Google+
Quote
NAVIGATION
PNHP RESOURCES

On The ‘Efficiency’ Of Managed Care Plans

Several decades’ experience has yielded scarce evidence supporting the efficiency of managed care, despite numerous claims that managed care has lowered health care costs.

By Kip Sullivan
Health Affairs, August, 2000

From the earliest days of the health maintenance organization (HMO) movement, many have asserted that HMOs or managed care plans save money and are more “efficient” than fee-for-service (FFS) plans. As a result, that notion has reached the status of folklore within a substantial portion of the health policy community.

This folklore is also widely accepted among politicians. President Bill Clinton announced during his 1998 State of the Union address that “managed care plans…save money.”1 Republican members of the National Bipartisan Commission on the Future of Medicare and other proponents of the premium-support approach to “modernizing” Medicare assert that managed care plans are more efficient than the FFS Medicare program.2 In this paper I argue that the evidence in support of the claim that managed care saves money is inconclusive. I begin with a description of two types of errors frequently made in descriptions of the empirical evidence. I then review what little evidence there is on the efficiency of managed care plans compared with FFS plans. I close by presenting a hypothesis that explains why health care inflation fell during the mid-1990s that does not rely on the as-yet undocumented assumption that managed care plans of any sort are more efficient than FFS plans.

Throughout the paper I use the word “efficient” in the narrow and misleading sense in which it is often used in the health policy literature— that is, to mean “lower priced” or “using fewer resources” as opposed to “providing higher value.” In other words, I ignore the question of whether managed care plans, if in fact they are priced lower, achieve that status by offering lower quality.

Two Common Errors

The claim that managed care plans are more efficien t than FFS plans typically rests on one or both of the following errors.

Error number one.
The first error is this: Lower medical care costs in managed care plans necessarily means that managed care plans’ total costs are lower. Since at least 1970, when the campaign to pass the HMO Act of 1973 began, managed care advocates have cited evidence that managed care plans reduce plan-level medical costs, either by reducing use of medical services, especially hospital services, or by extracting discounts from providers, especially hospitals. The issue here is not whether HMOs reduce utilizatio n rates below FFS levels, nor whether HMOs and preferred provider organizations (PPOs) extract discounts from providers and drug companies that FFS plans cannot get. It is well established that HMOs reduce utilizatio n rates and that managed care plans can compel their sup pliers to offer large disco unts. The issue is, rather, whether medical costs equal total costs. They clearly do not. Total costs are the sum of medical costs plus administrative costs plus profit or surplus. As I show below, evidence indicates that managed care has driven up administrative costs. Throughout the remainder of this paper I use the phrase “expen ditures on medical services” to refer to expenditures that do not include plan-level administrative costs or plan profits, and the phrase “total expenditures” to refer to the sum of expenditures on medical services plus expenditures on plan-level administrative function s plus plan profits. Total and medical expenditures can be measured either at the plan level or at a systemwide level (at the level of the market, state , region, or nation).

Examples of this type of error are legion. Paul Ellwood and colleagues, for example, used this argument in their seminal 1971 article calling on the N ixon ad ministration to e ndorse HMOs.3 Th e y claimed that HMOs were more efficien t than FFS providers, based on studies (which they did not cite) showing that HMOs provide fewer services. Nearly three decades later Sheila Smith and colleagues committed the same error.4 They asserted that the “sustaine d plateau in health spending…since 1993 reflects changes in financin g and delivery” and a shift away from the “fee-for-service system [to] managed care.” Unlike Ellwood and colleagues, Sm ith and colleagues offered a reference (and only one) for their claim: a 1994 literature review by Robert Miller and Harold Luft. But, as I indicate in my discussion below, this review found only that HMOs (not “managed care”) reduce utilizatio n rates (not total expenditures).

Error number two. The secon d common error is this: Lower HMO premiums necessarily means that HMOs are more efficient. The claim that “managed care plans are more efficient” has often been “docum ented” with eviden ce that HMO premiums are often lowe r than FFS premiums. The problem with this claim is, of course, that factor s that have nothing to do with efficiency, such as HMO cherry-pick ing (selectin g lower-risk persons as members) and cost shifting, may explain why HMO premiums tend to be slightly lower than FFS premiums. No studies of total spending controlling for these variable s have been done, at least not in the past fifteen years.

An example of this type of error is Alain Enthoven ’s flattering description of HMOs in his 1978 Shattuck Lecture.5 Contrary to Enthoven’s claim, the literature review he cited did not report that HMO premiums were lower because HMOs were more efficient.

Review Of The Spending Literature As Of 1994

In their 1994 review of studie s comparing the performance of managed care plans with FFS plans, Miller and Luft found two dozen reliable studies indicatin g that HMOs (not PPOs, point-of-service plans, “managed care,” or “changes in financin g and delivery”) reduce use of medical services below levels provided by FFS plans.6 These were almost entirely services offered by or in hospitals. But Miller and Luft could find no reliable studies showing that managed care plans reduce total spending. They could find “no peer-review ed literature [that] determined differences among plans in premium levels,” adjuste d for differences in enrollee health and other factors that could have affected premium levels. They could find no “peerreviewed studies [that] estim ated the impact of managed care plans on national or regional area health care exp enditures.” They described a few studies that showed that California HMOs and PPOs reduced hospital expenditures for all patients (managed care and FFS patients), but, it should go without saying, hospital expenditures do not equal total medical spending, much less total plan-level sp ending.

It is worth noting that Miller and Luft could find very little reliable evidence on PPOs (now, according to some surveys, the dominant form of managed care).7 It also is worth noting that the sole PPO observation they did report “showed significantl y higher total [medical] expenditures for PPO enrollees compared with indemnity plan enrollees.”8

Miller and Luft did offer this speculation: If HMOs provide fewer expensive services and “more comprehensive coverage” than FFS plans, “the findings suggest that HMOs provide care at lower cost than do indemnity plans.”9 Even if this speculation about HMOs is accurate , it does not warrant the conclusion that the entire managed care industry deserves the same praise. But this hypothesis has two serious defects. First, it ignores the myriad methods HMOs have invented to shift costs from themselves to other payers, methods that are less availabl e or completely unavailabl e to traditional FFS plans. Second, this hypothesis assumes that increased adm inistrative costs, for both HMOs and the providers HMOs deal with, have not offset savings achieved in the medical expenditures portion of HMO budgets. I review each of these defects below.

Cost shifting. The evidence bearing on the cost-shifting methods available to HMOs and, to a lesser extent, PPOs varies in quality, and I do not attempt to evaluate that evidence here. I merely observe that these cost-shiftin g methods are numerous and lucrative. They include using financial incentives that encourage HMO doctors to shift costs to other payers, including workers’ compensation programs and unpaid family caregivers; indirectly inducing drug companies and hospitals to shift costs to weaker payers, including FFS plans, by extracting large discounts from them; avoidin g their share of research expenses, charity care, and graduate medical education (GME) expenses even while being subsidize d by Medicare for GME costs; provoking enrollees to pay for care outside of HMO networks; enrollin g populations that are healthier than average (excessi ve payments to Medicare HMOs because of favorable selection may equal 15–20 percent of total Medicare payments to HMOs); billing Medicare for billions of dollars (equal to approximately 5–10 percent of total Medicare payments to HMOs) in administrative costs that should have been billed to private payers; and failing to reimburse Veterans Affairs (VA) hosp itals for services rendered by these hospitals to Medicare HMO enrollees.10 Until these externalize d costs are measured and placed back in the expenditures column of the HMO ledger, it is unwise to assume that HMOs are more efficient than FFS plans are.

Rising administrative costs. Managed care plans’ ability to cut their own medical costs and the absence of conclusive evide nce that they reduce total plan costs seem paradoxical. However, a logical explanation is at hand: The cost-control tactics known collectively as “managed care” may require additional employees at both the plan and provider levels. Because the literature on administrative costs of managed care plans and providers is so sparse, it is fair to say that the issue has never been on the health policy research agenda. What eviden ce we have indicates that adm inistrative costs have soared as managed care methods have spread.
The State o f Minn esota re ported that Minnesota HMOs increased their administrative spending per member by 403 percent between 1980 and 1991, while their spending per member on medical care rose only 255 percent. This rapid increase in administrative costs occurred at a time when HMO enrollment more than doubled, from 415,105 to 1,193,800.1 1 In most industries a doubling of production leads to a fall, not a large rise, in per unit administrative costs.

A similar pattern prevailed at the national level. During the period 1987–1993, a period in which managed care enrollment soared , private health insurance adm inistrative expenditures per person covered rose by 236 percent, while FFS premiums rose 70 percent and HMO premiums rose 71 percent.12

Other evidence indicates that the arrival of managed care has driven up providers’ administrative costs as well.13 David Himmelstein and colleagues demonstrated that employment of administrators, in both the insurer and the provider sectors, rose from 18.1 percent of total employment in the U.S. health care system in 1968 to 27.1 percent in 1993. Administrative personnel grew by 288 percent during this period, while the number of physicians, the great engine of medical inflation according to managed care advocates, grew by just 77 percent.14

It is not clear to what extent the additional expenditures on administrators generated by managed care methods offset the savings HMOs and PPOs have extracted from providers and patients, but it is obvious that this has occurred to some degree. United-Health Group’s late 1999 announcement that it will abandon prospective utilizatio n review because this form of “cost control” was costing three times what it was saving and other reports of high utilization review costs suggest that this form of managed care generates costs that exceed savings.15

A Summary Of The Literature After 1994

No new, reliable studies have been published since 1994 that warrant altering Miller and Luft’s 1994 conclusion that existing “literature on managed care plan performance does not provide policy-makers with adequate bottom-line estimates of expe n di ture differences per enrollee compared with indemnity plans.” A few more studies suggesting that managed care’s expansion in a given market can lead to marketwide reduction s in expenditures on the medical care component of total spending have appeared.16 However, as I have noted already, it is not appropriate to conclude from such stud ies that managed care plans reduce total spending.

Two studies on HMOs’ impact on total spending, and one study on the impact of HMOs and PPOs, have appeared since 1994. One study found virtually no effect, while the other two reported modest HMO (not PPO) effects.17 None of these three studies attempted to measure the impact of HMO cost shifting.

An Alternative Explanation For The Recent Inflation Lull

Prior to 1995 health policy researchers refrained from claiming that managed care was affecting national health care spending . Enthoven, for example, published a paper in 1993 entitled, “Why Managed Care Has Failed to Contain Health Costs.”18 However, all of that changed around 1995, when it became clear that the lull in premium and national exp enditure inflation was real (not nominal). By 1997 Enthoven and his colleague Sara Singer were asserting, “Since the early 1990s cost pressures have moderated significantly , and there is no explanation except competitive markets and managed care.”19 It is my impression that the number of unqualified and poorly documented assertion s that managed care cut costs rose exponentially in the peer-reviewed and lay literature after 1995.

Is it possible that the rapid spread of managed care methods had little to do with the lull in health care inflation? I offer an alternative hypothesis: that the mid-1990s lull was caused primarily by the short-term reactions of the industry to the near-simultaneou s occurrence of four events: (1) a downturn in the three-years-up, threeyears-down health insurance pricing cycle; (2) the delayed effect of the 1990–1991 recession; (3) the endorseme nt of managed compe tition models of health reform by the White House and nume rous state and federal politicians; and (4) the merger fever triggered by these political endorsements. The latter three phenomena deepened and lengthened what otherwise would have been a shallowe r and shorter downturn in the usual insurance-pricing cycle.

Health care inflation was torrid during the late 1980s primarily because the nation’s insurers were making up for losses suffered during 1986–1988.20 Under the usual cycle, premium inflation should have begun to drop around 1992. This it did. In fact, aided by a historic 50 percent drop in the underlying (economywide) inflation rate that began in 1991, premium inflation plummeted, from 10.9 percent in 1992 to 0.5 percent in 1996.21 The drop in premium and systemwide inflation was deepened by the delayed effects of the recession of 1990–1991. Evidence indicates that for several decades changes in national income have been a powerful determinant of changes in national spending on health care three to five years later.22

The unusual depth and length of the inflatio n lull was caused primarily by the endorsement of the theory of managed competition, en masse, by a substantial swath of the nation’s political and economic elite just as the lull was beginning, and by the merger fever these endorsements triggered. Between October 1992 and May 1993 managed competition was endorsed by presidential candidate Bill Clinton, the legislatures of Washington and Minnesota, numerous other state and federal politicians, numerous large employers, and many newspapers.

It is possible that these political events, especially President Clinton’s endorsement of “managed competition within a budget” and his threats of price controls on prescription drugs, had a direct impact on inflation . As Henry Aaron noted, the last time the nation enjoyed a similar respite from health care inflation was the late 1970s, which just happens to be the last time an American president, in this case Jimmy Carter, threatened to use price controls to contain health care spending.23 But the establishment’s embrace of managed competition worked its most significant effect on health care inflation by triggering conso lidation fever throughout the industry. The overnight transformation of managed competition from untried theory to de facto national policy set off, or at least contributed greatly to, a wild scramble by insurers and providers to position themselves as high on the future food chain as possible or, in the case of small fry, to remain within nibbling distance of the “reformed” food chain. This meant, in short, getting big quickly.

Several tactics for getting bigger were available. One was low-balling one’s prices and premiums in order to seize market share rapidly or, in the case of smaller fry, to survive and remain attractive to the larger fry who were getting bigger not only by low-balling but by merging or otherwise consolidating. Everyone knew that low-balling could not go on forever, but everyone also knew that their first priority was to be alive and as integrated as possible into the food chain when the merger/low-balling panic was over. Recouping losses would have to wait until these first priorities were met. The losses, predictably enough, began to materialize in 1996 and became intolerable by 1997. The inflation lull was over.

Managed care has come to stand for a wide variety of cost containment tactics. But, to my knowledge, no one believes that “managed care” includes political endorsements of managed competition, merger sprees, and reactions to the underwriting cycle and to changes in national income. If that is the case, then it is fair to say that the jury has yet to return a verdict on whether “managed care” is more efficient than the old FFS system. And the jury will remain out until empirical evidence appears demonstrating that managed care plans have figured out how to reduce spending on medical services without cost shifting and without driving up administrative costs. Until such empirical evidence is published, analysts should not assert that managed care plans are more efficient than FFS plans, and they should refrain from attributing the 1990s inflation lull to the spread of managed care.

Kip Sullivan, an attorney, has been an organizer and researcher with consumer groups since 1976. When he wrote this paper, he was research director of Minnesota COACT (Citizens Organized Acting Together), an organization representing 14,000 Minnesotans. He is now a member of the steering commitee of the Health Care Campaign of Minnesota.

This paper was funded by grants from the Public Welfare Foundation and the Jay and Rose Phillips Family Foundation.

PNHP note: The orginal article, which appeared in Health Affairs in the July/August 2000 issue, is available here.

NOTES

1. The text of President Clinton's 1998 address is available online at www.pub.whitehouse.gov/uri-res/I2R?urn:pdi://oma.eop.gov.us /1998/1/27/11.text.1 (27 January 1998).
2. See, for example, transcripts of meetings of the National Bipartisan Commission on the Future of Medicare during 1998 and 1999, available online at rs9.loc.gov/medicare/trans.html.
3. P.M. Ellwood et al., “Health Maintenance Strategy,” Medical Care 9, no. 3 (1971): 291–298. In this paper, devoid of footnotes, Ellwood and colleagues assert that
capitation “assures that [HMOs] deliver services more efficiently…than conventional providers” (p. 295). But the only evidence offered for this assertion is this sentence: “Rates of hospital utilizatio n and surgery are…lower for health maintenance organizations, and HMO subscribers require fewer physicians per capita and fewer hospital beds than is true for the United States as a whole” (p. 296). Since these claims are not documented by source citations, it is impossible to know whether the studies upon which these claims were based examined HMO and FFS patients with similar health status. In any case, Ellwood and colleagues do not argue, nor could they, that medical costs are an excellent proxy for total costs.
4. S. Smith et al., “The Next Decade of Health Spending: A New Outlook,” Health Affairs ( July/Aug 1999): 86–95.
5. A.C. Enthoven, “Cutting Cost without Cutting the Quality of Care,” New England Journal of Medicine (1 June 1978): 1229–1238. In this paper Enthoven states that Luft had conducted a review of “many comparison studies done since 1950” and that Luft had “concluded that [HMOs cut total expenditures ] on the order of 10 to 40 percent,” that these “savings…are mainly attributable to much lower hospitalization rates and greater economy and efficiency of [HMOs, not] out-of-plan utilization, differences in age and sex composition, previous health status, or government subsidies” (p. 1230). In fact, Luft’s review examined five studies, not “many”; all of these studies involved Kaiser Permanente; and in fact Luft concluded that factors other than HMO efficiency — including "self-selection" into, and “undertreat[ment]” by, HMOs—could have accounted for the HMOs’ lower premiums. H.S. Luft, “How Do Health-Maintenance Organizations Achieve Their ‘Savings’?” New England Journal of Medicin e (15 June 1978): 1336–1343. Enthoven repeated this mistake in his 1989 article with Richard Kronick. A. Enthoven and R. Kronick,
“A Consumer-Choice Health Plan for the 1990s: Universal Health Insurance in a System Designed to Promote Quality and Economy (First of Two Parts),” New England Journal of Medicine (5 January 1989): 29–37.
6. R. Miller and H. Luft, “Managed Care Performance since 1980: A Literature Analysis,” Journal of the American Medical Association (18 May 1994) : 1512–1519,1515.
7. S.H. Long and M.S. Marquis, “Comparing Employee Health Benefits in the Public and Private Sectors, 1997,” Health Affair s (Nov/Dec 1999): 183–193.
8. Miller and Luft, “Managed Care Performance since 1980,” 1515. The PPO study Miller and Luft were referring to was J. Zwanziger and R.R. Auerbach, “Evaluating PPO Performance Using Prior Expenditure Data,” Medical Care 29, no. 2 (1991): 142–151.
9. Miller and Luft, “Managed Care Performance since 1980,” 1517.
10. R.L. Van Horn, L.R. Burns, and D.R. Wholey, “The Impact of Physician Involvement in Managed Care on Efficient Use of Hospital Resources,” Medical Care 35, no. 9 (1997): 873–889; R.J. Butler, R.P. Hartwig, and H. Gardner, “HMOs, Moral Hazard, and Cost Shiftin g in Workers’ Compensation,” Journal of Health Economic s 16 (1997): 191–206; “Drug-Pricing Suit Settlement Gets Preliminary Approval,” New York Times, 9 May 1996, C8; Prescriptio n D rug Study: A Report to the Minnesot a Legislature on the Prescriptio n Drug Marke t (St. Paul: Minnesota Department of Health, April 1994); K. Hiduchenko , “Do Health Maintenance Organizations Control Costs or Shift Costs?” New England Journa l of Medicine (1 April 1993): 971; E. Moy et al., “Relationship between National Institutes of Health Research Awards to U.S. Medical Schools and Managed Care Market Penetration, ” Journal of the American Medical Associatio n (16 July 1997): 217–221; J.S. Weissman et al., “Market Forces and Unsponsored Research in Academic Health Centers,” Journa l of the American Medical Associatio n (24/31 March 1999): 1093–1098; P.J. Cunningham et al., “Managed Care and Physicians’ Provision of Charity Care,” Journa l of the American Medical Associatio n (24/31 March 1999) : 1087–1092; K. Davis et al., “Choic e Matters: Enrollees’ Views of Their Health Plans,” Health Affair s (Summer 1995): 99–112; “How Good Is Your Health Plan?” Consumer Reports (August 1996): 28–41; F.J. Hellinger, “Selection Bias in HMOs and PPOs: A Review of the Evidence,” Inquiry (Summer 1995): 135–142; Physician Payment Revie w Commission, Annual Report to Congress (Washington: PPRC, 1996), 255–279; R.O. Morgan et al., “The Medicare-HMO Revolving Door—The Healthy Go In and the Sick Go Out,” Ne w England Journa l of Medicin e (17 July 1997): 169–175; R. Pear, “HMO Study Finds
U.S. Overpayment: Medicare Is Spendin g Billions on ‘Highly Inflated’ Fees,” New York Times, 11 August 1998, A1; and L.J. Passman et al., “Elderly Veterans Receiving Care at a Veterans Affairs Medical Center while Enrolled in Medicare-Financed HMOs: Is the Taxpayer Paying Twice?” Journal of General Internal Medicine (April 1997): 247–249.
The study by Davis and colleagues found that “17 percent of managed care enrollees reported using services outside of their plan in the past twelve months” and that the out-of-plan users “reported an average of four out-of plan visits within the past year for services not covered by their plan” (p. 106). Out-of-plan use was clearly related to dissatisfaction with the plan. The Consumer Report s survey found that 18 percent of enrollees in managed care plans sought medical services outside their plan networks.
11. B. McKasy and M.E. Marschall, Study of Health Care Management Companies (Minnesota Departments of Commerce and Health, March 1993). The percentages are based on the author’s calculations based on HMO expenditure s per member per month, as reported in this document.
12. The percentage increase was calculated by the author based on “health insurance administrative costs per person” figures presented in Henry J. Kaiser Family Foundation, Trends and Indicator s in the Changing Marketplace: Chartbook (Menlo Park, Calif.: Kaiser Family Foundation, August 1998) , Exhibit 6.8. The percentage increases in FFS and HMO premiums were calculated by the author from data on premiums presented in A.B. Krueger and H. Levy, Accounting for the Slowdown in Employer Health Care Cost s (Cambridge, Mass.: National Bureau of Economic Research, January 1997).
13. S. Woolhandler and D. Himmelstein, “The Deteriorating Administrative Efficiency of the U.S. Health Care System,” New England Journal of Medicine (2 May 1991): 1253–1258; S. Woolhandler, D. Himmelstein, and J. Lewontin, “Administrative Costs in U.S. Hospitals,” New England Journal of Medicine (5 August 1993) : 400–403; J. Alexander and C. Lemak, “The Effects of Managed Care on Administrative Burden in Outpatient Substance Abuse Treatment Facilities,” Medical Care 35, no. 10 (1997): 1060–1068; and Kaiser Family Foundation/Harvard University School of Public Health, Survey of Physicians and Nurses (Menlo Park, Calif.: Kaiser Family Foundation, July 1999).
14. D. Himmelstein, J. Lewontin, and S. Woolhandler, “Who Administers? Who Cares? Medical Administrative and Clinical Employment in the United States and Canada,” American Journal of Public Health (February 1996): 172–178. The percentage increase in the number of administrative workers was calculated by the author using these authors’ data on three categories of administrative workers.
15. T. Bonfield, “HMO Change May Set Trend Doctors Want,” Cincinnati Enquirer, 6 November 1999; and T. Burton, “Firms That Promise Lower Medical Bills May Increase Them: Cost Police Pile Paper Work on Physicians, Hospitals, Which Pass on Expense,” Wall Street Journal , 28 July 1992, A1.
16. These include G.A. Melnick and J. Zwanziger, “State Health Care Expenditures under Competition and Regulation, 1980 through 1991,” American Journal of Public Health (October 1995): 1391–1396; L.C. Baker and S.K. Wheeler, “Managed Care and Technology Diffusion: The Case of MRI,” Health Affair s (Sep/Oct 1998) : 195–207; L.C. Baker, “The Effect of HMOs on Fee-for-Service Health CARE Care Expenditures: Evidence from Medicare,” Journal of Health Economic s 16 (1997): 453–481; and L.C. Baker, “Association of Managed Care Market Share and Health Expenditure s for Fee-for-Service Medicare Patients,” Journal of the American Medical Association (3 February 1999): 432–437.
17. Krueger and Levy, Accounting for the Slowdown in Employer Health Care Costs; P.J. Feldstein and T.M. Wickizer, “Analysis of Private Health Insurance Premium Growth Rates: 1985–1992,” Medical Care 33, no. 10 (1995): 1035–1050; and S.C. Hill and B.L. Wolfe, “Testing the HMO Competitive Strategy: An Analysis of Its Impact on Medical Care Resources,” Journal of Health Economics 16 (1997): 261–286.
18. A.C. Enthoven, “Why Managed Care Has Failed to Contain Health Costs,” Health Affairs (Fall 1993): 27–43.
19. A.C. Enthoven and S.J. Singer, “Markets and Collective Action in Regulating Managed Care,” Health Affairs (Nov/Dec 1997): 26–32.
20. R. Cunningham III and R.M. Cunningham Jr., The Blues: A History of the Blue Cross and Blue Shield System (Dekalb, Ill.: Northern Illinois University Press, 1997),
226.
21. P.B. Ginsburg and J.R. Gabel, “Tracking Health Care Costs: What’s New in 1998?” Health Affairs (Sep/Oct 1998) : 141–146.
22. T.E. Getzen, “Macro Forecasting of National Health Expenditures,” Advances in Health Economic and Health Services Research 11 (1990): 27–48; and J.P. Cookson and P. Reilly, Mode ling and Forecasting Health Care Consumption (Radnor, Pa.: Milliman and Robertson, 19 May 1994).
23. H.J. Aaron, “Thinking Straight about Medical Costs,” Health Affairs (Winter 1994): 8–13.