Health Affairs, 22, no. 6 (2003): 12-26 doi: 10.1377/hlthaff.22.6.12 © 2003 by Project HOPE	New Online   How Would Obama, McCain Cover The Uninsured?   Debating Cost Of Uninsured   Try Medicare-For-All   HA Blog Top 10

This Article Abstract Figures Only Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Shactman, D. Articles by Doonan, M. Search for Related Content PubMed PubMed Citation Articles by Shactman, D. Articles by Doonan, M. Related Collections Hospitals Health Spending

Supply & Demand

The Outlook For Hospital Spending

Abstract

 
Hospital use and spending greatly increased in 2001 and 2002, reversing a long-term trend. In this paper we contend that the forces driving current hospital expenditures are more likely to continue than they are to abate. If current trends continue, real hospital spending per capita will increase 75 percent between 2002 and 2012, and the demand for hospital beds will increase considerably. We discuss numerous forces that will contribute to spending growth, including technology, which is likely to continue to raise costs. We also find that hospital spending by baby boomers grew more rapidly than that of the elderly, a change in trend that could presage increased spending as this cohort moves into higher-spending age groups.

Nevertheless, most private and public analysts are predicting that recent trends in hospital and total health spending growth will moderate over the next ten to twenty years. Recent ten-year projections by the Centers for Medicare and Medicaid Services (CMS) are partly based on their reasoning that increases in total health spending will cause a resurgence in managed care that will reduce the growth in use of hospital and other health care services.² Based on CMS data, the Center for Studying Health System Change (HSC) reports that the growth rate of non-Medicare personal health care spending per capita will decline between 2001 and 2010.³ Some analysts predict that declining rates of disability will greatly reduce the growth in Medicare hospital and nonhospital spending in future years.⁴

In this paper we take a different view. Despite evidence that hospital use slowed through the first half of 2003, we contend that the factors that have driven demand for hospital services in the past few years are more likely to continue than to abate. Advances in medical technology, growth and aging in the population, and increases in the propensity of baby boomers and younger age cohorts to use health care services could cause hospital spending to greatly increase between now and 2012. Even the CMS projections, which we believe are too conservative, predict a 55 percent increase from 2000 to 2012. Given current capacity constraints, the increased utilization could bring about a surge in demand for construction of more inpatient and outpatient hospital beds and a concomitant demand for capital to finance this growth.

We begin by examining past trends in hospital spending. We then simply extrapolate current trends to illustrate what would happen to hospital spending if nothing changed. Examining these trends, we decompose some of the components of use and spending growth and estimate the impact on the need for hospital beds. Comparing current trends to projections made by the CMS, we pose the question, Which is more likely: a continuation of current trends or a reduction in many of the components of recent spending growth? We then identify a number of factors that could drive use beyond what many analysts predict.

Examining Past Trends

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
Although total hospital spending has grown continuously since 1960, the rate of spending growth has changed direction a number of times (Exhibit 1). For example, after rising at an increasing rate through the 1960s, the growth rate slowed through the 1970s and early 1980s, accelerated from 1985 to 1990, and then fell sharply from 1991 to 1997. The industry responded to the latest downturn in demand by merging and consolidating hospitals, transforming inpatient beds to other uses, and closing some facilities entirely. The number of community hospitals fell from 5,292 in 1992 to 4,908 in 2001, and the number of beds from 921,000 to 826,000 in the same period.⁵ Some analysts, such as Jeff Goldsmith, predicted that hospitals would no longer be the center of the health care system.⁶ Their use would continue to decline until they cared only for trauma patients and the very old with chronic medical care needs.

View larger version (20K): [in this window] [in a new window]   EXHIBIT 1 Annual Hospital Spending Growth, Actual And Trend, GDP Price Deflated, 1961–2001 SOURCE: Authors’ calculations based on S. Heffler et al., "Health Spending Projections for 2002–2012," 7 February 2003, www.healthaffairs.org/WebExclusives/Heffler_Web_Excl_020703.htm (10 July 2003); and Bureau of Economic Analysis, U.S. Department of Commerce NOTE: GDP is gross domestic product.

However, in the late 1990s the trend in inpatient spending reversed direction, and the growth rate for privately insured people changed from –5.3 percent in 1997 to + 7.1 percent in 2001.⁸ Per capita admission rates, which bottomed out in 1998, rose over the next four years.⁹ Although spending for outpatient care continued to rise, inpatient spending is now rising at a relatively faster rate, and the CMS predicts that the outpatient and inpatient growth rates will nearly converge by the year 2012.¹⁰ Equally important, the long-term downward trend in length-of-stay appears to have hit bottom, and it has even begun to increase in several states.¹¹

We do not know all of the reasons for the unexpected resurgence in inpatient use or for the leveling off in length-of-stay. Among the possible factors are weaker managed care, aging of the baby boomers, application of new technologies, and limits to the ability to reduce lengths-of-stay. Regardless, the change in in-patient/outpatient service mix and length-of-stay during the past several years has had a notable effect on the growth in hospital spending. We see this in the next section as we examine current trends.

Projecting Current Trends

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
Projection. We begin by making a simple straight-line extrapolation of hospital spending from 1999 to 2001, continuing the trend through 2012. Although this is not a prediction, it provides a useful perspective, telling us what would happen if the current forces generating the increase in hospital spending continue unchanged. To establish this trend line (data not shown), we began with the CMS data for per capita hospital spending in 1999–2001. We then added the impact of population growth, using the middle estimates from the Census Bureau’s Current Population Survey (CPS). To include the impact of population aging, we multiplied the Census Bureau’s age-specific population estimates by age-specific per capita spending rates derived from the Medical Expenditure Panel Survey (MEPS, 1997–1999 average), and then subtracted the amount attributable to population growth alone. The age-specific spending rates were held constant to isolate the impact of changes in age composition. To make our trend line consistent with the CMS projections, we adjusted our data by the gross domestic product (GDP) price deflator. Although we refer to this trend line as "real hospital spending," it includes the excess of hospital inflation over general (economywide) inflation.

Our extrapolation suggests that if current trends continue and nothing else changes, real hospital spending between 2000 and 2012 would increase by an average annual rate of 4.8 percent, or a total increase in real hospital spending of 75 percent over the twelve years. This is a substantial rate of growth, but even the more conservative projections of the CMS suggest that real hospital spending could grow by 55 percent over this time period. Under either scenario, hospital spending will grow much faster than GDP. By 2012 it will equal 5.6 percent of GDP according to our trend line or 4.9 percent according to the CMS. In comparison, hospital spending was 4.2 percent of GDP in 2000.

Decomposition. Of the 4.8 percent annual real growth rate in hospital spending, 0.5 percentage points (10 percent) is attributable to the aging of the population and 0.9 percentage points (18 percent) is the result of population growth. Of the remaining 3.5 percentage points, 1.4 percentage points (29 percent) are attributable to the excess of hospital inflation over the amount of inflation in the general economy.¹² The balance of 2.1 percentage points (43 percent) represents the use of new technologies and greater use of hospital care throughout the population. Although population aging is often cited as a reason for increased spending, it only accounts for 10 percent of spending growth over this period.

Estimating Demand For Hospital Beds

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
A critical question is what impact such increases in demand will have on hospital capacity. Can the current supply of hospital beds support such growth, or will the hospital industry need to reverse the decade-long reduction in bed capacity?

Recent estimates of the need for more hospital capacity vary. The Health Care Advisory Board predicted in its middle estimate that hospitals would need to add 40 percent more capacity between 2000 and 2010.¹³ Solucient LLC estimated that demographic changes alone would require 46 percent more acute care beds during 2002–2027.¹⁴ Assuming the same perspective as earlier, we asked how many additional hospital beds we would need if current trends continued and the forces driving hospital spending remained unchanged. Again, this is not a prediction, but it provides a rough estimate from extrapolating current trends.

Using linear regression, we trended forward revenues (as previously) and revenues per admission. We assumed that length-of-stay would remain constant and that the growth rates of inpatient and outpatient revenues would be approximately the same (as both were discussed earlier). We multiplied projected revenues by admissions per revenue to derive admissions and then multiplied that by length-of-stay to derive inpatient days. To calculate bed demand from inpatient days, we assumed that, on average, a bed would be used 75 percent of the time. Finally, since the hospital industry is not at full capacity, we calculated the number of beds hospitals would have to build if, overall, they could absorb 10–20 percent more bed demand before needing to add capacity.

"Labor shortages, especially those of nurses and various technicians, have been an important driver in hospital costs."

Under those assumptions, if hospitals can absorb 10 percent more capacity, they will have to increase the number of beds by 211,000, or 28 percent, over the period 2000–2012. If they can absorb 20 percent more capacity, they will have to increase the number of beds by 136,000, or 18 percent.

These are clearly rough estimates based on numerous assumptions. Nevertheless, they provide a sober warning: If recent trends were to continue unabated, hospitals would have to add substantial capacity, and the capital investment required to do so would be significant. Even if spending growth were to follow the more conservative CMS trend lines, hospitals would need 18 percent more beds, assuming that they could absorb 10 percent more capacity. Given even these rough estimates, the era of excess hospital capacity appears to have ended.

Comparing Current Trends To CMS Projections

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
Having examined the result of extrapolating current trends as an upper-bound estimate that assumes no change in current forces, we examine the CMS projections to ascertain if they seem more likely. We find that the CMS projections are dependent on sharp reductions in several of the cost components that make up current trends. These components include labor costs, the shift away from MCOs, hospital admissions, and total inpatient spending (Exhibit 2).

View larger version (19K): [in this window] [in a new window]   EXHIBIT 2 Medicare Trustees’ Projections Of Effect Of Various Hospital Cost Components, 2001–2004 SOURCE: 2003 Annual Report of the Board of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds (Baltimore: Centers for Medicare and Medicaid Services, 2003).

Shift away from MCOs. The managed care shift represents the annual change in spending because of the shift in population away from MCOs and their implementation of utilization controls. The Medicare trustees report a growth rate in hospital spending of 2.2 percent in 2001 and 2.4 percent in 2002 (each over the previous year) but then virtually eliminate any spending growth attributed to this component in 2003. However, well into 2003 we see no evidence of a return to more tightly managed care.

Admission rates. The admission incidence is the rate of inpatient admissions among Medicare beneficiaries. The trustees’ projections show a growth rate of 1.2 percent in 2003 (over 2002), but this factor is reduced to almost zero in 2004 and then turns negative.

In total, the Medicare trustees estimate that the growth rate of Medicare hospital payments (which are derived from these and numerous other components) will fall sharply, from 10 percent in 2002 to 4.9 percent in 2003, before recovering modestly to 5.8 percent in 2004. A key factor here is what will happen to the Medicare hospital payment rate. Although budgetary pressures will likely restrict future payment increases, Medicare hospital payments have never fallen much below cost. Since hospital cost increases are expected to continue, any reduction in this component of spending growth is likely to be limited.

In addition to projecting this steep reduction in the growth rate of Medicare hospital spending, the CMS projects the annual growth rate in total Medicaid spending (not just hospitals) to fall from 12.1 percent in 2002 to 8.8 percent in 2004–2008.¹⁷ Concurrently, it projects the growth rate in private hospital spending to fall from 7.7 percent in 2003 to 5.0 percent in 2012. Stephen Heffler and colleagues project "that a slowing economy will resurrect pressures—captured in part by the HMO proxy—to restrain hospital spending growth."¹⁸ This prediction may well turn out to be accurate, but one can observe that (1) a slowing economy ordinarily increases the growth of Medicaid spending; (2) we see no evidence of either an upswing in MCO membership or a return to aggressive utilization controls; and (3) the slowing of the economy is not likely to persist very much longer, if at all. In the longer term, predictions by the CMS actuaries are constrained by their methodology, which requires spending to increase at the rate of growth of GDP plus 1 percent. For the CMS baseline to be a more likely prediction than a simple projection of current trends, many of the components driving spending would have to change very quickly, all in the appropriate direction.

Other Factors Pointing To A Higher Trend

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
Previously we accounted for population growth, population aging, the general rate of inflation, and the additional amount of hospital inflation. In our decomposition, 43 percent of projected hospital spending growth was attributed to "the use of new technologies and the greater use of hospital care throughout the population." In the remainder of the paper we discuss three factors that drive this residual: baby boomers’ propensity to consume, medical technology, and health status.

Baby boomers’ propensity to consume. The conventional wisdom in the United States is that the elderly consume an increasingly large proportion of health care expenditures. Data show that from 1953 to 1987, spending on the elderly increased at a much faster rate than spending on the nonelderly.¹⁹ We now find evidence that this long-standing trend may have ended.

Unlike their parents and grandparents, baby boomers and younger age cohorts have been educated and reared in a society in which there is an expectation of a medical cure for almost any condition. It is therefore possible that their expectations and, hence, their propensity to consume medical services could exceed those of older generations. Furthermore, they have higher disposable incomes and better insurance coverage than past generations. This is significant because we know that in developed countries health care is a luxury good, and the long-term income elasticity of demand for health services is quite high.²⁰

To determine if the relative spending patterns between older and younger generations have changed, we measured the growth rate in spending for different age groups from 1987 to 1999. We adjusted data from the 1987 National Medical Expenditure Survey (NMES) to make them comparable to data from MEPS.

Comparative hospital spending. The data show that hospital spending by baby boomers (ages 31–50) over the eleven-year period grew at a rate 2.3 times faster than for older cohorts and nearly 2.5 times as fast as for the elderly (those over age sixty-five) (Exhibit 3). Spending by people up to age thirty grew close to four times the rate of people over age fifty, but their growth was dominated by spending on newborns—by far the fastest-growing spending segment of the population. The growth rate of the 21–30 age group (who use little hospital care) was 1.29 percent, just under 1.2 times the rate of the cohort over age fifty.

At this point, we do not know whether these adjusted survey data will be borne out by further research or, if they are, whether this trend is likely to continue. Furthermore, having just identified this phenomenon, we do not know what factors may have brought it about. However, we can cite three factors that are consistent with an increased propensity to consume on the part of the baby boomers.

Growth in total baby-boomer medical spending. First, our analysis of the NMES and MEPS data shows that the growth rate in total medical spending between 1987 and 1998 was also higher for the boomers and younger cohorts than it was for those over age fifty (Exhibit 3).

The growth rate in spending for the baby boomers exceeds that of the over-fifty age cohorts by 1.2 times, as does the growth rate in spending by people up to age thirty (Exhibit 3). This is important because one of the limitations of this study is the inability to control for the increasing shift of patients from acute to postacute care. David Cutler and Ellen Meara attributed much of the spending growth by the oldest old in the 1990s to increases in postacute care.²¹ These results show that even if postacute care is included, spending for the nonelderly increased faster than spending by the elderly during this period.

Prescription drug use. Second, drug use increased at a faster rate among the baby boomers than it did for older generations. Between 1987 and 1998 drug spending for the 31–50 age cohort increased at an annual per capita rate of 12.03 percent, compared with 10.87 percent for those over age fifty, a differential increase of 10.7 percent (Exhibit 3). In this case, spending by the 0–30 cohort increased at a slower rate than for those over age fifty, except among the 0–5 age group, for which spending increased by 10.99 percent. Increased drug use does not necessarily translate into greater hospital spending. However, it is consistent with a greater propensity to use more health care, which the baby boomers have exhibited even while their overall health status has improved.

Higher incomes. Third, we note that the median real income of people ages 35–54 rose 28 percent between 1981 and 2001.²² Health care is a luxury good, and the long-term income elasticity of demand for health services across Organization for Economic Cooperation and Development (OECD) countries has been estimated to be 1.3–1.6.²³ As the capabilities of medicine grow, and as surgical techniques become safer and less invasive, many more elective and "lifestyle" procedures can be undertaken in inpatient and outpatient hospital settings. These include such procedures as hip, knee, and other joint replacements; facial cosmetic surgery; breast enlargement and reduction; weight reduction; and a potential growing variety of fertility-related and genetic techniques. End-of-life care could become more intensive and aggressive with increased use of experimental treatments such as bone and stem-cell transplants. Furthermore, complex batteries of diagnostic procedures using imaging and genetic testing, even in the absence of illness, could become affordable and commonplace.

Potential impact on hospital spending. If, in fact, this trend of increased spending growth among the nonelderly continues, how would it affect hospital spending? We estimated the magnitude of this effect for the baby-boom cohort (ages 31–50 in 1998). To do this, we estimated the average per capita spending in 2020 under two scenarios. In scenario one we assumed that for each age group the rate of spending growth would be the same between 1998 and 2020 as it was between 1987 and 1998. In scenario two we assumed that the baby-boom cohort, over the next twenty years, would continue its past rate of spending growth, but this rate would be applied to the spending levels appropriate to the older age group.

Our computations show that scenario two adds an additional 0.46 percent to the average annual growth rate of per capita hospital spending. Most of this growth will occur in the later years, when all of the baby boomers are in the 51–70 age group. This figure may be too low, however, because it only reflects the impact of continued high spending growth rates of the baby-boom cohort. If younger cohorts also exhibit this effect, the impact on spending would be greater.

Caveats. There are limitations to our findings. First, as we stated earlier, our analysis does not adjust for the increased use of postacute care during the study period. Second, our analysis was dependent upon survey data and upon adjusting the NMES data to be comparable to the MEPS data. As we used the age-specific breakdown, our analytic power was limited by some small cells. Third, the even greater spending rate in the under-age-thirty category was disproportionately affected by spending on newborns. This will continue to be an important factor in driving the use of inpatient hospital services, but the enormous growth rate that occurred in the study period is likely to abate. Finally, new technologies could have been a stronger cost driver for younger age groups than for the elderly during this period. We think that this is unlikely, although it is more apt to be a possibility in the case of pharmaceutical spending.

"It is unlikely that technology will be curtailed, and it will continue to improve the length and quality of life."

Medical technology. Technology has been identified as the most important variable driving total health care costs.²⁴ Will technological innovation increase or decrease the use and cost of hospital services? This section explores two arguments. The first is that technology will be curative and reduce hospital use over the long term. The second is that technology is cumulative, enabling doctors and hospitals to do more and better things. In either case, it is unlikely that technology will be curtailed, and it will continue to improve the length and quality of life.

Technology will reduce hospital use. The first argument is that technological breakthroughs will keep people healthier and out of hospitals. Advances in disease management, biomedical science, and control of health risk factors will combine to reduce the incidence of death and disability, leading to less demand, particularly for inpatient hospital care.²⁵ Under this scenario, people will live longer, healthier, disease-free lives until they are very old. The theory, developed by James Fries in 1980, is referred to as the "compression of morbidity."²⁶ Although health care costs generally increase with age, they dip down for the very old as less intensive treatment is employed and, in some cases, replaced by hospice care.

Most medical breakthroughs introduced in the past fifty years have been "halfway technologies." They treat the symptoms of disease at considerable cost, particularly in the hospital. Examples include much of the current technology used to treat cancer, including chemotherapy. In contrast, "full technologies" either cure or prevent disease. The best examples are vaccines that eradicated certain diseases and prevented others. If new biomedical genetic technology "identifies the molecular mechanisms underlying disease entities," disease can be prevented and costly treatment avoided.²⁷ Promise can be seen in the completion of the human genome sequencing and the continued growth in budgets for the National Institutes of Health (NIH). Full technologies have the potential to greatly reduce hospital care.

Disease management and increased use of pharmaceuticals hold perhaps the best near-term promise to reduce hospital costs. Cindy Thomas and colleagues identify a number of cost-effectiveness studies that demonstrate hospital savings from drug therapy in the area of cancer, AIDS, congestive heart failure, and other diseases.²⁸ Although it is difficult to isolate and determine the impact of all prescription drugs, some breakthrough drugs have reduced total medical costs. J.D.Kleinke makes the case that pharmaceuticals are moving the health care system away from labor toward technology, creating increased efficiency and a "bargain for society."²⁹ "Increased spending on drugs that specifically manage disease, preclude or delay surgeries, or reduce hospital admissions and lengths-of-stay pay for themselves many times over," Kleinke writes.³⁰

Technology will increase hospital use. The second argument is that technology drives use and costs and will expand the need for hospital services. In this case, medical advances do not necessarily replace old technology but provide additional tools and methods of treatment. Computed tomography (CT) scans did not reduce the need for x-rays, and magnetic resonance imaging (MRI) scans did not eliminate the need for CT scans. Treatment for heart disease, the leading cause of death in the United States, provides a clear example of the cumulative nature of technology. Victor Fuchs shows that the rapid increase in the number of angioplasties did not lower the number of coronary bypass procedures between 1987 and 1995. Instead, the number of both procedures greatly increased.³¹ Even when the incidence of heart disease was reduced by a decline in smoking and the use of hypertension and cholesterol-lowering drugs, more intense treatment accompanied by improved outcomes increased the use of hospital services.

Medical technology, including new drugs and surgical procedures such as angioplasties and hip replacements, makes it possible to do more for each patient and to intervene with more patients.³² Improved technology enables more to be done in the hospital, not less. According to Fuchs, technology diffusion is generally slow at first, until greater confidence and familiarity lead to increased use.³³ Technology is then applied and adapted in ways not previously considered.

The promise of "full" technologies. The strongest argument for decreasing hospital utilization places faith in full technologies that cure disease. However, evidence suggests that despite great advances, dramatic cures are still unlikely in the near future. Biomedical advancements and gene therapy hold enormous promise. However, identifying the genes responsible for particular problems is a long way from developing therapies that prevent disease. One major hurdle is developing a delivery system to insert the genetic material into the target cells.³⁴ In addition, while currently untreatable diseases may eventually be treatable, there are likely to be adverse reactions, particularly in the early stages of diffusion, that will likely require intensive care. Although major scientific advancement can be realized with spectacular breakthroughs, the application of technology is generally a more incremental process of trial and error. Much of this will be done in hospitals.

Finally, there is no guarantee that "full technologies" will result in decreased hospital use over the life of a patient. Preventing early death can lead to increases in health cares costs as people age. Burton Weisbrod and Craig LaMay state that "the use of antibiotics to prevent death from infections can cause people to live longer and hence to die from heart disease and cancer, which typically entail greater costs."³⁵ Further, Kleinke states that liver transplants, which were enabled in part by the development of antirejection medication, are expensive and increase the life expectancy for tens of thousands with liver disease who will live to consume more health care services in the future.³⁶ Even the baby-boom seniors, with improved health status, will need hip replacements and knee and back procedures to maintain active lifestyles.

"Continued reductions in disability could be a force to reduce both hospital costs and use."

Net impact on hospital use. There will certainly be both cost-increasing and cost-reducing developments in technology during the next ten years. But what will be the net impact on hospital use? Arguments for a decrease in hospital use depend on the development of full technologies that do not yet exist. The development of new medications is likely to bring about some reductions in inpatient hospital stays. Nevertheless, technology diffusion and evidence that more and better services are being done in the hospital argue that technology will continue to expand the demand for hospital services as it has done for many years. We are still far from being able to turn biotech advances into full technologies, and getting there is likely to require increased hospital use. Furthermore, history tells us that preventing early deaths may require increased medical services down the road. On balance, we believe that technology will lead to greater use of hospital services, particularly in the short and medium terms, and the benefits to society will be significant.

Changes In health status. The final component we consider is changing health status, particularly of the elderly. Kenneth Manton and XiLiang Gu observe that the disability rate for the elderly declined by 1.6 percent per year from 1989 to 1994 and by 2.6 percent from 1994 to 1999.³⁷ Burton Singer and Manton suggest that continued declines will greatly reduce hospital and other health spending by the elderly and will thereby preserve the solvency of the Medicare trust fund until 2070.³⁸ We agree that continued reductions in disability could be a force to reduce both costs and use. Nonetheless, we offer several reasons why the extent of the impact predicted by Manton and his colleagues is unlikely.

Reduction in disability already incorporated into current trends. Manton and Gu state that disability has been declining in the United States since the turn of the century, and they specifically track the decline since the National Long-Term Care Survey in 1982.³⁹ However, despite this long-term reduction in disability, per capita medical and hospital spending has steadily increased, and health spending has risen as a percentage of GDP. If the rate of disability had not declined, current trend lines for health care and hospital spending would certainly be higher. But the effect of reduced disability has long been incorporated into these trends. Current trend lines should only be reduced if the prevalence of disability decreases at a rate that exceeds that of recent years. In fact, Manton and Gu find that the rate did decrease faster in the 1990s than in the 1980s. However, the possibility that rates will continue to decline faster and faster into the long-term future seems highly unlikely. What seems more likely is that the decline will continue, but the rate of decline will regress toward the long-term mean. If that is the case, the impact of declining disability on spending and use will actually be reduced.

Complex relationship between disability and health care use. The relationship between disability and health care use is neither simple nor unidirectional. Although reduced disability can often result in lower health spending, it can also be true that more health spending is needed to reduce disability. The hip and knee replacements, angioplasties, and cardiac catheterizations that could prevent or delay expenses related to disability are themselves costly, hospital-based procedures. Age-specific health status is improving, but it is difficult to tease out the upstream costs leading to those improvements. Reduced disability also shifts the locus of costs across health care providers so that there may be additional downstream costs. Hence, deinstitutionalization, which is a positive result from reduced disability, can shift additional costs from skilled nursing facilities and nursing homes to hospital outpatient and inpatient services. Stephane Jacobzone states, "In spite of the optimism that may arise from the positive trends in disability, the implications for health care services are more likely to be linked with increases in cost."⁴⁰

Impact of obesity on health status. Although Manton and others have documented reductions in the prevalence of disability in the United States, there is an important factor working in the opposite direction. Although there were only minor increases in obesity between 1960 and 1980, there was an astonishing increase of 60 percent from 1991 to 2000. Nearly one-fourth (23 percent) of the U.S. population is now classified as obese, and 33 percent are overweight.⁴¹ This has major health implications, since obesity is strongly associated with chronic illness and hospital use. Roland Sturm found that obesity is associated with a 36 percent increase in inpatient and outpatient hospital spending.⁴² Because the steep rise in prevalence occurred fairly recently, and because there is a time lag between the onset of obesity and the development of chronic conditions, it is likely that much of the impact of increased obesity has not yet been factored into current trends in hospital use and spending. Unlike the reduction in disability that has long been built into current trends, much of the impact of increased obesity will increase these trend lines in the future. And this is not a small effect. If 23 percent of the population spends 36 percent more on inpatient and outpatient hospital care, even over a moderate period of time hospital use will increase much more than is currently projected.

Summary And Concluding Remarks

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
The rate of growth of hospital spending began to rise in 1998, reversing a long-term trend. Contrary to the predictions of many analysts, we believe that spending and use are more likely to follow current trends than to abate. If current trends were to continue, hospital spending would increase by 4.8 percent annually between 2000 and 2012, a total real spending increase of 75 percent. This rate of growth would result in the need for much more hospital capacity.

A number of factors will drive the growth in hospital spending and use. Of the 4.8 percent real annual spending growth presaged by current trends, increased population will account for 0.9 percent. Although many identify aging as a major force, we estimate that in this time period aging will account for just 0.5 percent. However, we identify a factor we call the age-specific propensity to consume: the trend of baby boomers and younger generations to have higher growth rates in health and hospital spending than their older counterparts. If this trend continues as the baby boomers age into higher spending categories, it could cause a considerable increase in the demand for hospital services.

Technology remains the chief driver of hospital spending. Although full technologies and cost-reducing innovations are predicted for the future, for the most part they have not yet arrived. Hence, most new technological innovations in the next ten years are still likely to be cost increasing.

The one factor working against increased hospital spending is improved health status and declining disability across the U.S. population. However, we contend that the impact of improved health status has already been built into recent trends. Furthermore, the recent epidemic in obesity could mitigate improvements in health status over the next ten years.

Editor's Notes

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 
An earlier version of this paper was presented at the conference, "The American Hospital: What Does the Future Hold?" in Washington, D.C., 21 April 2003. The authors gratefully acknowledge research support from the Robert Wood Johnson Foundation, the research assistance of Curtis Florence and Grant Ritter, and the helpful comments of two anonymous reviewers.

David Shactman is a senior fellow at the Schneider Institute for Health Policy, Heller School for Social Policy and Management, Brandeis University, in Waltham, Massachusetts. Stuart Altman is the Sol C. Chaikin Professor of National Health Policy at the Schneider Institute; Efrat Eilat is a research associate and Michael Doonan is an assistant professor there. Ken Thorpe is the Robert W. Woodruff Professor and Chair at Emory University’s Rollins School of Public Health in Atlanta, Georgia.

NOTES

Top Examining Past Trends Projecting Current Trends Estimating Demand For Hospital... Comparing Current Trends To... Other Factors Pointing To... Summary And Concluding Remarks Editor's Notes NOTES

 

1. S.Heffler et al., "Health Spending Projections for 2002–2012," 7 February 2003, www.healthaffairs.org/WebExclusives/Heffler_Web_Excl_020703.htm (10 July 2003).
2. Ibid.
3. B.C.Strunk and P.B. Ginsburg, "Aging Plays Limited Role in Health Care Cost Trends," Data Bulletin no. 23 (Washington: Center for Studying Health System Change, September 2002).
4. B.Singer and K.G. Manton, "The Effects of Health Changes on Projections of Health Service Needs for the Elderly Population of the United States," Proceedings of the National Academy of Sciences (December 1998): 15618–15622.
5. American Hospital Association, Hospital Statistics (Chicago: Health Forum, AHA, 1998, 2003).
6. J.Goldsmith, "A Radical Prescription for Hospitals," Harvard Business Review (May–June 1989): 104–111; and J.Goldsmith, "Can Hospitals Survive? The New Competitive Health Care Market" (Homewood, Ill.: Dow Jones–Irwin, 1981).
7. Authors’ calculations from Health Care Financing Review (various issues) and from AHA, Statistical Guide (2003).
8. Strunk and Ginsburg, "Aging Plays Limited Role."
9. AHA, Hospital Statistics (various years).
10. Heffler et al., "Health Spending Projections for 2002–2012."
11. AHA, Hospital Statistics (2003).
12. Data derived from the 2003 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds (Baltimore: Centers for Medicare and Medicaid Services, 2003).
13. Health Care Advisory Board, "The New Economics of Care: Briefing for the Board and Health System Executives" (Washington: Advisory Board Company, Fall 2001).
14. Solucient LLC, "National and Local Impact of Long-Term Demographic Change on Inpatient Acute Care," November 2002, www.solucient.com/docs/Long_Term_Demo_Change.pdf (8 August 2003).
15. Data derived from the 2003 Annual Report of the Boards of Trustees.
16. P.Buerhaus, "Demographics of the Registered Nurse Workforce: Trouble Now, Big Trouble Ahead" (Paper presented to the Council on Health Care Economics and Policy, Washington, D.C., 14 December 2001).
17. Heffler et al., "Health Spending Projections for 2002–2012."
18. Ibid.
19. D.Cutler and E. Meara, "The Medical Costs of the Young and Old: A Forty Year Perspective," NBER Working Paper no. 6114 (Cambridge, Mass.: National Bureau of Economic Research, 1997).
20. W.J.Moore et al., "Measuring the Relationship between Income and National Health Expenditures," Health Care Financing Review 14, no. 1 (1992): 133–139.[Medline]
21. D.Cutler and E. Meara, "The Concentration of Medical Spending: An Update," NBER Working Paper no. 7279 (Cambridge, Mass.: National Bureau of Economic Research, 1999).
22. U.S. Census Bureau, "Historical Income Tables—People," 30 September 2002, www.census.gov/hhes/income/histinc/p10.html (8 August 2003).
23. Moore et al., "Measuring the Relationship."
24. V.R.Fuchs, "Health Care for the Elderly: How Much? Who Will Pay for It?" NBER Working Paper no. 6755 (Cambridge, Mass.: NBER, 1998).
25. C.Thomas et al., "Health Status, Technological Innovation, and Health Care Expenditures" (Paper presented to the Council on Health Care Economics and Policy, Washington, D.C., February 1999).
26. J.F.Fries et al., "Beyond Health Promotion: Reducing the Need and Demand for Medical Care," Health Affairs (Mar/Apr 1998): 70–84.
27. Thomas et al., "Health Status, Technological Innovation, and Health Care Expenditures."
28. Ibid.
29. J.D.Kleinke, "The Price of Progress: Prescription Drugs in the Health Care Market," Health Affairs (Sep/Oct 2001): 46.
30. Ibid.
31. Fuchs, "Health Care for the Elderly."
32. Ibid.
33. Ibid.
34. W.B.Schwartz, Life without Disease: The Pursuit of Medical Utopia (Berkeley: University of California Press, 1998).
35. B.A.Weisbrod and C.L. LaMay, "Mixed Signals: Public Policy and the Future of Health Care R&D," Health Affairs (Mar/Apr 1999): 116.
36. Kleinke, "The Price of Progress," 46.
37. K.G.Manton and X. Gu, "Changes in the Prevalence of Chronic Disability in the United States Black and Nonblack Population above Age Sixty-five from 1982 to 1999," Proceedings of the National Academy of Sciences (May 2001): 6354–6359.
38. B.Singer and K.G. Manton, "The Effects of Health Changes on Projections of Health Service Needs for the Elderly Population of the United States," Proceedings of the National Academy of Sciences (December 1998): 15618–15622.
39. Manton and Gu, "Changes in the Prevalence of Chronic Disability."
40. S.Jacobzone, "Coping with Aging: International Challenges," Health Affairs (May/June 2000): 213–225.
41. R.Sturm, "The Effects of Obesity, Smoking, and Drinking on Medical Problems and Costs," Health Affairs (Mar/Apr 2002): 245–253.
42. Ibid.

                      What's this?

This article has been cited by other articles:

				E. Kramarow, J. Lubitz, H. Lentzner, and Y. Gorina Trends In The Health Of Older Americans, 1970 2005 Health Aff., September 1, 2007; 26(5): 1417 - 1425. [Abstract] [Full Text] [PDF]

				B. C. Strunk, P. B. Ginsburg, and M. I. Banker The Effect Of Population Aging On Future Hospital Demand Health Aff., May 1, 2006; 25(3): w141 - w149. [Abstract] [Full Text] [PDF]

				J. M. Hoffman, N. D. Shah, L. C. Vermeulen, R. J. Hunkler, and K. M. Hontz Projecting future drug expenditures--2005 Am. J. Health Syst. Pharm., January 15, 2005; 62(2): 149 - 167. [Abstract] [Full Text] [PDF]