In 1209, when his soldiers were unable to distinguish the pious from the heretics, Arnaud Amalric ordered his men to “[k]ill them all [for] God will recognize his own!”1 To many individuals living in the 21st century, Amalric’s indiscriminate use of blunt force to massacre 20,000 Cathari may seem like ancient history. However, in the world of healthcare Amalric’s words seem to have become the unofficial mantra of reformers. The US spends $2 trillion a year on healthcare, or $10,000 per capita, while 47 million Americans cannot afford healthcare insurance coverage.2 Even before the current economic slump, the US could not afford its healthcare system. So, if the US is to remain competitive in a global economy, healthcare reform is no longer an issue; reform is coming and only the details remain to be resolved.
Unquestionably in the crosshairs of reformers is America’s over-consumption of healthcare services. According to 2006 World Health Organization (WHO) figures, the US spends more than twice as much per capita on healthcare as the UK or Japan.3 In particular, coronary interventions (coronary artery bypass grafting [CABG] or percutaneous coronary intervention [PCI]) are performed many times more frequently in the US than in Europe.4 The WHO’s data, which are corroborated by other reputable sources,5 suggest to healthcare reformers that coronary intervention is being over-prescribed in the US. However, lacking clinical acumen, reformers are hard-pressed to identify which patients could be managed medically and which require more invasive intervention.
Enter provider-specific report cards, a methodology ostensibly designed to improve patients’ autonomy by increasing market transparency.6 Recently, the Centers for Medicare and Medicaid Services (CMS) launched its own report card on its Hospital Compare website.7 Already receiving 2.5 million hits per month, Hospital Compare ranks private hospitals according to performance measures, reimbursement, and risk-adjusted mortality.8 On the other hand, the realpolitik of provider-specific report cards (including Hospital Compare) is to enlist the aid of market forces9 to ration healthcare.10 More specifically, report cards use proxies (e.g. surgical-site infection [SSI] rate) to measure ‘quality,’ and an adverse report card can ruin a hospital’s reputation. To reformers, fewer hospitals means fewer services will be provided. However, the rather blunt nature of these quality proxies ensures that any reward or punishment a hospital receives based on its report card performance will be granted more or less indiscriminately.
So, from the perspective of hospitals, report-card-driven healthcare reform is likely to be perceived as a ‘kill all of the hospitals for the market will know which hospitals provide quality care’ downsizing program. Accordingly, after reviewing the evidence for a surplus of hospitals, this article discusses how provider-specific reports can create reputational incentives capable of reducing the number of hospitals providing coronary intervention.
Closing Hospitals
Surplus Capacity. Overall, the hospital bed capacity in the US is 3.0 per 1,000 population.11 This figure, however, varies from state to state and does not correlate with a state’s population density. Thus, several low-population-density states (North and South Dakota, Mississippi, Montana, Nebraska, and West Virginia) and high-population-density states (New York, Illinois, and Florida) have hospital bed capacity rates that exceed the 3.0 per 1,000 population mark.
To a degree, the disproportionately high hospital bed capacity rate found in many low-population-density states is deceptive. Hays, Kansas (population of 130,000) is the home of a 158-bed medical center that provides invasive cardiac services.12 If only the community of Hays is considered, Hays probably has a surplus of hospital beds. However, when it is realized that the nearest competing medical center that provides invasive cardiac services is 150 miles away in Wichita, Kansas, it could be argued that western Kansas has a shortage of hospital beds for providing cardiac services.
Yet does Philadelphia really need five hospitals that offer cardiac transplantation,13 and does Chicago really need 41 hospitals that provide cardiac surgery?14 To healthcare reformers many high-population-density regions have a surplus of hospital beds. For example, the Berger Commission recommended that New York State (NYS) eliminate 4,200 hospital beds (7% of the market) by closing nine hospitals.15 Similarly, a New Jersey commission concluded that the Hackensack market had an “oversupply of beds” that resulted in “serious financial distress” for the state.16
Surplus bed capacity concerns healthcare payors because it has been correlated with the over-utilization of healthcare services.17 In fields as diverse as pediatrics,18 surgery,19 and critical care medicine,20 markets with surplus bed capacity have physicians hospitalizing patients more frequently. Consequently, markets with surplus hospital bed capacity also tend to be the markets with highest per capita healthcare spending.21 This is not to say that high per capita healthcare spending is fundamentally bad. If such markets could show that excessive healthcare spending resulted in better outcomes, reformers would be silenced because such information would support the conclusion that in healthcare ‘you get what you pay for.’ Unfortunately for providers, the literature of the last 30 years has failed to demonstrate that regions with higher per capita spending have better clinical outcomes than the regions with low per capita spending.22 Even when utilization data are examined at the level of individual hospitals, it is not possible to demonstrate a positive correlation between healthcare spending and clinical outcomes.23
Collectively, these observations suggest to healthcare reformers that substantial reductions in healthcare spending can be achieved—without sacrificing quality—by reducing the number of hospitals (beds). To illustrate this, consider the potential savings that could be achieved if the US were able to reduce its hospital bed capacity by 5% (or 2% less than the downsizing figure recommended by the Berger Commission). With 31% of healthcare spending going to hospitals,24 such a reduction in hospital capacity could save $40 billion. Moreover, while the data are not conclusive, the mere act of closing a hospital can have a salutary effect on healthcare. When an urban hospital closes, competing hospitals within a five-mile radius of the closing hospital begin to operate more efficiently such that their unit-adjusted cost per admission actually falls by 2–4%.
Thus, it should come as no surprise that the government has already created incentives to downsize the hospital market. Some of these incentives include:
- declining to pay for certain complications (thereby removing $4.8 billion/year from the hospital market);25
- changing reimbursement rules to influence hospitals that provide “post-acute care services to alter their level of participation in Medicare;”26 and
- cutting Medicaid reimbursement to safety net hospitals.27
However, now we reach the hard issue for healthcare reformers: which hospitals should be closed? In many communities the local hospital is the largest employer in the county,28 which makes it politically dangerous to directly attack and close a hospital by fiat.29 The Berger Commission’s use of a fiat decision-making process—regardless of how thoughtfully it was prepared—demonstrates just how much public backlash can be associated with fiat hospital closure. Politicians, who instinctively fear any public outcry as it may herald an unsuccessful re-election bid, cannot afford to leave their fingerprints on any legislation that would downsize the hospital market. Rather, a politically safer way to downsize the hospital market would be indirectly through the use of market forces. Leaving no fingerprints, the use of report cards to unleash market forces is favored by President Obama as a technique to reshape our healthcare market.30
Report Cards
History. In 2003, the Institute of Medicine (IOM) recommended that the US jettison its traditional administrative pricing of healthcare services.31 According to the IOM, a value-based purchasing (VBP) system not only would promote improvements in quality of care, but also would help control medical inflation. At its heart, VBP uses report cards to determine which providers should receive the carrot and those that will get the stick as their ‘reward.’ The carrot is the now familiar pay-for-performance bonuses that are given to providers who score well on their report cards. Conversely, providers who consistently fail to meet their performance measures are to be ‘removed’ from the market.32
Here we reach a paradoxical situation. A fundamental maxim of patient safety is that fear does not belong in the market,33 yet it is axiomatic that forcing providers out of the market is a formula for raising the market’s level of fear. To resolve this paradox it needs to be recognized that while the IOM has used patient safety as a pretext for healthcare reform, the cost-reduction subtext of the IOM’s publications makes it clear that patient safety reform is as much about saving money as it is about protecting patients.34 By using report-card-generated reputational incentives, market fears are assuaged because hospitals are given a choice: improve or exit the market.35
Shaping markets with reputational incentives is a great theory, but does it work? The definitive answer is not yet available because of the plurality of quality metrics and methodologies that are used to create report cards. The Joint Commission and several other patient safety organizations each has its own report cards, and none of these report cards has achieved universal acceptance.36 Moreover, the plurality of report cards is likely to confuse the public because hospitals are ranked differently by CMS’ Hospital Compare and US Newsweek’s Best Hospitals.37 As a plurality of methodologies makes comparison of report card rankings of hospitals difficult,38 “[s]erious issues have been raised about the accuracy, validity, meaningfulness, and interpretability” of hospital report cards.39
The Cardiac Surgery Model. Still, the market impact of reputational incentives can be demonstrated from studies of NYS’ cardiac services report card for hospitals.40 NYS’ cardiac services report card is a valuable model for studying the impact of report cards because:
- the metrics and methodologies employed have been more or less constant for two decades;
- it is the most studied report card model; and
- this model has been adopted by a number of states, including California, Massachusetts, New Jersey and Pennsylvania.
In response to a freedom of information lawsuit, NYS reluctantly began publishing provider-specific data in 1989.41 At first, NYS’ cardiac services report card appeared to have little market impact. Peterson, for example, found that after the publication of an adverse report card, high-mortality hospitals experienced only a transient drop in the volume of cardiac surgical procedures performed.42 These early studies also tended to suggest that financial incentives trumped reputational incentives. Between 1994 and 2005, the number of NYS hospitals providing low-volume cardiac services grew by 10%,43 even though low-volume hospitals tend to have higher mortality rates.44
Yet by 2005 the studies coming out of NYS demonstrated solidly that reputational incentives do have market impact. Although a positive report card rarely translates into a hospital gaining market share,45 survey data indicate that an adverse report card can tarnish a hospital’s reputation.46 Once a hospital’s reputation for cardiac services has been tarnished, Medicare data have shown that the hospital becomes more risk-averse. In particular, a hospital with a tarnished image tends to:
- ship more high-risk cardiac surgery cases to higher-volume institutions;47 and
- offer fewer ancillary cardiac procedures:48 during a recent three-year period, a 33% reduction in the number of CABGs performed in NYS was better explained by the state’s report card publication than by cardiologists substituting PCIs for CABGs.
In addition, markets that use reputational incentives have significantly lower hospital mortality rates because only the best-risk patients are offered coronary intervention.49 Fearing even the smallest loss of reputation,50 hospitals encourage cardiac surgeons to ‘cherry-pick’ patients51 and encourage cardiologists to offer PCI to fewer patients with acute myocardial infarction. Not surprisingly, by 2005 NYS could boast that it had the lowest mortality rate for cardiac surgery in the country.52
Some might argue that NYS’ excellent cardiac surgery mortality rate was the end result of process improvement, but this seems unlikely. Longitudinally collected data suggest that the complexity of CABGs performed in NYS is decreasing. However, if process improvement were really the force driving down mortality, it would be expected that the improved hospitals should be better able to handle the complex cases—something that has not been observed. Moreover, little money is available for process improvement. With one-third of hospitals operating with a negative profit margin53 and CMS continuing to reduce the profitability of providing cardiac services,54 discretionary funds are in short supply.
When these economic realities are coupled with the collapse of the bond market,55 it is unlikely that the financial health of hospitals will improve any time soon.56 This means that a hospital that receives low marks on a report card is unlikely to have the necessary cash to fund a quality improvement program to restore its image. So, in the future when a hospital’s reputation is sullied by a bad report card, its options are likely to be limited to shuttering or merging the program.57 Either way, the reputational incentives associated with the publication of provider-specific report cards encourages the downsizing of cardiac services markets.
Who Benefits? So, who benefits when NYS publishes its annual cardiac surgery report cards? An obvious beneficiary is any hospital located in a state that does not publish report cards. The Cleveland Clinic, for example, experienced a 31% increase in its volume of (high-risk) CABG procedures after NYS began publishing report cards.58 Yet, as we go forward, hospitals in states that do not utilize reputational incentives will not be as fortunate.
As a national report card, Hospital Compare will allow patients located on either side of a state line to discover a hospital’s quality metrics. So, even if a hospital is located in a state that does not publish provider-specific report cards, Hospital Compare provides an incentive for all hospitals to think twice about to whom it offers coronary intervention. Similarly, hospitals that are dependent on performing a certain volume of cardiac services to stay afloat will increasingly find this strategy to be perilous. Any hospital that is strategically dependent on performing a certain volume of coronary interventions could have their finances upset by an adverse report card.
This, of course, is exactly what federal government wants. Every time a provider declines to offer a patient either PCI or CABG, the government saves money. Viewed in this light, while report cards may improve market transparency, it is also clear that report cards function as a device to ration healthcare. For politicians and reformers, the beauty of unleashing market forces with Hospital Compare is that the cause and effect of report-card- induced market downsizing is blurred. As market forces do not leave the cachet of any governmental agency, it will be the hospitals and physicians who have declined to offer patients for coronary intervention that will be exposed to the public’s ire. Meanwhile, the government will happily take full credit for creating Hospital Compare to expose low-quality providers.
Implications
Now imagine your 65-year-old father has two-vessel coronary disease. While experts in the field debate the optimal treatment for this condition,59 your father’s cardiologist recommends that he receive a CABG, which can be performed at either a traditional hospital or a heart hospital. When your father consults Hospital Compare, he discovers, with three exceptions, that the traditional and heart hospitals have almost identical statistics. The three exceptions are:
- last year the heart hospital performed twice as many CABGs as the traditional hospital;
- the SSI rate at the heart hospital is 3% greater than at the traditional hospital; and
- the cost of the CABG is substantially cheaper at the heart hospital.
Indeed, your father discovers that on the Hospital Compare website, heart hospitals are usually the least expensive provider of cardiac services in their local market. Figuring that you get what you pay for, your father selects the traditional hospital, hoping to avoid an SSI. However, to your father’s chagrin, the traditional hospital declines to offer him an operation. He later undergoes a CABG without complication at the heart hospital.
Once it is accepted that report cards are a ‘kill-them-all’ scheme for rationing healthcare, all of the actors in the above hypothetical situation behave logically. As most patients are not trained in statistics, they do not realize that a 3% difference in the incidence of SSIs at two hospitals is unlikely to be statistically significant. The Hospital Compare website, which has several footnotes for qualifying data, makes no attempt to provide patients with information on statistical inferences. Consequently, there is no way for your father to learn from Hospital Compare that the differences in SSI rate between the traditional and heart hospital are likely to be due to chance alone. Nor does Hospital Compare disclose the fact that the risk of an SSI has as much to do with a patient’s comorbidities as it does with the measured quality metrics of antibiotic administration, maintenance of normothermia during the peri-operative period, and post-operative glucose control.60
Indeed, it is precisely because of your father’s chronic renal insufficiency (CRI) that the traditional hospital turned him down for revascularization. Operating at a much lower case volume, the traditional hospital can ill afford bad outcomes, including an SSI, which occurs more frequently in patients with CRI. Wishing to avoid bad publicity, the traditional hospital has placed pressure on its single cardiac surgeon not to offer a CABG to any patient with CRI.
While this policy may be successful in maintaining the hospital’s image for cardiac services in the short term, over the long term this policy is likely to have a deleterious impact. As the hospital becomes increasingly more risk-averse, its volume of cardiac services continues to contract. In a hospital dependent on performing a certain volume of CABG cases to cross-subsidize other services, the traditional hospital will become less capable of carrying out its mission of providing all services to all patients. There is also a catch-22: by offering fewer patients coronary intervention, the hospital becomes even less capable of handling the impact of an adverse outcome. This fact then sets off another round of risk avoidance behavior, resulting in further volume contraction.
In contrast, operating at a higher case load, the heart hospital in the hypothetical situation is in a better position to handle the risks associated with your father’s CRI. That is, in a reportcard- monitored market the higher-volume heart hospitals can profit because they can better handle the risks of providing coronary intervention. Still, for heart hospitals, the reputational incentives highlighted in this hypothetical situation have two implications.
First, in a market where patients are increasingly engaged in comparison shopping, the delivery of quality coronary intervention at a good price is unlikely to maintain market share. As time goes on, traditional hospitals will emulate heart hospital behavior. Second, heart hospitals should lead the way in educating patients on the pitfalls of report card interpretation to engender patient goodwill—a ‘quality metric’ that is not measured on the Hospital Compare website.
Consider your father. It is a stress-riser for any patient to undergo a CABG at their second-choice hospital. So, using Hospital Compare to comparison-shop for a surgical procedure, it is not unreasonable to believe your father was made more anxious. While it is hard to place a dollar value on such anxiety, it is nonetheless a real hardship for the patient. So, it seems likely that any hospital that takes the time to educate patients on the pitfalls of report card interpretation will only profit from the cultivation of goodwill.
Conclusion
Report cards may be a blunt instrument for downsizing the hospital market, but few healthcare economists doubt the effectiveness of reputational incentives. Nor is the government likely to retreat from publishing provider-specific data. Next year the government is planning to spend $1.1 billion on comparative healthcare effectiveness research.61 Armed with these data, the government will determine which coronary interventions are ‘medically indicated.’
Once that happens, watch for a new quality metric, ‘inappropriate treatment,’ to be posted on Hospital Compare. Can you imagine what patients will think when they read that 10% of the CABGs performed at the local hospital were deemed ‘inappropriate treatment’? Sounds far-fetched? Perhaps. Yet the important point here is that hospitals with strategic plans to provide coronary interventions should amend those plans to include not only the resources to provide high-quality care, but also to include funding for public education on report card interpretation.
The author wishes to thank Ms Rosemarie A Adkins, librarian for the Eastern Kansas VA Healthcare System, for her assistance in gathering reference material.