1 See Linda A. Bergtold, Benefit Design Choices Under Managed Competition, Health Aff., Supp. 1993, at 99, 100-01. President Clinton's reform proposal includes in the minimum benefit package some items that large numbers of people would actually rather not buy, such as extensive benefits for mental illness and dental care. Cf. Pauly, Mark V., The Rational Nonpurchase of Long-Term-Care Insurance, 98 J. Pol. Econ. 153 (1990)Google Scholar. According to one account, “an official with [DHHS] participating on the White House task force, noted that the standard package of health benefits the White House has developed ‘includes literally everything … we're talking about a package comprehensive enough it will make your eyes cross.’ “ See Clinton Health Care Reform Plan Will Not Include Wage, Price Controls, Breaux Says, Daily Lab. Rep., Apr. 27, 1993, at 79. Statements by other Task Force members suggest that they incline toward this approach precisely because it would discourage competition in benefit packages and thus make more likely a uniform national benefit package across all demographic groups. Robert Pear, Health Planners’ Radical Idea: Same Coverage for All, N.Y. TIMES, Apr. 15, 1993, at A18.

2 For a vigorous attack on the value of consumer preference, see Frankford, David M., Privatizing Health Care: Economic Magic to Cure Legal Medicine, 66 S. Cal. L. Rev. 98 (1992)Google Scholar.

3 Hall, Mark A. & Anderson, Gerard F., Health Insurers’ Assessment of Medical Necessity, 140 U. Pa. L. Rev. 1637, 1645, 1659 (1992)Google Scholar.

4 In fact, many people today choose the least costly of the insurance options offered by their employer for just such reasons.

5 See supra note 1.

6 Of course, this objection is not avoided by plans that require employers to provide the coverage, since the employer's increased costs would otherwise be available for employee wages. Government provision of the services would of course require increased tax revenues. If collected from those who are receiving the excess health care insurance the case is indistinguishable from market purchase. Increasing taxes on third parties, on the other hand, would be difficult to justify where their purpose is to provide a service that recipients might rationally decline to buy if they had the necessary resources.

7 We use these and other numbers only for purposes of rough illustration without intending to maintain that the actual levels are accurate or desirable. We use 1000 as the size of the group because it is a round number that seems large enough to suggest some actuarial assurance that the experience in any particular year is unlikely to be skewed by the unusual healdi problems of a few subscribers.

8 Our example uses a larger pool in diis case to suggest statistical predictability for a more restricted range of services.

9 Henry J. Aaron & Wiluam B. Schwartz, The Painful Prescription: Rationing Hospital Care 17-20 (1984).

10 See Commission of Employer-Based Health Benefits, Division of Health Care Services, Institute of Medcine, Employment and Health Benefits: A Connection at Risk 101-05 (Marilyn J. Field & Harold T. Shapiro eds., 1993) [hereinafter Commission of Employer-Based Health Benefits].

11 This problem was shown dramatically when Oregon decided to contain Medicaid costs by creating priority lists of medical procedures. See generally W., John Thomas, The Oregon Medicaid Proposal: Ethical Paralysis, Tragic Democracy, and the Fate of a Utilitarian Health Care Program, 72 Or. L. Rev. 47 (1993)Google Scholar. The anguished and prolonged debate over the arrangement of different procedures on this list could not escape the reality that no satisfactory line can be drawn because the entire conception of the task is flawed. So long as the rationing plan starts with the premise that every procedure is either fully in or fully out, the result will be a system that covers treatments of very limited benefit in some cases while denying coverage for other treatments that in some situations are highly beneficial.

12 A solution to this problem would be beneficial to the insurance companies as well as to the insured. Insurance contracts today generally require reimbursement for all appropriate medical care in covered categories, and “appropriateness” is generally defined as care that is necessary and nonexperimental. Hall & Anderson, supra note 3, at 1640-41. The vagueness of these terms has made them dysfunctional. Courts, when faced with a lack of consensus in the medical and research communities about a treatment's appropriateness, consistently side with the discretionary judgment of the patient's treating physician over an insurer's self-interested denial of coverage. Ferguson, John H. et al., Court-Ordered Reimbursement for Unproven Medical Technology: Circumventing Technology Assessment, 269 JAMA 2116, 2120 (1993)Google Scholar. As a consequence, the coverage terms in existing health insurance, both public and private, tend to pay for all care in the designated blocks of service that the treating physician considers to be of any medical benefit, no matter how small and how uncertain, and no matter how expensive compared with the available alternatives. Hall & Anderson, supra note 3, at 1649-51.

13 David Hadorn, Necessary Care Guidelines, in Basic Benefits and Clinical Guidelines (D. Hadorn ed., 1992); Robert, Brook, Health, Health Insurance, and the Uninsured, 265 JAMA 2998 (1991)Google Scholar; Havighurst, Clark C., Practice Guidelines for Medical Care: The Policy Rationale, 34 St. Louis U. LJ. 777, 796 (1990)Google Scholar.

14 The Managed Competition Act of 1992, H.R. 5936, 102d Cong., 2d Sess. (1992).

15 Deborah, Garnick et al., Can Practice Guidelines Reduce the Number and Costs of Malpractice Claims?, 266 JAMA 2856 (1991)Google Scholar; Edward, Hirshfeld, Should Practice Parameters Be the Standard of Care in Malpractice Litigation?, 266 JAMA 2886 (1991)Google Scholar.

16 See Commission of Employer-Based Health Benefits, supra note 10, at 103.

17 See Brook, supra note 13; Hadorri,’ supra note 13; P., Kalb, Controlling Health Care Costs By Controlling Technology: A Private Contractual Approach, 99 Yale L.J. 1109 (1990)Google Scholar.

18 Anderson, Gerard F. et al., Medical Technology Assessment and Practice Guidelines: Their Day in Court, 83 Am. J. Pub. Health 1635, 1638 (1993)Google Scholar.

19 Institute of Medcine, Controlling Costs and Changing Patient Care? The Role of Utilization Management (Bradford H. Gray & Marilyn J. Field eds., 1989).

20 See Grumet, Gerald W., Health Care Rationing Through Inconvenience: The Third Party's Secret Weapon, 321 New Eng. J. Med. 607 (1989)Google Scholar.

21 Technically, of course, insurance companies do not grant or deny access, but merely decide whether to pay for the access. For most people, of course, these two are often the same, especially in the case of the costly procedures that demand the most focused attention in these reviews.

22 John Wennberg & Alan Gittelsohn, Variations in Medical Care Among Small Areas, Sci. Am., Apr. 1982, at 120. See generally David M. Eddy, Variations in Physician Practice: The Role of Uncertainty, Health Aff., Summer 1984, at 74.

23 Regarding the size of these risk pools and the amounts of coverage, see supra notes 7 & 8.

24 The appropriate cut-off points would clearly vary from procedure to procedure depending upon factors like the relative cost and the importance of the information that might be obtained. Consider, for example, a diagnostic procedure that would increase from 95 to 99 percent the physician's confidence in a diagnosis. In many cases this increased confidence will not be worth the cost of an expensive procedure. On the other hand, if in a particular case diagnostic error is likely to have serious health consequences for the patient, then even this small improvement in the diagnostic accuracy may be worth the cost.

25 Consider the current dispute over whether T.P.A. (tissue plasminogen activator), a genetically engineered substance costing $2,400 a dose, or streptokinase, an older drug derived from bacteria at $240 a dose, is preferable treatment for patients experiencing heart attacks. See Lawrence K. Altman, A Costlier Heart Drug Also Proves Better, N.Y. TIMES, May 1, 1993, at A7. The most recent study concludes that one percent more patients survive with the more expensive drug. Id. Under a BRP policy, hospitals would receive strong, if not conclusive, guidance for resolving the “agonizing decision” of which drug to use if subscribers budgeted an explicit amount for acute treatment of heart ailments.

26 Veatch, Robert M. & Spicer, Carol M., Medically Futile Care: The Role of the Physician in Setting Limits, 18 Am. J.L. & Med. 15 (1992)Google Scholar.

27 An alternative approach would ask these insurers to define their benefit packages in units of service rather than dollars, but this would make cross-insurer comparisons much more, difficult, and it would deter such insurers from adopting technological innovations that resulted in better care by employing fewer and more expensive services. It is better to require insurers to define budgets in monetary terms. On the other hand, insurers may still market these plans with reports of prior years’ experience as to the number of units of service provided per dollar spent.

28 This is likely to occur because people with established physician relationships are both less willing to accept a plan that limits their choice of physician and more likely to have a history of illness. A similar phenomena might arise under BRPs in that individuals with a history of undergoing costly procedures might disproportionately select risk tiers that provide generous budgets for those procedures.

29 This problem may be at least partially self-correcting under BRPs since some of the less healthy subscribers in the high-cost/high-risk group will find it advantageous to switch to lower cost risk pools populated primarily by healthier patients because they would then face less competition for the services, even if fewer in total number. At equilibrium, each risk pool would tend to have a mix (although not necessarily an equal mix) of both healthy and sick subscribers. This would not hold true to the same extent if tiers were defined by large differences in service categories.

30 Thus, insurers would set the treatment criteria for each risk tier in a BRP system as if the subscribers to it were a random selection of health risks from the population. If, as we suggest, each company is required to offer a full selection of risk tiers, then the result, if the subscribers in fact sort themselves among tiers non-randomly with respect to health status, will a risk budget surplus in the tier attracting healthier subscribers and a risk budget shortfall in the tier attracting less healthy subscribers. At that point the company could be allowed to apply the surplus from the first group to the shortfall in the second.

31 These concepts are presently evolving under the federal law known as ERISA that draws from trust law to govern claims disputes under employer-based insurance. See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 110 (1989).

32 Hall & Anderson, supra note 3, at 1666.

33 Although we do not wish to explore this topic here, we observe that varying tort standards should apply to non-treatment decisions rendered at different funding tiers. See Hall, Mark A., Health Care Cost Containment and The Stratification of Malpractice Law, 30 Jurimetrics J. 501 (1990)Google Scholar; Siliciano, John A., Wealth Equity, and the Unitary Medical Malpractice Standard, 77 Va. L. Rev. 439 (1991)Google Scholar. However, within each tier, the standard of care should be influenced by the customs that prevail among similarly funded plans.

34 See supra note 28.

35 See Gilbert, Welch, Should the Health Care Forest Be Selectively Thinned by Physicians or Clear Cut by Payers?, 115 Annals of Internal Med. 223 (1991)Google Scholar; Marc A. Hall, The Law, Ethics and Economics of Physician Bedside Rationing (unpublished manuscript on file with author).

36 See, e.g., Epstein, Richard A., Why Health Care is Special?, 40 Kan. L. Rev. 307 (1992)Google Scholar. We observe, however, that while Americans believe overwhelmingly that everyone is entitled to decent health care, they also believe that no one should be barred from purchasing as much health care or health insurance as they wish. So long as some people wish to buy more than others, multiple tiers above a decent minimum are essential if both principles are to be honored. Id.