++
Health care now consumes about $1 in $6 spent in the United States, up from $1 in $12 in 1960. This increase in spending is not a problem in and of itself; growth in a sector of the economy can be highly desirable if it reflects the development of valuable new technologies (eg, computers). However, in the case of health care there are good reasons to question whether increases in spending have been consistently worthwhile. One reason for concern is that much of health care spending is covered by insurance. Insurance helps ensure that high-cost, unpredictable health care is available to people when they need it. However, insurance also causes individuals to consume health care even when it might not be otherwise worthwhile for them. Moreover, because patients may not easily understand many of the complex decisions about their care, often spending may not align well with their underlying preferences. While many efforts are being made to better inform and empower patients to participate in their medical decisions, health care providers, hospitals, insurers, and policy makers must make decisions that patients cannot fully participate in and that influence patient care. Social institutions, such as licensure, accreditation, public reporting of outcomes, and market competition, may also help ensure that these parts of the health care system act more effectively as agents of patients.
++
Indeed, while there is evidence that the value of increased health care spending on average has far exceeded the cost, there is also evidence that a sizable fraction of medical interventions provide little health benefit. The United States spends more per capita on health care than any other country, but many countries achieve comparable or better population health outcomes. These cross-national differences cannot be attributed completely to differences in the health care systems across countries. Nevertheless, the efficiency with which resources are used within these health care systems is a critical concern when seeking to maximize the health outcomes given available resources. Tools such as medical cost-effectiveness analysis are often used to assess how the cost of care compares to the health benefits produced. One metric for assessing cost-effectiveness is the incremental cost-effectiveness ratio, which is calculated by dividing the change in costs resulting from an intervention by the change in life years lived, usually adjusted for quality of life. Whereas the decisions that people make about their own health suggest that people may value a 1-year increase in life expectancy at about $50,000 to $200,000, some commonly used medical interventions cost more than $1 million per life year saved. Many interventions may not produce any health benefits, regardless of costs. The emerging field of comparative-effectiveness research seeks to compare the effectiveness of health care interventions to determine when specific interventions should be utilized. While the appropriate role of costs and cost-effectiveness analysis in comparative effectiveness research is controversial, the concept that both costs and effectiveness of medical interventions are important policy concerns is widely accepted.
+++
APPROACHES TO CONTROL HEALTH CARE EXPENDITURES
++
A wide range of approaches have been proposed to control health care expenditures. These include market-based approaches such as making patients and providers more sensitive to costs by limiting what services will be covered or how much of their costs will be paid. Consumer-side market incentives at the point of care, such as copayments, have been proven to reduce expenditures in numerous contexts, including the RAND Health Insurance Experiment. However, the ability to control the spending of high-cost consumers through copayments is limited by the high economic burdens these create for these individuals and their families. This limitation is especially important because health care costs are highly concentrated; more than 90% of all health care costs are borne by the top 50% of users. Within Medicare, the top 5% of users constitute 50% of spending. Since hospitalized patients are typically among these high-cost users, consumer-side cost sharing often has little effect on hospital spending because hospitalized patients often will have reached caps on out of cost payments.
++
Another approach to controlling health care spending is to reduce payments to providers. The extent to which this is possible is limited by the need for providers to maintain their financial solvency, but it will generally reduce expenditures in the short run and place downward pressure on long-run cost growth if it is maintained. Nevertheless, in a fee-for-service environment, reducing payments to providers may cause providers to engage in “demand inducement” in which they increase the quantity of services provided to make up for lost revenue. This is one reason that many health economists believe that controlling health care costs will require paying for all of health care with fixed payments to cover all of a person’s care over a defined period of time (capitation). Indeed, the single most important cost containment approach for hospitals has been the use of prospective payment for hospital care, in which hospitals are paid a fixed amount of money for a hospitalization for a given condition. We return to these ideas in the following section. Market-based approaches remain important under capitation because competition among providers may encourage providers to be more efficient to offer greater value to patients. Even in more regulated systems such as single-payer national health insurance systems, such competition between providers may increase quality and allow reductions in costs over time.
+++
THE ORGANIZATIONAL STRUCTURE OF HOSPITALS
++
Most hospitals in the United States are not-for-profit institutions operated by staff employed by the hospital and acting under the supervision of a board that typically includes community leaders, physicians, patients, and others representing the diverse stakeholders that have interests in the operation of the hospital. To maintain their not-for-profit status and associated tax advantages, not-for-profit hospitals cannot make profits and must demonstrate evidence of the benefits they create for their community. Such community benefits can be hard to define, and even not-for-profit hospitals may be heavily influenced to make decisions that benefit hospital managers and staff more than patients. For-profit hospitals do not have the same community benefit obligations as not-for-profit hospitals and are less likely to provide types of care that are not profitable. However, the differences in services between for-profit and not-for-profit hospitals are often less than one might expect. Even government safety net hospitals, such as county hospitals, may face economic incentives to provide care for conditions that receive better insurance reimbursement. In general, payment rates in the United States are highest for private insurance, followed by Medicare, then Medicaid. Under the Emergency Medical Treatment and Active Labor Act (EMTALA), hospitals must ensure the provision of care for persons without insurance who face immediate life-threatening conditions. As an unfunded mandate, EMTALA has adverse effects on the financial well-being of some hospitals that serve low-income communities and the availability of emergency and hospital services in those communities.
++
The compensation of physicians varies across hospitals, with physicians in many for-profit and not-for-profit hospitals being paid through the professional fees that they bill, while physicians in government hospitals, academic medical centers, and hospitals owned by staff-model health maintenance organizations (HMOs) are more often paid a salary. Often, physicians working fee-for-service have unpaid institutional obligations, such as caring for uninsured patients or serving on committees, while physicians paid by salary will also have incentives for productivity. Thus, the differences in these means of paying physicians may not always be as large as they would seem.
++
Nevertheless, understanding the organizational structure of a hospital is essential to understanding how various policy changes may affect hospital care. For example, hospital managers will tend to want hospital resources to be very heavily used, while physicians may prefer that the hospital has excess capacity so that they and their patients can more easily access care. Decisions regarding the adoption of costly new technologies may have similar dynamics among physicians and hospitals. One exception is when hospitals can be paid more for care that is provided using an expensive new technology. In these cases, in areas with many hospitals, hospitals may compete with each other to attract physicians by buying costly new technologies in what has been termed a “medical arms race” (Zwanziger and Melnick, 1988). This contrasts with the typical expectation that competition will drive down costs as more efficient organizations drive less efficient ones out of business. In recent years, as payment policies have been less generous, hospital competition has been associated with lower costs.
+++
THE PAYMENT OF HOSPITALS
++
Before the 1980s, almost all hospitals were retrospectively paid fee for service for the care that they provided. This meant that when the patient received more care, for example by staying an extra day in the hospital, the hospital would receive more payment. Since rates typically covered the cost of hospital care plus some profit, hospitals had incentives to provide more care to increase profits or produce surplus revenue in excess of costs to help them achieve other parts of their mission. Not surprisingly, hospital costs grew rapidly.
++
In the 1980s, Medicare changed its reimbursement policy for hospital care to the Medicare Prospective Payment System (PPS), which paid hospitals a fixed amount for the hospital care of a specific group of conditions, which Medicare termed a “Diagnosis-Related Group” (DRG). With hospital length of stay as a major driver of costs, hospitals could make more money if they reduced length of stay to allow them to care for more patients with a fixed amount of resources, and both hospital length of stay and costs fell dramatically. PPS raised several concerns, however, including the possibility that hospitals would admit healthy patients who did not have to be hospitalized but would have short lengths of stay and be very profitable, and the possibility that hospitals would discharge patients too quickly, resulting in bad outcomes. To address these issues, Medicare established peer review organizations (PROs) to monitor for inappropriate utilization and low quality care. While there has been some evidence of inappropriate utilization and a tendency to discharge patients “sicker and quicker” (Kosecoff et al, 1990) under PPS, the PROs and subsequent utilization and quality initiatives suggest that the net cost advantages of PPS are large compared to these concerns. It should be noted that PPS has also been shown to occasionally reduce care and impair outcomes for the sicker people within any diagnostic grouping. This might be predicted because these sicker-than-average patients in a diagnostic group become less profitable under prospective payment. This may be one reason that prospective payment has never been fully prospective. For example, PPS does not provide the same reimbursement for the treatment of a heart attack regardless of whether bypass is or is not performed. To do so would be to produce financial penalties for performing bypass that likely would be harmful for patients. Similarly, PPS provides some extra “outlier” payments for patients with exceptionally long hospital length of stay relative to other patients with similar diagnoses.
++
PPS is also especially important to hospitalists because it increased the incentives for hospitals to reduce length of stay, making the potential benefits of hospitalist efforts to shorten the average length of stay more attractive to those hospitals. PPS also raised the average acuity of hospitalized patients, making the physical presence of hospitalists in the hospital more important. Thus, Medicare PPS probably contributed to the rapid growth of hospitalists. That said, there is also strong evidence that the number of hospitalists grew because they offered primary care physicians (PCPs) a way to avoid costly trips to the hospital to see an increasingly smaller number of hospitalized patients as primary care practice increasingly came to focus on preventive care. This suggests the potential for exciting new models of care in which a new type of hospitalist physician provides primary care and hospital care for a highly selected group of patients at high risk of hospitalization.
+++
EFFECTS OF HOSPITALISTS ON HOSPITAL COSTS
++
Shortly after the term hospitalist was defined in the mid-1990s, reports began to be published describing reductions in hospital length of stay and costs due to hospitalists. Reviews of the subsequent literature have largely confirmed these early findings, though not all studies have found savings. Some factors have been suggested to affect the reductions in hospital costs from hospitalists, including hospitalist experience, which may be a particular concern because of the relative youth of the field and relatively high rates of burnout and turnover among hospitalists.
++
While several of these studies rely on natural experiments in which patients are assigned to hospitalist or nonhospitalist physicians based on a predetermined call schedule, a limitation of many of these studies is that patients may be assigned nonrandomly to physicians based on patient attributes that may be difficult to control for and may predict hospital resource utilization. This is especially a concern when attempting to understand the effects of hospitalists on costs for patients who would otherwise be cared for in the hospital by their own PCP. All of the studies with a design resembling random assignment are limited to academic medical centers, where the comparison is between having a hospitalist attending who is not the patient’s PCP or a nonhospitalist attending who is also not the patient’s PCP. To the extent one believes that receiving hospital care from a physician who has an ongoing relationship to the patient (such as the patient’s PCP) may be valuable, this limitation of the literature is a major one. There are also relatively few studies that characterize well how hospitalists may affect resource utilization after hospital discharge.
++
There is also a literature suggesting that hospitalists may produce resource savings when they serve as comanaging physicians with surgical or medical specialists. Such models are supported by the fact that both the hospitalist and the subspecialist can bill for their services in most settings. Whether such models would be economically attractive in settings in which both physicians could not bill for their time is not clear.
++
It is worth noting that many studies of the effects of hospitalists on resource utilization focus on length of stay. This is typically appropriate because length of stay is the largest determinant of hospital costs; most of the costs of hospital care are fixed costs and most hospitals operate close to capacity. However, in cases in which hospitals are not close to capacity and costs are fixed, or if staffing is easily varied, reductions in length of stay may not have the same implications for the cost per case as they more typically would. Correctly estimating the costs of care and savings or costs from specific programs in the context of hospitals is extremely difficult. Activity-based accounting systems, which attribute all the costs in the hospital to specific activities of patient care that can then be allocated across patients, are considered state of the art in estimating hospital costs. However, these systems can be applied in many different ways across institutions that require that the resulting estimates of costs be considered with caution.
+++
SYSTEM EFFECTS OF HOSPITALISTS
++
In addition to the direct effects of hospitalists on costs through the patient care they provide, hospitalists may also have effects on costs by helping to develop systems that address important clinical and economic needs of hospitals. These may include roles in designing and implementing utilization review and bed flow positions, health information system innovations such as computerized order entry, and efforts to reduce costly errors or improve adherence by developing quality indicators that are increasingly tied to economic incentives. Numerous studies suggest that medical errors are costly and that costly practices that might be reduced by improved information systems are common in hospital care. A growing literature also suggests that specific strategies to reduce errors and/or influence practice patterns through health information systems may produce cost savings. Such findings are neither universal nor completely compelling at this point, but it is likely that better implementation of these systems may produce more compelling results in the future, and hospitalists are very well situated to play such a role in that implementation. Economic incentives to reduce readmission are another major example of an area in which hospitalists are well situated to help a hospital respond to incentives to reduce resource use. In teaching hospitals, hospitalists may also improve the education of house staff in cost-effective approaches to patient care, which would reduce resource use by those residents even when they are no longer working with hospitalists. Even in settings in which PCPs continue to come to the hospital to direct patient care on a daily basis, hospitalists may decrease length of stay and costs by providing needed clinical expertise at times when the PCP cannot be present.
+++
ACCOUNTABLE CARE ORGANIZATIONS AND THE FUTURE OF HOSPITALISTS
++
While retrospective reimbursement of health care provides attractive incentives for the provision of needed care, it can lead to excess provision of care. As a result, there is great interest in moving toward models in which payment is made prospectively for the care of an individual over a defined period of time (Fisher et al, 2007). Medicare PPS was a step in that direction, but it did not include costs incurred after discharge, potentially increasing the incentive for early discharge and increased rates of readmission. Similarly, the fact that PPS does not include physician fees in the payment for hospital care can increase incentives for consultation, the use of comanagement models, and even extended hospital length of stay from the physician perspective.
++
The Patient Protection and Affordable Care Act (PPACA) (United States. Congress. House., 2010) includes a spectrum of strategies to address these potentially perverse incentives, with penalties for rapid readmissions as a first step and payment of a fixed fee to care for a patient over a defined period of time at the other end of the spectrum. Under PPACA, Medicare is already charged to implement reductions in inpatient payment rates for hospitals that exceed targeted readmission rates. Other efforts under PPACA included the Medicare Acute Care Episode (ACE) demonstration project, which sought to test the effects on resource use and outcomes of providing a single payment to hospitals to cover all the hospital, physician, and other post-acute care costs for the 30 days before and 30 days after an acute episode of illness. In addition to requiring study of the effects of such “bundled payment” strategies, PPACA also mandates that Medicare establish infrastructure to begin to establish accountable care organizations (ACOs) that will assume full responsibility for the care of patients over a defined period of time in exchange for a fixed payment. The economic advantages of such a payment system in terms of cost reduction, integration of care, and avoidance of the perverse incentives for excess use created by retrospective fee-for-service systems are suggested by the successes of some staff model HMOs in these domains. The early results of Medicare’s ACO efforts have been mixed; it remains to be seen whether the successes of systems based on capitation can reduce costs while promoting quality of care.
++
Payment reforms under PPACA may have important implications for hospitalists. Already, hospitals are increasingly looking to hospitalists and others to develop new systems to reduce readmissions. Bundled payment systems that combine hospital and professional fees could also change the incentives for comanagement models by eliminating the ability to obtain additional revenue by having multiple physicians treat a hospitalized patient on a given day. Bundled payments that include professional fees might also lead to complex negotiations between hospitalists and hospitals about payment for hospitalists. Full capitation models would reduce the extent to which Medicare physician fee schedules drive physician reimbursement and might lead to further changes in the relative earnings of specialists and generalists. Although the effects of such changes are difficult to predict, it seems likely that specialty reimbursement would be particularly reduced. PCPs might also find themselves facing increasing competition from nonphysician providers who might provide elements of primary care at lower cost. Together, these pressures might drive increasing numbers of young physicians into Hospital Medicine, which would probably put downward pressures on hospitalist earnings. Increases in health insurance coverage due to health care reform and the aging of the baby boomers might increase demand, somewhat offsetting these forces. In addition to these macro-level effects on hospitalist earnings, individual hospitalists may find their earnings strongly affected by the extent to which they can provide evidence of their ability to reduce costs and improve outcomes.
+++
THE NEW VALUE CASE FOR HOSPITALISTS
++
To date, hospitalists have largely sought to establish their value case based on their ability to reduce inpatient costs, generate inpatient revenue, and improve hospital performance on established inpatient quality indicators that may be tied to increased patient demand or that rewards under pay for performance. However, the expected changes in reimbursement toward greater bundling and capitation will force hospitalists to build value cases based on new metrics, such as the total cost of care, reductions in readmissions, and improved long-term outcomes. Efforts to reduce cost of care are likely to take many forms, including programs to reduce readmission, skilled nursing facility use, and emergency department use after discharge, disease management programs, and palliative care programs that may reduce costs at the end of life. The history of similar efforts suggests that programs in these areas are likely to vary in their efficacy. For example, the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT), which was intended to improve care at the end of life for patients at high risk of death, did not improve care, outcomes, or resource use. More recently, the results of the Massachusetts General Hospital Care Management for High Cost Beneficiaries Project suggested savings of $3600 per year for an aggressive case management system designed to support the care of high-cost Medicare beneficiaries. These studies will likely be followed by many more in the coming years that will aim to test new strategies to improve care. Many of these strategies are likely to include either public reporting of outcomes or financial incentives for hospitals to meet specified quality indicators or economic objectives. To date, public reporting and pay-for-performance programs have shown mixed results, again underscoring the importance of carefully evaluating these programs as they are implemented. However, such evaluations have encountered many challenges, including measurement issues, and the potential to more easily improve outcomes by changing patient mix rather than truly improving care. Under PPACA, a Patient Centered Outcomes Research Institute was established with a mission to assess the comparative effectiveness of alternative health care strategies, including new models of care. In addition, the Center for Medicare and Medicaid Services (CMS) created a CMS Innovation Center to expand its ability to develop and test new models of care. Hospitalists play important roles in many of these models of care.