Doing Better by Doing Less: Approaches to Tackle Overuse of Services

Timely Analysis of Immediate Issues January 2013 Robert A. Berenson and Elizabeth Docteur

This analysis presents what we fewer admissions, and There is variation in the level of know about the provision of shorter lengths of stay in inappropriate use by type of service, medically inappropriate and comparison with many developed and it is evident that some services unnecessary services that drive up countries.2 According to many are subject to a great deal of spending without indicators of preventive service use, overprovision. A recent review of making a positive impact on U.S. performance is only average or the research literature found rates at ’ health outcomes. It also below average.3 But there are which particular therapeutic describes approaches that have certain areas, including some procedures, tests, or medications already been used to address this relatively high-tech and high-cost were performed or prescribed when issue—with limited success. We services such as imaging and clinically inappropriate ranged from suggest that broader payment cardiac surgery, where the United a low of 1 percent to a high of 89 reforms are needed to minimize States appears among the most percent.7 For example, a 2007 study incentives to overdiagnose and prolific users.4 found that 60.8 percent of colon overtreat and to better support the cancer screenings undertaken were other approaches. Extrapolating from studies focusing medically inappropriate, while a on particular conditions or services, 2005 study found that 27 percent of Introduction some analysts have estimated that as prescriptions for much as a third of U.S. health care low back pain did not meet The United States is an outlier in spending is unnecessary and threshold standards for appropriate terms of its per capita spending on 5 wasteful. This estimate includes service use. health care, surpassing other the provision of medical services developed countries by a and the prescription of While any overuse of services considerable margin. High prices that are medically inappropriate—in drives up spending, inappropriate and fees in the U.S. health sector other words, health care from which service use is particularly important have been identified as important the derives no medical because it has the potential to harm drivers of these spending variations. benefit or for which the potential patients. One in every four patients The complexity and fragmentation harms exceed the potential benefits. admitted to the hospital is of our system also plays a role in More than a quarter of all wasteful prescribed an inappropriate generating higher-than-average spending in health care—an , sometimes leading to administrative costs and in creating estimated $210 billion out of $765 adverse drug reactions that are an environment in which waste, billion in wasteful spending in responsible for 20 percent of 1 fraud, and abuse can thrive. 2009—is attributed to overuse of inpatient deaths.8 Other examples of services that are medically harmful overuse include radiation The volume and mix of health care inappropriate or otherwise exposure from imaging scans such services provided in the United unnecessary, which includes as CT scans, elective C-sections States also play a role in explaining services that are provided more performed for convenience, and higher health spending, although the frequently than warranted and prescribing aggressive treatment story is complicated. By many higher-cost services that are no options to those with terminal measures of service volume, the better than lower-cost alternatives.6 illnesses without disclosing the United States is not an outlier. In likely futility of such interventions. fact, we have fewer physician visits,

Reducing the overuse of just seven What do we know about Based upon their findings, services known to be subject to high inappropriate service Korenstein and colleagues rates of inappropriate use could save concluded that inappropriate use is between $33 billion and $62 billion use? often a problem for the services 9 annually. Although inappropriate service use included in the published articles, has been studied by researchers although there is wide variation in The potential of achieving cost since at least the late 1970s, rates of overuse documented by the reductions while substantially measuring it is technically research. At the same time, for the improving health care quality and challenging and costly. As a result, vast majority of procedures, tests, outcomes can only tantalize policy- there is still a lot we do not know and medications in use today, no makers in this era of rapidly rising about its incidence and the reasons studies have assessed the extent to health spending and very strong for its persistence. Despite the which they are overused in practice. constraints on the financing side. limited evidence base, the evidence Because of limitations in the scope Meanwhile, current support for that we do have suggests that there of the research, there are important experimentation with new forms of is a great deal of costly overuse of gaps in our understanding that limit health service delivery and services that has defied efforts to opportunities to reduce health care concurrent changes in payment and address the problem. The problem is spending without adversely regulatory oversight make the difficult to study for reasons affecting public health. present a most opportune moment in discussed below, and those same which to consider the prospects for The review demonstrated that new reasons also make implementing ensuring that those changes help, and costly procedures were policies to tackle the problem rather than hinder, efforts to reduce particularly underrepresented in the difficult. 11 inappropriate service use. research literature. In part, this What has the research shown? dearth is explained by shortfalls in This analysis focuses on what we comparative effectiveness research know about the provision of After reviewing the research for many services, particularly new medically inappropriate and other literature on inappropriate services, ones. Filling these gaps in the unnecessary services, which drive Korenstein and colleagues evidence base would need to be up health spending without making uncovered 172 articles measuring prioritized if this research is to serve a positive impact on the health overuse that were published as a resource for targeting potential outcomes of patients. We begin by between 1978 and 2009 and met opportunities for enhancing the reviewing the literature on the minimum quality standards in terms cost-efficiency of spending on 10 extent of medically inappropriate of methodology and other criteria. health care services. overuse, continue by describing Of these, 53 articles addressed approaches used to tackle therapeutic procedures, 38 Certain services, including inappropriate service use, and then concerned diagnostic tests, and 81 antibiotic use and several coronary discuss the implications of current pertained to medications. Within procedures, have been subject to and proposed provider payment each area, a relatively small number repeated study over time. For some methods on the provision of of procedures, tests, and medicines of these services, such as carotid inappropriate services and in had been studied. For example, 59 endarterectomy, the rate of spurring excess volume of services. studies (more than one-third of the inappropriate use has declined 12 In the review of payment total) addressed antibiotics for considerably. In the case of approaches, we discuss upper respiratory tract infections, 17 inappropriate antibiotic use, there is opportunities to modify the various studies looked at coronary evidence of reductions in overuse, payment methods to address the angiography, and 7 studies but a high level of overuse persists, provision of medically inappropriate investigated use of upper despite significant outreach and services and overuse of unnecessary endoscopy. In sum, only 18 unique education efforts. This finding services, as more broadly construed. therapeutic procedures, 24 indicates that publication of diagnostic tests, and 13 medications research documenting overuse is not were evaluated in terms of the necessarily sufficient to ensure incidence of inappropriate use. changes in practice patterns.

Timely Analysis of Immediate Health Policy Issues 2

Why do we know so little about this particularly ones that are relatively claims data, continue to limit its critical aspect of health care new or in use for new indications, broad application. quality? and mean that practice standards may reflect differences in expert Do geographic areas where service Korenstein and colleagues pointed opinion where the science is use is relatively high have more to a number of reasons why unresolved. These standards require inappropriate services? inappropriate use of services is updating to be consistent with the In a recently published review of relatively less studied in comparison 15 current state of knowledge. For the literature, Keyhani and with underuse of medically example, recent studies determined appropriate services.13 Measuring colleagues identified only five that the accepted practice of raising studies addressing the question of overuse via the assessment of the end-stage renal disease patients’ rate of provision of medically whether geographic areas with high blood counts to levels that are rates of service use had higher rates inappropriate services is technically normal in the general population challenging. It requires defining of inappropriate care. They lead to serious, unforeseen events concluded that the limited available circumstances in which a particular such as heart attacks.16 service is inappropriate for use with evidence fails to support the patients with certain characteristics Further, for every indication in hypothesis that inappropriate use of or under certain circumstances, which use of a service can either be procedures is a major factor explaining geographic variations in generally through a process that labeled clearly appropriate or 18 involves reviewing relevant inappropriate, there are many more intensity or cost of care. Most of scientific evidence and developing for which the benefits are unknown, the studies reviewed found that consensus of clinical experts to unclear, or uncertain, leaving geographic areas with high rates of produce guidelines and performance physicians and patients to navigate a service use and areas with relatively standards. Furthermore, because vast gray area by relying on some low rates of service use had similar administrative data sets lack the combination of judgment, instinct, rates of inappropriate use of type of information and level of experience, and tradition. This gray services. specificity required, assessing area stymies both health services Work by researchers affiliated with inappropriateness generally requires researchers in their efforts to assess the RAND Corporation has been resource-intensive review of the prevalence of inappropriateness very prominent in establishing the medical records. In addition to these and health policy-makers in their information base in this area.19 technical considerations, efforts to target the problem for Chassin and colleagues found that identification of inappropriate use is reduction. We also lack research on geographic areas where Medicare controversial and sensitive, in part whether overuse of services varies beneficiaries were 2.3 times more because there are often stakeholders by race or ethnicity, and we know likely to receive a coronary with strong economic interest in little about the impact the overuse angiography had only modestly defending a particular procedure or of services has on health care higher rates of inappropriateness medication. This is the case with disparities. than was found in areas where use vertebroplasty, a surgical procedure of the procedure was dramatically on the spine that continues to be Although the issue of inappropriate overuse received significant lower (82 percent versus 71 widely performed despite findings percent). RAND work has also questioning its medical benefit.14 research attention in the 1980s and 1990s, the challenges inherent in found that areas of the United States For many services, if not most, this type of research have led to a did not have higher rates of measuring the extent to which they shift toward more population-based inappropriate service use than areas are subject to inappropriate use has studies, which demonstrate of Canada, despite having higher rates of provision of certain not been attempted because of substantial geographic variations in 20 limited research dollars to support service use without differences in services. 17 the effort, even though the pay-off outcomes. While there has been While further research to explore might be far greater than the cost. some recent renewal of interest in the relationship between high Prominent gaps in scientific appropriateness research, its volume and the rate of inappropriate knowledge make the development challenges, particularly the need to service use is needed, the available of evidence-based standards review detailed clinical information evidence has certain implications difficult for many types of services, that is unavailable in administrative for policy-makers. If it is true that

Timely Analysis of Immediate Health Policy Issues 3

the rate of inappropriate service efforts to impose standards of screening colonoscopies in people provision is similar across areas practice through over age 74. Because there are with different overall levels of plans’ utilization review processes specific procedure codes that service use, it is likely that the or even through public program distinguish between screening and volume differential can be coverage policies have met strong diagnostic colonoscopies, it is explained largely as a difference in resistance from patients, medical possible from claims data to use of so-called discretionary professionals, and the drug and calculate a rate of inappropriate services—services with relatively medical device industry screening colonoscopies for those low benefit-to-risk ratios (e.g., stakeholders. over 74—anything over zero is elective knee and hip replacements inappropriate. Of course, this all-or- for arthritis) or with indications for Approaches used to nothing measure assumes that use that are uncertain, including tackle inappropriateness colonoscopies are coded correctly. services for which provider and patient preferences play a decisive Efforts of policy-makers to tackle Most of the time, the clinical role. Thus, policy-makers could inappropriate service use have, to information needed to establish an reduce incentives for provision of date, come up short. Below we appropriateness measure is not both inappropriate and discretionary discuss general approaches that available from claims data. This services, as well as duplicative have been tried, including shortfall is important in that quality services reflecting poor care monitoring, reporting, and efforts to measurement is central to most coordination or case management, affect demand and supply, and we efforts to set baseline standards and to increase efficient use of resources consider why these efforts failed. to motivate and ensure and reduce cost pressure. improvement in health care quality, Efforts to monitor the rates of through approaches such as Why does overuse of services provision of inappropriate services informing patient choice and persist? Although measurement of health establishing administrative or regulatory rewards and sanctions. Possible reasons why overuse of care quality at the provider and inappropriate or unnecessary health plan levels has become an Efforts to educate physicians to services persists in spite of some increasingly common part of health comply with evidence of initiatives by policy-makers and care in the United States today, appropriateness health care administrators to target routine measurement of the rates of the problem include lags in the provision of inappropriate services Medical specialty societies, in an diffusion of new scientific evidence; is not a focus of these efforts. By effort to fulfill their role of incentives that promote overuse of one estimation, only four of the 39 educating their members to stay services, including financial quality measures in the 2011 current with the body of knowledge incentives (e.g., payment, Healthcare Effectiveness Data and the public expects professionals to ownership of equipment or Information Set (HEDIS) of quality have at their command, have long practices), legal incentives (e.g., measures that are commonly used devoted resources to develop malpractice), and administrative by payers and other parties to assess evidence-based care guidelines, care quality explicitly address including identifying specific incentives (e.g., quality standards); 21 limitations in available measures overuse. clinical indications for performing diagnostic and treatment and public reporting of information One reason is that the determination on appropriate use; patient demand, interventions. Some, including the of appropriateness requires review American College of Physicians, the including demand that is fuelled by of clinical information, which direct-to-consumer advertising of American College of Cardiology, generally cannot be ascertained and the American College of pharmaceuticals; professional and from claims data. The result is that cultural biases that favor employing Radiology, go further to identify the few appropriateness measures clear contra-indications to a treatment even if there is doubt as that have been adopted do not to its potential benefits; and the performance of interventions—that require clinical information but is, to specify inappropriate use. default standard in health care that rather can rely on demographic an intervention is presumed information that is available on Typically, the process of setting effective until proven not to be in claims. For example, recent appropriateness criteria begins with particular circumstances. Past guidelines call for cessation of a thorough literature review. But

Timely Analysis of Immediate Health Policy Issues 4

since the available literature is of Internal Medicine Foundation circumstances (e.g., usually not definitive in addressing and Consumer Reports in the first immunizations). all of the clinical variations that phase of the Choosing Wisely guidelines have to address, the campaign.22 Through this initiative, Implementing value-based benefit review is followed by an expert each society has developed a list of designs that assign higher cost- consensus process to set the actual five tests, treatments, and services sharing rates to services of lower guidelines. Other entities, such as that are commonly used in that value has proved challenging. The plans and specialty and for which there was a value of an MRI, for example, authorities like the U.S. Preventive judgment that the use was often not depends on the specific clinical Services Task Force, also develop necessary for patients. Examples circumstances in which the scan is evidence-based guidelines, which include unnecessary CT scans, used. So it is one thing to believe may provide advice about antibiotics for an acute sinus that in aggregate there are many inappropriate use. infection, and stress imaging tests inappropriate MRIs being during annual checkups. All of performed, but hard to define a Different users of evidence may these services were deemed to be policy to reduce inappropriate use come to somewhat different provided at times when they offer as there are also many MRIs that are conclusions about the implications no benefit to the patient or may appropriate and even essential. of the research findings on actually cause harm. It remains to There is no easy way to appropriateness of specific services, be seen if mounting a campaign to operationalize a priori a higher co- especially when the literature is not promote discussions between payment (e.g., for the weekend definitive. Payers, for example, are physicians and patients about the athlete with knee discomfort), more likely to interpret the findings merits of particular interventions, because of the discretion needed for more strictly than providers, rather than the more traditional individual cases. requiring a higher threshold of approach of promulgating The relative sparseness of patient evidence to support use of a guidelines for physician use, proves diagnostic or treatment intervention. education and information more successful in reducing campaigns focusing on the problem There are many facets involved in a inappropriate service use. of overuse may reflect both market comprehensive appropriateness Efforts to affect demand for failures and conflicts of interest, determination. For example, it may inappropriate services although recent investments in areas not be possible to determine how such as comparative effectiveness frequently a screening test should be Prior to the recent launch of the research and informed shared performed, even if the screening test Choosing Wisely campaign, efforts decision-making may offer promise has been proven useful and to activate patients to protect for the medium- and long-term appropriate. The result of these and themselves against overuse have future. However, it may require other considerations is that there been negligible, with the exception considerable effort to convince may be variation in criteria and in of instituting cost-sharing patients that more is not always guidance about what constitutes arrangements that give patients better when it comes to service use, inappropriate, as opposed to incentives to limit their use of as recent controversies relating to uncertain, use. services. After studies of the impact new and more restrictive guidelines of increased across-the-board cost for appropriate use of breast and Although physicians have a duty to sharing showed that patients tended prostate cancer screenings have adhere to the professional standards to reduce use of both appropriate demonstrated. that evidence-based practice and inappropriate care,23 interest has guidelines help establish, it is clear grown in moving to value-based Efforts to use administrative levers from the literature that benefit designs in insurance to reduce inappropriate services professionalism and the efforts of policies. Value-based benefit The same methodological problems professional societies have not designs increase cost-sharing for succeeded in substantially reducing that make it difficult to study services of uncertain benefits and appropriateness also make it the provision of inappropriate those prone to overuse (e.g., services. To increase attention to the difficult to target policy remedies to imaging), while reducing cost- this problem, particularly those that issue, nine specialty societies sharing for services that are viewed recently joined the American Board depend on defining standards and as high value under most patient overseeing compliance with

Timely Analysis of Immediate Health Policy Issues 5

evidence-based, clinical practice are uncertain or are believed to be circumstances—to influence clinical guidelines or criteria derived from minimal. Following reports of decisions and support efficient such guidelines. private health plans’ use of delivery that avoids overuse and radiology benefit managers to reduces the use of services offering Prior authorization, a process review and approve requests for low benefits, relative to risks, rather through which providers seek advanced imaging, the Medicare than on approaches that rely on advance approval from a payer that Payment Advisory Commission measurement of inappropriate the service will be paid for when recently recommended the targeted services or enforcing adherence to provided to the patient, is the application of prior authorization in standards. administrative approach most Medicare for office-based referrals widely used by payers to try to for advanced imaging.27 How payment methods reduce the provision of affect overuse inappropriate services. The Given the administrative complexity approach tries to address individual involved, the changing clinical Although research on the effects of patients’ specific, clinical evidence of what works and when, payment methods on volume of circumstances, usually not through and the large amount of uncertainty services has demonstrated that fee- full review of medical records but in the gray area, prior for-service tends to incentivize with review of the clinical authorization’s role in reducing more use of services than does information most relevant to the inappropriate overuse is somewhat capitation,28 the recent research appropriateness of the clinical limited. Yet this form of utilization literature is surprisingly sparse on intervention requested.24 Based management can certainly the question of how payment upon some early reports of success complement other approaches methods affect the provision of when used to assess the addressing overuse. inappropriate services. appropriateness of a patient’s proposed hospitalization,25 prior In summary, the various approaches One recent study found evidence to authorization was expanded to described above either have not suggest that physicians who were outpatient surgical procedures, and been used or have been used paid on the basis of capitation were then again to routine ambulatory without much impact on reducing less likely to indicate that they care referrals from one physician to overutilization. Few quality would provide discretionary another for fairly routine evaluation measures exist to measure overuse, services, as compared with 26 while patient education and physicians who were paid fee-for- and management services. 29 However, its pervasive, often information campaigns focusing on service. This study’s findings are intrusive application made it ripe for the problem of overuse have been consistent with the broad criticism by physicians and patients, half-hearted at best. Meanwhile, perceptions that fee-for-service contributing to the managed care professional societies’ efforts to rewards not only more service use backlash. Health plans have educate their members to reduce but also excessive volume of retrenched and now apply prior overuse are still fairly nascent and services, including both authorization more selectively. have yet to demonstrate a unnecessary and inappropriate Common applications include significant impact—to some extent services, and are supported by elective surgery, referral for because of the power of fee-for- observed anecdotal examples of real advanced imaging (such as MRI and service incentives as described medical practice phenomena. For CT scans), and as part of the below. Prior authorization has been example, the abusive overdosing of management of pharmaceutical successful to limit overuse for the anti-anemia drug erythropoietin formularies. particular clinical services but to raise blood counts took place becomes intrusive and counter- under generous fee-for-service When applied selectively to high- productive when used outside of its payments. The practice ceased cost, discretionary services where sweet spot. immediately when the drugs were objective information can be bundled into the dialysis payment, reviewed by qualified third parties, This reality suggests the need to because payments were fixed prior authorization can play a role in rely more on payment incentives— regardless of whether and how reducing inappropriate services, combined with investment in much of the erythropoietin drugs although with less success for production and dissemination of were used.30 services for which actual benefits evidence on what works under what

Timely Analysis of Immediate Health Policy Issues 6

Studies that have attempted to replaced. However, how physicians issue brief on the topic, “Expect the establish a relationship between respond to fee-for-service incentives Unexpected? Physician Responses payment method and service use, depends on whether or not a to Payment Changes.” From their whether inappropriate or not, are particular service is included for reading of the accumulating challenging because many payment in the fee schedule and on evidence, the authors argue that accompanying characteristics of a how profitable particular services cutting fees leads to higher payment method impact clinical are. A growing body of evidence utilization when the targeted decision-making.31 Of course the has documented excesses associated services account for a large share of generosity of the payment could with current fee-for-service practice, physician income.35 However, if the influence provider behavior. For including abusing physician self- services involved do not account for example, under capitation, a low referral of imaging and other substantial income, physicians payment level considered noninvasive office testing, selecting reduce their output of the services inadequate by physicians might lead the highest cost (and most whose fees were cut, plausibly to stinting on services and altered profitable) intervention for reducing both appropriate and referral patterns, whereas higher treatment of prostate cancer, and inappropriate services.i payments might not. But the context performing inappropriate major in which the payment method spine surgery on patients with In the Medicare Fee Schedule, applies should also strongly affect chronic low back pain.32 Again, payments for services such as test clinician decision-making; the while the literature does not actually and imaging interpretations and particular culture of the provider document that inappropriate minor procedures often far exceed organization in which the clinician services occur more commonly in the resource costs, making them practices can cause any particular fee-for-service than in capitated or highly profitable and, therefore, payment method to behave salaried practice, it is clear that at provided to excess. Other services differently. least some physician self-referral and patient care activities that are behavior would be financially self- uncompensated or relatively poorly As the following discussion makes defeating without fee-for-service. paid, such as engaging in care clear, provider responses to payment incentives can be quite The precise fee levels for specific i In the Deficit Reduction Act of 2005, nuanced. Understanding those services can also play a role in Congress decided to pay physician practices nuances can help design payment incentivizing overuse of particular for imaging services no more than what is paid approaches that would address the services. Longstanding research has to hospital outpatient departments. Fees were problem of unneeded services and shown that physicians respond to reduced significantly, as much as 40 percent possibly the provision of reductions in fees by increasing the for a common MRI of the brain. As the inappropriate services—especially services affected by the payment Government Accountability Office documented in the first year with the reduced the provision of care in the gray reductions. In the context of the fees, the approach generated first-year savings area that lacks evidence-based Medicare Fee Schedule, CMS of 13 percent for the affected imaging services, guidelines and prior authorization actuaries have calculated a largely from the direct reduction in prices paid. rules. There, payment incentives behavioral offset to partly account (See: Trends in Fees, Utilization, and can play an important role in for expected volume increases that Expenditures for Imaging Services before and influencing decisions, for better or physicians generate in response to after Implementation of the Deficit Reduction worse. We next review some basic fee reductions or freezes.33 Act of 2005. Washington: Government issues in how commonly used and However, other research, some Accountability Office, 2008, www.gao.gov/products/GAO-08-1102R.) The proposed payment approaches can quite recent, suggests that volume of the affected services continued to affect the volume of services in physicians in fact respond to fees rise but at a lower rate than in the most recent general and inappropriate overuse in more like other economic actors do. prior years. There was no behavioral response particular. That is, if a service becomes less to make up for pay reductions with volume. profitable, the incentive to produce Now, years later, the volume of imaging Fee-for-service it declines.34 services paid for under the Medicare fee schedule has flattened out, albeit still at a high There is a growing consensus that In short, the physician response to level with many, presumably unneeded, fee-for-service represents payment fee-for-service payment levels is inappropriate services. The point is that even for volume, regardless of actually not so simple, as well fee-for-service can be modified to reduce appropriateness, and needs to be summarized in the title of a recent provision of volume, some of which likely represents inappropriate overuse.

Timely Analysis of Immediate Health Policy Issues 7

planning activities and phone a large number of efficiently payment level does one have to consultations, are not prone to provided episodes. worry about distribution of overuse. These distortions are not surpluses to the various parties inevitable, although there are With episode-based payment, the providing care in the bundled inherent practical and political behavioral response differs for episodes. challenges in getting administered discretionary and nondiscretionary prices correct. services. As a general rule, The reason for the virtual approach providers cannot induce demand for is that maintenance of traditional Episodes and bundled episodes maternity care or treatment of a fee-for-service claims and payments fractured femur—they are is viewed by many as operationally The episode-based payment nondiscretionary. Yet, providers can and politically easier for providers approach involves a single payment and do induce demand for to manage and accept. However, covering all services provided over discretionary services, such as those this approach may have negative a defined period of time. In this found in the appropriateness consequences on provider behavior discussion, we distinguish episodes literature review cited earlier, to generate inappropriate services. from bundled episodes, since including joint replacements, heart bundling implies merging payment and large vessel procedures, spine The objective of bundling is to streams that had previously been surgery, and deliveries by Cesarean break down provider silos to paid separately to different section.37 promote more cooperation and providers. In the case of Medicare’s coordination among the various diagnosis-related group (DRG) The concern is that with episode- providers that deliver care. A major payments, which are a form of based payment, discretionary challenge in this approach is that episode-based payment, the period services will increase in volume— bundling across providers is of time for a hospitalization is from sometimes representing operationally difficult.38 Indeed, one to three days prior to admission inappropriate overuse—offsetting many consider it much more through the hospital discharge. In the savings achieved by challenging administratively than Maryland, which runs an all-payer internalizing the costs of producing capitation. rate-setting program for , the services to the provider. In the period of time covered by a short, some inappropriate services Our interest here is in how bundled single payment to the hospital has will decrease and others—the episodes address the problem of been extended to 30 days post- interventions that create the inappropriate overuse of services, discharge for many diagnoses.36 episode—might increase. particularly discretionary Another common example of an procedures. It is notable that some episode payment is the global Bundled episodes. With bundled analysts have not even raised the payment made to surgeons for episode-based payments, where the concern that bundled episodes have various payment streams to fee-for-service-like incentives for major surgery. The single fee-for- 39 service payment covers the cost of different providers are merged into generating unneeded services. the actual operation and routine one, the bundling can be actual (i.e., There is an apparent assumption post-operative care for up to 90 a single payment is made to a that incentives for overprovision of days after surgery. recipient provider who then is interventions when paid with responsible for distributing it to the bundled episodes (i.e., up-coding to The theoretical merit of paying for constituent providers who provide claim payment for a more complex an episode of care is that the services inside the bundled episode) episode) are no worse than the payment amount is fixed, regardless or it can be virtual (i.e., the separate incentives to prescribe unnecessary of the quantity and mix of services payments continue but with an services within fee-for-service. For actually provided inside the episode accounting of how much the actual example, a recent projection of period. The costs of care are payment distributions varied from substantial Medicare savings under internalized to the providers, who the target payment amount for the a regime of bundled episode essentially are at financial risk for bundled episode, as the basis for payments did not consider the spending inside the episode, though determining surpluses or deficits, possibility that the volume of exceptions are made for outlier which become the responsibility of bundles could be different from that cases. This means that the basic the bundled entity). Only if costs are under the baseline, fee-for-service incentive for providers is to provide less than the bundled episode use patterns.40

Timely Analysis of Immediate Health Policy Issues 8

Although there is little evidence one those changes involve a decrease or is measuring the rate of stent way or the other, given the rare increase in inappropriate services placement to coronary application of bundled episode would need to be separately catheterizations to try to identify payments to date, a plausible assessed through medical record inappropriate overuse of coronary argument can be advanced that review. Further, it is possible that artery stent placement. In this case, bringing providers together to some of the approaches cited the denominator in calculating collaborate on care for specific earlier, such as requiring prior comparative rates is not the acute care episodes could actually authorization, adherence to population served (information that increase volume to levels that could evidence-based appropriateness is not available from claims data), counteract any spending reductions criteria, or shared decision-making but rather the population of patients from the new internalized incentives with patient decision aids43 could be who have undergone a cardiac to reduce the cost of the bundled used to constrain further overuse of catheterization (data that is derived episode. Hospitals and specialists the bundled episodes for which from claims). Yet, this creative can brand and market service lines payment is claimed. ratio-based attempt to identify with the hope of attracting patients overuse is the exception that proves from other providers and creating Pay-for-performance the rule. Namely, in the absence of a new demand for the service line As is true with other types of quality population denominator, product. Such a strategy produces problems (e.g., underuse) the performance rates as the foundation non-price competition that raises potential for using pay-for- for measuring appropriate provision costs by increasing demand for performance methods to discourage of services generally are not technically oriented interventions, the provision of inappropriate and available. Further, even if there thereby contributing to a “medical were a population denominator, it 41 unnecessary services is limited by arms race.” In the last decade, accountability issues. As noted would still be necessary to do case- hospitals have done just this— above, there are very few measures mix adjustment for the population’s developing and marketing profitable of appropriateness in current use. health status, another challenging service lines by closely affiliating This is not for lack of interest. requirement. with specialist physicians essential Rather, there are inherent to hospitals’ service line products.42 Also, without a population limitations in being able to develop denominator, one could still try to Ultimately, whether bundled valid measures relying on measure the rate of performance of episode payments raise the volume administrative data sources in a fee- particular services in relation to of inappropriate services even more for-service payment environment. clinical indications for that patient. than under straight fee-for-service For the most part, individual But as noted earlier, claims data will be determined empirically. The patients are not assigned or formally used to assess performance Innovation Center at CMS and some associated with particular practices generally lack the clinical private health insurers have begun or delivery systems, so there is no information needed to permit an experiments with bundled episodes, relevant population denominator on accurate assessment of mostly involving hospitalization for which to establish norms for rates of appropriateness. an acute condition involving a services provided. procedure. Those tests will help tell The challenge of accounting for There may be some ability to relate variations in provision of us whether the incentives of rates of one service to another. For bundled episodes reduce discretionary services was well example, the Maryland Health captured in an article reviewing the discretionary services within Services Cost Review Commission episodes and whether they increase difficulties of basing payment ii policy on measurement of spending the number of episodes. Whether and health outcomes. The article supplies as physicians and the hospital ii In 2009, CMS began implementing the acute consolidated their negotiations with vendors. noted: “Hospital performance care episode demonstration of bundled At the same time, Hillcrest reported about 30 measurement will be most biased if payments for physician and hospital services percent increases in volume for the orthopedic the decision to admit varies furnished to patients as part of hip and knee and cardiac procedures. Whether that large systematically across hospitals and replacements and cardiac procedures. In its volume increase represents redistribution of regions. Such variation may reflect early results, Hillcrest Medical Center in Tulsa cases from other facilities or increased total greater illness levels in the reported cost savings, especially from interventions for beneficiaries in the area population. However, higher increased bargaining power for equipment and remains to be determined. (See endnote 27.)

Timely Analysis of Immediate Health Policy Issues 9

utilization also may be due to a financially desirable) and those that to move from fee-for-service but are lower threshold for admission, take place within an episode (and not willing or able to make the driven for example by higher per are therefore financially move to global or even partial capita supply of physicians and undesirable). In short, global capitation. Under the Medicare hospital beds. In such hospitals, capitation penalizes volume of Shared Savings Program that was sufficient slack capacity allows services, whether appropriate or not. created by the , physicians to work ‘down an Indeed, a major policy concern is initially fee-for-service to the appropriateness curve’ in their that organizations and their constituent provider members of the admission decisions” [emphasis constituent members will ACO will continue according to added].44 The stated implication is underserve patients because the existing Medicare payment rules, that unless one can account for incentives are too strong in the i.e., fees to physicians and different appropriateness thresholds opposite direction from fee-for- diagnosis-related group payments for admission, measures of hospital service. This concern has led to for inpatient hospitalization. At the spending and quality outcomes are proposals to mix fee-for-service and end of an accounting period, the likely to be inaccurate.45 capitation to try to balance actual spending attributed to the incentives (i.e., partial capitation)46 ACO is compared to a target Shared savings and global payment or to adopt other approaches that spending amount, with any Global payment or global capitation mitigate the extent of financial risk, surpluses split between Medicare is a population-based payment softening the incentives to withhold and the ACO. The ACO is approach in which a fixed per- needed services without eliminating responsible for determining how person, per-month prospective the incentive to economize that is surpluses (savings to Medicare) will payment is made to an organization inherent in global payment. be distributed among the members of the ACO. responsible for providing services to Adherence to professional standards individuals who elect or are is supposed to serve as an important In this one-sided risk approach, and assigned to receive care from the brake on some amount of in contrast to global payment, there providers in that organization. With inappropriate underuse just as it is no financial penalty if actual global capitation payment (in surely restrains some overuse under spending turns out to be more than contrast to and fee-for-service. Further, the clinical the target amount. In the Shared professional capitation), most of the domains with the greatest number of Savings Program, there is also an services that comprise the payer’s valid clinical quality measures are option for ACOs to assume two- benefit package are included in the in the area of primary and sided risk, an approach that would global payment. Sometimes secondary prevention, permitting a be required by year four of the prescription drugs, mental health complementary use of global program. With two-sided risk, both benefits, and other specialized payment and performance measures surpluses and deficits would be benefits may be carved out of the to detect some underuse. shared with Medicare. This two- global payment. Unfortunately, there are not good sided approach starts on the road to Global capitation is the payment measures to detect underuse at the global payment in that the marginal model that most fundamentally other end—for patients with serious incentive for the organization is to changes fee-for-service incentives health problems. For example, some reduce volume. patients might be best served with that reward provision of However, in considering shared inappropriate services. Simply put, referral to clinicians and/or facilities that are not part of the provider savings to ACOs under the CMS unneeded procedures and hospital program, one needs to recognize stays constitute what financial network receiving the capitated payment. However, we cannot that the incentives for the ACO and officers call “profit centers” under those for the ACO’s constituent fee-for-service and its variations, easily measure the failure to make an appropriate referral. providers might be quite different, whereas they are “cost centers” as in the case of managed care under global payment. In contrast to Shared savings. Shared savings organizations and their provider episode-based payment, under within an accountable care network constituents. Cash flow to global payment there is no organization (ACO) is being the constituent providers continues distinction between services that promoted as a possible transition based on fee-for-service. So for a create an episode (and are therefore approach for organizations seeking specialist achieving high incomes

Timely Analysis of Immediate Health Policy Issues 10

by doing discretionary procedures with standard fee-for-service cash in electronic health records and or a hospital trying to keep beds flow may undermine the objective tools for shared decision-making occupied by generating additional of the new payment approach to can make information and evidence admissions, the financial incentive economize on provision of services. more accessible to providers and to generate volume far exceeds the patients. Where possible, given the incentive to share in possible Conclusion methodological constraints of savings achieved by the collective measuring inappropriate service Failing to do whatever we can to ACO. It is not surprising then that use, we need to amplify use of reduce the use of medically in the Physician Group Practice quality measurement and inappropriate and unnecessary Demonstration, which was the monitoring, and we need to increase services is indefensible at a time of model for the ACA’s Shared the amount of data given to mounting pressures on health Savings Program, participating providers that shows their provision financing. But experience shows organizations focused on reducing of overused services relative to their that aligning incentives to squeeze spending associated with peers. out waste and improve the management of particular chronic efficiency of health services diseases, especially congestive heart For many services, particularly provision will not be easy. A failure, but not on reducing the commonly used and relatively low- sophisticated and multifaceted volume of diagnostic and cost services for which prior solution is required. therapeutic procedures that are the authorization is infeasible, as well as for all services of unknown or core income and revenue generators To get at inappropriate service use, uncertain benefit, administrative for important specialist physicians we need to invest in researching and policy tools will not suffice and and hospitals, including joint what works under what could be counter-productive. In replacements, spine surgery, cardiac circumstances and how alternative these cases, the incentives procedures, and other procedures treatment approaches compare to established through payment are found in the literature to be one another. That information is 47 likely to be of critical importance. provided inappropriately. sorely lacking, due to an Further, it is likely that the various underinvestment in research that is The incentives will actually vary educational efforts and only beginning to be remedied. based on the structure of the ACO. administrative approaches will be While research on effectiveness will In true multispecialty group more successful if coupled with always lag behind technology, the practices, fee-for-service incentives altered payment incentives.48 gap between what providers do and to do more should be moderated what researchers know can be substantially, even in a shared Aligning payment incentives with bridged, and strategic prioritization savings payment arrangement based desired outcomes can only help, and can identify the most promising on fee-for-service. It depends on the may in some cases be essential to areas for investment in further method the ACO chooses to reducing the enormous problem of research. distribute its global payments. But resources squandered through wasteful practice decisions. On the where the providers in the ACO But research alone will not suffice. cusp of important changes in essentially receive pass-through fee- There is a need for actors at all payment methodology, policy- for-service payments, with only levels to increase efforts to make makers need to design payment surpluses against target spending use of the available evidence. The methods in ways that reduce amounts subject to the group’s best approaches need to be incentives for provision of excess distribution formula, it is hard to see determined, whether they include service volume and maximize how the incentives on the individual gentle tools, such as establishing rewards for provision of beneficial, clinician vary much from those of campaigns to inform providers and appropriate care. fee-for-service. The ultimate effect patients, or stronger yet sometimes would likely be similar to that with cruder approaches like incentives to bundled episode payment, in which use clinical guidelines through cash flow to individual providers benefit design and coverage continues and only the savings are decisions, or finer scalpels, such as considered to be the group’s to prior authorization rules for certain divide up. In both cases, the procedures with high rates of practical approach of not interfering inappropriate use. New investments

Timely Analysis of Immediate Health Policy Issues 11

Glossary

Bundled episodes—A payment approach in which a single payment is made to cover the cost of services delivered by multiple providers over a defined period of time to treat a given episode of care (e.g., a knee replacement surgery or a year’s worth of diabetes care). Episode-based payment—A payment approach in which a single payment is made to cover the cost of services delivered by a single provider over a defined period of time to treat a given episode of care. Fee-for-service—A payment approach in which health care providers receive a separate fee for each service they deliver. Fee schedule—A comprehensive list of fees used by either a private or public health insurance plan or payer to reimburse health care providers on a fee-for-services basis. Financial risk—When an entity assumes liability for the financial loss that could occur if actual costs exceed expected revenues. Global capitation—A single payment made to a provider organization to cover the cost of a predefined set of services delivered to a patient (e.g., an amount paid per member per month to cover the cost of all of a patient’s health care needs). In many cases, the provider organization is responsible for reimbursing other providers for care they deliver to the patient. Partial capitation—When a payer pays for some types of services on a capitated basis (e.g., by contracting with a group of providers to deliver all of their enrollees’ outpatient care) and pays for other services on a fee-for-service basis (e.g., reimbursing any hospital in their network for inpatient care delivered to their enrollees). Shared savings—A payment approach whereby a provider or provider organization shares in the savings that accrue to a payer when actual spending for a defined population is less than a target amount. (Typically, performance targets on quality measures must be met to qualify for shared savings.) If actual costs are higher than projections, there are no financial repercussions for providers, unless shared losses are also part of the payment agreement.

______

Adapted from: Delaware Healthcare Association. “Delaware Healthcare Association Glossary of Health Care Terms and Acronyms.” Dover, Del.: Delaware Healthcare Association, www.deha.org/Glossary/GlossaryA.htm (accessed December 2012).

Timely Analysis of Immediate Health Policy Issues 12

The views expressed are those of the authors and should not be attributed to the Robert Wood Johnson Foundation, or the Urban Institute, its trustees, or its funders.

About the Authors and Acknowledgments Robert A. Berenson, MD, is an institute fellow at the Urban Institute, and Elizabeth Docteur is an independent health policy consultant. The authors thank Elizabeth McGlynn, director of the Kaiser Permanente Center for Effectiveness and Safety Research, for her helpful comments on this paper. This research was funded by the Robert Wood Johnson Foundation. About the Urban Institute The Urban Institute is a nonprofit, nonpartisan policy research and educational organization that examines the social, economic, and governance problems facing the nation. About the Robert Wood Johnson Foundation The Robert Wood Johnson Foundation focuses on the pressing health and health care issues facing our country. As the nation’s largest philanthropy devoted exclusively to health and health care, the Foundation works with a diverse group of organizations and individuals to identify solutions and achieve comprehensive, measurable, and timely change. For 40 years the Foundation has brought experience, commitment, and a rigorous, balanced approach to the problems that affect the health and health care of those it serves. When it comes to helping Americans lead healthier lives and get the care they need, the Foundation expects to make a difference in your lifetime. For more information, visit www.rwjf.org. Follow the Foundation on Twitter www.rwjf.org/twitter or Facebook www.rwjf.org/facebook.

Timely Analysis of Immediate Health Policy Issues 13

Notes

1 Angrisano C, Farrell D, Kocher B, et al. Accounting for the Cost of Health Care in the United States. McKinsey Global Institute, 2007, www.mckinsey.com/insights/mgi/research/americas/accounting_for_the_cost_of_health_care_in_the_united_states (accessed December 2012). 2 Health at a Glance 2011: OECD Indicators. Paris: OECD, 2011, www.oecd.org/health/healthpoliciesanddata/49105858.pdf (accessed December 2012). 3 ibid. 4 Squires D. The U.S. in Perspective: A Comparison of Twelve Industrialized Nations. New York: The Commonwealth Fund, 2011, www.commonwealthfund.org/~/media/Files/Publications/Issue%20Brief/2011/Jul/1532_Squires_US_hlt_sys_comparison_12_nations_intl_brief _v2.pdf (accessed December 2012); Health at a Glance 2011. 5 The Healthcare Imperative: Lowering Costs and Improving Outcomes: Workshop Series Summary. Washington: Institute of Medicine, 2010. 6 ibid. 7 Korenstein D, Falk R, Howell E, et al. “Overuse of Health Care Services in the United States: An Understudied Problem.” Archives of Internal Medicine, 172(2): 171-178, 2012. 8 Scott I and Jayathissa S. “Quality of Drug Prescribing in Older Patients: Is There a Problem and Can We Improve It?” Internal Medicine Journal, 40(1): 7-18, 2010. 9 Bentley TG, Effros RM, Palar K, et al. “Waste in the U.S. Health Care System: A Conceptual Framework.” Milbank Quarterly, 86(4): 629-659, 2008. 10 Korenstein, et al. 11 ibid. 12 ibid. 13 ibid. 14 Swensen S, Kaplan G, Meyer G, et al. “Controlling Healthcare Costs by Removing Waste: What American Doctors Can Do Now.” BMJ Quality & Safety, doi:10.1136/bmjqs.2010.049213, 2011. 15 Lawson EH, Gibbons MM, Ingraham AM, et al. “Appropriateness Criteria to Assess Variations in Surgical Procedure Use in the United States.” Archives of Surgery, 146(12): 1433-40, 2011. 16 Pfeffer MA, Burdmann EA, Chen CY, et al. “A Trial of Darbepoetin Alfa in Type 2 Diabetes and Chronic Kidney Disease.” New England Journal of Medicine, 361(21): 2019-2032, 2009; Singh AK, Szczech L, Tang KL, et al. “Correction of Anemia with Epoetin Alfa in Chronic Kidney Disease.” New England Journal of Medicine, 355 (20): 2085-2098, 2006. 17 Keyhani S and Siu A. “The Underuse of Overuse Research.” Health Services Research, 43(6): 1923-1930, 2008. 18 Keyhani S, Falk R, Bishop T, et al. "The Relationship Between Geographic Variations and Overuse of Health Care Services: A Systematic Review." Medical Care, 50(3): 257–61, 2012. 19 See for instance: Chassin MR, Kosecoff J, Park RE, et al. “Does Inappropriate Use Explain Geographic Variations in the Use of Health Care Services? A Study of Three Procedures.” Journal of the American Medical Association, 258(18): 2533-2537, 1987; Leape LL, Park RE, Solomon DH, et al. “Does Inappropriate Use Explain Small-Area Variations in the Use of Health Care Services?” Journal of the American Medical Association, 263(5): 669-672, 1990. 20 McGlynn EA, Naylor CD, Anderson GM, et al. “Comparison of the Appropriateness of Coronary Angiography and Coronary Artery Bypass Graft Surgery Between Canada and New York State.” Journal of the American Medical Association, 272: 934-940, 1994. 21 Korda H. Medical Overuse: More Care Is Not Necessarily Better Care. Washington: Altarum Institute, 2012, altarum.org/forum/post/medical- overuse-more-care-not-necessarily-better-care (accessed December 2012). 22 Cassel C and Guest J. “Choosing Wisely: Helping Physicians and Patients Make Smart Decisions About Their Care.” Journal of the American Medical Association, 307(17): 1801-1802, 2012. 23 Chernew ME and Newhouse JP. “What Does the RAND Health Insurance Experiment Tell Us About the Impact of Patient Cost Sharing on Health Outcomes? American Journal of Managed Care, 14(7): 412-414, 2008. 24 Gray B and Field M (eds). Controlling Costs and Changing Patient Care? The Role of Utilization Management. Washington: Institute of Medicine, 1989. 25 Wickizer T. “The Effect of Utilization Review on Hospital Use and Expenditures: A Review of the Literature and an Update on Recent Findings.” Medical Care Research and Review, 47(3): 327, 1990. 26 Berenson R and Harris M. “Using Managed Care Tools in Traditional Medicare: Should We? Could We?” Law and Contemporary Problems, 65(4): 139-167, 2002. 27 Report to the Congress: Medicare and the Health Care Delivery System. Washington: Medicare Payment Advisory Commission, 2011, medpac.gov/documents/Jun11_EntireReport.pdf (accessed December 2012). 28 Berenson R, Sunshine J, Deb A, et al. The Effect of Provider Payment Systems on Quality, Cost and Efficiency, and Access: A Systematic Literature Review. Warsaw: InterQuality Research Project, European Union, 2012; Robinson J. “Theory and Practice in the Design of Physician Payment Incentives.” Milbank Quarterly, 79(2): 149–177. 29 Shen J, Anderson R, Brook R, et al. “The Effects of Payment Method on Clinical Decision-Making: Physician Responses to Clinical Scenarios.” Medical Care, 42(3): 297-302, 2004.

Timely Analysis of Immediate Health Policy Issues 14

30 Sack K, “Unintended Consequence for Dialysis Patients as Drug Rule Changes,” New York Times, May 11, 2012, www.nytimes.com/2012/05/11/health/policy/dialysis-rule-changes-followed-by-transfusion-increases.html (accessed December 2012). 31 Report to Congress on Incentive Arrangements Offered by Health Maintenance Organizations and Competitive Medical Plans to Physicians. Washington: U.S. Department of Health and Human Services, 1990. 32 “Chapter 8: Addressing the Growth of Ancillary Services in Physicians’ Offices.” Report to the Congress: Aligning Incentives in Medicare. Washington: MedPAC, 2010, www.medpac.gov/documents/jun10_entirereport.pdf (accessed December 2012); Jacobs BL, Zhang Y, Skolarus TA, et al. “Growth Of High-Cost Intensity-Modulated Radiotherapy For Prostate Cancer Raises Concerns About Overuse.” Health Affairs, 31(4): 750-759, 2012; Waldman P, “Prostrate Patients Suffer as Money Overwhelms Best Therapy,” Bloomberg, November 6, 2012, www.bloomberg.com/news/2012-11-06/prostate-patients-suffer-as-money-overwhelms-optimal-therapy.html (accessed December 2012); Brox JI, Nygaard OP, Holm I, et al. “Four-year Follow-up of Surgical Versus Non-surgical Therapy for Chronic Low Back Pain.” Ann Rheum Dis., 69: 1643-1648, 2010. 33 Codespote S, London W and Shatto J. Physician Volume and Intensity Response. Baltimore: CMS, 1998, www.cms.gov/Research-Statistics-Data- and-Systems/Research/ActuarialStudies/downloads/PhysicianResponse.pdf (accessed December 2012). 34 Hadley J, Reschovsky J, Corey C, Zuckerman S. “Medicare Fees and the Volume of Physicians’ Services.” Inquiry, 46(4): 372-390, 2009. 35 Jacobson M and Newhouse J. “Expect the Unexpected? Physicians’ Responses to Payment Changes.” Washington, DC: National Institute for Health Care Management Foundation, November 2010, http://nihcm.org/pdf/EV-JacobsonNewhouseFINAL.pdf (accessed December 2012). 36 Sommers A, White C and Ginsburg P. “Addressing Hospital Pricing Leverage Through Regulation: State Rate Setting.” Washington: National Institute for , 2012, www.nihcr.org/1tl92 (accessed December 2012). 37 Gruber J and Owings M. “Physician Financial Incentives and Cesarean Section Delivery.” RAND Journal of Economics, 27(1): 99–123, 1996. 38 Hussey P, Ridgely MS and Rosenthal M. “The PROMETHEUS Bundled Payment Experiment: Slow Start Shows Problems in Implementing New Payment Models.” Health Affairs, 30(11): 2116-2124, 2011. 39 Bailit M, Burns M and Painter M. Bundled Payment Across the US Today: Status of Implementations and Operational Findings. Newtown, CT: Health Care Incentives Improvement Institute, 2012, www.hci3.org/sites/default/files/files/HCI-IssueBrief-4-2012.pdf (accessed December 2012); Cutler D and Ghosh K. “The Potential for Cost Savings Through Bundled Episode Payments.” New England Journal of Medicine, 366: 1075- 1077, 2012. 40 Cutler and Ghosh. 41 Robinson JC and Luft HS. ‘‘Competition and the Cost of Hospital Care, 1972–1982.’’ Journal of the American Medical Association, 257(23): 3241-3245, 1987. 42 Berenson R, Bodenheimer T and Pham H. “Specialty-Service Lines: Salvos in the New Medical Arms Race.” Health Affairs, 25(5): w337-w343, 2006; Casalino L, Devers K and Brewster L. “Focused Factories? Physician-Owned Specialty Facilities.” Health Affairs, 22(6): 56-67, 2003. 43 O’Connor AM, Llewellyn-Thomas HA and Flood AB. “Modifying Unwarranted Variations in Health Care: Shared Decision Making Using Patient Decision Aids.” Health Affairs, doi: 10.1377/hlthaff.var.63: VAR63-VAR72, 2004. 44 Huesch M. “Payment Policy Based on Measurement of Health Care Spending and Outcomes.” Journal of the American Medical Association, 303(23): 2406-6, 2010. 45 Siu AL, Sonnenberg FA, Manning WG, et al. “Inappropriate Use of Hospitals in a Randomized Trial of Health Insurance Plans.” New England Journal of Medicine, 315(20): 1259-66, 1986. 46 Newhouse J. “Patients at Risk: Health Reform and Risk Adjustment.” Health Affairs, 13(1): 132-146, 1994. 47 Report to Congress: Physician Group Practice Demonstration Evaluation Report. Washington: U.S. Department of Health and Human Services, 2009, www.cms.gov/Medicare/Demonstration-Projects/DemoProjectsEvalRpts/downloads/PGP_RTC_Sept.pdf (accessed December 2012); Report to Congress: Physician Group Practice Demonstration First Evaluation Report. Washington: U.S. Department of Health and Human Services, 2006, www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Reports/downloads/Leavitt1.pdf (accessed December 2012); Trisolini M, Aggarwal J, Leung M, et al. The Medicare Physician Group Practice Demonstration: Lessons Learned on Improving Quality and Efficiency in Health Care. New York: The Commonwealth Fund, 2008, www.commonwealthfund.org/~/media/Files/Publications/Fund%20Report/2008/Feb/The%20Medicare%20Physician%20Group%20Practice%2 0Demonstration%20%20Lessons%20Learned%20on%20Improving%20Quality%20and%20Effici/1094_Trisolini_Medicare_phys_group_practi ce_demo_lessons_learned%20pdf.pdf (accessed December 2012). 48 Ghomrawi HMK, Schackman BR, and Mushlin AI. “Appropriateness Criteria and Elective Procedures – Total Joint Arthroplasty.” New England Journal of Medicine, 367(26): 2467-2469, 2012.

Timely Analysis of Immediate Health Policy Issues 15