Providing Insights that Contribute to Better Health Policy
Insurance Coverage & Costs Costs The Uninsured Private Coverage Employer Sponsored Individual Public Coverage Medicare Medicaid and SCHIP Access to Care Quality & Care Delivery Health Care Markets Issue Briefs Data Bulletins Research Briefs Policy Analyses Community Reports Journal Articles Other Publications Surveys Site Visits Design and Methods Data Files
Looking Behind the Numbers: What's Driving Health Care Costs
Statement of Paul B. Ginsburg, Ph.D.
Committee on Education and the Workforce
Thank you Mr. Chairman, Congressman Andrews and members of the subcommittee for inviting me to testify about rising costs in employer-sponsored health insurance. I am Paul B. Ginsburg, President of the Center for Studying Health System Change (HSC). HSC is an independent nonpartisan policy research organization funded solely by The Robert Wood Johnson Foundation and affiliated with Mathematica Policy Research. Our longitudinal, nationally representative surveys of households and physicians and site visits to 12 U.S. communities provide a unique perspective on developments in the private health care markets and their impacts on people. Although we seek to inform policy with timely and objective analyses, we do not lobby or advocate for any particular policy position.
As an economist, I have personally studied health care cost trends since the late 1960s. At HSC, I have published an article each year since 1996 in the journal Health Affairs synthesizing various cost trend data to help policymakers understand the underlying dynamics in the U.S. health care system.1 The 1996 article convinced many people that the decline in cost trends at the time was a real one. Later articles documented the return of rapidly rising costs. In recent years, information collected by HSCs site visit team has been an important complement to the quantitative data to provide insight into the factors behind rising costs. I will draw heavily on that information in this testimony.
Although the most reliable surveys of employers 2002 premium increases have not yet been released, the increase is likely to be in the area of 13 percent, up from 11 percent for 2001. This figure probably understates the size of the increase, because it does not reflect the increases in patient cost sharing employers incorporated into their benefit plans in 2002. These double-digit increases come at a time when corporate profits are down and the average hourly wage is rising by only 4 percent. So you can see the magnitude of the problems this creates for both employers and employees.
To help you understand these cost trends and what is behind them, I will cover three areas in this testimony:
Cost Trends and Premium Trends: The Insurance Underwriting Cycle
Insurance premium trends often diverge from the trends in what insurers actually pay out in benefits, what I refer to as "underlying costs." For example, in 2001, premiums for employment-based coverage increased by 11 percent, while the underlying costs, or spending on care, increased by only 8.7 percent.2 However, five years ago, the reverse was the case, with premium increases trending below the underlying costs.
Economists refer to this as the "health insurance underwriting cycle." The differences between premium and cost trends are the result of two factors. First, the inevitable errors insurers make in projecting future costs at the time that premiums are set, and second, the cyclical expansion and contraction by health insurers in response to the industrys profitability. When the insurance business is unusually profitable, companies move to expand market share by competing on price, recognizing that they may lose money in the short-run but make it up later with a larger market share. When business is unprofitable, companies raise premiums faster than costs even if it means losing some market share.
For the last few years, premium increases have exceeded cost increases as the insurance industry moves to restore profitability. We have seen many examples of insurers exiting unprofitable markets but have not yet seen entry into new markets, which would be a sign of a turn in the underwriting cycle. Inevitably, such a turn will occur, probably within a year or two. Such a change will provide some relief to employers and employees, who will no longer face premium increases exceeding cost trends.
Key Components of the Cost Trend
To better understand the trend in underlying costs, it is useful to examine the various spending components, including inpatient hospital care, outpatient hospital care, physician services and prescription drugs. We can also look at changes in prices paid to health care providers and changes in the quantity of services provided per covered person.
Since 1995, prescription drugs have been the most rapidly growing component of health care costs. By 1999, drugs were the most important contributor to overall cost growth, with the rate of growth in drug spending per person reaching more than 18 percent. Since then, the rate of growth has slowed somewhat, but it is still in the mid-teens. More recently, increases in hospital costs have replaced prescription drugs as the most important driver of overall cost growth because a much higher share of health spending is for hospital care.
Spending trends for inpatient hospital care have made a remarkable turnabout in recent years. Each year from 1994 to 1998, inpatient hospital spending per insured person actually declined. In 1999, however, this ended and inpatient spending trends have increased each year since. For 2001, per capita inpatient hospital spending increased by 5.6 percent, an enormous turnabout from a 5.3 percent spending decrease in 1997. Hospital outpatient spending and physician spending trends followed similar patterns, although spending levels for these services never went into negative territory like inpatient spending. Overall, cost trends declined in the mid-1990s and then began to increase in 1997. When inpatient and outpatient hospital spending are combined, they now account for almost half of the increase in overall health care spending.
The pattern of spending trends slowing through the mid-1990s and then accelerating starting in 1997 is seen in both prices and quantities. Using the hospital component of the Producer Price Index as the measure of price, we have estimated that 38 percent of the hospital spending increase for 2001 is due to higher prices for care and 62 percent is due to higher utilization of services. The service use component has been growing particularly rapidly in recent years. The use of physician services is also increasing, but physician price trends have been level.
Factors Behind the Cost Trends
Important drivers of health care costs at this time include advances in technology, increases in per capita income, the retreat from tightly managed care, provider consolidation, and shortages of nurses and other skilled personnel. Surprisingly, the aging of the population is a relatively small driver of cost trends.
Medical Technology. Advancing medical technology is far and away the most important driver of costs. This includes not only the introduction of new services or equipment, but also new applications of existing services and equipment made possible because of advances in medical science. Prominent academic studies estimate that over the long run, advances in technology explain between one-half and two-thirds of the increase in costs in excess of general inflation.3 Much of this provides important benefits to patients, saving lives and avoiding disability. But together with valuable technologies, we also get technologies that turn out to do patients more harm than good and some where the benefits are very small in relation to the additional cost. Although advances in technology have different impacts on costs at different times, I am unaware of any evidence that technology advances are contributing more or less to costs than in the past.
Income Growth. Income growth is another important driver. Over the long run, as societies incomes grow, they tend to spend a disproportionate amount of the income increase on medical care. Studies have shown that in the short run, changes in per capita income affect trends in health costs, but with a substantial lag.4 Thus, the very rapid growth in incomes during the late 1990s is likely to be contributing to todays high health care cost growth. And these cost increases come at a time when the economy is sluggish.
Less-Restrictive Managed Care. HSCs site visit research has documented a pronounced trend away from tightly managed care.5 Health plans have dropped authorization requirements for hospital admissions, referrals to specialists and the use of expensive diagnostic procedures. Many patients can now see a specialist without first going to a primary care physician. Provider networks are now much broader, giving enrollees a much wider choice of hospitals and physicians. Plans are less likely to contract with providers on a capitated basis, a method that gives providers incentives to economize on service use. External review mechanisms have been made available to enrollees who disagree with plan decisions about medical necessity. Although some of these changes came from mandates, much if it happened in response to demands by employers and consumers for a less restrictive insurance product.
These changes have added to costs. Reduction of authorization requirements has probably led to more utilization of services. Many observers believe that authorization requirements had a sentinel effect, with physicians not requesting authorizations for services that they expected would be turned down. We have seen anecdotal evidence of sharp increases in the number of MRI screens after health plans dropped requirements for authorization.
Broader provider networks are likely to have contributed to a shift in negotiating power away from health plans and toward hospitals. Today, plans feel compelled to have all of the prominent hospitals in a community in their networks. As hospitals figured this out, they were able to demand substantial increases in payment rates from health plans.6
Other Factors. Hospitals consolidated a great deal during the 1990s. As a result, most communities now have fewer hospital systems, with many smaller communities now having only one or two. This, in turn, increased hospitals leverage with health plans. Although we have not seen quantitative estimates of the impact of consolidation, its potential to raise costs is clear.
Shortages of nurses and other skilled personnel also are driving up costs through higher wage rates. According to data from the Bureau of Labor Statistics, average hourly wages in hospitals increased by 5.9 percent in 2001, compared to 3.2 percent in 2000. This is higher than the 4 percent average wage increase in all industries.
Many analysts believe that the aging of the American population is an important driver of health care costs. We have been analyzing this and find that while a driver, aging is a relatively small one. Using data on health care spending per capita by people of different ages and data on the changing age distribution of the population, preliminary estimates suggest that at this time, aging of the working-age population contributes about 0.7 percentage points to the cost trend. Viewed in relation to the 2001 cost increase of almost 9 percent, aging is a relatively small driver. I would note, however, that aging contributes more to the cost trend than it did 10 years ago, when it contributed only 0.1 percentage points. Per capita spending rises sharply around age 55, and the leading edge of the baby boom generation is now reaching this age.
Implications of Rapidly Rising Health Care Costs
When spending rises for most goods and services, policymakers attitude toward it is neutral because of our belief in consumer sovereignty. But health care is quite different because most health care is financed through health insurance because of the uncertainty about when someone will need very expensive care. This reliance on third-party payment blunts consumer incentives to economize on the use of care. And rising premiums cause problems for the employers and governments that pay for health insurance. So, policymakers have good reason to be concerned about rising health care costs.
Unlike a housing purchase, for example, where a consumer can tailor the purchase to what they are willing to spend and can afford, consumers have much less ability to adjust their health spending to their ability to pay. For the most part, we have a single standard of what should be done for people who have various illnesses. This means that if we want people with lower incomes to have access to care, they need to be subsidized, either through pools that employers establish (where the employer makes the same contribution regardless of the workers earnings) or through government programs.
Rising health costs affect peoples ability to afford health insurance. When insurance premiums rise faster than workers wages, fewer people obtain employment-based health insurance. This happens through small employers deciding not to provide coverage to their employees and employees deciding not to take up employer coverage because the employee contribution is too high. If health care cost trends continue to exceed increases in wage rates by a large margin, this could result in substantial loss of employer-based health insurance.
Finally, rising health care costs also pose a problem for the federal and state governments, which finance 45 percent of national health expenditures, mostly through Medicare and Medicaid. With public revenues staying at a relatively constant percentage of national income, growth in outlays for these programs in excess of growth in income that is taxed poses particular strains on public budgets. States are facing these strains now in an acute manner, as Medicaid outlay growth exceeds the trend in state revenues by a large margin. The strain will become acute for the federal government when the baby boom generation begins to become eligible for Medicare.
While I have reviewed many of the drivers contributing to the largest jump in health care costs in a decade, I want to close by touching on a core factor that is behind much of this. In the United States, our culture emphasizes that people should get all beneficial medical care, regardless of cost. This works against attempts to discourage the development of treatments in which the benefits are uncertain or known to be small. Until the public becomes more aware of cost-quality trade-offs, rising health care costs will continue to strain the resources of government purchasers, employers and consumers.
1 Strunk, Bradley C., and Peter J. Cunningham, "Treading Water: Americans Access to Needed Medical Care, 1997-2001," Center for Studying Health System Change Tracking Report No. 1, March 2002.