The Dangers of Hospital Closures Ch 2

c. Evidence on hospital closings’ and bed supply’s link to hospital costs This section summarizes our analyses of seven types of evidence on the links between hospital closings’ and hospital bed supply (on one hand) and the cost of hospital care. First, the financial benefits of closing hospitals have been systematically over-sold. Walter McClure, whose early work was one of the foundations of the policy of closing hospitals to save money, essentially repudiated his own quantitative analyses after observing how closings and consolidations actually worked in practice. As just noted, he saw that the financially and politically vulnerable hospitals were being closed, and that more care was being given in costly and intensive hospitals. Second, the financial harm and the harm to access have been systematically downplayed. One reason why it has been possible to downplay harm has been the consistent refusal to carefully analyze the actual effects of closing on patient outcomes. The adverse and disproportionate impact of closings on black/African-American citizens (and on other patients who are vulnerable to deprivation of needed care), owing to their greater reliance on hospitals for ambulatory care in the ER and especially in the hospital outpatient department, has been largely ignored. So has the systematic and persistent racial association with hospital closings themselves, as will be seen in the next section. Third, closings and proposed closings in Washington, D.C. and in Los Angeles have been accompanied by misleading and, sometimes, outright inaccurate claims and promises. In Washington, for example, the city government claimed that evidence from other cities supported the workability of a proposed free-standing emergency room, but this purported evidence was simply inapplicable.21 Nonetheless, that proposal keeps resurfacing. Promises of support for appropriate primary care, ambulatory care, and other vital preventive services swirl around proposals to close hospitals. These promises are seldom redeemed. Promises that substitute inpatient care would be adequate, available, and accessible also abound. They are seldom fulfilled.22 Fourth, the association of higher ratios of beds per 1,000 people to higher hospital cost per person—across states or over time—is very weak, as the data and multivariate analyses in the previous section have shown. Even worse, in practice, hospital costs keep rising even as hospitals and beds are closed. As McClure showed in 1979, the less costly hospitals are more likely to close. And even if there were there an association across states or over time, that does not mean that lowering the beds per 1,000 people will result in lower hospital cost per person. Physicians and hospitals of a city, region, or state respond and adapt to changes in the number of beds in ways that can keep revenue flowing at high rates. The hospitals that generally remain open tend to be larger teaching hospitals that already provide more intensive patterns of care. A wide range of forces shape hospital care and hospital costs, as our full report will show. For example, many doctors earn incomes by providing care in hospitals. Doctors’ decisions control, shape, and influence almost 90 percent of personal health spending. Closing hospitals does not directly affect doctors’ behavior. They may respond to a closing by relocating, changing their patterns of care, and otherwise adapt—all in ways that sustain or even increase the flow of revenues to surviving hospitals, and that safeguard their own incomes. In other words, hospital costs result from a complex set of relations. Closing beds is a mechanical action, one that does not acknowledge or respect this underlying complexity. Not surprisingly, it fails to save money. Fifth, public actions to close hospitals—whether through planning in Detroit or through a combination of targeted planning and overall tight financial squeezes in New York—tend to accelerate the closing of the less costly non-teaching hospitals. Sixth, the Finger Lakes experience in New York State shows that closing a hospital can certainly make sense under the right circumstances—especially when methods of paying hospitals reward efficient delivery of care and when hospitals and their communities are oriented toward meeting patient needs. These conditions liberate hospitals—and their physicians—to steer by clinical compasses, not financial compasses. In particular, these conditions remove incentives to boost revenue and cover costs by hiking volume or intensity of care. Seventh, as the AIDS experience in New York City reminds us, need for hospitals beds can ebb but also flow. In the years ahead, the growing number of older Americans will require more hospital beds. Many of the beds closed today will have to be rebuilt tomorrow, probably at very great cost. Costs of building a new hospital have passed $500,000 per bed and sometimes approach $1,000,000 per bed.23 In a decade, and especially in urban areas where costs are already high, $2,000,000 per bed is likely. Consider, then, that building a new 500-bed hospital to replace beds closed in recent years would cost $1 billion around the time the baby boomers begin passing age 70. d. Hospital closings in New York State and New York City, 1969-1980 Faced with unexpectedly high Medicaid cost, New York State restricted growth in payments to hospitals starting in 1969. This financial squeeze threatened many hospitals statewide. In response, hospitals demanded a more targeted cost control effort—closing some hospitals to spare the rest. In 1969, New York State was the first to control increases in rates of payment to hospitals. As Caress and Kotelchuck report, this was directly associated with the launch of Medicare and, especially, of Medicaid. (This section of our report rests heavily on a 1977 paper by Caress and Kotelchuck.)24 New York’s legislature cut spending directly, acting in 1969 to declare over one million recipients to be ineligible for coverage. But even this was not enough. In 1969, the legislature also acted to freeze Medicaid’s rates of payment to hospitals for three years. After January of 1970, a prospectively set daily rate was imposed on Medicaid and Blue Cross, which together provided 60 percent of hospitals’ revenues from patient care. Hospitals responded, in part, by increasing length-of-stay very slightly even though average stays fell slightly nationally. Believing unused beds to be costly, the state obliged hospitals to shut some 6,000 acute beds between 1972 and 1977. Caress and Kotelchuck cite Dowling’s estimate that the state’s legislative ceiling on daily hospital payments saved Medicaid and Blue Cross over $200 million in total between 1970 and 1974. This was also not enough. Even these strong steps proved inadequate to hold hospital costs in New York State to affordable levels. High hospital costs and a weak economy pressed New York State to take sharper action. In 1976, an emergency Medicaid law empowered the state budget director to approve all reimbursement rates. Medicaid payment rates were essentially frozen. A “publicly authorized but privately-constituted planning body dominated by powerful voluntary interests” sought to direct cost control activities.25 This body, a task force of the Health and Hospitals Planning Council, proposed selected hospital closings in place of the uniformly applied freeze on Medicaid payment rates. “’The Task Force believes that a sounder approach to decreasing expenditures is to eliminate entirely expenditures for services in institutions which are not needed to meet the public’s requirements for services.’”26 The Task Force called for 27 hospitals to be closed. Most of the beds slated for elimination were in public hospitals. Twenty small for-profit and non-profit hospitals were to be shut. The state’s health commissioner was authorized to close hospitals thought to be unnecessary owing to excess bed availability. Commissioner Whalen in August of 1976 announced a “’hit list’ of 13 New York City hospitals he deemed to be ‘unnecessary.’”27 All were on the Task Force’s list. In November of 1976, the state’s Health Planning Commission decided that New York State thenceforth needed 3.75 beds per 1,000 people, not 4.1 beds per 1,000. New York City therefore moved from being considered under-bedded to being over-bedded. Beginning in 1975, New York City’s federally-financed health systems agency (HSA) had been planning its own approach to bed-reduction. The HSA proposed closing two large non-profit hospitals and cutting beds dramatically at a third. “The HSA argued that ‘closure solely of small voluntary and proprietary hospitals could be counterproductive— the patients from these institutions might well simply be redistributed among the relatively more expensive hospitals providing a more intensive and unnecessary level of care.’”28 Larger voluntary hospitals disliked this idea. The final HSA plan removed mention of excess beds. In January of 1977, the governor and mayor declared they would appoint a “health ‘czar’” to close hospitals. “According to Stephen Berger, Director of the Emergency Financial Control Board, the state’s agency for controlling New York City finances, the czar ‘would take the heat and . . . public officeholders would be in a position to step back.’” 29 Exhibits C and D clearly indicate that, owing to a combination of regulatory threats and pressures, and tight payments to hospitals, a very substantial number of hospitals and beds was indeed closed in New York City. McClure has also noted that New York State was able to exert modest financial pressure on hospitals by tightening Medicaid and Blue Cross payment rates in the 1970s. Writing in 1979, he claimed that “capacity is so excessive in New York that even this modest financial pressure has been sufficient to forces a number of weaker hospitals over the edge.” But he also noted that “Presently only very weak hospitals with weak constituencies are being forced out by rate controls. When large prestigious hospitals start to go over the edge (the ones that generate the bulk of expenditures), it may not be so easy to maintain tight rates.”30 The number of acute care beds in New York City continued to slide downward decade by decade. This slide was slowed for a time in the later half of the 1980s and early 1990s, when AIDS—combined with crack cocaine use and related violence and trauma31—led to a shortage of acute hospital beds in New York in many places and times. Late in 1987, for example, the state’s health commissioner authorized a rise of up to 500 beds in the city’s allowed capacity.32 This is roughly in accord with the prediction by Green and colleagues in 1987, who projected that treatment costs for patients with AIDS would rise from 3 percent of hospital costs in New York City in 1986 to 8 percent in 1991.33 Exhibit C Number of Acute Care Hospitals, New York City, 1937-2006 160 139 140 124 123 124 119 120 100 85 80 2006 estimated 69 62 60 55 40 20 0 1937 1946 1950 1960 1970 1980 1990 2001 Exhibit D 45,000 Licensed Beds in New York City Acute Care Hospitals, 1937- 2001 38,917 40,000 39,860 34,972 35,000 30,126 31,733 32,473 31,453 30,000 25,431 25,000 20,000 15,000 10,000 5,000 0 1937 1946 1950 1960 1970 1980 1990 2001 2006 e. A new wave of hospital financial distress sparks a second wave of demands to close hospitals, 2000-forward Since about the late-1990s, as in the late-1960s, widespread hospital financial distress has again sparked demands to close hospitals in New York State. As was shown in Exhibit A, New York State hospitals’ operating margins were secondworst in the nation in 2004. This is a long-standing problem, one whose genesis is fairly well understood. For many years, state government set rates of payment to New York hospitals. Then, hospital payments were deregulated. Around this time, the federal government acted to substantially slow the growth of Medicare payments to hospitals. Also around this time, managed care organizations in New York grew large enough to exert financial leverage over hospitals. Hospitals were squeezed tightly in many or even most parts of the state. As in the 1970s, many hospitals sought relief. As in the 1970s, they hoped to find it through the closing of some other hospitals. As in the 1970s, Stephen Berger emerged as a central proponent of the idea that closing hospitals would save money overall, and that it would help some hospitals to survive. The first seems untrue. The second is possible. But would the right hospitals survive? 3. Evidence from a 52-city study of race, efficiency, and hospital closings We have been studying closings among 1,200 hospitals in 52 cities since the 1930s. Decade after decade, clear patterns have emerged: 9 9 9 9 Surprisingly, greater efficiency does not help hospitals survive. Larger teaching hospitals are much more likely to survive. Hospitals in black/African-American neighborhoods are less likely to survive. Hospitals with more money in the bank are more likely to survive. This section focuses on efficiency. Advocates of closings promise that they will save money. Because the hospitals that close are typically as efficient as those that survive, though, it is hard to hope to save money by closing hospitals. Efficiency does not help hospitals survive. The fittest are not more likely to survive. This relationship holds consistently in all decades studied. This relationship holds when considering efficiency alone and also when controlling for other factors (see Appendix 2 for the regression results). Given the tendency of hospitals with more endowment per bed to remain open, the observed pattern might better be called “survival of the fattest.” No free market justifies hospital closings or survival. The disproportionate survival of larger teaching hospitals is not result of winning an efficiency contest or a contest to satisfy sovereign consumers’ desires in a freely competitive market. Nothing close to a free market exists for hospital care. Hospitals that do better financially, or hospitals that are larger and more seemingly indispensable, often assume that they deserve to survive. Their leaders—and public officials anxious to appear active in fighting high health costs—often assume that financially weaker hospitals deserve to close. In reality, some hospitals are financially weaker even though they seem just as efficient as those that do better financially. They may not be able amass enough revenue to cover their costs—perhaps because they can’t attract enough doctors to admit enough patients, and perhaps because the patients they do attract include many who are covered by Medicaid or who are uninsured and therefore don’t generate enough revenue to cover the legitimate costs of efficient care. In hospital care, no competitive market exists. Therefore, there is no mechanism to render a fair judgment that financially failing hospitals deserve to close. And many governments—obsessed by the persisting hope of saving money by closing hospitals—have paid insufficient attention to assessing which hospitals, of what types, and in what locations are actually needed to protect the health of the public. Without a functioning free market, careful government plans and regulations, or fair payment to hospitals and doctors for each patient in need, hospitals today try to survive under conditions of growing financial and political anarchy. There is surprisingly little published evidence on the comparative costs of closed and surviving hospitals.34 Shepard performed one empirical analysis of the cost associated with a hospital closing. He found that the closing of a small community hospital probably resulted in a slight increase in area-wide hospital costs. In large part, this was because a great share of the displaced patients who obtained care elsewhere were served in a costly teaching hospital. Shepard suggested that “for service discontinuations to generate savings, recipient hospitals must be low in costs, the outflow rate [drop in admissions] must be large, and the marginal cost ratio must be low.”35 We have found that, decade by decade, larger hospitals generally and teaching hospitals specifically are likelier to survive in the 52 cities studied. Many of these teaching hospitals apparently suffer high average costs, adjusted for case mix.36 This may be attributable to superior quality, inefficiency, or other factors. In 2005, we compared the costs at Brooklyn’s St. Mary’s Hospital with those of surviving nearby alternative hospitals. Specifically, we examined the efficiency of St. Mary’s in 1990—the last year for which we now have data—with the efficiency of Wycoff Heights, Brookdale, Interfaith-St. John’s, Downstate, Woodhull, and Kingsbrook. The results are presented in Exhibit E. St. Mary’s was appreciably less costly—more efficient—in 1990 than all but one of the hospitals most likely to deliver substitute care after it closed. The efficiency measure controls for case mix—severity of illness. Indeed, the alternative hospitals’ average case mix-adjusted cost per adjusted discharge of $5,116 was fully 17.4 percent higher than that of St. Mary’s. The measure of efficiency is the one used consistently in this report, cost per adjusted discharge, as adjusted for case mix index (one measure of severity of illness). Exhibit E also compares the efficiency of St. Mary’s with that of Manhattan’s Presbyterian Hospital and N.Y.U Medical Center. Those two Manhattan teaching hospitals were 39 percent and 59 percent more costly, respectively, than St. Mary’s, again controlling for severity of illness. Exhibit E Case Mix-Adjusted Cost per Adjusted Admission, 7 Brooklyn Hospitals and 2 Manhattan Teaching Hospitals, 1990 $7,000 $5,000 $6,040 $5,753 $5,352 $5,533 $6,000 $4,854 $5,027 $6,902 $4,348 $4,181 $4,000 $3,000 $2,000 $1,000 . C tr n ed . N .Y .U .M es b Pr K in gs br o yt er ia ok dh ul l W oo at ns t ow D t. J th -S rf ai In te e oh n' s le da B ro ok of f W yc St .M ar y' s $0 When patients are displaced from a closed community hospital or a smaller teaching hospital and obtain care at a surviving larger teaching hospital, the cost of their care might rise if they receive greater amounts of care similar in type to that provided earlier, a richer mix of care, or less efficient care at the substitute hospital.37 Some teaching hospitals serve higher proportions of poor patients, whose care is likely to cost more.38 Displaced patients’ cost might also rise if they are charged higher prices to crosssubsidize teaching hospitals’ more esoteric services (if the latter are priced below cost). Even if short-run marginal costs at surviving hospitals are low, what matters eventually is the long-run average cost. Solid comparative information on average costs for urban hospitals is not yet available. But McKay and Dorner report that while surviving rural hospitals located near closed rural hospitals experienced higher volumes, this was not associated with either reduced average cost or improved financial margins.39 Garber and colleagues compared costs of faculty and community services at one teaching hospital, and found that the former’s DRG-adjusted costs were 11 percent higher.40 Frick and others reported teaching hospital care to be 60 percent more costly than non-teaching hospital care, with only one-quarter of the excess attributable to case mix difference.41 In both instances, higher costs may have been associated with greater severity of illness. Since African-American and Latino patients are displaced disproportionately by hospital closings, they and their insurers may be especially likely to suffer higher costs when they are served by surviving larger teaching hospitals. Today, it’s important to note that about 9 in 10 beds in New York City are in teaching hospitals, compared with about three-fourths across all of the 52 cities we’ve been tracking. This means that closing hospitals is very likely to mean closing teaching hospitals. And this, in turn, points to the importance of distinguishing among the various types of teaching hospitals—those large and prestigious institutions with many residents and close affiliations with medical schools, and other, smaller teaching hospitals. Among the teaching hospitals, it is the less prestigious institutions that are likelier to be targeted for closing. And these hospitals typically have lower costs. D. CONCLUSIONS— HOSPITAL CLOSINGS AND SAVING MONEY The evidence overwhelmingly indicates that closing hospitals will do little or nothing to slow health care cost increases. Indeed, it is more likely that costs will actually increase—if, as is widely expected, the types of hospitals that close in the future resemble the community hospital and smaller teaching hospitals that closed in the past—and if the patients displaced by closings actually obtain needed care at surviving hospitals. For three decades, the policy of closing hospitals has been held out as a painless, efficient, and even noble way to cut hospital costs by cutting ostensibly wasted care and liberating funds to be used, somehow, to finance more primary care services. Proponents of closings have wrapped themselves in the mantle of the public interest and have cleverly castigated those who oppose closings as selfish protectors of special interests, as unreasoning obstacles to more efficient and equitable delivery of care. But if hospital closings don’t save money, and if they tend to benefit the more powerful surviving hospitals, perhaps it is the advocates of closings who are selfishly violating the public interest in affordable care for all New Yorkers. Having more hospital beds does little to cause high health care costs today. Decades of evidence indicates that even if more beds did contribute to higher costs (and more beds seem to be only a minor contributor), closing hospitals would still not save money and can even lead to increased costs. The speedier rise of hospital costs per person since about 1975, while beds per 1,000 people were cut, should be called a macro-failure of policy. We suggest that a continued focus on the hospital bed supply per 1,000 people will allow costs to continue to increase. Focusing on closing hospitals to cut beds distracts attention from taking serious action to save money by attacking the sources of waste in health care in New York State. The way to contain cost is by containing cost, not by cutting beds.