manage and be accountable for the quality and costs of care by supporting the creation of population health programs and mitigating the effects of adverse selection on total-cost-of-care measures.60 On the other hand, hospitals, faculty practice groups, and community partners must then come to an agreement about how to allocate financial gains and losses.62
Nonetheless, academic medical centers also can draw on several resources and attributes that will facilitate their adaptation to value-based payment as well as ensure ongoing leadership in the development of strategies to improve the quality, outcomes, and costs of care. Faculty at academic institutions are often motivated by a desire to advance care in their fields through innovation, and promotion criteria create strong incentives to invest time and energy in advancing the science of providing high-value care. Many faculty researchers have training and experience in the measurement of care quality and health care outcomes as well as in advanced statistical analysis, skills that can be repurposed for quality improvement initiatives.63
Because trainees are still learning the norms of practice and have not yet developed notions about how systems of care should be designed, they may be receptive to such improvement efforts. Many graduates of training programs choose to join the faculty and, based on several studies of primary and hospitalist physicians, some resident physicians and recent graduates appear to exhibit higher performance on quality metrics and better outcomes than more experienced physicians do.64–66 Sophisticated electronic health record systems can support new needs for clinical documentation, facilitate the integration and analysis of data, and isolate specific sources of variation in clinical outcomes and costs.60,62 These systems also create opportunities to develop new ways of measuring the severity of illness, the complexity of patients’ social needs, and the use of resources.67 Clinicians in training today grew up using computers and are likely to see electronic health record systems as useful tools rather than as impediments to productivity.68–71 Partnerships with community organizations can help to address the social factors that commonly contribute to poor outcomes and higher costs.
Implications for the Education of Health Professionals
For academic centers to navigate the transition to value-based care, and for value-based care to become ingrained in the health care system over the long term, trainees in diverse fields will need to acquire a deep and meaningful understanding of systems design, population health, quality of care, appropriate utilization, social determinants of health, and costs. How to best incorporate these topics into training continues to be a work in progress, and many questions remain:72–75
- How can these principles be incorporated into training without drawing time and attention away from essential clinical knowledge?
- What stage in training is best to learn the principles of health systems science?
- Will trainees appreciate the urgency to understand these principles and incorporate them into their practice?
- Given the dearth of academic faculty with training in these fields, what is the best approach to ensure rapid and widespread uptake?
- What competencies in health systems science should be established?
- What accreditation or licensing bodies will ensure that these competencies are attained?
- Should the outcomes of medical education include performance on measures of quality and value?
- Where in the organizational structure of academic medical centers will activities related to population health reside?
Many institutions are starting to create curricula and hands-on learning opportunities for resident physicians, nurses, and other clinicians to engage in or lead efforts to improve quality and value. The American Medical Association has published a concise and well-written text, Health Systems Science, which includes fundamentals on systems of health care delivery, interprofessional practice, and strategies for improving quality, safety, and value.76 The High Value Practice Academic Alliance, a national, cross-institutional collaboration for high-value care, is working to disseminate innovations and training opportunities across institutions.77
Examples of specific initiatives include resident-led efforts at multiple major academic centers to reduce unnecessary test-ordering.78,79 New York University has established a Department of Population Health to lead research, education, and training activities as well as to advance health-system innovations within its own care delivery system and more broadly across the city.55 Both Penn State College of Medicine and the University of California, San Francisco, School of Medicine have integrated training in health systems science into undergraduate medical education, with classroom-based didactics and experiential learning opportunities. For example, medical students serve as patient navigators, work with patients to identify barriers to care, partner with quality-improvement initiatives, and are embedded in interprofessional clinical teams.75
At Cedars-Sinai Medical Center, residents with greater interest in health systems science are embedded in projects with the Quality Council, which oversees major quality-improvement initiatives with support from top executives. Through these projects, residents provide peer-to-peer education, work with experts on organizational guidelines, and collaborate on solving problems that range from diagnostic errors to documentation of patient treatment preferences. At Cedars-Sinai, the newly established Master in Health Delivery Science degree program is training academic faculty as well as midlevel managers in diverse hospital departments, giving them the ability to simultaneously redesign existing systems of care and to pass their newly acquired knowledge to new generations of clinicians in training.
A Call to Action at and on Behalf of Academic Medical Centers
Teaching hospitals serve their communities in diverse and essential ways, yet their financial health and, therefore, their effectiveness are likely to be strained by new value-based payment policies. Public and private insurers should evaluate the implications of these policies and should consider modifications that preserve the long-term viability, mission, and effectiveness of teaching institutions. At the same time, major academic centers must continue to embrace their responsibility to drive forward operational and research efforts to improve the quality of care, reduce unnecessary expenditures, and train future generations of clinicians to thrive in a value-based payment environment. Academic centers must become as dedicated to advancing operational and clinical efficiency as they have been to advancing the science of medicine.