Health Tracking MarketWatch Health Benefits In 2004: Four Years Of Double- Digit Premium Increases Take Their Toll On Coverage Five million fewer jobs provided health insurance in 2004 than in 2001, this new analysis finds. by Jon Gabel, Gary Claxton, Isadora Gil, Jeremy Pickreign, Heidi Whitmore, Erin Holve, Benjamin Finder, Samantha Hawkins, and Diane Rowland ABSTRACT: This paper reports changes in employer-based insurance during the past year and since 2001. From spring 2003 to spring 2004, premiums increased 11.2 percent (compared with 13.9 percent last year). Since 2001, premiums have increased 59 percent, employee contributions have grown by 57 percent for single coverage and 49 percent for family coverage, and the percentage of workers covered by their own employer s health plan has fallen from 65 percent in 2001 to 61 percent in 2004. The worst of the current round of premium inflation appears to be over, but employers plan to increase employee cost sharing next year. Employer-sponsored health insurance covers 161 million Americans under age sixty-five and nearly 12 million elderly persons. 1 One year ago in this journal we reported that the cost of job-based health benefits had increased at the most rapid rate since 1990 (13.9 percent). 2 Employees out-ofpocket expenses were rising, as their monthly contributions for premiums, deductibles, copayments, and coinsurance increased. This paper reports on changes in job-based health insurance from spring 2003 to spring 2004. To gain a clearer perspective about how the market has been changing, we also highlight cumulative changes between 2001 and 2004. We find that premiums have risen 11.2 percent, still at a double-digit pace but below the record pace of 2003. The percentage of small firms (3 199 workers) offering health benefits has fallen from 68 percent in 2001 to 63 percent in 2004. The percentage of workers covered by their own employer s health plan has fallen from 65 percent to 61 percent over the same period. Jon Gabel (jgabel@aha.org) is vice president, health systems studies, at the Health Research and Educational Trust (HRET) in Washington, D.C. Gary Claxton is a vice president of the Henry J. Kaiser Family Foundation, also in Washington; Isadora Gil is a policy analyst there. Jeremy Pickreign is a statistician at the HRET, where Heidi Whitmore is deputy director, health systems studies. Erin Holve is a senior policy analyst at the Kaiser Family Foundation in Menlo Park, California. Benjamin Finder is a research assistant at the foundation s Washington, D.C., office. Samantha Hawkins is a research assistant at the HRET. Diane Rowland is the foundation s executive vice president. 200 September/October 2004 DOI 10.1377/hlthaff.23.5.200 2004 Project HOPE The People-to-People Health Foundation, Inc.
MarketWatch Study Data And Methods This is the sixth year of the annual Henry J. Kaiser Family Foundation/Health Research and Educational Trust (Kaiser/HRET) Survey of Employer Health Benefits. Core elements of the questionnaire were included in earlier surveys conducted by the Health Insurance Association of America from 1987 to 1991 and KPMG Peat Marwick from 1991 to 1998. The survey collects information on a firm s largest indemnity, health maintenance organization (HMO), preferred provider organization (PPO), and point-of-service (POS) plan as well as overall statistics on employer offerings and employee participation. Using computer-assisted telephone interviews, National Research LLC conducted telephone interviews with employee benefit managers from January to May 2004. The 2004 survey consists of 1,925 randomly selected firms that completed the entire survey, of which 1,378 also participated in the survey in 2003. Prior surveys indicate that firms not offering benefits are less inclined to participate. Therefore, we asked those firms that declined to participate in the full survey one question, Does your company offer or contribute to a health insurance program as a benefit to your employees? A total of 1,092 additional firms answered this question. Estimates of the percentage of firms offering health insurance to their workers use the sample of firms that responded to the full survey as well as the firms that did not complete the full survey but answered the one offer-rate question (n = 3,017). Kaiser/HRET selects its sample from a listing of U.S. firms compiled by Dun and Bradstreet. Employers range in size from three to hundreds of thousands of workers and include private and public firms. The sample is stratified by firm size and industry. The overall response rate for the survey was 50 percent. Because firms are randomly selected, it is possible to use statistical weights to extrapolate from the sample to national, regional, firm-size, and industry averages. Data are identified in the text and exhibits as representing the percentage of either firms, workers, or insured workers. Weights can be used to extrapolate to firm- or employee-based figures. In calculating weights, we first determined the basic weight and then applied a nonresponse adjustment, and then a poststratification adjustment. We used the Statistics of U.S. Businesses compiled by the U.S. Census Bureau as the basis for the poststratification adjustment. Most of the findings, and in particular those that involve dollar amounts, are presented as a percentage of workers with insurance who are affected. Some findings, in particular those involving firms behavior (such as the decision to offer coverage) or attitudes, are presented as a percentage of firms. All findings discussed in the text are significant at the.05 level unless otherwise noted. Departing from our usual practice, in this paper we take a closer look at changes in jobbased insurance over a three-year period (2001 2004). Small changes in key measures of performance such as cost sharing, coverage, and offer rates are often statistically insignificant from year to year. Yet when these measures move in the same direction each year, the cumulative changes may be not only statistically significant but also substantial. Findings Health insurance premium increases. Premiums increased 11.2 percent from spring 2003 to 2004, making 2004 the fourth consecutive year of double-digit premium increases. 3 Premium increases outpaced the economywide rate of inflation and increase in workers hourly earnings by almost nine percentage points. 4 During the past four years, the cost of health insurance has increased 59 percent. The year 2004 represents the first year since 1996 that premium increases slowed compared with the prior year. Based on the historical pattern of premium increases displayed in Exhibit 1, this suggests that we may have seen the worst of the current round of health care inflation. Premium increases vary considerably across firms. One-fourth of employees work for firms where premium increases were less than 5 percent, while 28 percent work for HEALTH AFFAIRS ~ Volume 23, Number 5 201
Health Tracking EXHIBIT 1 Increases In Employer Health Insurance Premiums Compared With Other Indicators, Selected Years 1988 2004 firms where increases exceeded 15 percent. Premium increases outpaced the estimated growth in medical claims expenses. Data from Milliman USA reported in this journal show that estimated claims expenses rose 7.4 percent in 2003. 5 Hence, underwriting profits of insurers, defined as profits before investment income, grew substantially over the past year. 6 Cost of coverage. The average monthly cost of single coverage rose to $308 ($3,695 annually) for single coverage and to $829 ($9,950 annually) for family coverage in 2004 (Exhibit 2). The total premiums paid by small firms (3 199 workers) for single coverage and for family coverage are statistically equivalent to premiums paid by larger firms. The average contribution that small firms make toward single coverage also is statistically equivalent to the average contribution made by large firms. Large firms, however, make a significantly greater contribution toward family coverage than do smaller firms; cost sharing also is greater in small firms. HMO premiums remain lower in 2004 than the premiums for PPO plans, the most common type of coverage (Exhibit 2). This comparison is unadjusted for differences in the health status of the covered population, covered benefits, cost sharing, and geographic location. Employee cost sharing. Contributions for family coverage rose from $201 in 2003 to $222 in 2004, an increase of 10 percent (Exhibit 3). The percentage of insured workers in plans where the employer pays 100 percent of the premium continued to fall in 2004: Between 2001 and 2004, the percentage fell from 34 percent to 21 percent for single coverage and from 14 percent to 7 percent for family coverage. Employees in 2004 on average pay 16 percent of the cost of single coverage and 28 percent of the cost of family coverage (data not shown). Contrary to expectations, out-of-pocket costs for deductibles and copayments did not increase appreciably for most employees in 2004. The single-coverage deductible for innetwork PPO services is $287 (statistically 202 September/October 2004
MarketWatch EXHIBIT 2 Monthly Premiums And Employee Contribution Levels For Single And Family Coverage, 2004 Premium ($) Employee contribution ($) Category Single Family Single Family All plans 308 829 47 222 Plan type Conventional HMO PPO POS 318 288 317 302 800 792 851 818 39 46 48 45 198 223 224 218 Region Northeast Midwest South West 316 314 302 302 871** 857** 802** 802 53 48 44 43 196** 193** 258** 219 Firm size (workers) 3 24 25 199 200 999 1,000 4,999 5,000 or more 323 303 313 311 303 820 806 827 863** 832 47 40 44 51 49 264** 293 227 194** 184** SOURCE: Henry J. Kaiser Family Foundation/Health Research and Educational Trust. NOTES: Statistical significance denotes difference from all plans. HMO is health maintenance organization. PPO is preferred provider organization. POS is point-of-service plan. **p <.05 unchanged from $275 in 2003), and the deductible for out-of-network services is $558 (essentially unchanged from 2003). However, since 2001, PPO deductibles when using innetwork providers have risen about 40 percent (and about 36 percent for out-of-network services). Copayments for office visits continue to drift higher across all types of plans: The percentage of covered workers in a plan with a $20 copayment rose to 27 percent in 2004, up from 19 percent last year. Consistent with previous years, cost sharing for employees remains higher in small firms (3 199 workers) than in large firms (200 or more workers). For PPOs, the average deductible for preferred providers for small firms is $420, compared with $232 for large firms. In POS plans, the comparable average deductibles are $427 for small firms and $75 for large firms. In HMO plans, the comparable average deductibles are $119 for small firms and $17 for large firms. Workers with PPO coverage also pay somewhat higher copayments: The mean PPO copayment for services from preferred providers in small firms is $19, compared with $17 in large firms. The percentage of covered workers who face separate hospital cost-sharing arrangements rose from 44 percent in 2003 to 53 percent in 2004. In the 1990s separate hospital cost-sharing deductibles or copayments were rare. 7 Nearly 30 percent of employees with job-based coverage face either a separate annual deductible or copayment for hospital services, with an average required payment of $222. Thirteen percent of covered workers must pay coinsurance when hospitalized, with an average coinsurance rate of 16 percent. Another 5 percent face both types of cost sharing. Prescription drug expenses. Virtually all employees with job-based insurance continue to receive coverage for prescription HEALTH AFFAIRS ~ Volume 23, Number 5 203
Health Tracking EXHIBIT 3 Average Monthly Contribution, Percentage Of Premiums Paid By Covered Workers For Single And Family Coverage, And Average Deductible By Plan Type, Selected Years 1988 2004 1988 1993 1996 2000 2001 2002 2003 2004 Monthly worker contribution Single Family Percent of premiums paid by worker Single Family Deductibles Conventional, individual Conventional, family HMO, individual HMO, family PPO, in network PPO, out of network POS, in network POS, out of network $ 8 52 $34 124 $37 122 $ 28** 135 $30 149 $ 39** 178** $42 201** $47 222** 11% 29 20% 32 20% 27 14%** 26 14% 26 16% 28 16% 27 16% 28 $163 375 106 177 b b $222 495 170 289 b b $264 594 180 313 71 324 $248 580 175 340 70 352 $239 598 204** 409** 92 407 $295 700 251** 466 54** 409 $384 785 30 65 275 561** 113** 442 $414 861 44 80 287 558 210 575 SOURCES: Henry J. Kaiser Family Foundation/Health Research and Educational Trust Survey of Employer-Sponsored Health Benefits, 2000, 2001, 2002, 2003, and 2004; KPMG Survey of Employer-Sponsored Health Benefits, 1993 and 1996; and Health Insurance Association of America (HIAA), 1988. NOTES: Plans with lower enrollment, such as point-of-service and conventional plans, have large variation in their estimates, which means that large differences may not be significantly different. Statistical significance denotes difference from previous year shown. HMO is health maintenance organization. PPO is preferred provider organization. a Data were not collected for HMO plans from 1998 to 2002. b Data were not collected for POS plans in 1988 and 1993. **p <.05 drugs. To control expenses, employers have redesigned their drug benefits, with a sizable majority making use of multi-tier cost-sharing formulas and mail-order discount plans. Sixty-nine percent of covered workers are in plans that use three- or four-tier cost sharing, wherein the patient s copayment or coinsurance is dependent on the type of drug prescribed and the availability of cheaper generic or brand-name options. In 2000 only 27 percent of covered workers were enrolled in plans that used a three-tier cost-sharing formula (we did not ask about four-tier arrangements before 2004). The average copayment for a nonpreferred drug (for which a generic or cheaper brand-name alternative is available) is $33 in 2004, up from $17 in 2000 and $29 in 2003. The average copayment for a generic drug is $10 in 2004, compared with $9 in 2003 and $7 in 2000. Copayments for generic and preferred drugs are higher in small firms than in large firms. Four-fifths of covered workers are enrolled in a plan that has a mail-order discount arrangement. These arrangements typically reduce employees cost sharing if they obtain their drugs through the plan s mail service. Perhaps because of the backlash against managed care, employers appear to be more comfortable using financial incentives rather than strict rules to get their employees to use less costly drugs. While a substantial majority of employers use cost-sharing incentives to encourage consumption of lower-cost drugs, only 19 percent of covered workers are in a plan that requires them to use a generic drug when it is available. Changes in covered benefits. Despite the recent round of health care inflation, employers remain reluctant to eliminate specific covered benefits. When asked if covered benefits increased, decreased, or remained the same, benefit managers representing 79 percent of covered workers indicated that benefits remained the same (other than changes in 204 September/October 2004
MarketWatch cost sharing), a figure almost identical to that of 2003. Fifteen percent of insured workers work for firms that reduced benefits, and 6 percent of insured workers work for firms that increased benefits. Coverage for selected preventive benefits remains high and is rising. More than 95 percent of workers have coverage for adult physicals, well-baby care, annual obstetric/ gynecological exams, and prenatal exams. More than 99 percent of insured workers have prescription drug coverage, and 98 percent have coverage for inpatient and outpatient mental health benefits. Coverage for oral contraceptives has increased from 67 percent in 2001 to 89 percent today. Coverage for alternative medicine also has grown since 2002. Acupuncture coverage has risen from 33 percent to 47 percent of covered workers, and chiropractic coverage has risen from 79 percent to 87 percent. 8 Workers in small firms are less likely than workers in large firms to have coverage for acupuncture (41 percent versus 50 percent) and chiropractic care (79 percent versus 91 percent). Plan enrollment and employer shopping. Enrollment in conventional, HMO, PPO, and POS plans changed little from 2003 to 2004 (Exhibit 4). The majority of workers with job-based coverage belong to a PPO plan (55 percent); HMOs constitute 25 percent of enrollment, followed by POS plans with 15 percent. In 1996, the high-water mark for tightly managed care, HMOs claimed the largest market share at 31 percent, followed by PPOs at 28 percent. The fact that there was little change in enrollment across plan types does not mean that employers were complacent. More than half (56 percent) of firms offering coverage shopped for a new health plan in 2003, and of the firms that shopped, 31 percent changed insurance carriers and 34 percent changed types of health plans offered. Of shopping firms, 29 percent reported that they have been using their existing carrier for less than one year, and EXHIBIT 4 Health Plan Enrollment Among Covered Workers, By Type Of Plan, Selected Years 1988 2004 HEALTH AFFAIRS ~ Volume 23, Number 5 205
Health Tracking only 13 percent reported using the same carrier that they had six years ago. 9 Interest in new plan arrangements. There has been considerable interest in consumer-driven health plans over the past few years. A common approach pairs a plan with a relatively high deductible (such as $1,000) with a savings account option. About 10 percent of firms reported offering a high-deductible plan to their employees in 2004 (up from 5 percent in 2003), including 20 percent of the largest firms (5,000 or more employees). Thirteen percent of workers in all firms (and 19 percent of workers in large firms) that offer health benefits work for employers that offer a high-deductible plan to at least some workers. Despite the growth in the number of firms offering such plans, however, only 3.5 percent of firms offering a high-deductible plan also offer a health reimbursement arrangement (HRA) or other savings account in conjunction with the high-deductible plan. When asked about the likelihood of offering a high-deductible plan with a savings arrangement within the next two years, firms employing 13 percent of covered workers indicated that the firm is very likely to offer an HRA-type plan in the next two years; firms employing 26 percent of covered workers reported that they are somewhat likely to do so. This level of interest, particularly among the largest firms, suggests that these arrangements could grow considerably in the near future. Availability of coverage. The offer rate among small employers continues to drift downward, with just 63 percent of small firms offering coverage in 2004, compared with 65 percent in 2003 and 68 percent in 2000 and 2001 (Exhibit 5). 10 The offer rate among large employers remains high at 99 percent (data not shown). Since 2001 the percentage of workers (including those in firms not offering health benefits) covered by their employers health plan has declined from 65 percent to 61 percent (Exhibit 6). This decline occurred largely among small firms. Assuming a constant number of workers across the period, we estimate that the number of jobs with health insurance fell by five million or more between 2001 and 2004. 11 This change does not translate directly into a loss of workers with health insurance because approximately one-third of covered workers have access to a second policy through another member of their family and EXHIBIT 5 Percentage Of Small Firms Offering Health Benefits, By Firm Size, Selected Years 1996 2004 206 September/October 2004
MarketWatch EXHIBIT 6 Percentage Of Workers Covered By Their Employer s Health Benefits, Among Firms Offering And Not Offering Health Benefits, By Firm Size, 2000 2004 some workers may pick up coverage in the individual health insurance market or through a public program. 12 Similar to previous years, in 2004 small firms not offering coverage cite the cost of insurance (79 percent) and the belief that employees have alternative sources of available coverage (36 percent) as very important reasons for not doing so. Other factors affecting the decision by small firms not to offer coverage include the ability to attract and retain good workers without offering coverage (31 percent), high turnover (13 percent), and administrative hassle (9 percent). When employers do offer coverage, often not all workers are eligible to take advantage of the benefit. For example, among firms offering health coverage, 48 percent of workers are employed by firms that extend eligibility to part-time workers, and only 6 percent are in firms that make it available to temporary employees. Overall, 80 percent of workers in firms offering health benefits are eligible for coverage, and 82 percent of eligible workers take up coverage from their firms. This means that one-third of workers in firms offering insurance are not enrolled in the employer s plan, because they either are not eligible or do not take it up (likely because of the cost or because they have other coverage available). Employers attitudes and views about the future. Employers remain somewhat skeptical about the effectiveness of different forms of cost containment. Only a few employers rate any approach as likely to be very effective (15 percent for disease management, 11 percent for consumer-driven health plans, and 9 percent for tightly managed networks and higher employee cost sharing). A fairly large percentage (32 47 percent) continue to believe that these approaches are somewhat effective in controlling costs. Firms with more than 5,000 workers, in contrast, are enthusiastic about the ability of disease management programs to control costs, with 32 percent considering them very effective and 50 percent rating them as somewhat effective. Large employers remain more likely than small employers to say that they plan on increasing employee costs next year. More than 80 percent of large employers say that they are very or somewhat likely to increase employee contributions, compared with 44 percent of small employers. Large employers also are more likely to say that they are very or somewhat likely to raise drug cost sharing: 55 percent for large firms, compared with 38 percent for small firms. Some analysts have questioned whether a decline in the take-up of family coverage may be a factor in the growth of the number of people without insurance. In 2004 we asked employers a series of questions about their views and practices related to family coverage. We HEALTH AFFAIRS ~ Volume 23, Number 5 207
Health Tracking first asked whether the percentage of employees electing family coverage had risen or fallen in the past several years. Majorities of both small and large firms reported that the percentage has remained constant. We also asked firms offering coverage about policies that may be seen as discouraging employees either from electing coverage at all or from electing family coverage. Twelve percent of employers reported that they vary the required contribution for family coverage depending on whether the spouse has access to coverage from another source, and 3 percent reported that they provide additional compensation or benefits to employees who elect single rather than family coverage. Seventeen percent of employers provide additional compensation or benefits to employees who forgo health benefits altogether. We did not ask about the type or size of the incentives offered. Although it appears that relatively few employers are engaged in explicit practices that could discourage the take-up of family coverage, more than 40 percent of all firms, including more than 72 percent of large firms, reported that they are very or somewhat likely to increase the percentage of the premium that employees must contribute for family coverage. Among large firms, 41 percent reported that they are very likely to increase the premium share. Employers were also asked to choose which of two statements best reflects their attitude about providing health benefits: (1) It is important that the firm pay a significant portion of the cost of health benefits for our workers and their families ; or (2) It is important that the firm pay a significant portion of the cost of health benefits for our workers, but the primary responsibility for funding the costs of family members lies with the worker. The responses differ significantly by size of the employer, with large employers largely selecting More than 40 percent of all firms are very or somewhat likely to increase the percentage of the premium that employees must contribute for family coverage. the first expression (73 percent versus 26 percent) while small employers are more evenly split (43 percent versus 53 percent). These employer views are likely to influence the success of policy approaches for expanding health insurance coverage through small employers. Outlook For The Future During the past few years, in the face of poor economic conditions and sharply rising insurance costs, employersponsored health insurance has demonstrated surprising stability. Since 2001, premiums have risen 59 percent, employee contributions for single coverage have risen 57 percent, and contributions for family coverage have risen 49 percent. Yet the overall percentage of firms offering health insurance and the percentage of workers covered in firms both offering and not offering coverage have exhibited only moderate annual declines. Slight annual changes nevertheless have a significant and substantial impact on workers over time. The offer rate among small firms has dropped from 68 percent to 63 percent since 2001. 13 Further, the number of jobs with health insurance has fallen by five million or more since 2001. Four-fifths of this decline is among employees working in small firms. One possible explanation for coverage not dropping more dramatically is that employers are paying for higher premiums by reducing their workers earnings. Workers productivity, measured in terms of output per hour in the business sector, grew by more than 13.3 percent between spring 2001 and spring 2004, while increases in workers earnings actually slowed by one-half during the same time period. Some slowing of earnings is undoubtedly attributable to a weak labor market, but the high cost of health insurance also may be contributing to slower earnings growth. Other trends in job-based insurance are also apparent. More employers are familiar 208 September/October 2004
MarketWatch with and offer consumer-driven health plans in 2004 than in 2003. A growing number of workers are employed in firms with the option to enroll in such a plan. Increased interest in high-deductible plans suggests that this trend will continue. The forecast for employer-sponsored health insurance is mixed. The annual rate of increase in underlying health care expenses has declined from 10 percent in 2001 to 7.4 percent in 2003. 14 Hence, the worst of the current round of inflation may be over. On the other hand, monthly contributions for health insurance, deductibles, and copayments have risen in absolute dollars over the past four years. Most importantly, after four years of double-digit premium increases, the cost of health coverage has risen to levels that have further eroded the already tattered employer-based system. NOTES 1. P. Fronstin, Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 2003 Current Population Survey, Issue Brief no. 264 (Washington: Employee Benefit Research Institute, December 2003). 2. J. Gabel et al., Health Benefits in 2003: Premiums Reach Thirteen-Year High as Employers Adopt New Forms of Cost Sharing, Health Affairs 22, no. 5 (2003): 117 125. 3. Premium-increase figures are based on respondents answers to two questions: How do the total costs for family coverage compare with what they were one year ago? and What percentage did costs for family coverage increase (decrease) since last year? There are no adjustments in the premium-increase figures for benefit adjustments. 4. The Consumer Price Index (CPI) rose 2.3 percent, and workers hourly wages increased 2.2 percent for this same period of time, according to the U.S. Bureau of Labor Statistics. 5. B. Strunk and P. Ginsburg, Tracking Health Care Costs: Trends Turn Downward in 2003, Health Affairs, 9 June 2004, content.healthaffairs.org/cgi/content/abstract/hlthaff.w4.354 (1 July 2004). 6. J.M. Grossman and P.B. Ginsburg, As the Health Insurance Underwriting Cycle Turns: What Next? Health Affairs (forthcoming). 7. Surveys from the 1990s did not ask about the presence of such cost-sharing arrangements because these arrangements were so uncommon. 8. The Henry J. Kaiser Foundation and the Health Research and Educational Trust annual survey rotates the questions on covered benefits each year. Consequently, it is not possible to compare 2004 figures with a standard base year. 9. In fairness, consolidation and exit of many insurers and managed care organizations may be a factor in determining this 13 percent figure. 10. The change in offer rate by all small firms is not statistically significantly different between 2003 and 2004, but it is significantly different from 2000 to 2004 and from 2001 to 2004 (p <.10). 11. We calculated this loss-of-jobs figure as the difference in the number of workers multiplied by the coverage rate for 2001 and 2004. The coverage rate includes employees in firms that do and do not offer coverage. Our estimate does not account for the decline in employment between 2001 and 2004, although data from the Bureau of Labor Statistics show that employment declined from 132.2 million in April 2001 to 131.0 million in April 2004. The manufacturing sector, a sector with a high rate of coverage, lost 2.9 million jobs during this period. Accounting for declines in employment would raise our estimate of the number of jobs with health insurance lost. Our estimate also assumes that the mix of employment in industries did not change (no firmsize/industry-mix data are available for 2004). People in firms with fewer than three employees and the self-employed are also excluded from the estimate. 12. Estimates of dual coverage are from the Medical Expenditure Panel Survey (MEPS) and were provided by Joel Cohen of the Agency for Healthcare Research and Quality (AHRQ). 13. The change is significant only at the.10 level. 14. Strunk and Ginsburg, Tracking Health Care Costs: Trends Turn Downward in 2003. HEALTH AFFAIRS ~ Volume 23, Number 5 209