This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

H E A L T H  T R A C K I N G : 
M A R K E T W A T C H
W E B  E X C L U S I V E   

11 June 2003  
  

Tracking Health Care Costs: 
Trends Stabilize But Remain High In 2002

Hospital spending continues to drive overall health care spending trends,
fueled by rising hospital price inflation.


by Bradley C. Strunk and Paul B. Ginsburg

ABSTRACT:

Health care spending per privately insured person increased 9.6 percent in 2002, a slight reduction from the 10 percent increase in 2001. This is the first time in five years that the spending trend did not accelerate. Nonetheless, health care spending grew nearly four times faster than the U.S. economy grew in 2002. Growth in hospital spending accounted for the largest portion of the overall increase (51 percent) for the second straight year. Moreover, hospital price inflation—which accelerated significantly in 2002—accounted for a larger share of hospital spending growth in 2002 than in 2001. Premium increases accelerated again in 2003, despite 2002’s slight deceleration of the overall spending trend.

During the past few years double-digit premium increases have greeted employers offering and employees taking up health insurance.¹ A recent poll indicated that more Americans are worried about health care costs than about losing their job, paying their rent or mortgage, or being a victim of a terrorist attack.² Past research has demonstrated that a steady, multi-year acceleration in the rate of growth of underlying health care costs is largely to blame for the increase in private health insurance premiums.³ While growth in prescription drug costs was a major component of the acceleration in health care cost growth of the late 1990s, increases in spending on hospital care have accounted for the largest portion of cost growth more recently.⁴

This paper updates past analyses of trends in the health care costs that underlie private health insurance premiums. We focus on underlying health care costs because trends in such costs are the most important determinant of long-term trends in premiums. As such, cost trends also influence employers’ ability to offer health insurance to their workers, employees’ out-of-pocket costs, and their decisions about accepting an offer of insurance from their employer. Research has clearly linked rising health insurance premiums (in excess of wage growth) to rising rates of uninsurance.⁵

Data And Methods

We use a variety of data sources to analyze health care cost trends and the implications of such trends for private health insurance premiums and consumers’ out-of-pocket spending. We chose data sources based on their ability to provide reliable information with a short time lag, as we have done in prior analyses.

Cost trend data. To examine recent cost trends, we used the Milliman USA Health Cost Index (HCI), which measures the health care spending increases that underlie private health insurance premiums.⁶ This index is designed to reflect the claims trends experienced by private insurers for a typical policy.⁷ As such, it only measures spending on health services that tend to be insured: inpatient and outpatient hospital services, physician services, and prescription drugs. Although Milliman USA removes Medicare spending from the HCI, its inability to remove spending by Medicaid and uninsured patients is an important limitation of the HCI’s ability to track private spending trends. Nevertheless, a past comparison of the HCI and the National Health Accounts (NHA), compiled by the Centers for Medicare and Medicaid Services (CMS), indicated that the HCI is a good measure of private health care costs.⁸ We used the HCI instead of the NHA because the HCI is available with a shorter time lag.

One further difference between the HCI and the CMS data is the classification of ambulatory facilities that are not owned by hospitals. The HCI includes such facilities in its hospital outpatient category, while the CMS includes them with physician services. Since spending in these facilities is growing rapidly, we would expect the HCI to have a higher rate of increase for outpatient hospital services and a lower rate of increase for physician services than reflected in the CMS numbers.

To gain insight into the factors driving growth in hospital spending, we broke down the spending trend into its price and quantity components. We used the “all other payers” series of the Bureau of Labor Statistics (BLS) Producer Price Index (PPI) for general medical and surgical hospitals to measure changes in hospital prices for privately insured patients. This hospital PPI reflects negotiated payments rather than billed charges. Because of its specific methodology, the hospital PPI is not usually affected by changes in length-of-stay, but it is vulnerable to being influenced by change in resource intensity per case.⁹

Changes in hospital quantity (utilization and length-of-stay) were calculated indirectly as the residual of the Milliman USA HCI for hospital services (inpatient and outpatient combined) and the hospital PPI. Because we use this methodology, our quantity index is subject to the same limitation as the HCI— that is, it includes changes in prices and utilization for Medicaid and uninsured patients. Data for one of these factors—hospital prices for Medicaid patients—are available, and we have estimated that this factor’s effect on the annual changes in our quantity index ranges from a 0.7-percentage-point downward bias to a 0.1-percentage-point upward bias across our study period.¹⁰ Nevertheless, it is important to note that, despite these small biases, the “story” with respect to hospital utilization and length-of-stay would not change. No data are available to measure, and therefore remove, trends in utilization and length-of-stay for Medicaid and uninsured patients, but their effects are likely to be small as well.

We broke down the trend in spending on physician services similarly. We used the BLS’s “offices of physicians” PPI series to measure changes in the price of physician care. This physician PPI includes the effect of Medicare conversion factor updates; however, we removed this effect (using data from the NHA to calculate the share of physician care spending attributable to Medicare) to make the series more reflective of price changes that affect the privately insured. Changes in the quantity of physician care used were again calculated as the residual of the HCI and the PPI.

Finally, we used data on payroll costs for hospitals to understand changes in their largest operating cost factor. These data, compiled monthly by the BLS through a survey of non-farm establishments and known as the Employment, Hours, and Earnings (EHE) series, are useful for their reliability and short time lag. We report BLS payroll data on a per capita basis because this makes them directly comparable to the HCI and to data on premiums (what is charged to cover an individual or family) and therefore the most relevant measure for policymakers.¹¹

Premiums and cost sharing. We examined findings from two employer surveys, the Towers Perrin 2003 Health Care Cost Survey and the Eighth Annual Washington Business Group on Health (WBGH)/Watson Wyatt Worldwide Employer Survey of Trends in the Health Care Marketplace, to glean insights into health insurance premium trends for 2003.¹² Both surveys relied on a convenience-sample strategy. We regard convenience samples as less problematic for estimating the trend in premiums than for estimating the level of premiums as long as the question is about the increase in (rather than the level of) premiums. The Towers Perrin survey included responses from 358 employers representing approximately 4.6 million workers and retirees. Participating employers were asked to report their 2003 per capita premium costs for insured health plans and premium equivalents for self-insured plans. The WBGH/Watson Wyatt Worldwide survey included responses from 434 employers with 1,000 or more full-time workers, representing more than seven million workers. Participating employers represented all major industrial sectors and U.S. geographic regions.

The main limitation of these surveys for our purposes is that the samples do not include small employers. For this reason, we focus most attention on how the 2003 increase compared with the 2002 increase. Since premium increases for small employers have been larger than those for large employers in recent years, the estimates cited in this report probably slightly underestimate the true increase across employers of all sizes.¹³ Also, neither survey provided an estimate of the role that increased cost sharing and reduced benefits played in the premium trends (“premium buy-down”). To fill this gap, we obtained information from a Wall Street analyst for the managed care industry on buy-down percentages reported to her by the major publicly traded insurers.¹⁴

Health Care Spending Trends

Total health care spending per privately insured person increased 9.6 percent in 2002 (Exhibit 1). Not only was this increase large by historical standards, it was also considerably larger than the rate of growth of the overall U.S. economy (2.7 percent per capita in nominal terms). Nevertheless, 2002 marked an important development in the spending trend: It was the first time in five years that a one-year increase in spending did not exceed the previous year’s increase. More specifically, health care spending in 2002 grew by 0.4 percentage points less than it did in 2001, whereas over the 1996–2001 period the spending trend steadily accelerated by a factor of five, from 2 percent to 10 percent. The slightly slowing cost trend in 2002 could signal that some of the driving forces behind the ever-larger increases in health spending over the past half-decade began to wane in 2002.

The slight deceleration in the overall spending trend was the result of decelerating trends in all four categories of total spending: hospital inpatient and outpatient care, prescription drugs, and physician care. We examine each of these categories in turn below.

Hospital spending. Spending on hospital inpatient services per privately insured person increased 6.8 percent in 2002. Meanwhile, concomitant spending on hospital outpatient care increased 14.6 percent and remained the fastest-growing category of health care spending for the second year in a row. Both increases, as for overall health spending, were very large by historical standards. For example, hospital inpatient spending actually experienced annual declines in the mid 1990s by as much as 5.3 percent, and hospital outpatient spending increased by as little as 7.7 percent per year during the past ten years.

Nevertheless, the 2002 increases were smaller than the 2001 increases. Spending on inpatient care grew by 0.3 percentage points less than it did in 2001. More importantly, the trend for outpatient care decelerated to a greater degree: 1.7 percentage points less than in 2001. Indeed, the slowdown in the outpatient trend accounted for nearly all of the slight slowdown in the overall spending trend.

These downward trends notwithstanding, there was no change in the degree to which growth in spending on hospital care contributed to the overall health spending increase. About 51 percent of this overall increase was attributable to higher spending on hospital inpatient and outpatient care—the same percentage as in 2001 (Exhibit 2). The contribution of hospital spending growth to the overall cost trend remained considerably higher than that of any other spending category.¹⁵

After climbing in 2001, the trend for hospital utilization slowed in 2002 (Exhibit 3). Hospital utilization (as measured by our residual hospital quantity index) increased 5.7 percent—2.3 percentage points less than in 2001. This is likely the result of two factors. One, which we examine later, is a sharp increase in cost sharing for people enrolled in private insurance that began in 2002. Since increased cost sharing leads to lower use of services, it likely slows the rate of growth in use of services. The second factor is the completion of the adjustment to more loosely managed care. Over the past few years the managed care industry has scaled back its use of various tools to tightly manage enrollees’ care.¹⁶ If this retreat was in fact responsible for the surge in usage in 2001 (as we have suggested in the past), the completion of this transition would cause rates of growth in utilization to moderate.

The effect of a decelerating hospital utilization trend on the hospital spending trend was, however, partially offset by a rising acceleration in hospital price inflation. Hospital prices (for both inpatient and outpatient care, as measured by the hospital PPI) increased 5.1 percent in 2002. This increase was 1.5 percentage points larger than the increase in 2001 and the largest one-year increase since the BLS began tracking negotiated prices for hospital care in 1993. As a result, hospital price inflation became a more important contributor to the increase in hospital care spending, whereas in 2001 its role was much less important than that of the increase in utilization.

Evidence from the fourth round of the Center for Studying Health System Change’s Community Tracking Study (CTS) site visits suggests that hospitals continue to use the formidable negotiating leverage they regained over health plans in the late 1990s to demand large payment rate increases.¹⁷ Hospitals’ demands are in part an effort to reverse the effect of discounted rates during the mid-1990s.

Hospitals also might be seeking to pass through more of the wage increases they have granted to address labor shortages, particularly for nurses. The large price increase in 2002 could reflect a delayed passing through of the very sharp increase in wage rates in 2001 (Exhibit 4). Since hospitals negotiate payment contracts with health plans prospectively, unexpected growth in wage rates in one year would affect reimbursement rates negotiated in subsequent years.

Hospital payroll cost data for 2002 indicate that payroll costs continued to increase at a high rate, but that rate came down slightly compared with 2001. Payroll costs grew by 7.9 percent in 2002, as a result of a 5.5 percent increase in the average hourly wage and a 2.2 percent increase in total hours worked. Nevertheless, the gap between the average hourly wage for hospital workers and that of workers in all industries actually expanded, because wage rate growth for hospital workers decelerated at a slower pace (probably due to the continued shortages of nurses and other hospital personnel) than did wage rate growth in the general economy.

Prescription drugs. Spending on prescription drugs per privately insured person rose 13.2 percent in 2002. While 2002 marked the first time in many years that the trends for inpatient and outpatient hospital care decelerated, the trend for drug spending has decelerated for three straight years now. This differs from the five-year period preceding 2000, when the drug spending trend accelerated each year and reached a peak increase of 18.4 percent in 1999. At that time, prescription drugs were by far the fastest-growing category of health spending (as for the entire 1995–2000 period).

Ever since health spending trends began accelerating in 1997, much attention has been focused on the role of drug spending in the upswing. Indeed, in the first few years of that upswing, rapid drug spending growth was an important factor driving up total health care spending. In 1999, for example, growth in drug spending accounted for about the same portion (34 percent) of overall growth in health care spending as did hospital and physician care (Exhibit 2), even though drug spending makes up a much smaller portion of the total health care bill. By 2002, however, drug spending growth accounted for only 22 percent of the spending trend. Although this is still greater than its proportion of health care spending in total, its relative contribution to rising health spending has diminished considerably.

A number of factors could explain the slowing in the drug spending trend. Three-tier copayment structures for prescription drugs continue to gain a foothold in employers’ health benefit offerings, and copayment differences between the tiers are increasing.¹⁸ Also, the pace of technological innovation in the area of prescription drugs continues to slow, as evidenced by the fact that the U.S. Food and Drug Administration (FDA) approved only fifteen new drugs in 2002, compared with an average of thirty-one new drugs a year during the previous five years.¹⁹ Finally, a number of costly drugs have recently gone off patent.

Physician care. Spending on physician care per privately insured person increased 6.5 percent in 2002, which was just slightly less than the 2001 increase. Also, it was the slowest-growing spending category; in fact, it grew by less than half the rate of hospital outpatient care and prescription drug spending. Over the past ten years the trend for physician care spending has been the steadiest of any category; it went from a low of 1.6 percent in 1995 to a high of 6.7 in 2001. The 2002 increase in spending on physician care accounted for 27 percent of the growth in overall health spending.

The 2002 increase in physician care spending was attributed to increases in both price and utilization, although the latter played a much larger role. The price of physician care (as measured by the physician PPI) grew 1.2 percent in 2002—1.1 percentage points smaller than the 2001 increase. Note that this number applies to non-Medicare patients, so it does not reflect the 5.4 percent cut in Medicare rates in 2002. Meanwhile, use of physician care grew by 5.3 percent—1.0 percentage points larger than the 2001 increase. Over the longer term, however, neither the trend in physician prices nor the trend in use of physician care has exhibited a consistent pattern (data not shown).

Health Insurance Premium Trends

Findings from the Towers Perrin and WBGH/Watson Wyatt Worldwide employer surveys suggest that premiums for employment-based insurance continued to trend upward in 2003. The Towers Perrin survey reports that premiums for active employees increased 15 percent, on average, in 2003, compared with 13 percent in 2002 and 12 percent in 2001. Moreover, the 2003 increase was the largest in at least a decade. The WBGH/Watson Wyatt Worldwide survey reports that the median premium increase was 15 percent, compared with 14.7 percent in 2002 (the twenty-fifth-percentile increase was also larger in 2003 than it was in 2002, but the seventy-fifth-percentile increase was not).

Premium increases have been larger than underlying cost increases for a number of years now. This is a characteristic of the “hard” phase of the health insurance underwriting cycle, when insurers raise premiums more rapidly than underlying costs to make up for past financial losses, a practice known as “catch-up pricing.” During this phase, insurers focus on restoring and solidifying profitability rather than gaining market share. Cost increases in 2000 and 2001 might have exceeded what insurers had predicted in those years, which might have delayed the planned recovery of profit margins and stretched out the current phase of the underwriting cycle. But the end of the string of successively higher cost trends in 2002 could bring the next turn in the underwriting cycle closer.

Implications For Consumers

Consumers appear to be facing a second round of sizable increases to cost-sharing requirements in 2003. Recent reports of insurers to Wall Street suggest that employers “bought down” their insurance premiums by an average of roughly 3 percent in 2003. This means that insurance premiums would have increased 3 percent more than they did had employers made no changes to their benefit structures. This increase in cost sharing comes on the heels of a 2–3 percent buy-down in 2002.²⁰ Employers’ ability to pass along sizable increases in cost sharing to their employees is likely a reflection of the continuing softness of the U.S. economy and rising unemployment that has led to a loosening of labor markets.

While employers raised cost-sharing requirements to control rising premiums, they made little change to the proportion of the total premium that employees are required to pay. According to the Towers Perrin employer survey, active employees are paying 19.3 percent of the cost of single coverage and 22.2 percent of the cost of family coverage in 2003, compared with 18.7 percent and 21.7 percent, respectively, in 2002.²¹ Indeed, the effect of increased cost sharing on out-of-pocket spending will be much greater than any increases to employees’ premium share.

The fact that employers targeted cost sharing rather than employees’ share of premium to control costs reflects a clear strategy. Employees can avoid some or all of the increase in cost sharing and reduce their out-of-pocket spending by responding to the incentive and using fewer services. Moreover, some respondents participating in the CTS site visits have perceived that first-dollar coverage in health benefits is problematic because it does not moderate the use of services. In addition, limiting increases in employee contributions to premiums will help avoid further reductions in employees’ take-up of health insurance.

Outlook For The Future

The story regarding cost trends in 2002 was decidedly mixed. On the one hand, a long period of steadily accelerating growth in the health care costs that underlie private health insurance premiums finally came to an end. On the other hand, the cost trend remained nearly as high as it was in 2001, when the cost increase was the largest in more than a decade. Moreover, premium increases for private health insurance continued to accelerate despite the stabilization of the underlying cost trend.

The outlook for the future is equally mixed. A number of forces could lead to further deceleration of the health care cost trend, but some developments on the horizon might also drive the cost trend upward again. Two forces that could contribute to deceleration are the growth in consumer cost sharing and the sluggish U.S. economy. As consumers bear a larger portion of their health care bill, they will likely reduce their demand for health care services. The impact of a given increase in cost sharing could even increase over time as consumers develop more experience in economizing on health care. Also, prior research has demonstrated that growth in the economy influences health care cost trends, but with a substantial lag.²² For these and other reasons, researchers at the CMS recently forecast that the trend in private personal health care spending will slow throughout the rest of this decade.²³

Nevertheless, several forces could exert upward pressure on the cost trend. Many states may choose to address large budget deficits by cutting Medicaid spending. To do so, state policymakers might reduce payment rates to providers for services delivered to Medicaid patients. Affected providers might then attempt to make up for lower Medicaid revenue by seeking offsetting increases from private payers (although economists disagree on the extent to which providers engage in such “cost shifting,” if at all). The federal government also could further constrain provider payment rates for Medicare patients, which could lead to additional shifting of costs onto private payers. Many providers across the country are building new general acute care and specialty care facilities. They may seek to recover some or all of the cost of these facilities through higher payment rates, and the increased provider capacity could induce demand for services to fill the new facilities.

Despite the uncertainty about future cost trends, conditions do seem ripe for the underwriting cycle to turn soon—and slow the premium trend—if the cost trend does not increase in 2003. The health insurance industry is now experiencing strong profitability in general. This will eventually set off a new round of price competition as plans begin to enter new markets and shift their strategic focus from improving profitability to growing market share. However, in our site visits we have not yet observed significant new market entry or aggressive attempts to grow market share, so it is unclear if the turn in the underwriting cycle would come in 2004 or later.

A turn in the underwriting cycle will not, however, bring about a major slowdown in premium increases; this can only be accomplished by a major slowdown in underlying cost trends. Until this happens, employers and employees will continue to face the many negative consequences of high cost growth, and uninsurance will likely continue to rise.

The authors are grateful to John Cookson of Milliman USA for permission to use the Health Cost Index. Comments from an anonymous reviewer contributed important insights. The authors gratefully acknowledge the Robert Wood Johnson Foundation for its financial support.

NOTES

1. J. Gabel et al., “Job-Based Health Benefits in 2002: Some Important Trends,” Health Affairs (Sep/Oct 2002): 143–151.
2. Henry J. Kaiser Family Foundation, “Health Care Worries in Context with Other Worries,” Kaiser Health Poll Report, March/April 2003,
www.kff.org/ healthpollreport/templates/detail.php?page= 9&feature=hsw (14 May 2003).
3. B.C. Strunk, P.B. Ginsburg, and J.R. Gabel, “Tracking Health Care Costs: Growth Accelerates Again in 2001,” 25 September 2002,
www.healthaffairs.org/WebExclusives/Strunk_Web_Excl_092502.htm (18 March 2003).
4. Ibid.
5. R. Kronick and T. Gilmer, “Explaining the Decline in Health Insurance Coverage, 1979–1995,” Health Affairs (Mar/Apr 1999): 30–47.
6. Often the terms “costs” and “spending” are used interchangeably. Conceptually, the primary interest is in costs, which reflect the resources devoted to health care that are not available to produce other goods and services. Practically, most available data, including the HCI, reflect spending, or what is paid for health services by those who purchase them (or received by providers of health services). Costs and spending differ when the payment is greater or less than the resources that go into providing the services.
7. The index that Milliman USA provides to its clients is intended to assist insurers in forecasting their claims payments and comparing them with those of others. It simulates trends in claims for a “standard” private health insurance policy with a $250 deductible. The trend in such an index would slightly overstate the actual trend in spending because the standard policy would pay for a slightly higher proportion of expenditures each year. To avoid this problem, Milliman USA has provided us with a version of the index that reflects a hypothetical policy with no deductible.
8. Strunk et al., “Tracking Health Care Costs.”
9. See B. Catron and B. Murphy, “Hospital Price Inflation: What Does the New PPI Tell Us?” Monthly Labor Review (July 1996): 24–31.
10. See the Producer Price Index for general medical and surgical hospitals, Medicaid patients series, at www.bls.gov/ppi/home.htm (8 May 2003).
11. For additional detail, see Strunk et al., “Tracking Health Care Costs.”
12. Towers Perrin, 2003 Health Care Cost Survey: Report of Key Findings (New York: Towers Perrin, 2003); and Watson Wyatt Worldwide and Washington Business Group on Health, Creating a Sustainable Health Care Program: Eighth Annual Washington Business Group on Health/Watson Wyatt Survey Report, 2003, www.watsonwyatt.com/research/resrender.asp?id=w-640&page=1 (18 March 2003).
13. See Kaiser Family Foundation and Health Research and Educational Trust, Employer Health Benefits: 2002 Annual Survey, September 2002, www.kff.org/content/2002/3251/3251.pdf (14 May 2003), and previous survey reports for estimates of premiums increases for small versus large employers.
14. Roberta W. Goodman, Merrill Lynch, personal communication, 16 March 2003.
15. The share of health care spending growth accounted for by growth in spending on hospital care was somewhat higher than the share of health care spending accounted for by spending on hospital care. Therefore, these differences are not simply the result of the fact that spending on hospital care is proportionally much higher than spending on some other categories, such as prescription drugs.
16. See, for example, D. Draper et al., “The Changing Face of Managed Care,” Health Affairs (Jan/Feb 2002): 11–23; R.H. Hurley et al., “A Longitudinal Perspective on Health Plan–Provider Risk Contracting,” Health Affairs (July/Aug 2002): 144–153; and S. Felt-Lisk and G.P. Mays, “Back to the Drawing Board: New Directions in Health Plans’ Care Management Strategies,” Health Affairs (Sep/ Oct 2002): 210–217.
17. For more detail on the methods of the CTS site visits, see C.S. Lesser, P.B. Ginsburg, and K.J. Devers, “The End of an Era: What Became of the ‘Managed Care Revolution’ in 2001?” Health Services Research 38, no. 1, part II (2003): 337–355. Regarding hospitals’ leverage, see B.C. Strunk, K. Devers, and R.H. Hurley, Health Plan–Provider Showdowns on the Rise, Issue Brief no. 40 (Washington: Center for Studying Health System Change, June 2001).
18. Gabel et al., “Job-Based Health Benefits in 2002.”
19. R. Dobson, “Lack of New Drugs Is Reaching Crisis Point, Says Review,” British Medical Journal (18 January 2003): 119.
20. Strunk et al., “Tracking Health Care Costs.”
21. The Towers Perrin survey includes responses from mostly large employers, which usually pay a larger portion of the cost of family coverage than small employers pay. This explains why the Towers Perrin estimate of employees’ share of family coverage is lower than estimates in other surveys that include all employers.
22. S. Heffler et al., “Health Spending Projections for 2002–2012,” 7 February 2003, www.healthaffairs.org/WebExclusives/Heffler_Web_Excl_020703.htm (18 March 2003).
23. Ibid.

Bradley Strunk is a health research analyst at the Center for Studying Health System Change in Washington, D.C. Paul Ginsburg is the center's president.

©2003 Project HOPE–The People-to-People Health Foundation, Inc.