The Dangers of Hospital Closures Ch 1

Closing Hospitals in New York State Won’t Save Money But Will Harm Access to Health Care I. CLOSING HOSPITALS DOES NOT SAVE MONEY A. THE PROBLEM OF HIGH HEALTH COSTS High health care costs and high hospital costs have long plagued New York State. Indeed, New York State has the second-highest health care costs among the 50 states and therefore the second-highest in the world. In 2004, personal health spending per New Yorker was 25 percent above the U.S. average. Hospital spending per New Yorker was 34 percent above the national average. If New Yorkers had spent on hospital care at the national per person average, savings in 2004 would have totaled $10.7 billion. At the same time, New York had 21 percent more hospital beds per 1,000 people than the national average. Health care costs have continued to rise in the nation and in New York despite efforts to use both public regulation and market forces to control them. Indeed, none of the cost control attempts of the past three decades has worked well. Today, informed Americans have great reason to worry that the public and private dollars needed to finance constantly growing health care spending simply can’t be found. This strengthens pressures to somehow contain costs. In New York State, hospital financial margins are among the tightest in the nation. It is not clear whether this stems from high costs, from low revenues, or from some combination of the two. As in the 1970s, when hospitals’ revenues were also tight, thoughts turn to closing some hospitals in hopes of relieving the financial pressure on those that would survive. B. THE APPEAL OF HOSPITAL CLOSINGS Many people have long believed that having a great number of hospital beds is one of the main causes of high hospital costs. From this follows a belief that closing beds will save substantial sums. Since the 1970s, New York State has periodically sought to slow the rise in health care spending by closing hospitals. These attempts have been accompanied by public declarations that the state had too many hospitals and too many beds, that extra beds wasted money, and that—done carefully—closing beds could save money without appreciable harm to access to care. As will be seen shortly, the evidence indicates that the link between hospital bed supply and costs of hospital care is very weak. Still, many efforts to contain health care costs have focused on hospital closings. There are at least eight main reasons for this focus. Some are public-spirited but others are more selfish. First is logic. New York’s hospital costs are indeed high. And so is the state’s ratio of hospital beds per 1,000 people. Long-standing literatures on hospital bed supply and costs seem to suggest that closing beds and closing hospitals would save money. Second is feasibility. It seems to be politically possible to actually close hospitals. The hammer of closing hospitals is a tool that can be swung, with the force of legislation behind it, to hit the nail of certain hospitals. And that might be viewed as a visible political and a policy success—something that was believed to slow the rise in health care spending was actually done. Few other politically feasible steps to contain costs seem to be available. It follows that the hospitals likeliest to be forced to close would tend to be those with the least political power, allowing closings to be accomplished with relatively little political struggle. Third is symbolism. The very weakness of the policy in containing costs probably appeals to those who want to be seen to be doing something about high health care costs—without actually slowing spending growth in ways that might anger hospitals and others who don’t really want spending increases to be slowed. Fourth is self-interest of some hospitals. The most powerful and prestigious hospitals seldom expect that they would be affected by proposals to close hospitals. Rather, their competitors might be closed, thereby increasing the flow of patients to the surviving hospitals (making for higher volume and revenue at surviving hospitals) while also boosting the market power of the surviving hospitals. This would allow them to raise their prices, and also enable them to borrow money at lower interest rates. Fifth is the insulation of the decision-makers. Those who decide which hospitals must close seldom expect that the hospitals that serve them or their families—or their physicians—will be among the victims. Understandably, it’s easier to try to close other people’s hospitals. Sixth is self-interest of some bond-holders. Those who hold the bonds of hospitals slated for closing, or who insure those bonds, might benefit from a process of planned closings of hospitals that includes provisions to pay off the bonds of those hospitals—bonds that the hospitals might not have been able to pay themselves. Seventh is hope of linkage to primary care. Some who desire to see higher spending on primary care services and prevention hope or imagine that closing hospitals will somehow liberate funds that can be devoted to addressing pressing primary care and prevention needs. Many advocates of this position see hospitals as inefficient institutional dinosaurs. Eighth is high-mindedness. Hospital closings are sold as essential to the public interest in cost control, which should prevail over what is portrayed as the private or parochial interests of hospitals and of communities that selfishly fight to retain unneeded hospitals. C. THE FAILURE OF HOSPITAL CLOSINGS TO SAVE MONEY These eight reasons help to explain why some people would try to close hospitals in hopes of saving money. But evidence is more important than motives or hopes, however appealing. It is more important to assess whether closing hospitals has saved money in the past or could be expected to save money in the future. The current campaign to close New York hospitals does not seem to have been informed by careful or reasonable analysis of 9 the evidence on the causes of high hospital costs in New York State, 9 why previous efforts to save money by closing hospitals failed, or 9 alternative methods of containing hospital costs. This is somewhat surprising, given Stephen Berger’s involvement in the New York State effort of the 1970s to close hospitals and his central role in the new campaign. The best that can be expected from the current campaign is that few hospitals are closed, and that these hospitals are located in areas that will continue to be well-served, meaning that this effort does little harm and possibly a little good. Even if public efforts to close hospitals proceed in this way, they are a waste of time, at best, because they steal energy, time, and attention from identifying the real causes of high health costs, devising methods to attack those causes, and developing politically feasible ways to employ those methods. Wasting time could work to delegitimize other efforts by governments that might actually help to safely restrain health care spending. At worst, the campaign would be counter-productive—if it closes relatively low-cost and good-quality hospitals in lower-income communities at risk of under-service. The campaign would thereby accelerate the rise in health care spending and the drop in equity of access to health care. That’s because the hospitals targeted for closing will probably be those with the least money and located in the areas where they are most needed. Unfortunately, the evidence indicates that closing hospitals has not worked to save money and will not work to save money. This evidence comes in three main packages. 9 The first is the detailed analysis of hospital costs in New York State. 9 The second is the evidence from hospital closing efforts over the past 30 years. 9 And the third is the evidence from analyses of hospital closings in 52 U.S. cities since the 1930s. 1. Why Are New York State’s Hospital Costs So High? One reason why closing hospitals and beds is ineffective is that, even if over-bedding were a major source of high costs, a growing share of hospital costs are associated with outpatient care. The inpatient share of New York’s hospital’s revenue is now only about two-thirds, down from 90 percent in 1970. It is revealing to compare hospital resources, use, and costs in New York State with those of the nation as a whole. Selected comparisons are summarized in Exhibit A. Hospital cost per person in New York is 34 percent above the U.S. average, ranking 3rd among the states. But New York ranks only 15th in beds per 1,000 people. To begin to understand why hospital cost per person is so high in New York State, it is helpful to exclude some of the potential culprits. These have been widely but wrongly discussed. 9 Empty beds are not a culprit. New York ranks second-highest nationally in occupancy rate—the share of beds that are actually used. Even more important, the cost of building the beds themselves is a sunk (fixed) cost. There is no added cost unless they are occupied and staffed. Empty beds are not staffed—no hospital can afford to do so. Therefore, there’s not much reason to worry about the cost of empty beds. 9 A high rate of ER use is not a culprit. New York is only 3 percent above the U.S. average rate of ER visits per 1,000 people. What factors do seem to be important in explaining high hospital costs in New York? 9 Hospital patient-days per 1,000 people are very high, 42 percent above the U.S. average. 9 Only a small share of the extra patient-days is attributable to more admissions or higher surgery rates (10 percent and 9 percent above the national averages, respectively). 9 Instead, the extra patient-days are overwhelmingly attributable to longer stays (30 percent above the national average). New York’s long average stays don’t present a big opportunity to cut hospital costs. That is because savings from cutting length-of-stay tend to be very low. Suppose a cut in the number of hospital beds in New York State. Hospitals are likely to respond by cutting average length-of-stay, since it has remained relatively high in New York State. But this will do little to reduce the cost of hospital care, since days that can be cut are typically recuperative, and these are inexpensive for the hospital to provide. 9 New Yorkers rely heavily on hospitals to provide ordinary outpatient department (OPD) or clinic visits. They do so at rates 40 percent above the national average. 9 Hospitals in New York employ salaried doctors and dentists at double the national rate. 9 The share of patients served in major teaching hospitals, members of the Council of Teaching Hospitals (COTH) is more than double the national average share. Exhibit A Characteristics of Hospital Resources, Use, and Costs, New York State versus the Nation, 2004 NYS rank among states USA New York NYS as % of USA $1,639 $2,196 134% 3 672.8 119.5 5.6 958.6 131.0 7.3 142% 110% 130% 4 13 7 67.0% 78.9% 118% 2 Beds/1,000 2.75 3.33 121% 15 Surgeries/1000 93.3 101.5 109% 24 383.5 1,562.9 394.8 2,184.4 103% 140% 23 13 0.25 0.29 0.50 0.80 199% 279% 6 1 COTH % beds COTH % admissions 18.5% 20.7% 39.5% 45.3% 213% 218% 4 4 Operating margin Total margin 3.58% 5.18% -0.41% 1.32% -11% 25% 49 47 Cost/person Patient-days/1,000 people Admissions/1,000 people Average length-of-stay Occupancy rate ER visits/1000 OPD visits/1000 Physician + dentist FTEs/1,000 Resident FTEs/1,000 Notes: Costs per person are hospital costs divided by population. Rates per 1,000 people are the hospital data per 1,000 people. ER is emergency room. OPD is hospital outpatient department. FTE is full-time equivalents. Sources: American Hospital Association, Hospital Statistics, 2006 Edition, Chicago: American Hospital Publishing, 2006, Tables 3, 6, and 8. Outpatient care. New Yorkers’ heavy reliance on outpatient (OPD) care is a substantial source of higher cost, and one that can only be addressed by serious efforts to make primary care and specialty services much more widely available outside hospitals. Simply closing hospitals to cut OPD use might save money but is irresponsible and medically dangerous. Also, as will be shown in Part II, the harm would fall disproportionately on patients who are already vulnerable to denial of needed care— black/African-American and other New Yorkers. In the U.S. black/African-American reliance on hospitals’ ERs and especially OPDs to see a doctor is double that of whites. Cutting New Yorkers’ visits to doctors in hospital OPDs to national rates would eliminate 3.5 million OPD visits by black/African-American residents of the state, we estimate. And if hospital closings were concentrated disproportionately in black/African-American communities, the loss of visits would be appreciably higher. Similar losses are likely to prevail among Latino/Hispanic, Asian, and other groups of patients who are vulnerable to deprivation of needed care. Major teaching hospital care. High reliance on larger teaching hospitals—major medical centers—is associated with high hospital costs in New York State. New Yorkers’ reliance on larger teaching hospitals (members of the national Council of Teaching Hospitals, COTH) is more than double the national average. Supposed that New York State sought to have hospital patients to be served in COTH members at the national average. That would require shifting some 624,000 admissions out of larger teaching hospitals and into smaller teaching hospitals or community hospitals annually. We have calculated the effects on costs of closing 2,000 beds at non-teaching hospitals and shifting those admissions to major teaching hospitals. We estimate that hospital costs would rise by $1.5 billion annually as a result. The calculations that support this estimate are derived from a multivariate regression analysis, whose methods and findings are explained in detail in Appendix 1. Which of these factors have greater impacts on cost of hospital care in New York State? Results of a multivariate analysis of predictors of hospital costs per person We have performed a multivariate regression analysis to try to tease out the main independent predictors of high hospital costs across the 50 states. We compiled data on all 50 states on a number of hospital, physician, population, and other characteristics. Five independent variables formed associations with hospital cost per person that would be found by chance no more often than five times in one hundred. That is, they had a statistical significance of p = 0.05 or lower. COTH share of admissions in the state was most important, followed by OPD visits per 1,000 residents of the state, beds per 1,000 residents, the ratio of practicing office-based physicians per 10,000 residents, and labor share of total expenses. The regression equation as a whole was highly statistically significant. More important, it was practically powerful. The five independent variables explained fully 81.3 percent of the difference in hospital cost per person across the 50 states (R2 = 81.3 percent). A variable’s semi-partial correlation squared measures the share of the total variance in the dependent variable, hospital cost per person, that is uniquely accounted for by a particular independent variable. We see that the five variables, together, sum to almost 0.44—44 percent—of the total variance.3 COTH hospitals’ share of admissions accounted for about 12 percent of the total variance. So did the rate of OPD visits/1,000 residents. Hospital beds/1,000 residents accounted for 11 percent. The ratio of practicing office-based physicians to population accounted for about 5 percent, and labor share of total hospital expenses accounted for about 4 percent. This analysis strongly suggests that efforts to reduce hospital cost per person by trying to cut hospital beds/1,000 people would not be very effective. After controlling for other characteristics, it appears that hospital spending per person is not very strongly influenced by hospital beds/1,000 people. Closing beds is likely to be accomplished by closing community hospitals or smaller teaching hospitals. Some of these may even formally be COTH members. (Nationally, COTH members tend to be limited to the largest academic medical centers. New York State has 39 COTH members, fully one-seventh of the national total of 274. This suggests that some of New York’s COTH members may not be large academic medical centers.) If the patients displaced by the closing succeed in obtaining substitute care, their new hospitals are likely to be costlier than their old ones. The regression equation allows us to estimate just how much costlier. The boxed text tracks this estimation process. Shifting beds, increasing costs This analysis strongly suggests that efforts to reduce hospital cost per person by trying to cut hospital beds/1,000 people would not be very effective. After controlling for other characteristics, it appears that hospital spending per person in a state is not very strongly influenced by hospital beds/1,000 people. Further, the close correlation between COTH share of hospital admissions and cost per person indicates that if beds are closed at non-teaching hospitals, the teaching hospital share of hospital admissions will rise. This could be expected to result in an increase in hospital costs per person in a given state. This size of this relation can be estimated with reasonable precision. As discussed in Appendix 1, COTH share of hospital admissions had a Beta of 0.458635, controlling for the other independent variables in the regression equation. Thus, a one percent rise in the share of admissions to teaching hospitals will predict a rise of 0.458635 percent in hospital cost per person statewide. Hospital admissions totaled a bit over 2.5 million in New York State in 2004. And some 45.3 percent of admissions were to COTH hospitals (Exhibit A). Statewide, we have calculated, admissions per hospital bed averaged almost exactly 43 admissions in 2004.4 Suppose, then, that 2,000 beds are closed at non-COTH hospitals and that the patients formerly admitted to these hospitals are shifted to COTH hospitals.5 We are thereby shifting some 86,000 admissions. Teaching hospitals admitted some 1,139,000 patients in New York State in 2004. This rises to 1,225,000, or 48.7 percent of admissions as a result of the closing of 1,000 beds in non-teaching hospitals. That’s a rise of 7.6 percent in teaching hospitals’ share of admissions (48.7/45.3 = 1.076). Consequently, hospital cost per resident of New York State rises by 7.6 * 0.458635 (the Beta), or almost 3.5 percent. A 3.5 percent rise in hospital spending per person would have been $76.70 in 2004. This would have translated into an actual increase of $1.5 billion in 2004’s statewide hospital spending—raising spending from $42.0 billion to $43.5 billion. 2. Evidence from Efforts to Close New York and U.S. Hospitals over the Past 30 Years We have examined the evidence on hospital closings that has been compiled since the mid-1970s. This evidence does not support the hope that closing hospitals will save money. Similarly, it does not support the hope that closing hospitals will allow improved financing of badly-needed doctor care and other primary care. a. Roemer’s Law, Saving Money by Closing Whole Hospitals, Crowded Lifeboats, and Bondholders’ Hopes Six specific clusters of ideas and analysis have sparked and supported efforts to save money by closing hospitals, closing beds, or restraining growth in beds. They are: i. ii. iii. iv. v. vi. Capital spending shapes care delivery generally, and a bed built is a bed filled, specifically. The health planning perspective that reducing capacity and consolidating services can both save money and improve quality—without harming access to care. The managed care perspective that HMOs needed and used fewer beds, and a more general view that variations in patient-days per 1,000 people were not associated with differences in quality of care. The perspective that saving substantial sums through bed-reduction is best accomplished by closing entire hospitals. The crowded lifeboat perspective, which makes some hospitals happy to close their competitors. The bondholder or bond-insurer perspective that regards closing of some hospitals favorably. This short report now examines the origins, evidence, and interests supporting the first, fourth, fifth, and sixth of these ideas. i. Capital spending shapes care delivery generally, and a bed built is a bed filled, specifically In the late 1950s, Roemer and Shain proposed that a greater supply of hospital beds led to greater use.6 This proposition is sometimes phrased informally as “a bed built is a bed used.” This is sometimes called “Roemer’s Law.” In 1983, Ginsburg and Koretz examined the empirical evidence on Roemer’s Law. Reviewing a number of analyses, they concluded that an increase of 10 percent in beds per 1,000 people would typically result in about a 4 percent rise in use of hospitals by Medicare patients. Use was measured in patient-days per 1,000 Medicare enrollees.7 A little more than seven years after Ginsburg and Koretz published their analysis, the General Accounting Office examined studies conducted from 1979 to 1983 and concluded that there is only a weak relation between bed capacity and volume of use of hospital care.8 The GAO report estimated that “in 1983, a reduction of one bed per thousand population would have reduced admissions by almost 3 percent (equivalent to about $5 billion in costs). . . .” We have examined this finding. Please consider that a reduction of one bed per thousand people in 1983 would have meant a cut from about 4.2 beds per thousand to about 3.2 beds per thousand—a reduction of 24 percent. Cutting costs by $5 billion that year (out of $136.3 billion) would have meant a reduction of 3.7 percent. Cost would have fallen only about one-seventh as much as beds.9 But both of these meta-analyses considered the relation between bed supply and bed use while Medicare was still reimbursing hospitals’ costs. On 1 October 1983, Medicare abandoned cost-reimbursement and began paying hospitals by a formula set in advance— by the discharge, as clustered into hundreds of diagnosis-related groups. Once hospitals came to be paid by the discharge, they had a financial incentive to cut length-of-stay. As length-of-stay began to fall, beds emptied. This suggests that costreimbursement provided the nutrient-rich financial environment in which Roemer’s Law had some explanatory value. Abandonment of cost reimbursement may have largely repealed Roemer’s Law. iv. The perspective that saving substantial sums through bed-reduction is best accomplished by closing entire hospitals. In 1976, McClure urged closing entire hospitals to close beds to save money. Only three years later, he concluded that the types of hospitals likely to be closed—whether by financial pressures through reimbursement or markets, or by planning pressures through governments—were unlikely to save money. He focused instead on the types of hospitals that remain open, on their intensity, and on the cost of their care. McClure’s first report is far better-remembered today. His second report, which essentially repudiates the core recommendations of the first, is generally ignored. This is regrettable. McClure in 1976 prepared a powerful report in support of cutting beds to save money and urging that the greatest savings would be won by closing entire hospitals. McClure’s main findings from 1976 are worth citing in detail because they received a great deal of attention when they were published: 10 • “Hospital capacity in the United States could be reduced, conservatively, by at least 20% or more without harm to the health of the American people.” • “Similar excess capacity, both under-utilized and unnecessarily-utilized, appears to be present in high intensity service facilities.” • “Retiring entire hospitals produces substantially more savings than an equivalent reduction closing portions of several hospitals.” • “Savings from hospital capacity reduction will be gradually eroded unless hospitals are restrained from excessively increasing the labor and capital intensity of the remaining beds.” Repudiation. Only three years later, McClure published a second report. It examined case studies of hospital closings and conversions. McClure’s new findings, which received much less publicity, merit at least as much attention as his better-known work. That’s probably because they call into question the entire saving-money-by-closinghospitals enterprise, one that McClure had just strikingly endorsed. Essentially, McClure discovered that the hospitals that could actually be closed easily were inexpensive, so closing them did not save much money. Costs could rise as care was relocated to more intensive surviving facilities. These are the salient new findings:11 • “The most likely candidates for hospital closure appear to be financially marginal, older hospitals with weak constituencies. . . . Unfortunately such hospitals contribute relatively little to rising national hospital expenditures. Nationally, 12% of hospitals account for 50% of hospital expenditures, and 50% of hospitals account for 90% of expenditures. Under financial pressure it seems likely that the smaller, financially weaker hospitals will be pushed over the edge first, especially if they have no constituency that can rescue them. Under planning pressure—especially a public process that must inevitably respond as much to political muscle as to technical merits—it seems especially unlikely that the larger, costlier hospitals will succumb. Thus no matter what tools are used, an emphasis on closing capacity is likely to force out the smaller, weaker hospitals. This will require much effort but will only make a small dent in the expenditures problem.” (Emphasis supplied.) • “Even a reduction in hospital beds and patient days did not appear to slow the annual increase in hospital expenditures in the community; apparently, increased service intensity more than offset the reduction in bed use and bed capacity.” (Emphasis supplied.) • “It thus appears that the profligate growth of intensity capacity (i.e., the capacity to use labor and capital more intensively in a single bed or bed day) is as much or greater a danger to expenditure containment than bed capacity. Attention to intensity capacity may be more vital than attention to bed capacity. The best measure of capacity is expenditures. It is the capacity to generate expenditures that we are trying to reduce. It appears that beds are a very poor measure of capacity. It is not clear, for example, that closing two smaller hospitals, with no tertiary capability or aspirations, in favor of a larger hospital, with less total bed capacity but more intensity capacity, confers any net cost containment gain. (If intensity capacity is already adequate in the community, there is no quality gain either.) Rather, the more intensive hospital can use the larger patient base to further increase intensity capacity.” (Emphasis supplied.) Many people who still hope to save money by closing hospitals remain captives of McClure’s 1976 report and are not aware that he essentially repudiated it only three years later. v. The crowded lifeboat perspective, which makes some hospitals happy to close their competitors. Hospitals in a state or region sometimes point to past hospital closings and current financial distress of surviving hospitals as reasons to increase rates of payment to hospitals. But the powerful hospitals that often dominate state or regional hospital trade associations seldom fight for objective or balanced analysis of which hospitals are needed. Generally, powerful hospitals are willing to accept the closing of smaller and weaker hospitals.12 Hospitals that are financially and politically powerful—owing to some combination of size, teaching status, prestige, reputation and perceived quality, service to politically influential people, and high revenue associated with a desirable payer mix—have been unlikely to close. Survivors tend to be sanguine about the loss of their weaker competitors. If some hospitals close, and if those closings can be attributed to managerial incompetence, bad planning, and other blameworthy behaviors, then the survivors can applaud themselves for their superior competitive prowess. Winners often imagine that a putative freely competitive market sanctifies their successes. Since hospital revenue is often finite, more powerful hospitals are willing to see their weaker competitors slip overboard from the health care lifeboat (whose supplies of food and water are not sufficient for its many passengers), and may even be willing to give them a push. The closing of some hospitals means more patients for the survivors. At the same time, fewer hospitals means more bargaining power for the survivors, other things equal. In New York State, where hospitals’ financial margins are low, on average, some hospitals apparently hope that the current hospital closings/rightsizings process will benefit them. “The state intends to close under-used hospitals and improve the economic health of those that remain. . . . ” 13 During two distinct periods of fiscal crisis, powerful New York hospitals supported a policy of hospital closing, apparently because they saw that policy to be helpful—or even essential—to protect their own institutions. The first period was triggered by state government’s regulatory squeeze on Medicaid and Blue Cross payments to hospitals during the 1970s. The state took this action to hold down its own Medicaid obligations during the economic downturn that began with the first oil shock of 1973. 14 The second period began early in the 20th century. Large teaching hospitals—among them Westchester County Medical Center, St. Vincent’s, Mt. Sinai, and Nassau County Medical Center—lost large sums. The rate of hospital closings accelerated in many parts of the state. A combination of forces seems to have weakened hospitals’ finances. One is the federal Balanced Budget Act of 1997, which slowed Medicare payment increases to hospitals. A second was the repeal of the state’s hospital rate setting law, which had protected payments to hospitals. A third was the squeeze on hospitals’ revenues by some managed care plans. A fourth was the apparent high levels of borrowing, building, and debt accumulated by many hospitals. 15 Bad strategic and managerial decisions seem to have been problems at some institutions.16 But it seems that certain hospitals are likely to be protected even if they made bad managerial decisions or otherwise were suffering financially. “The consensus view is that some hospitals will be spared, no matter how poorly they fare financially. These include those that are far from other hospitals, are closely tied to medical schools, have very busy emergency rooms, or are rated as Level I trauma centers. . . .” 17 With hospital payments in New York State largely deregulated, and with more hospitals to be closed in the future, the surviving hospitals will have greater market power. That would allow them to raise prices—as has been happening elsewhere. Without either effective regulation or effective market competition, New York State would enter an era of growing hospital anarchy, with more care concentrated in fewer and costlier teaching hospitals that are newly empowered to raise prices. Hospital closings, sold to the public as a vehicle for saving money, could then be used to enable higher spending. This would not be the first time that a public policy achieved ends different from those publicly declared by its sponsors. vi. The bondholder or bond-insurer perspective that regards closing of some hospitals favorably. Similarly, those who worry about New York hospitals’ difficulty in selling bonds without public subsidies might wish to boost hospitals’ operating margins so that they can obtain lower interest rates without public support. This concern emerges loudly from the pages of the 2004 Berger Commission’s third report to Governor Pataki.18 Those who now hold some hospitals’ bonds but fear they may not be repaid might welcome a public buy-out of their investments. Similarly, those who worry about the capacity of New York’s major teaching hospitals and strong community hospitals to borrow at low interest rates might prefer to see an improvement in hospitals’ operating margins. The 2004 Berger Commission was clearly concerned by New York State hospitals’ heavy reliance on Federal Housing Administration insurance to obtain strong credit ratings and thereby borrow through the Dormitory Authority of the State of New York (DASNY). That report noted that “over 60% of FHA-insured [hospital] debt nationwide is in New York State.”19 The needs of stronger New York State hospitals for more generous operating margins and for greater ability to borrow at low interest rates without public insurance are real. In today’s health care market, hospitals often feel that they must be able to borrow to invest in new buildings and equipment to continue to attract the doctors who admit the patients who pay the bills. But it would be ironic and peculiar to boost the finances of the stronger hospitals by closing the weaker hospitals—especially since, as McClure warns, this is likely to increase total costs. If New York State moves to close hospitals and thereby builds surviving hospitals’ market power, while at the same time refusing to contain hospital costs by promoting careful cooperation among hospitals or by regulating hospitals’ revenues, then it has deliberately set down all three of the main tools for containing costs. It is willfully advancing down the road toward hospital anarchy. Higher spending on fewer surviving hospitals would mean that fewer New Yorkers will be served, but at higher cost. b. Beds close but hospital costs keep rising rapidly The nation’s number of acute hospital beds peaked around 1985 at just over one million beds. In the two decades from 1985 to 2004, the number of beds fell by about 193 thousand, or almost one-fifth. Since the nation’s population rose by 23 percent during the same two decades, it is not surprising that the number of acute care beds per 1,000 people plummeted from 4.21 in 1985 to 2.75 in 2004—a cut of 34.7 percent. This means that the most aggressive proposals of the 1970s—for a bed-to-population ratio of 3.7 per 1,000 people or less—have been much more than realized. Even so, hospital costs have generally risen at faster rates over time, as is shown Exhibit B. Remaining advocates of closing hospitals might assert that costs would have risen even faster had beds not been closed. Well, they have to say something. Exhibit B Hospital Cost/person (Consumer Price Index-adjusted) and Beds/1,000, U.S. 1946 - 2004 4.50 $1,800 4.00 $1,400 $1,200 TRENDLINE 3.50 $1,000 $800 3.00 Beds/1,000 $/person, CPI-adjusted $1,600 $600 $400 2.50 $200 $0 2.00 1946 1950 1955 1960 1965 Beds/1,000 1970 1975 1980 $/person 1985 1990 1995 2000 2004 Linear ($/person) Exhibit B shows that there is no association between hospital cost per person and beds per 1,000 people over the years from 1946 to 2004. 20 Beds per 1,000 Americans rise and fall over time. But cost per person rises fairly steadily. Importantly, cost per person rises relatively rapidly during the years after 1985, when beds per 1,000 people falls substantially. In Exhibit B, this is captured by the actual rise in spending per person—the green line— rising above the trendline. The trendline reflects the increase in spending per person over the full 58 years from 1946 to 2004. The correlation between inflation-adjusted hospital cost per person and beds per 1,000 people over time is Rp = -0.2973, R2 = 8.8%, significant at p = 0.324. Change in one variable over time predicts only about 9 percent of the change in the other. What is salient here is not the statistical significance but the inverse or negative relationship between beds and costs: Over time, more beds per thousand people are associated with lower costs per person.