Printer Friendly

Health insurers' assessment of medical necessity.

One can hardly imagine a more difficult choice than that faced by the district court in Rollo v. Blue Cross/Blue Shield.(1) In the judge's own words:

I was called upon to decide whether eight year old Tishna Rollo could live or whether she must die, a humbling and sobering decision. Tishna, I was told, had virtually no chance of surviving the relapsed Wilms' tumor [of the kidney] from which she is suffering and Blue Cross/Blue Shield had denied coverage for autologous bone marrow transplant ("ABMT") with accompanying high dose chemotherapy, a treatment which could well prolong and quite possibly save her life and which, concededly, provided her only realistic hope of either. The University of Nebraska Medical Center ... would not admit her without coverage, her parents could not afford to pay the projected $130,000--$140,000 costs of the treatment, and there was simply no time to wait, for Tishna's "window of opportunity" would soon close. The coverage issue had to be litigated and litigated quickly.(2)

Blue Cross denied coverage because the insurance policy covering Tishna Rollo explicitly excluded "experimental" procedures, and Blue Cross had concluded that the use of autologous bone marrow transplant (ABMT)(3) to treat Wilms's tumor had not yet been proven medically appropriate.(4) The "experimental" exclusion common in health insurance policies responds to a growing concern that most current medical procedures were adopted without ever having been tested rigorously and that at least some of the procedures commonly used today have limited or no medical value.(5) In the view of many health policy analysts, one reason so many unproven medical procedures have diffused into common medical practice is that insurers have been unwilling or unable to deny payment for care whose clinical efficacy is still in doubt.(6)

By requiring clinicians to prove that new procedures are efficacious before they are covered, the hope is that existing resources will be better allocated to maximize the health status of the overall population.(7) Thus, efforts to spend health dollars more wisely may ultimately enable the approximately 35 million Americans without health insurance to obtain coverage,(8) foster adequate prenatal care,(9) and make long-term-care insurance affordable.(10) Nevertheless, the judge in Rollo decided that Blue Cross should pay for the treatment.(11) Sixteen other courts over the past two years have also ruled that ABMT is not experimental, and they have ordered private insurers to pay the costs of administering it to terminally ill cancer patients, most often for metastatic breast cancer.(12) In about a dozen similar cases, however, judges have ruled that the use of ABMT is still experimental and denied coverage.(13)

These dramatic rulings have been viewed by the popular press and the health policy community alike as an extraordinary development.(14) From a legal perspective, however, these rulings are merely the latest in a long series of ordinary contract disputes over the interpretation of terms such as "medical necessity" or "experimental," which determine the coverage of health insurance policies.(15) For the most part, these previous contract disputes involved treatment at the periphery of traditional medicine with only modest amounts of money at stake.(16) Now, however, the stakes are much higher on both sides. The treatments now being questioned are very expensive new procedures that if performed on a widespread basis could leave more people uninsured and drive up costs for those who remain covered. These treatments, however, may also offer an individual patient's only hope of survival. Thus, bone marrow transplant cases present one of the core problems that perplex this country's health care financing policy: should we rely on rigorous empirical evidence before making collective decisions to spend our limited health care resources, or should we continue to rely on the individual physician's and patient's judgment about the appropriate treatment in each case?

This Article examines whether courts should allow both public(17) and private insurers to control health care costs by denying payment for specific services based on the insurers' judgments of medical appropriateness. Our primary focus in on private insurers; however, there is significant overlap with public sector activities because private insurers rely on governmental assessments to guide their decisions, and because the statutory terms of coverage under public programs resemble the coverage terms in private insurance contracts. Therefore, although we concentrate on the case law that interprets private insurance contracts, the Article's conclusions and policy implications extend to the public sector.

First, we set in perspective the most recent litigation by describing how the cases have evolved from the interplay of litigation and insurers' responses to prior cycles of judicial decisions. Next, we develop a normative model of who should decide questions of medical appropriateness for the purposes of insurance coverage and what criteria they should employ. In particular, we critique the role the courts have played in making and reviewing medical appropriateness determinations. In the final section, we set forth a proposal for reforming the process of determining medical appropriateness. We specify a new medical necessity review mechanism and analyze how this initiative is likely to fare in the present legal climate.

Our focus is on the policy dimensions of how insurers currently, and in the future are likely to, make medical appropriateness determinations. As such, this analysis differs from other articles that focus on the doctrinal elements of the case law(18) or addressing collateral legal issues such as the potential malpractice or antitrust liability exposure that results from making medical appropriateness determinations.(19) Our principal concern is with the core contractual questions of whether and under what terms the courts will allow insurers to play any significant role in assessing medical judgment through their coverage policies.

Other articles in this Symposium address more explicitly the theme of rationing medical care--that is, withholding useful treatment based solely on its cost. Because our primary focus is somewhat different, we prefer not to frame our discussion in terms of rationing. We consider whether insurers may exercise independent judgement on questions of medical benefit or comparative cost effectiveness (concepts we define with more precision later). We do not necessarily advocate that insurers should play an explicit rationing role. To the extent this Article is conceived of in terms of rationing, however, the issue properly framed is whether potential subscribers should be allowed to seek more superior forms of rationing than those that presently exist, not whether insurer rationing should be permitted at all. The traditional coverage terms of public and private health insurance, as those terms have been interpreted by the courts, presently ration by rendering insurance unaffordable for entire categories of medical care (nursing home, mental health, and dental, for example), or by rendering health insurance entirely unaffordable for thirty-five million people. Even if allowing public and private insurance to exercise greater oversight of medical appropriateness were viewed as tantamount to insurer rationing (which is not necessarily the case), this could only improve the existing blunt form of rationing.

I. POLICY CONSIDERATIONS UNDERLYING OLD AND NEW JUDICIAL RULINGS

Although the ABMT cases seemed to emerge overnight, they are a logical outgrowth of decades of litigation involving private insurers' coverage decisions. Understanding this history and the policy issues underlying this large body of case law helps set the present controversy in its proper framework. This analysis will show that the inclination of judges to adopt every conceivable argument in favor of coverage has essentially precluded insurers from exercising any meaningful oversight of medical appropriateness.

A. Retroactive Denials of Coverage

Because the hospital industry gave birth to modern private health insurance in the United States, it is not surprising that private insurers were initially very deferential to both hospitals and physicians.(20) At first, insurance policies contained no explicit medical necessity limitations or review mechanisms. Instead, they covered, within defined monetary and service limits, all care ordered by any physician.(21) To the extent that private insurers restricted coverage in any form, it was through design of their benefit package. For example, a package might have covered only hospital care or only specified diseases.(22) Insurers first began questioning the judgment of individual physicians in the 1960s when they were asked to pay for the use of hospital facilities for such purposes as weight reduction or resting up from a fall.(23) Even these mild protests were generally unsuccessful in court, however.(24) Judicial deference to the practicing physician in this period reached its pinnacle in Duncan v. J.C. Penny Life Insurance Co.,(25) where the court required coverage for two three-week periods of hospitalization for a husband and wife for bruises and sprains, under facts that "strongly indicate[d] a motive on the part of the Duncans ... to reap financial gain."(26) The court ordered the insurer to pay even though the Duncan's doctor admitted that the care could have been administered at home and five other doctors agreed that hospitalization was medically unnecessary.

In response to these rulings, insurers began to revise their contracts by inserting an explicit requirement of "medical necessity."(27) This, they thought, would give them the flexibility to deny potentially unnecessary medical care or care that might actually be harmful to patients. They continued, however, to lose their challenges with regularity, even under the most extreme facts.(28) For instance, they were required to pay for alternative cancer therapies such as laetrile and "immuno-augmentative" treatment that were outlawed in the United States but delivered in Mexican and Caribbean clinics.(29)

By the end of the 1970s, many insurers had adopted two new contractual revisions in response to this additional round of losses:(30) first, they specified that medical necessity is to be determined in the insurer's judgment,(31) and second, they explicitly excluded payment for "experimental" or "investigational" procedures.(32) The recent bone marrow transplant cases, as well as decisions overruling insurers' denials of coverage for more routine treatment,(33) demonstrate, however, that these contractual revisions still have not convinced the courts to accept insurers' assessments of medical appropriateness.

When deciding in favor of the patient, the courts routinely profess to limit the scope of their ruling to the unique facts in each case, as if to justify in their own mind that a generous ruling will not extend beyond the particular case.(34) The courts continually offer the assurance that if the contract had been properly phrased, the insurer would have prevailed. When insurers follow this advice, they still lose when courts, confronted with the same suggested modifications, find the contract "ambiguous" or contrary to the policyholder's "reasonable expectations."(35) For example, the Duncan court found the contract defective because it failed to require "medical necessity,"(36) even though countless other courts have found this term hopelessly ambiguous and therefore unenforceable.(37) When insurers first refused to pay for unapproved drugs, the courts chastised them for failing explicitly to exclude coverage for "experimental" drugs,(38) yet this is the very language courts are now refusing to enforce in the ABMT cases.(39)

These and other contradictions in the courts' rulings create the appearance of a judiciary driven by a single-minded concern for extending coverage to patients in desperate situations at all costs.(40) In addition to this humanitarian objective, the courts have been concerned about the perceived unfairness of a retroactive denial of coverage after a patient has relied on his physician's advice and incurred a bill for treatment later found by the insurer to be inappropriate.(41) These concerns led to the position that, in deciding what is "medically necessary" or "experimental," the doctor's decision should control.(42) After all, the insurance company has a conflict of interest; once it collects its premium dollars, anything it pays out in claims affects its profit margin. This judicial attitude has rendered unreliable any agreement that allows the insurer to police medical decisionmaking.(43) Since this limitation has been enforced covertly through several cycles of strained contract interpretation rather than by explicit announcement, its validity and implications have not been fully considered.(44)

B. Prospective Denials of Coverage

1. The Emergence of Prospective Utilization Review

Until recently, the judicial constraint on market contracting had relatively minor ramifications. Insurers issued only a few denials at the periphery of conventional medical practice involving fairly modest sums. The medical establishment was largely unaffected, health insurance was still sold profitably, and most employers were generally unconcerned about the cost of employee health benefits.(45)

These idyllic times have ended. During the 1980s, a number of studies questioned the appropriateness of many of the procedures commonly ordered by physicians.(46) Other studies found wide variations in hospital admission rates across geographic areas that could not be explained by demography, health status, economic status, or other relevant factors.(47) The general consensus of the researchers was that much of the variation in medical practice could only be explained by a discretionary "medical practice factor."(48) These studies encouraged insurers to begin reviewing the appropriateness of medical procedures more closely and in advance of treatment, a technique know as "prospective utilization review." Both prospective and retrospective utilization review gained momentum in the 1980s for several other reasons. When Congress changed the system for paying hospitals under the Medicare program in 1983, it established Peer Review Organizations (PROs) to monitor the appropriateness and quality of medical care provided to Medicare beneficiaries.(49) At about the same time, large private employers began to self-insure and implement their own screening criteria for medical effectiveness.(50) Private insurers, attempting to maintain market share by demonstrating that they could control health care costs, also began instituting prospective utilization review programs. Suddenly, from all directions, physicians experienced much greater scrutiny of their treatment decisions than ever before.(51)

As the utilization review programs evolved, a significant difference emerged between the public and private sector programs. The public sector continued the retrospective review of claims, allowing the denial of payment after the service was performed. The private sector, responding to the courts' concern over the perceived unfairness of retroactive denials,(52) instituted prospective review wherever possible.(53) Under prospective utilization review, the determination of medical necessity is made prior to treatment. Various methods for prospective utilization review quickly emerged from the usually sluggish indemnity market,(54) testifying to the strength of the competitive pressures placed on insurers by employers faced with skyrocketing health insurance costs. For example, from 1984 to 1988, the proportion of conventional indemnity insurance that employed pre-admission certification, a form of prospective review, increased from five to seventy-three percent.(55)

A typical precertification review system operates as follows. The insurance contract requires the subscriber or physician, except in emergencies, to obtain permission before entering the hospital or undergoing certain expensive outpatient procedures.(56) The treatment request is usually reviewed initially by a computer algorithm that flags certain requests for further clinical review.(57) Then a nurse, applying fairly rudimentary screening criteria, reviews these cases to determine which ones require further physician review.(58) Physician reviewers, who rely on published studies of medical effectiveness as well as their own clinical experience, then apply their independent judgment of medical appropriateness, usually after consulting with the treating physician in cases of disagreement.(59) In most such cases, the two doctors reach some accord; the overall percentage of denied requests is only one to two percent.(60)

2. Judicial Enforcement of Prospective Review

Prospective review was intended to alleviate the perceived unfairness of retroactive denials of coverage.(61) Instead of having this effect, however, prospective review creates in the minds of some judges a heightened sense of hardship since providers demand a source of payment before undertaking treatment. As one court explained: "A mistaken conclusion about medical necessity following retrospective review will result in the wrongful withholding of payment. An erroneous decision in a prospective review process, on the other hand, in practical consequences, results in the withholding of necessary care, potentially leading to a patient's permanent disability or death."(62) In cases of terminal illness, courts often assume the worst-case, life-or-death scenario prior to treatment: that the patient will certainly die without the requested treatment and that the treatment will definitely save the patient's life.(63) These extreme assumptions, coupled with the payment demands of providers, force most courts to confront these cases in a preliminary injunction context. This legal posture requires the court to balance the equities between the parties in a manner that inevitably favors avoiding the possible loss of life over the insurers' monetary loss.(64)

Viewing the choice from this perspective, it is not surprising that insurers continue to lose despite extremely attenuated grounds for coverage. For instance, in Bradley v. Empire Blue Cross & Blue Shield(65) the court enjoined Blue Cross from classifying ABMT treatment for an HIV-positive patient as experimental.(66) The court found Blue Cross likely to lose on the merits, despite evidence that the patient's doctor was the only physician in the country who had ever employed this treatment and despite the patient's signing of a "clinical investigation consent form" that "emphasize[d] the research aspect of the procedure."(67) Almost by definition, treatment subject to certain forms of clinical investigation must be considered of unproven effectiveness. Before patients are randomly selected to participate in the non-treatment arm of the study, ethical requirements demand that an independent, broadly composed "human subjects research review board" (also called an "institutional review board" or "IRB") determine there to be as good a chance that the therapy will hurt the patient as help.(68) In Bradley, the court found the procedure non-experimental even though the level of confidence had not even progressed to this preliminary "cointoss" stage of confidence.(69)

In summary, insurance coverage disputes have evolved through three cycles of litigation, each one characterized by the courts' unwillingness to accept and enforce the very contractual terms suggested in the prior stage. Courts continually fail to see beyond the heart-rending facts of the immediate case to the reality that the present strained ruling in favor of coverage will be applied by other courts even if the contract is revised in the suggested manner. Consequently, the parties to the health insurance contract are frequently precluded from enforcing the terms they have chosen to define the limits of coverage. The following section explores the impact of judicial interference in the health insurance market.

C. The Social Cost and Deterrent Effect of the Courts' Decisions

While it is easy to understand why a judge would do everything possible to find coverage in a life-threatening situation, this visceral response ignores the long-term implications of inhibiting insurers from questioning treating physicians' decisions.

1. Pricing Policyholders Out of the Market

The judicial inclination to find all conceivable grounds for coverage imposes considerable costs on policyholders beyond the mere price of the treatment in the litigated cases. One cost is the deterrent effect these rulings have on insurers' cost containment efforts. A series of negative decisions is likely to convince insurers to extend coverage in all similar cases, even though the subscribers may not have desired to pay for this additional coverage at the outset. In effect, courts are mandating coverage that informed consumers in the private marketplace would have chosen not to purchase, a phenomenon known as "judge-made insurance."(70) Like legislatively mandated insurance,(71) judge-made insurance has an associated cost.(72) It increases the price of health insurance for everyone, which forces some consumers at the margin out of the health insurance market altogether.(73)

Even if an insurer continues to deny certain claims, the expense an uncertainty of litigation impose considerable additional costs on the insurance product.(74) The expense of litigation is self-evident. The uncertainty component results from the unpredictability of how courts will interpret contractual terms. Insurers must either include this additional uncertainty of judicial interpretation in their risk calculations when pricing the insurance product, or they must write their coverage in a less desirable fashion to achieve the degree of predictability necessary to price the insurance product accurately.(75) Thus, refusing to honor the plain meaning of insurance contracts results in consumers being unable to purchase the scope of insurance coverage they desire.(76)

2. The Deterrent Effect on Cost Containment Efforts

Another social cost of these rulings is that they deter insurers from assuming a more aggressive role in containing health care costs. The difficulty of enforcing payment denials for inappropriate care has made insurers notiously submissive in applying medical appropriateness standards.(77) Until recently, insurers have been willing to defend their denials of coverage only at the outer extremes of medical practice. Even with their current greater resolve, insurers deny only one to two percent of claims reviewed under prospective utilization review,(78) despite the growing body of literature suggesting that a much greater portion of the procedures currently being covered are inappropriate.(79)

These rulings have also deterred insurers from choosing the criteria to assess medical appropriateness. The medical technology assessment literature identifies three levels of evaluation for medical procedures, drugs, and devices. They can be found (1) safe and efficacious, (2) cost effective, and/or (3) cost beneficial. The first level, which is employed by the Food and Drug Administration to evaluate all new drugs and devices, asks only whether the procedure is safe and provides some medical benefit, however small.(80) The second level asks whether this medical benefit is superior to what can be achieved by other procedures at equal or lesser cost. Only the third level asks whether a net increase in medical benefit is worth the cost.(81)

Health insurers have historically operated only at the first level, seeking to justify their coverage denials primarily as a means to protect patients from harmful or fraudulent care.(82) They have not even asked whether a beneficial therapy might be more cheaply performed without sacrificing any benefit, let alone whether a marginally increased benefit is simply too expensive to be worthwhile. They have asked only that some medical benefit be demonstrated, leaving the choice of less versus more effective and less versus more expensive modalities to the doctor and patient.(83) Even with this extreme trepidation, they consistently lose in court.

Although insurers are considering incorporating cost in their assessment of medical effectiveness,(84) this means only that they might move to the second tier of assessment, not that they are taking on the much more controversial role of paying for services only if the benefits exceed the costs.(85) Whether they can make even this modest move depends, of course, on how successful they have been in the first tier. Since insurers have been remarkably unsuccessful, it is certain that their cost containment efforts will continue to be cautious at best.

This deterrent effect is heightened by the extraordinary penalties that can attach to payment denials the courts find to be erroneous. Through reasoning unprecedented in other areas of law, courts have imported elements of tort law into these contractual disputes by imposing punitive damages on insurers for breaches of contract found to be in bad faith,(86) while at the same time, lowering the ordinary tort standards for determining when allegations of bad faith may be presented to the jury.(87) Insurers may now face malpractice liability when their prospective coverage denials are followed by a deterioration in the patient's condition, or even by the patient's failure to improve.(88) Although the Employee Retirement Income Security Act (ERISA)(89) may preempt much of this law, leaving plaintiffs primarily to their contract measure of compensation,(90) ERISA also has a penal element allowing the plaintiff's court costs and attorney fees to be imposed on the insurer. Moreover, now that utilization review is occurring on a prospective basis, the consequential damages that are available from a breach of contract remedy likely could include the same measure of harm present in a tort action.(91)

Because the courts reflect the attitude of society generally,(92) this description of the case law and its consequences is not intended to argue that these rulings lack popular support. We only observe that these extreme rulings impose costs that courts frequently ignore. If insurers must fully litigate each coverage decision in each jurisdiction under the wording of each plan for each employer, they will continue to be relegated to a role of writing blank checks to the medical profession. As a result, the courts, rightly or wrongly, impose considerable private and social costs by deterring insurers from assuming a more active role in slowing the diffusion of medical technologies that are not efficacious, cost effective or cost beneficial. The following section analyzes more systematically whether these rulings serve goals worth the sacrifices or whether they instead subvert broader public policy objectives.

II. SHOULD INSURERS DETERMINE MEDICAL APPROPRIATENESS?

A. A New Role For Insurers

Until the mid-1970s the primary health policy objective was to expand health care services. By adopting professional norms to govern coverage decisions, the insurance industry was conducive to this developmental era. This harmonious, expansionary attitude continued when the Medicare and Medicaid programs were enacted in 1965.(93) Some time in the mid-1970s, however, the emphasis shifted towards controlling health care costs.(94) Concerns over improving access to health insurance for the uninsured, expanding coverage to the marginally insured, and providing long-term-care coverage have been put on hold until health care costs can be brought under control.(95) In this new environment, the failure of insurers to take a more aggressive role in questioning medical practices has led to the criticism that they are not using their tremendous market power to control health care costs and to reduce inappropriate medical care.(96)

Cost containment became a policy objective for several reasons. Policymakers within state and the federal governments saw the portion of their budget going for medical services increase rapidly, constraining their ability to fund other programs.(97) Corporate executives grew concerned that high health care costs were harming their international competitiveness.(98) As governments and private industry attempted to control costs by imposing cost sharing, the consumer became more aware of the cost of health care. Unions now frequently strike in response to corporate proposals to increase the level of cost sharing.(99) All of this is occurring within the backdrop of health services research findings that a significant portion of medical care is inappropriate and is delivered in widely varying patterns that cannot be explained by variations in health status or other factors.(100)

According to most observers, a driving force behind the increase in health care spending is new technology.(101) While there have been attempts to control new technology through regulations, such as certificate of need controls(102) and hospital payment reform,(103) the growing consensus is that more rigorous assessments of the actual health outcomes of alternative treatment modalities will also be an effective mechanism for eliminating unnecessary or inappropriate medical procedures.(104) The federal government established a new federal agency, the Agency for Health Care Policy and Research, with a mandate to conduct medical effectiveness studies and issue detailed practice guidelines.(105) It is anticipated that the results will be used to inform physicians and patients about appropriate medical care and to assist public and private insurers in developing coverage policy.(106)

The need for this activity is ever more urgent in light of the relentless pace of medical advances. Biotechnology products such as erythropoietin and interferon are rapidly becoming available, with a huge potential market and an enormous total price tag.(107) In the future, an unforetold range of genetic therapies is likely to become available(108) at a potentially extraordinary additional cost. Often these new technologies are provided in addition to, not as substitutes for, existing technologies.(109) To use these new technologies wisely, it will be necessary to determine when they are the most cost effective. Unfortunately, none of the existing regulatory processes such as Food and Drug Administration approval mandate such an analysis.(110) Therefore, public and private insurers are the most likely institutions to mandate a rigorous evaluation of the appropriateness of new medical products.(111)

Many independent observers view insurers' review of medical appropriateness as a positive development.(112) Insurers can use their data and direct contact with consumers to design a contract that reflects a tradeoff between price and the scope of covered services. Insurers are situated to make these welfare-optimizing decisions because they operate in a marketplace that penalizes them for failing to balance the customers' desire for cost containment with the desire for access to necessary medical services. Their old role of reinforcing the professional preferences of medical providers has been replaced by a new role that expresses the view of consumers in the marketplace.(113)

B. Objections to Insurers' New Role

While there may be agreement in the abstract that insurers should be given a larger role in determining medical appropriateness,(114) there is considerable disagreement over what that role should be. Specifically, can insurers be trusted to make decisions in their subscribers' best interest? The two principal objections to insurers reviewing medical appropriateness are (1) that, by virtue of their proprietary interest in the premiums they have already collected, insurers have a conflict of interest that precludes them from making a neutral, unbiased decision; and (2) they lack the knowledge and expertise to make medical treatment decisions wisely. In the following sections we discuss the merits of these objections.

1. Competing Conflicts of Interest

a. Physicians' Conflict of Interest

One reason subscribers may want to give their insurers authority to review their doctors' determinations of medical necessity is that, otherwise, doctors enjoy an unconstrained ability to determine their own payments, with obvious consequences for the price of the insurance product.(115) Imagine what our food bills would be if we wheeled an empty shopping cart up to the grocer each week and asked him what he thought it would be "dietarily appropriate" to eat.(116) Although ethical and clinical standards deter physicians from ordering completely unnecessary tests, and coinsurance and inconvenience decrease patients' willingness to undergo excessive testing and medical procedures, ample data suggests that physician financial incentives are nevertheless a significant determinant of treatment behavior.(117)

Similar incentives affect doctors called to testify in coverage disputes. Even if they do not benefit directly from the particular patient in question, doctors may have a strong professional interest in seeing the particular technology more widely disseminated.(118) Many physicians have adopted the ethic in insurance matters that the ends of treatment justify even a borderline-fraudulent means of obtaining payment.(119)

b. Constraints on Insurers' Conflict of Interest

On the other hand, giving the insurer absolute authority over coverage decisions could create an even stronger conflict of interest operating in the direction of withholding treatment. The insurer, after all, stands to profit from not paying claims, particularly after it has enrolled the subscribers and collected the premiums. This point is beginning to occupy much of the health benefits litigation under ERISA.(120) A considerable number of safeguards, however, are already in place to prevent the insurer from acting in an opportunistic fashion.

First, over half the claims that insurers process are for self-insured clients.(121) For these clients, the insurer has no immediate economic conflict because it is paid only for its administrative services and is not at risk for claims paid. In fact, since the insurer is commonly paid a percentage of the value of the claims that it honors, the financial incentive is to pay more claims. Plaintiffs might argue that there is still an incentive to deny claims because a self-insured employer is more likely to renew an administrative services contract if the insurer saves the employer money.(122) The employer, however, is more likely to be concerned about employee dissatisfaction if the claims administrator denies too many claims.(123) Therefore, there is no more basis for using contract law to police the employer's choice of the claims administrator than there is to second-guess its decision of how much health insurance to buy in the first place.

The concern over financial self-interest is perhaps more apparent for self-insured employers who make their own in-house medical necessity determinations. Far from viewing the incentives in this light, however, decisions under ERISA have underscored the propriety of economically conservative claims administration by employer plan administrators. This perspective is derived from the realization that claims paid out of a pool of money dilute the funds available to the other employees. Thus, ERISA law, far from viewing employers as unfairly influenced by financial incentives, sees them as fiduciaries charged with managing the assets of the plan for the use of all beneficiaries.(124) This fiduciary perspective makes explicit the cost/benefit tradeoffs involved in medical decisions--the same tradeoffs that courts too often fail to recognize.(125) As a consequence, the only conflict the employer faces is that between a single claimant and a pool of beneficiaries, the very conflict that should be foremost in the insurer's mind when assessing medical appropriateness.(126)

The concern over insurers' financial incentives becomes more forceful under conventional indemnity arrangements where a third party insurer bears the insurance risk. Nevertheless, it is still not the case that insurers enjoy absolute, unreviewable discretion; several countervailing forces constrain their latitude. First, some at-risk insurers do not make their own medical necessity decisions. Those contracting for utilization review services with outside firms turn this decision over to disinterested reviewing personnel.(127) These utilization review firms either apply their own criteria or purchase review criteria developed independently.(128) Additional assurance of neutrality is provided by the fact that, under the usual practice, claims are denied for medical inappropriateness only if they are reviewed by a physician consultant who is a licensed practitioner exercising his own independent medical judgment.(129) Most of these physicians are compensated on an hourly, salaried, or piecework basis that does not reward the number of denials.(130)

c. Market Discipline, Professional Standards, and Judicial Review

Even for those financially self-interested review decisions made in-house by risk-bearing insurers, insurance purchasers have three additional safeguards that help allay conflict of interest concerns. One safeguard exists through the marketplace, the second through the professionally oriented standard that governs insurers' decisions, and the third through the courts.(131)

The present health insurance market is highly competitive.(132) It is dominated by large employer group purchasers who bargain aggressively over short-term contracts. The health insurance market contrasts markedly in this regard with life insurance, which is composed more of individual contracts covering a much longer period.(133) Insurers who attempt to make excessive profits will quickly see their clients switch companies or move to self-insurance. Moreover, overly profiteering health insurers are subject to the same labor market discipline that constrains self-insured employers who hire claims administrators.(134) The number of labor strikes attributable to health benefits concerns demonstrates the sensitivity of corporate benefit managers to fair coverage determinations.(135) As a result, insurers are acutely aware that a well-publicized dispute over an inappropriately denied claim might cause them to lose the next renewal of their contract.

A second avenue of protection arises because of the professionally oriented standard governing insurers' coverage determinations. This standard refers primarily to prevailing medical practice and is administered in the final analysis by licensed physicians.(136) Other commentators suggest quasi-conspiratorial explanations of why health insurers uniformly define their coverage in terms that appear to delegate the reimbursement decision to the same medical community that is being paid.(137) But a less sinister, more worthwhile purpose than enriching physicians may be at play. A professional standard is chosen as a means of striking a sensible compromise between the need for flexibility in a health insurance product and the need for some passably objective standard that sets limits to the coverage. Because the universe of possible adverse health conditions is far too numerous and varied to define a precise schedule of payments for specified compensable events, the parties to the insurance contract resort to generic concepts of illness and medical necessity to define coverage.(138) The need to counteract the treating physician's resulting conflict of interest, however, requires a mechanism to police the physician's treatment-ordering decision. One such mechanism is to place the decision in the hands of the insurer. This alignment strategically balances the competing conflicts of interests of the insurer against the doctor by requiring the insurer to justify its denials under a physician-oriented standard.(139)

A final avenue of protection is the courts. No matter how much discretion a contract gives the insurer, courts will never dispense with a level of review entailing at least a showing of minimum rationality and substantial evidence,(140) concepts that are well-articulated in administrative law and constitutional due process jurisprudence.(141) This requirement is imposed by the obligation of good faith and fair dealing that attaches to every contract.(142)

2. Judicial Deference to the Chosen Decisionmaker

Courts have gone beyond this limited standard of review, however, to decide the issue of medical appropriateness independently under a de novo standard of review. Although some judges have perceived the complexity of the choice between the opposing conflicts of interest and therefore have not prohibited insurer oversight,(143) many judges still do not honor the parties' choice of decisionmaker. Instead, they take it upon themselves to decide the question of medical necessity, despite the contract's explicit delegation of this authority to the insurer.(144) Although, judicial resolution appears to be an attractive solution to an intractable choice between imperfect decisionmakers, additional reflection reveals that private parties might view it in their best mutual interests to keep these decisions out of the courts.

a. The Courts' Difficulty with Clinical Questions

The prior discussion has detailed the costs imposed by litigation, both in terms of the process itself and the deterrent effect that it creates.(145) More important, however, is the validity of the outcome; the judicial system lacks the expertise to decide many scientific questions.(146) Because judges and juries are drawn from the lay population, they must rely on expert testimony presented in an adversarial setting. This process tends to distort the evidence presented and, hence, the accuracy of the outcome. In contrast with scientific processes, litigants in the adversarial system carefully choose their witnesses with tactical advantages in mind. Witnesses that are opinionated and dogmatic are favored over those that have a more balanced view of the competing merits.(147) Moreover, because winning, not truth-finding, is the ultimate objective, litigants resort to tactics that actively undermine truthfulness from a scientific perspective, such as exploiting the demeanor of the opposing scientists and launching ad hominem attacks on their personal credibility. These structural problems create serious obstacles to judges and juries diving scientific fact from scientific fiction.(148)

b. Judicial Bias Toward Individual Versus Group Preferences

Even if litigation were costless and decisions rendered by a clinically trained judge with access to all the relevant data, there is another reason why the parties to the insurance contract might not prefer judicial determinations of medical necessity. The courts, when presented with a single case, are likely to be influenced by the dire condition of the patient.(149) Superficially, this perspective seems correct,(150) but further reflection reveals that a broader perspective should be considered. The proper perspective can be illuminated using four interrelated conceptual principles: (1) philosopher John Rawls's "veil of ignorance;" (2) insurability versus medical necessity as the decision norm; (3) the ex ante versus the ex post perspective; and (4) adjudicative versus legislative facts.

Coverage disputes are most appropriately viewed as an insurance-purchasing decision by a pool of subscribers, not a medical-treatment decision made by an individual patient. The denial of coverage does not prevent the doctor from rendering care;(151) it merely determines that, in the insurer's judgment, the subscriber pool has chosen not to pay for the particular treatment. Thus, where the parties leave the scope of coverage undefined, a particular case is rationally decided by asking only what range of treatment options the purchasers would have chosen to insure at the time they signed up, not what treatment they want to receive now that the insurance has been paid for and their illness is manifest.

Moral theory has adopted the metaphor of the "veil of ignorance," introduced by philosopher John Rawls, to determine the proper framework for deriving principles of distributive justice. The core of Rawls's insight is that, when we determine principles of justice, we must be blind to our particular circumstances in life so that "no one is able to design principles to favor his particular condition."(152) In the present context, one who decides health insurance coverage should be blind to the consideration of whether he will personally become ill in the future. Determining medical necessity from the perspective of a sick patient lifts the veil, whereas making this determination from the perspective of a person contemplating an insurance purchase meets the ideal conditions imposed by Rawlsian concepts of justice.

In the ABMT cases, for instance, the relevant question is not whether a terminally ill cancer patient for whom all other therapies have failed would want all possible treatments performed; the relevant question is whether a cross section of healthy subscribers would pay their share of the cost of these treatments to provide for the unlikely event that any one of them were to fall ill.(153) It is not always the preference of purchasers to insure for all health care that provides any benefit, or even for all health care that proves to be cost beneficial.(154) Some very beneficial care may be well worth paying for out-of-pocket at the time of illness but may not be efficient to insure for in advance.

Taking this insurance-purchasing perspective is particularly revealing in the context of the exclusion of "experimental" treatment. Courts facing the application of this term to new therapies for terminally ill patients have ruled that any possibility of success makes the procedure medically appropriate because the patient is sure to die and all other therapies have proven ineffective.(155) But medical appropriateness is not the sole reason for excluding coverage of experimental treatment. There are serious concerns about the insurability of experimental care simply because it constitutes new, unanticipated treatment not reflected in the actuarial data used to price the insurance policy. Although purchasers of health insurance might choose to pay an extra premium to cover rapidly evolving medical technologies, they may also determine that the price is not worth the insurance benefit. The expense of the newest technologies(156) coupled with the uncertainty of their costs, destroys the central efficiency of insurance pools.(157) Therefore, consumers may decide they are better off bearing the risk themselves (even if this means foregoing new treatments altogether) until new technologies are in sufficiently widespread use that their actual effect on the pooled health care risks can be known with greater certainty and incorporated into actuarial projections more precisely.(158)

Courts are institutionally ill-suited to apply this insurability perspective to coverage disputes. By the very nature of the adjudicative process, they are presented with an individual sick patient who is already insured, not a pool of healthy patients deciding what insurance to purchase. Economists, who label these the ex post versus the ex ante perspectives, observe that efficient legal rules and contractual arrangements cannot be generated from the ex post litigation perspective.(159) For instance, in the extreme case of a terminally ill patient, potentially an infinite amount of money is worth even an infinitesimally small chance of recovery because without the treatment the patient will not be around to enjoy whatever is saved by foregoing treatment.(160) Because the ex post perspective imposed by case-specific judicial review distorts the courts' analysis of whether the benefits of treatment outweigh their costs,(161) the parties to the insurance contract have a compelling reason to withdraw their disputes from the judicial arena. As Peter Huber has observed,

[S]table insurance requires unemotional assessment of risk and disbursement of payments, with the temperament of an actuary and a bookkeeper, treating people as statistics. The driving force in the law today is sympathy and emotion in the individual case. Legal rules rooted in a spirit of compulsion, and applied emotionally case by case, are profoundly inimical to insurance.(162)

Legal scholars underscore the institutional limits of case-specific adjudication by using the somewhat more analytical terms of "legislative facts" or "polycentrism." Adjudication is best suited for disputes in which all affected parties can be brought into a single arena and the relevant facts relate to those parties as individuals. Typically, disputes between two parties to a contract are ideal subjects for adjudication. Health insurance disputes differ, however, because the dispute turns upon the broad, societal effects of disseminating a medical technology whose medium-to-long-term costs are borne by a pool of subscribers. When these sorts of issues are presented to public regulators, administrative law scholars have labeled them "legislative" rather than "adjudicative" issues and have advocated the use of non-adversarial processes for their resolution.(163) The same notion is captured by Lon Fuller's concept of "polycentrism," as applied to private disputes. In his seminal article The Forms and Limits of Adjudication,(164) Fuller identified as the principal limitation to adjudication those polycentric disputes whose resolution is akin to building a bridge:(165)

There are rational principles for building bridges of structural steel. But there is no rational principle which states, for example, that the angle between girder A and girder B must always be 45 degrees. This depends on the bridge as a whole. One cannot construct a bridge by conducting successive separate arguments concerning the proper angle for every pair of intersecting girders. One must deal with the whole structure.(166)

The cost/benefit trade-offs inherent in medical technology assessment and the principles of insurability that underlie optimal insurance contract design, partake heavily of these same aspects of polycentrism.(167)

3. The Market Solution

To paraphrase Winston Churchill, insurer review of medical appropriateness may be the worst conceivable coverage determination mechanism, except for all the others. Viewed simply, a potential insurance purchaser is faced with a less-than-utopian choice: a treating physician with an incentive for ordering too much treatment, and a reviewing physician with an incentive to pay for too little. Nevertheless, the purchaser enjoys the protection that neither physician, despite his bias, may operate outside the range of acceptable medical practice.(168) Within that range, the choice of whose decision should control is dictated by economics. One purchaser might prefer a more expensive product that defers absolutely to her doctor, while a second purchaser, recognizing the broad range of medicine that is untested and of uncertain benefit, might decide that the best use of his insurance dollar is to insure against only the least expensive acceptable course of treatment. This can be accomplished only if the courts are willing to entrust the decision to insurers and are willing to give the insurer the benefit of the doubt in the gray area where there is a legitimate difference of opinion on effectiveness.

Because there is no ideal decisionmaker with a pure, undistorted perspective of the scope of health insurance coverage, the courts lack any public policy basis for interfering with the choice selected by the parties to the insurance contract, either by explicit or implicit judicial fiat. This point is perhaps best illustrated by Health Maintenance Organizations (HMOs). HMOs must, by their very nature, integrate the coverage decision with risk-bearing financial constraints. Because an HMO is both a treating institution and an insuring institution, its coverage decisions are necessarily affected by its financial interests. The law does not question the properiety of HMO physicians playing this dual role;(169) indeed, public policy actively promotes this economizing option.(170) Yet it is precisely this option that some existing precedents overtly renounce and that others silently undermine by their strained interpretation of the medical necessity and experimental treatment language common to conventional insurance contracts.

III. SOLVING THE MEDICAL APPROPRIATENESS PROBLEM

We now change the focus from general policy analysis to particular contractual and legal issues. We will address how the principles discussed in Sections I and II can be implemented in actual contract language and how the contractual alternatives are likely to fare under existing legal doctrine and judicial predilections.

A. Possible Contract Revisions

A variety of contracting mechanisms are available to define the limits of health insurance coverage. While all are potentially of some use, this Section will demonstrate that none of them completely eliminates the need for the parties to leave a portion of the covered services unspecified. The best option for resolving disputes in this undefined area is to specify a decisional process as an alternative to litigation. Thus, as in so many other areas of the law, the ultimate solution lies more in the procedure than in the substance of the matter.

1. Specify Excluded Procedures

Lee Newcomer has outlined three contractual mechanisms for coping with the judicial enforcement problems presented by medically unnecessary and experimental exclusions.(171) The first is to list specific exclusions for all services that are deemed to be unsafe, unnecessary, or uninsurable. As he and others recognize, however, this laundry list approach offers only a partial solution. First, there is the hurdle of judicial inconsistency. Although courts sometimes profess their desire for the parties to use this technique, in reality they have shown the same antipathy toward specific service exclusions as they have toward general coverage terms.(172) For instance, in Ponder v. Blue Cross, the court held that an exclusion of "temporomandibular joint syndrome" was expressed in language that was too technical and inconspicuous to be enforceable.(173) Other courts have complained that more general terms are too "ambiguous" to enforce.(174)

In addition, state insurance commissioners sometimes frustrate this drafting technique by insisting that procedure-specific exclusions be qualified by language that covers the excluded procedure whenever it is "medically necessary."(175) This regulatory oversight entirely defeats the purpose of having an explicit exclusion in the first place. Moreover, even where insurance regulators impose no such obstacles, the long lead time involved in redrafting contract forms and submitting revisions for regulatory approval renders it infeasible to undertake frequent revisions of highly detailed policy terms to incorporate up-to-the-minute advances in medical practice and effectiveness information.(176)

Another difficulty with the laundry list approach is the practical problem of including all materially relevant coverage terms in each summary of the policy. Some courts have "rejected . . . out of hand" the argument that language in a group policy's master plan may be incorporated by reference into the summary plan booklet given to the individual members.(177) Another practical problem with the laundry list approach is that, while it may be possible to identify clearly unwarranted procedures or specific applications of useful procedures where care is unwarranted in specific circumstances, no listing can possibly be detailed enough to cover all the permutations that could possibly occur.(178) Health status and patient preferences are simply too varied to practice medicine entirely by cookbook or computer.(179) Even if this were possible, it is impossible to conceive of physicians practicing under multiple, competing cookbooks governing each of the dozens of different insurance plans in which their patients may be enrolled.(180) Therefore, the laundry list technique will inevitably be subject to legal and practical limitations.

2. Specify the Review Criteria

The second contracting technique is to specify the criteria that will be used to assess medical necessity or experimental status. For example, Blue Cross has been criticized by the courts because it does not incorporate into its contracts the more detailed "technology assessment criteria" that it uses to make coverage decisions.(181) Courts have naively urged insurers to specify the "particular threshold of statistical success" that a new treatment must demonstrate to be considered nonexperimental.(182) This is not realistic, however, because most procedures diffuse into medical practice without a clinical trial.(183) Even where effectiveness is precisely quantified, there is no magical "threshold of statistical success" that would be appropriate in all circumstances.(184)

A more realistic alternative is to specify the quality of evidence that is necessary to satisfy the coverage standard, such as data reported in the peer reviewed literature.(185) In addition, as we outline in more detail below,(186) refining a structure first proposed by Kalb,(187) insurance contracts could specify at what level an assessment is to occur-safe and efficacious, cost effective, or cost beneficial. Even this approach is problematic, however, because different reviewers will reach contradictory conclusions due to differences in their data, their methodology, or their frames of reference. For instance, some medical procedures are cost-effective relative to one treatment mode but not to another, or are cost-beneficial according to some criteria of benefit but not according to other criteria.(188) Inevitably, these qualitative assessment criteria, no matter how precisely framed, are subject to interpretational disputes and are susceptible to judicial manipulation in favor of coverage.(189) Specifying the substantive appropriateness criteria more precisely is only a starting point; we must continue to search for additional methods to limit judicial intervention.

3. Specify the Technology Assessments

The third suggested contracting alternative is to incorporate by reference the medical technology assessment decisions made by specified organizations.(190) Numerous governmental and private organizations actively participate in technology assessment.(191) In principle, it would be a simple matter for a contract to state that it will not cover a procedure that is determined by any of these organizations to be unsafe, unproven, ineffective, or outmoded. Alternatively, the contract could state that procedures will be covered if and only if they are provided in conformity with appropriateness indicators articulated by one of these organizations.(192)

Although this technique has greater potential than specifying specific procedures or specific criteria, it also faces obstacles of both judicial enforcement and practicality. Some judges are reluctant to allow insurers to delegate their assessment to organizations that are not parties to the contract, even if they are closely related insurance trade associations.(193) These courts believe that it is unfair to deny the subscribers a chance to confront the ultimate decisionmakers with conflicting evidence and argument,(194) or they reason that incorporating by reference the present and future decisions of outside organizations fails to give the subscriber fair notice of the scope of their coverage.(195)

On a practical level, deference to external technology assessment organizations is not a complete solution. First, this mechanism will not address many existing and new technologies simply because the available resources are too limited to apply formal assessment techniques to the vast array of new and promising medical procedures, devices, and drugs in all their possible uses and combinations, let alone the many thousands of existing treatments that are untested or for which tests are out of date.(196) Second, the number of organizations performing technology assessments combined with the variety of criteria likely to be used suggest the possibility that two organizations performing technology assessments could reach opposite conclusions. This will lead to a debate over whose methodology is superior.

B. A Prototype Coverage Determination Process

The overriding defect in these three contracting options is that they are directed primarily toward the substance of the decision, not the process for making the decision.(197) Regardless of where the parties settle along the spectrum of specificity, from procedure-specific exclusions to generic criteria for appropriateness, an unsatisfactory arbiter of the inevitable disputes could frustrate the design of the contract. Because the parties could never anticipate all of the inventive objections that courts are capable of posing to the wording of substantive coverage terms, and because courts are ill-equipped to resolve medical appropriateness disputes, the only resort is to specify in the contract an alternative mechanism for making unanticipated and unspecified coverage decisions. This Section proposes a new coverage determination mechanism and the following Section analyzes how it is likely to fare in the courts.

A prototype coverage process would contain these specifications:

(1) Particular treatments the parties know in advance they do not wish to cover.

(2) Standards to determine under what circumstances unspecified treatments are covered.

(3) An entity to apply these standards in making prospective assessments of medical treatments. These assessments would be general rulings that apply to all patients.

(4) The same or a different entity to make case-specific applications to particular patients of the specified exclusions in (1), the standards in (2), and the general rulings in (3).

(5) Criteria for determining when sufficient additional information requires the entity to reconsider its general (3) or specific (4) rulings.

(6) The process to be used in (3), (4) and (5), and an agreement that the determinations these processes produce are within the sole discretion of the specified entities and are binding on all parties.

(7) An agreement that the role of the courts is limited to assuring that the processes actually followed are a reasonable interpretation of what the contract specifies, and an agreement that the primary remedy for any defect in process is to have the case properly reconsidered using the correct procedures, not for the court to decide the issue itself.

This skeletal articulation of our proposal obviously leaves many issues to be determined by the parties. We can offer an opinion on some of them in the following sections; but inevitably there will be room for considerable variation in the specifies that individual employers and insurers choose to negotiate.

1. General Coverage Decisions

The most critical question is who should conduct the assessment process noted in (3). Traditionally, an insurance company's medical director makes the final coverage decision when medical necessity is in dispute. It is the medical director who frequently discusses the scope of coverage with the client during contract negotiations and who sets coverage policy in combination with the client and other members of the insurance company. For reasons developed in the previous section, many purchasers will continue to prefer a medical appropriateness review mechanism that is supervised, if not actually conducted, by the insurer. Conflict of interest problems, however, as well as the inefficiency resulting from hundreds of individual medical directors conducting redundant technology assessments, suggests the necessity of a fresh approach. Several commentators have advocated that the insurance industry (either through individual companies or through their various insurance trade associations) fund an external technology assessment organization(198) to perform some of these functions.(199) These proposals are being actively considered by the insurance industry.(200)

For this organization's decisions to solve the contracting problems just surveyed, there must be a tight fit between the wording of its decisions and the wording of the contractual standards specified in subsection (2). This process would be facilitated if the organization announced in advance a taxonomy according to which it will assess each technology. The parties to a particular contract then may choose the particular criteria they wish to govern their coverage decisions from this precise vocabulary of standards. For instance, the organization could undertake to classify each drug, device, or procedure into one of the following categories(201) (arrayed in ascending order of restrictiveness):

A. Unsafe;

B. Safety in doubt;

C. Safe but ineffective, less effective than available alternatives, or effectiveness in doubt;

D. Safe and effective but not cost effective, that is, less costly alternatives achieve equal results;

E. Cost effective (superior results not available elsewhere for lower costs);

F. Cost beneficial, that is, increment in increased effectiveness is worth the cost.(202)

These terms could be qualified by whatever appropriateness indicators or case-specific detail the organization chooses to impose.

To illustrate more concretely, suppose the assessment organization ruled that adult liver transplants are generally medically accepted (for defined levels of liver disfunction, tissue compatibility, age of patient, etc.); they are cost effective only if performed at transplant centers that average at least two dozen procedures a year; and they are not cost beneficial at any location, for patients over 17. An insurance contract that specified coverage at level D (safe and efficacious) would pay for the procedure (if the appropriateness indicators are met) at any location, but one that covered at level E (cost effective) could insist that the patient go to one of the designated transplant centers. An insurer that specified coverage at level F (cost beneficial) would not cover adult liver transplants at all.

This consolidated technology assessment mechanism does not impose a monolithic outcome on all insurance contracts. First, each insurance company or each sector of the insurance industry (Blue Cross, HMOs, or commercial insurers), is free to establish its own external review organization. Second, each contract will specify its own criteria for review, which may vary among categories of medical services.(203) The ultimate substance of the coverage is left to individual contract negotiation. As a result, this proposal is likely to encounter less resistance under the court rulings that preclude insurers from incorporating by reference material into the contract that is not physically distributed to policyholders. The assessment organization's decisions are not incorporated wholesale; rather, this organization is specified as a process for interpreting the contract, and its decisions control only to the extent that they invoke the standard of coverage that is chosen in the contract. This constitutes no more reliance on extra-contractual material than, for example, a contract that refers to trade custom and usage or to the rules of the American Arbitration Association.

In order for this process to work, it will be necessary to continually update technology assessments. As specified in subsection (5) of the prototype, organizations would be required to revise their assessments as new information becomes available. An assessment that a technology is "experimental" at one time must be modified if additional research or clinical findings validate (or repudiate) its effectiveness.(204) Similarly, if information becomes available that conventional therapy is no longer medically appropriate, it would be necessary to reclassify such therapy as outmoded. To avoid disputes over these matters, the contract should specify a standard for when various assessments must be revisited.(205)

2. Case-Specific Decisions

The preceding component of our proposal is concerned principally with the general, prospective assessments contemplated in subsection (3) of the prototype. As subsection (4) specifies, however, a mechanism is also needed for case-specific rulings. Each coverage decision under an insurance contract will require a decision on at least one of the following issues:

1. Whether the patient is eligible and whether the service is specifically excluded as in subsection (1);(206)

2. Whether an existing assessment ruling determined in subsection (3) applies and whether its terms are met;

3. Whether treatment is covered when there is no existing ruling that applies;

4. Whether changed circumstances warrant reconsidering a previous denial for a particular patient.

It is possible for the same entity to make both the rule-oriented decisions and the case-specific applications,(207) but this may not be the preferred choice. Different skills may be required to make general and specific assessments and, to the extent that the industry chooses to consolidate the prospective decisions in one or a few entities, it becomes impractical to funnel case-specific decisions through a central body. This is particularly important when those decisions are based on the terms of different insurance contracts. This approach would also heighten antitrust concerns.(208) Therefore, it may be necessary that case-specific decisions continue to be made as they currently are, either through claims processing and medical director review internal to the insurer or through utilization review services contracted to independent organizations.

Another option is to employ alternative dispute resolution mechanisms such as arbitration or formal grievance procedures,(209) or more democratic, electoral processes (such as those common in organized labor governance committees which have representatives from both insurers and subscribers).(210) Democratic dispute resolution is particularly attractive because it mitigates the conflict of interest of the insurer or self-insured employer at the same time that it preserves the group perspective of a pool of insured subscribers.

Whichever case-specific review mechanism the parties choose, it should give the patient recourse to protest a denial of coverage, and it should require one or more stages of physician review before the denial becomes final.(211) Again, precisely how this is accomplished, and with what safeguards, depends on the specifies of an individual contract, as shaped by requirements imposed by judicial review.(212)

C. The Judicial Standard of Review

The level of scrutiny the courts will apply to these decisions depends on what model of judicial review courts adopt. Three somewhat competing and somewhat overlapping models are available to draw upon: trust law, administrative and constitutional law, and binding arbitration.

1. The Trust Law Model

Decisions under ERISA invoke a trust law model to review the decisions of health plan administrators.(213) This private law model, however, is not applicable to the problems of insurance coverage because it is premised on a trustee administering benefits to a single, or small number, of beneficiaries.(214) Therefore, trust law does not offer much guidance on how to resolve conflicts among the beneficiaries or between one beneficiary and the group.(215) For instance, trust law, taken literally, would not allow a risk-bearing insurer to have any role in claims administration because trust law does not merely scrutinize conflicts of interest, it bars them altogether as a preventive measure to avoid any possibility that the trustee's loyalty to a beneficiary will be diverted by other considerations.(216) The very nature of administering a pool of benefits for a number of beneficiaries requires just the opposite: the trustee must at some level pursue the interests of the group (which a risk-bearing insurer or employer represents) over the individual claimant.(217)

Fortunately, the trust law model is capable of rehabilitation by reference to public law doctrines developed in administrative and constitutional law. These doctrines are much more attuned to these conflicts because they deal explicitly with the underlying distributional issues. The Supreme Court has shown itself amenable to this approach by refraining from taking trust law to its literal extreme; it has ruled that an insurer's conflict of interest is not disabling but merely becomes one factor to consider in applying an "arbitrary and capricious" or "abuse of discretion" standard of review.(218) By no small coincidence, these are terms of art well established in administrative law. Thus, rather than reformulating trust law to fit the needs of an insurance dispute, the courts would be better advised to consider explicitly an administrative law model. A shift to established administrative law would provide more guidance than searching in vain through trust law precedents or inventing new doctrine whole cloth, as some courts have attempted unsatisfactorily to do.(219)

2. The Administrative Law Model

An administrative law model for reviewing private insurers' medical appropriateness decisions provides two helpful guideposts: (1) an arbitrary and capricious standard of review to determine matters of substance, and (2) principles of constitutional due process to determine matters of procedure.

a. Arbitrary and Capricious Standard of Review of the Substance of Medical Appropriateness Decisions

The administrative model would view private insurer medical appropriateness decisions through the lens of a governmental agency vested with quasi-legislative and quasi-adjudicative authority. The entity that makes the general assessments is treated, in essence, as a private legislature "elected" by the parties to the contract. The entity that makes the case-specific applications of these coverage policies is seen as a private tribunal. Under this model, the role of the courts is limited to the appellate level of review that ensures only that the decision is not arbitrary and capricious and that it is supported by substantial evidence. These review standards require the court to give insurers the same level of respect that is given to trial courts and government agencies: the insurer's decision must be affirmed even if the court disagrees with it, so long as it is not irrational.

This standard would apply regardless of whether the insurer itself makes the assessment or delegates it to an external entity. Although the external entity may be less biased, courts should recognize that parties may actually desire the financial incentives that the insurer brings to bear on the question of appropriateness.(220) Therefore, as long as the insurer has made the decision following the procedures and standards specified by the purchaser, the courts should presume that any perceived conflict was intentional and therefore does not justify heightening the level of review.(221)

The arbitrary and capricious standard is well suited to the subject matter of determining whether a medical appropriateness decision falls within the wide range of established practice patterns.(222) Because the arbitrary and capricious standard acknowledges the possibility of more than one correct answer, it allows courts to review complex decisions without demanding an unattainable degree of certitude.(223) Therefore, it more accurately reflects the reality of clinical practice.

Another advantage of the arbitrary and capricious standard is that it alleviates the uncertainty effects of litigation that impose substantial costs on the parties to the insurance contract. Applied correctly, this standard of review requires the courts to limit themselves to the record of evidence that was before the insurer at the time it decided rather than to hear testimony independently.(224) This elimination of the need to re-prove each decision in court should enable many challenges to be decided on summary judgment since compliance with this appellate review standard is a question of law that can be determined favorably to either party based purely on a documentary record.(225)

The following decisions illustrate the operation of the arbitrary and capricious standard properly applied. In Jones v. Laborers Health & Welfare Trust Fund,(226) the court deferred to the decision by a board composed of both management and union employees that hyperthermia treatment for breast cancer is not effective, despite conflicting evidence from the woman's physician.(227) In Johnson v. District 2 Marine Engineers Beneficial Ass'n,(228) the court found no error in denying coverage for a liver transplant, observing that "we must resolve the ambiguity [of the term "experimental"] in favor of the trustees' interpretation if it is reasonable."(229) In contrast, the court in Schroeder v. Blue Cross & Blue Shield United,(230) ruled that the insurer could not discontinue paying for daily home care services for an extremely debilitated elderly couple for whom the services were necessary to keep them out of a nursing home.(231) Although coverage was debatable, Blue Cross had paid for the services in the past and, when the contract was renewed, the couple had specifically rejected an amended contract and had agreed to a fifty percent increase in their premium to remain at home. The court properly found that the prior course of dealing and the circumstances surrounding the amendment foreclosed Blue Cross from exercising its own independent discretion in determining medical necessity.(232)

For this vision of judicial review to work, the parties may need to undertake several contractual adjustments to respond to various idiosyncracies of applying the arbitrary and capricious standard to medical decisions. One adjustment relates to the scope of the issues that are typically subjected to arbitrary and capricious review. Under administrative law, this deferential review applies only to factual and policy questions, not to questions of law.(233) Thus, in the governmental framework, the courts are to construe de novo the meaning of the statutory terminology that governs the agency.(234) The private analogue is that the meaning of the contract is a question of law for the court to determine de novo.(235) The Supreme Court has clarified, however, that where the statute commits its own construction to the discretion of the agency, courts are to give heavy deference to the agency in determining questions of law.(236)

On its surface, this statutory model has a perfect analogue in the construction of insurance contracts that explicitly give the insurer the discretion to interpret "medical necessity" and "experimental." Indeed, it is precisely such language that the Supreme Court has recently held to be critical to invoking the arbitrary and capricious standard of review under ERISA, which governs all insurance disputes arising out of the workplace.(237) This simple schematic--de novo review except where the contract specifies deference to the insurer--is clouded, however, by two complicating factors. First, whether the contract in fact gives the insurer this degree of discretion is a matter of interpretation that the courts take upon themselves, with conflicting results.(238) Second, the Supreme Court has introduced in the ERISA cases the complication that the degree of deference is to be modulated where the insurer has a conflict of interest.(239) Precisely what this means is a matter of considerable confusion in the lower courts, but one decision has gone so far as to suggest that the basic conflict inherent in an insurer's retention of risk requires the invocation of de novo review in all cases.(240) It is for this reason that we recommended the assessment entity be independent of the insurance company (or, for self-insured companies, of the employer).

A related complication of the arbitrary and capricious model is the recognition in administrative law that courts apply this standard with varying levels of scrutiny, according to their degree of suspicion of agency motives and expertise.(241) Scholars have labeled as "hard look review" the standard that courts employ when they intensify their review of agency decisions beyond their usual deference.(242) This exercise of the judicial prerogative inherent in the subjectivity of the terms arbitrary and capricious is not disabling, however, if it is used sparingly and is invoked explicitly rather than tacitly.(243) However, the prior cycles of consistently extreme rulings followed by ineffective contract revisions counsel much greater judicial restraint in selecting those cases that deserve more scrutiny.(244)

b. Due Process and Minimally Acceptable Procedures for Medical Appropriateness Determinations

The arbitrary and capricious standard is limited, however, by its application principally to matters of substance, not procedure. Procedural questions are generally considered by administrative law to be subject to de novo review under varying process standards. This allows the courts considerable leeway to frustrate--through intensive scrutiny of procedural fairness--those decisions they disagree with based on substance, but which the arbitrary and capricious standard prevents them from reversing outright. It has been noted that the most common exercise of "hard look" review is for matters of procedure rather than substance,(245) a phenomenon well-known in insurance coverage disputes. There are two popular techniques. One, which is illustrated in the ABMT cases, is to define the scope of the issue so narrowly that current technology assessments cannot possibly answer the question.(246) The absence of a technology assessment for patients with the plaintiff's precise condition, in the same age group, with the same sex, and the same medical history leaves the court free to conduct its own assessment or to allow the patient to win by default. The second technique is to find that the insurer's investigation was incomplete. In two notorious cases, courts have awarded large punitive damages against insurers who denied small claims based on incomplete medical histories.(247) In one decision, the court ruled that "it was Aetna's responsibility to marshal all of the medical facts ... before its refusal to pay,"(248) an impossibility in the very limited time available for prospective utilization review. In the second case, the court reached the same decision despite the insurer's repeated requests to the physician to forward all pertinent information.(249)

This weakness in the arbitrary and capricious standard can potentially be corrected either by the parties specifying in great detail the process to be followed in determining medical appropriateness and the remedies for breach of this process, or in the absence of such specificity, by the courts adopting an explicitly process-based model for judicial review. A constitutional due process model is the most suitable candidate for resolving process-based disputes surrounding medical appropriateness assessments. This model serves two purposes: it provides a framework for determining fair procedure when the parties fail to specify one in their contract, and it sets a minimum floor of procedural fairness against which the contract is to be tested for public policy and fundamental fairness concerns when it does specify a procedural mechanism.

Most conventional indemnity health insurance contracts contain no specification of how insurers are to go about assessing medical appropriateness.(250) They do not reveal who is to decide, what records are to be consulted, and what rights the patient has to object and respond to the initial decision. Even if parties were to address these questions more explicitly, as is suggested above, it remains impractical to craft a procedural code that will anticipate every nuance and procedural nicety that courts could require. Inevitably, then, the contract will leave a procedural void,(251) and the courts will quickly fill the empty space with additional requirements that insurers must fulfill.

Without a set of concrete principles that bind the courts' discretion in imposing these procedural hurdles, this game can be played endlessly. An established reference point is readily found in the existing body of constitutional due process jurisprudence,(252) particularly those cases relating to disability income and health care benefits under the Social Security Act. A full exegesis of these principles is beyond the scope of this Article. A summary reveals, however, a number of precedents well suited to the present subject matter. Disability income benefits may be denied without providing a formal hearing prior to termination.(253) In the proceedings following termination or denial of disability income, Medicaid, or Medicare benefits, the federal courts have held that it is constitutionally sufficient to employ an inquisitorial rather than adversarial process,(254) to consult a physician retained by the agency who has not personally examined the patient,(255) to limit the patient's personal presentation of evidence to a documentary record coupled with a toll-free phone line,(256) to have different levels of review for different amounts in controversy,(257) and to preclude judicial review altogether.(258)

Moreover, where the decision in question is a general rule applicable to an entire class of patients rather than specific to an individual case, due process does not require any hearing rights at all for the affected individuals because their interests are reflected through their representation in electoral, quasi-legislative processes.(259) Finally, courts impose few constraints on the highly discretionary decision of whether to employ a quasi-legislative or a quasi-adjudicative process, that is, whether to decide through general rules or through case-specific analysis.(260) By no means do these rulings constitute an ideal procedural code, but at least they set a baseline of reasonableness or public policy acceptability against which insurance contracts can be judged. If these procedures suffice for governmental deprivation of constitutionally protected property and liberty, surely they suffice as the minimum requirements for private contracts.

In applying these principles of "flexible due process," it will be necessary for courts to recognize that they are not free to impose a "gold standard" of ideal investigation and decision, as they have done in the past.(261) Invoking the constitutional due process standard connotes a minimum standard of fairness. Courts cannot continue to object that the reviewing physician did not have the precise qualifications of the patient's physician(262) or that the insurer did not investigate every nuance.(263) Giving the patient and his physician an opportunity to make their case, and requiring that there be some evidence to support the insurer's final decision, ensures that whatever procedure the parties in fact choose to follow will comfort with minimal fairness.(264) Beyond this, the courts should not "upset the review process [for which] the parties have bargained."(265)

A critical component of the due process model is the explicit recognition that the proper remedy for a procedural defect is reconsideration by the insurer, not automatic coverage plus punitive damages. To date, courts systematically have denied to insurers the same opportunity that government agencies, and, indeed, the courts themselves enjoy to correct their procedural errors. Not only have judges seized on purely procedural flaws to impose decisions on the merits, they have used these flaws to extraordinarily amplify damages under bad faith law. An explicit contractual provision that specifies the remedy for procedural breaches, as well as a due process model for judicial review, should help to restrain the worst excesses of this judicial activism.(266)

The most difficult of the due process requirements for insurers to fully satisfy will be the requirement of an unbiased decisionmaker, for, as discussed above, the risk-bearing insurer harbors an inherent economic conflict of interest.(267) A return to the administrative law analogy will inform the courts how this conflict should be analyzed. Judicial analysis of potential bias in agency decisionmakers has determined that conflicting incentives at the institutional level are not disabling, for most agencies pursue conflicting tasks within their broad ambit of responsibility.(268) For instance in adjudicatory proceedings, the agency as an institution acts as both prosecutor and judge, a blatant compromise of adversarial justice. Nevertheless, courts have accepted these conflicting roles as long as the individual decisionmakers are not themselves afflicted with obvious, actual bias.(269) Likewise, in the insurance context, if the actual reviewing physicians who determine medical necessity decide according to their independent medical judgment and are not explicitly compensated for refusing coverage, their judgment should be viewed as sufficiently insulated to pass at least bare minimum requirements of procedural fairness.(270)

3. The Binding Arbitration Model

Despite the theoretical attractiveness of an administrative/constitutional law model, this approach to resolving health care coverage disputes may prove too burdensome and uncertain to be attractive to contracting parties. This model still follows an adversarial process, which might be viewed as ill-suited to the matter at hand, and it requires intricate contractual provisions that have not been tested in the courts. Therefore, the parties might view an entirely nonjudicial model of dispute resolution as the most attractive.

Several such models are available,(271) but the one deserving the most attention is binding arbitration. It has received the most experience in the health insurance context.(272) The contract could specify a small number of public representatives and physicians from the community to form a permanent arbitration panel.(273) In cases where the applicability of the general assessments is in doubt, or when no general assessment has been performed, these arbitration panels would review the information in the particular case, examine the insurance contract to assess what type of coverage was intended, and review the available clinical information before making a coverage determination. If additional technical expertise is required, the panel would be permitted to hire outside consultants to advise them on the clinical facts involved. The clinical and coverage determinations of the panels would be binding on both parties.

A binding arbitration model may be a superior way for the parties to express their desire for non-judicial resolution because this language is more likely to place the courts in a mindset that is aware of the public policies in favor of private, contractual remedies. Therefore, an arbitration model may provide the greatest degree of deference for both the substance and the procedure of coverage determinations, whether of a quasi-legislative or a quasi-judicial nature.(274) Moreover, depending on the composition of the resolution mechanism, this may be the most informed and least self-interested mechanism, and it may be the one that is most oriented toward the ex ante, insurance-purchasing perspective on the coverage issues.(275)

CONCLUSION

Health insurance coverage disputes are subject to a complex interplay among courts, insurers, physicians, and patients. Courts, out of deference to treating physicians, are refusing to respect the mechanism the parties have chosen to define the scope of coverage, forcing them to contract in ways they prefer not to, and even then refusing to enforce the provisions other courts have imposed. Two forms of market failure result: pricing purchasers out of the market altogether, or forcing them to buy more expensive insurance products than they desire. We have proposed a mechanism that we hope will resolve this impasse: to contract explicitly for a process of resolving disputes over medical appropriateness rather than to further define the substance of the criteria applied by the courts.

We do not naively assume that this process by itself can solve all the vexing allocation problems that confront our nation's policymakers. We also recognize that case-specific utilization review ultimately may not prove to be the most effective cost containment tool.(276) But the courts are not the proper institution for deciding which type of managed care-utilization review, provider selectivity, financial incentives directed to patients and physicians, or a combination thereof--is best. The proper role of the courts is not to insert themselves into the center of this debate, but to stand on the sidelines to referee the processes that are followed, with an objective, non-result-oriented view of what the parties have agreed to within prevailing ethical, social, and economic constraints.

(1) No. 90-597, 1990 U.S. Dist. LEXIS 5376 (D.N.J. Mar. 22, 1990).

(2) Id. at *1-*2.

(3) ABMT is a therapy that allows much higher doses of chemotherapy and radiation to be administered to a cancer patient than is ordinarily possible because of their toxic effects on bone marrow, which supports the immune system. ABMT involves temporarily removing and freezing a portion of the patient's own bone marrow (in contrast with allogenic donation by family members, or unrelated donation, for which there are greater problems of donor availability and patient/ donor matching). The marrow is then reinfused after the high dose chemotherapy or radiation is administered. ABMT has emerged over the past few years as a lastgasp measure to stop cancers that have not responded to more conventional therapies or that have "metastasized," that is, entered the blood or lymph system and thus threatened to cause damage far beyond their initial location. See id. at *3-*5.

The most frequently litigated application of ABMT has been for metastatic breast cancer. It is a dangerous, painful, and expensive therapy whose success rate varies according to the particular cancer involved and its precise stage, as well as by the treatment protocol followed. The therapy itself may cause the patient's death in a long-term success has not yet been demonstrated for many cancers. See id. at *1 n.1; see also Bucci v. Blue Cross-Blue Shield, 764 F. Supp. 728, 730 (D. Conn. 1991); Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 588 (E.D. Va. 1990); Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131, 133 (D.N.J. 1989); Bruce E. Hillner et al., Efficacy and Cost-Effectiveness of Autologous Bone Marrow Transplantation in Metastatic Breast Cancer, 267 JAMA 2055 (1992); Perry C. Panantonis, Comment, Experiemental Exclusions: Are Insurance Companies Really Protected?, 9 N.Y.L. SCH. J. HUM. RTS. 217, 219-20 (1991); Robert Bazell, Topic of Cancer: Autologous Bone Marrow Transplant Policy, NEW REPUBLIC, Dec. 31, 1990, at 9.

The debate over the efficacy of ABMT is epitomized by conflicting medical reports about the health of former presidential candidate Paul Tsongas. At first, his doctors reported that ABMT successfully treated his lymphoma (cancer of the immune system). See Lawrence K. Altman, Doctors Say Cancer Therapy has Tsongas in Good Health, N.Y. TIMES, Mar. 5, 1992, at A1. More recently, however, his doctors conceded a relapse following the ABMT treatment, which other doctors interpret as meaning that the ABMT treatment did not effect a cure. See Lawrence K. Altman, Doctors Now Say that Tsongas Suffered a Recurrence of Cancer, N.Y. TIMES, Apr.22, 1992, at A1.

(4) We will use the term "medical appropriateness" to refer generically to medical necessity, accepted medical practice, cost effective care, cost beneficial care, and the host of other verbal permutations that describe a normative assessment of what medical treatment should be performed. Sometimes, we use "medical necessity" interchangeably to connote this same generic concept of appropriateness because it is the term that most health insurance contracts adopt. Later in the Article, we will suggest the adoption of more precise language to describe assessing appropriateness at different levels. See infra text accompanying notes 201-02.

(5) See U.S. BIPARTISAN COMM'N ON COMPREHENSIVE HEALTH CARE (THE PEPPER COMMISSION), A CALL FOR ACTION: FINAL REPORT 41 (1990) [hereinafter PEPPER COMMISSION REPORT] (estimating that only 10-20% of the medical procedures used today have been subjected to randomized clinical trials--the most conclusive method of determining if a procedure is medically effective). A recent study found, for example, that a medical procedure used for the past 50 years has no statistically significant clinical value. See Wallace J. Epstein et al., Effect of Parenterally Administered Gold Therapy on the Course of Adult Rheumatoid Arthritis, 114 ANNALS INTERNAL MED. 437 (1991).

(6) See infra text accompanying notes 96 & 112.

(7) Victor Fuchs was one of the first economists to suggest that excessive use of sophisticated medical technology may do little to improve overall health status. See VICTOR FUCHS, WHO SHALL LIVE? HEALTH, ECONOMICS, AND SOCIAL CHOICES 94-95 (1974).

(8) As many as 35 million Americans do not have health insurance. See M. Eugene Moyer, A Revised Look at the Number of Uninsured Americans, HEALTH AFFAIRS, Summer 1989, at 102, 102; see also PEPPER COMMISSION REPORT, supra note 5, at 22. The Pepper Commission's and others' proposals for universal access are hampered by the political infeasibility of committing more public funding resources.

(9) Only 60% of American Indian, Mexican-American, African-American, and Puerto Rican mothers receive prenatal care in the first trimester of pregnancy compared to 80% of White, Cuban, and Asian mothers. See NATIONAL CTR. FOR HEALTH STATISTICS, HEALTH, UNITED STATES, 1990, at 9 (1990). Moreover, infant mortality rates (the proportion of deaths in the first year of life) are 50% higher for American Indians and 40% higher for Puerto Ricans than for non-Hispanic whites. See id. at 11.

(10) The Pepper Commission developed a proposal for long-term-care insurance whose fully implemented cost is $42.8 billion. See PEPPER COMMISSION REPORT, supra note 5, at 16. Congress has deferred action on the proposal until a funding source can be found.

(11) See Rollo, 1990 U.S. Dist. LEXIS 5376, at *2-*3 ("Tishna's parents, who knew that they had medical insurance and quite understandably believed that when their daughter became ill Blue Cross/Blue Shield would pay for the treatment of illness, should not have had to come to this court.").

(12) See Adams v. Blue Cross & Blue Shield, 757 F. Supp. 661 (D. Md. 1991); Bucci v. Blue Cross & Blue Shield, 764 F. Supp. 728 (D. Conn. 1991); White v. Caterpillar, 765 F. Supp. 1418 (W.D.Mo. 1991); Kulakowski v. Rochester Hospital Service Corp., No. 91-CV6505T (W.D.N.Y. Dec. 17, 1991); Clark v. K-Mart Corporation, No. 91-1431 (W.D. Pa. Sept. 11, 1991); Reiff v. Blue Cross & Blue Shield, No. 90-C-1030-E (N.D. Okla. Feb. 11, 1991); Cole v. Blue Cross & Blue Shield, 738 F. Supp. 42 (D. Mass. 1990); Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586 (E.D. Va. 1990); Stewart v. Hewlett-Packard Co., No. 90-875-A (E.D. Va. Aug. 7, 1990); Thomas v. Blue Cross & Blue Shield, No. 90-10831-H (D. Mass. Apr. 5, 1990); Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131 (D.N.J. 1989); Miller v. Blue Cross & Blue Shield, No. 91-E-411-B B (N.H. Super. Ct. Oct. 2, 1991); Terninko v. Blue Cross & Shield, no. 91-E-517- (N.H.Super. Ct. Sept. 24, 1991); Bradley v. Empire Blue Cross & Blue Shield, 562 N.Y.S.2d 908 (N.Y. Sup. Ct. 1990); Speed v. Prudential Health Care Plan, No. 90-1470 (Va. Ch. Nov. 19, 1990). As of this writing, three dozen additional ABMT cases have been settled or are pending in the federal courts. See BLUE CROSS AND BLUE SHIELD ASS'N, INVESTIGATIONAL EXCLUSIONS: A SUMMARY OF THE CASE LAW at VI-1 to VI-8 (1992).

For other reviews of this body of litigation, see Frank P. James, The Experimental Treatment Exclusion Clause: A Tool for Silent Rationing of Health Care?, 12 J. LEGAL MED. 359 (1991) (proposing a clearer definition of "experimental"); Paul J. Molino, Reimbursement Disputes Involving Experimental Medical Treatment, 24 J. HEALTH & HOSP. L. 329 (1991) (same); Jennifer Belk, Comment, Undefined Experimental Treatment Exclusions in Health Insurance Contracts: A Proposal for Judicial Response, 66 U. WASH. L. REV. 809 (1991) (supporting these rulings); Panantonis, supra note 3 (criticizing these decisions). See generally Julia F. Costich, Note, Denial of Coverage for "Experimental" Medical Procedures: The Problem of De Novo Review under ERISA, 79 KY. L.J. 801 (1991) (discussing ABMT along with other experimental treatments).

(13) See Harris v. Blue Cross & Blue Shield, 729 F. Supp. 49 (N.D. Tex. 1991); Schnitker v. Blue Cross & Blue Shield, No. CV-91-0-412 (D. Neb. Nov. 11, 1991); Farley v. Benefit Trust Life Ins. Co., No. 90-761 (E.D. Mo. Oct. 17, 1991); West v. Blue Cross & Blue Shield, No. 91-582 (D. Or. Oct. 1, 1991); Holder v. Prudential, No. W-89-CA 172 (W.D. Tex. Nov. 28, 1990); Sweeney v. Gerber Prod. Co. Medical Benefits Plan, 728 F. Supp. 594 (D. Neb. 1989); Thomas v. Gulf Health Plan, 688 F. Supp. 590 (S.D. Ala. 1988); Green Hosp. v. United States, 23 Cl. Ct. 393 (1991); Evans v. HMO Colorado, No. 91CV3797 (Colo. Dist. Ct. June 14, 1991); Whittle v. Blue Cross/Blue Shield, No. 87-CP-15-625 (S.C. Ct. C.P. Colleton Country Jan. 17, 1990); Hurowitz vs. Blue Cross & Blue Shield, No. N-7849-4 (Va. Cir. Ct. May 19, 1989) (transcript of hearing on motion for preliminary injunction).

The position that ABMT is still experimental may have been strengthened by a recent National Institutes of Health decision to conduct a randomized clinical trial of women with metastatic breast cancer seeking ABMT; randomized clinical trials can be conducted only if the medical effectiveness of the procedure is in doubt. See Curt Suplee, Blue Cross Agrees to Fund Breast Cancer Experiment: Women to Undergo Bone Marrow Transplants, WASH. POST, Nov. 13, 1990, at A1

(14) See Lee N. Newcomer, Defining Experimental Therapy-A Third-Party Payer's Dilemma, 323 NEW ENG. J. MED. 1702 (1990); Franklin M. Zweig & Seymour Perry, Health Care Goes to Court: Judges Need Access to Impartial Medical Expertise, WASH. POST, July 17, 1990, Health Section, at 6.

(15) See infra text accompanying notes 20-44. This case law is summarized in JOHN A. APPLEMAN, INSURANCE LAW AND PRACTICE subsection 705.35, 709.75 (1981 & Supp. 1991); Alan Bloom, Interpretation of Insurance Policy Coverage in the Case of Ambiguity--Or How Big is the Consumer's Piece of the Rock?, 3 WHITTIER L. REV. 177 (1981); James S. Cline & Keith A. Rosten, The Effect of Policy Language on the Containment of Health Care Cost, 21 TORT & INS. L.J. 120 (1985); James S. Cline & Keith A. Rosten, The Effect of Policy Language on the Containment of Health Care Cost: A Footnote, 21 TORT & INS. L.J. 653 (1986); Paul E. Kalb, Controlling Health Care Costs by Controlling Technology: A Private Contractual Approach, 99 YALE L.J. 1109 (1989); Grace P. Monaco & Rebecca L. Burke, Insurer as Gatekeeper: Handling Claims for Unproven Methods of Medical Management, 18 FORUM 591 (1983); Grace P. Monaco & Rebecca L. Burke, Insurer as Gatekeeper-Part Two: Policy Obstacles in Unproven Methods of Litigation, 20 FORUM 400 (1985) [hereinafter Monaco & Burke, Gatekeeper Part Two]; Annotation, What Services, Equipment, or Supplies are "Medically Necessary" for Purposes of Coverage Under Medical Insurance, 75 A.L.R. 4TH 763 (1990).

(16) The only other area of serious dispute has been major organ transplantation (heart, lung, liver, pancreas). See, e.g., Johnson v. District 2 Marine Eng'rs Beneficial Ass'n, 857 F.2d 514 (9th Cir. 1988) (upholding denial of coverage for liver transplant). Most private insurers pay for such procedures, but when they do not, these coverage denials usually do not produce litigation because they are based on explicit contractual exclusions of the particular procedures rather than on general "medically unnecessary" or "experimental" exclusions. See Thomas Musco, Insurance Coverage for Organ and Tissue Transplants in HEALTH INSURANCE ASSOCIATION OF AMERICA RESEARCH BULLETIN R1489 (Nov. 1989); Joel E. Miller, The Private Insurer Response to Advanced Health Care Technology: The Case of Organ Transplants, in PEDIATRIC BRAIN DEATH AND ORGAN/TISSUE RETRIEVAL: MEDICAL, ETHICAL, AND LEGAL ASPECTS 331 (Howard H. Kaufman ed., 1990).

(17) The federal government's interest in assessing medical appropriateness is demonstrated by the creation of a new agency within the Public Health Service, the Agency for Health Care Policy and Research, whose mandate is to determine and then promote effective medical care. In 1991, approximately $60 million was allocated to support research teams to develop guidelines for appropriate medical practice. See HSR REPORTS (Ass'n. of Health Servs. Res.), Mar. 1991, at 3. These guidelines are intended to influence Medicare and Medicaid coverage decisions as well as physicians' practice patterns. See 42 U.S.C. subsection 299-299c (Supp. 1989).

(18) For examples of such articles, see those cited supra note 15.

(19) For a sampling, see AMA, LEGAL IMPLICATIONS OF PRACTICE PARAMETERS (1990); John D. Blum, An Analysis of Legal Liability in Health Care Utilization Review and Case Management, 26 HOUS. L. REV. 191 (1989); Sandra J. Byrnes, Corporation's Institution of Health Care Utilization Review: Legal Risks, 33 MED. TRIAL TECH. Q. 478 (1987); Clark C. Havighurst, Applying Antitrust Law to Collaboration in the Production of Information: The Case of Medical Technology Assessment, LAW & CONTEMP. PROBS., Spring 1983, at 341; William A. Helvestine, Legal Implications of Utilization Review, in CONTROLLING COSTS AND CHANGING PATIENT CARE? THE ROLE OF UTILIZATION MANAGEMENT 169 (Marilyn J. Field & Bradford H. Gray eds., 1989); Nathan Hershey, Fourth-Party Audit Organizations: Practical and Legal Considerations, 14 LAW MED. & HEALTH CARE 54 (1986); Martin Rose & Robert F. Leibenluft, Antitrust Implications of Medical Technology Assessment, 314 NEW ENG. J. MED. 1490 (1986).

Not only are the malpractice and antitrust issues secondary to whether insurers may engage in medical necessity review to begin with, but the legal barriers they construct are less imposing ones. Most analysts agree that, as long as insurers act independently in making their determinations, antitrust considerations are of little consequence unless a particular insurer enjoys strong market power. See Rose & Leibenluft, supra, at 1493; Havighurst, supra, at 364-69. Malpractice considerations are potentially of greater concern, except for the fact that much of what falls under this tort-law rubric in ordinary patient-physician treatment relationships is displaced by contract law, given the extensive contractual underpinnings of the patient-insurer relationship. Thus, the most significant malpractice case to date, Wickline v. State, 239 Cal. Rptr. 810 (Cal. Ct. App. 1986), concerned a governmental review mechanism that had no contractual status. The court held that California was not liable in damages resulting from the denial of coverage to a Medicaid recipient. When the same court confronted a private review mechanism, it distinguished Wickline, observing that the contract between the parties governed the insurer's responsibilities. See Wilson v. Blue Cross, 271 Cal. Rptr. 876, 879 (Cal. Ct. App. 1988), review denied, No. SO17315, 1990 Cal. LEXIS 4574 (Oct. 11, 1990). Although the court, as a consequence, ruled that Blue Cross was subject to greater liability than Medi-Cal (the California Medicaid agency), this holding was specific to the factual circumstances of Wilson, which included a contract that contained no clear authorization for Blue Cross to conduct precertification review. Since most contracts will contain this explicit authorization, it is a certainty that in future cases contract law will have far more immediate application to the respective rights and responsibilities than tort law.

(20) According to a classic book on the health insurance industry written three decades ago:

[T]here is little evidence that most of the industry has any real interest in developing effective controls over the costs of medical care.... The links between Blue Cross and the hospitals, and Blue Shield and the medical societies, are well known. Less visible but, in some subtle ways, even more powerful is the long-standing alliance between the [commercial] insurance industry and organized medicine. The [commercial] insurance companies are highly reluctant--on legal, historical, and political grounds--to assume any such responsibility. HERMAN M. SOMERS & ANNE R. SOMERS, DOCTORS, PATIENTS AND HEALTH INSURANCE 414-15 (1961); see also PAUL STARR, THE SOCIAL TRANSFORMATION OF AMERICAN MEDICINE 290-300, 306-10, 327-34 (1982) (describing the establishment of Blue Cross by hospitals and Blue Shield by physicians). For descriptions of the gradual adoption of cost control through coverage restriction by Blue Cross/Blue Shield and commercial insurers, see SYLVIA LAW, BLUE CROSS: WHAT WENT WRONG? 59-63, 93-102 (1974); Barbara Greenberg & Robert A. Derzon, Determining Health Insurance Coverage of Technology: Problems and Options, 19 MED. CARE 967, 971-73 (1981); Clark C. Havighurst, Professional Restraints on Innovation in Health Care Financing, 1978 DUKE L.J. 303, 338; Fred J. Hellinger, Controlling Costs by Adjusting Payment for Medical Technologies, 19 INQUIRY 34, 38-41 (1982).

Because Blue Cross/Blue Shield and commercial insurance providers administer claims payment under Medicare, whose terms of coverage mirror those in standard insurance contracts, the same attitude carries over into the realm of public insurance.

(21) See, e.g., Mount Sinai Hosp. v. Zorek, 271 N.Y.S.2d 1012 (N.Y. Civ. Ct. 1966) (involving such a contract). Surprisingly, this language still persists in some contemporary contracts. See, e.g., McLaughlin v. Connecticut Gen. Life Ins. Co., 565 F. Supp. 434 (N.D. Cal. 1983); Schroeder v. Blue Cross & Blue Shield, 450 N.W.2d 470 (Wis. Ct. App. 1989).

(22) See Note, Controlling Health Care Costs Through Commercial Insurance Companies, 1978 DUKE L.J. 728.

(23) According to one account, Blue Cross's efforts to require physicians merely to certify medical necessity and length of stay in order crack down on fraud were considered "drastic measures" and "desperate moves." SOMERS & SOMERS, supra note 20, at 416.

(24) For instances, insurers were forced to pay for treatment in the following cases: Aetna Life Ins. Co. v. Sanders, 193 S.E.2d 173 (Ga. Ct. App. 1972) (intestinal bypass surgery for refractory obesity); Myerson v. Associated Hosp. Serv., 314 N.Y.S.2d 834 (N.Y. Sup. Ct. 1968) (hospitalization for observation of adbominal pains); Zorek, 271 N.Y.S.2d 1012 (three week hospitalization for dietary regime for extreme obesity).

(25) 388 So. 2d 470 (La. Ct. App. 1980).

(26) Id. at 472. The couple had duplicative coverage under nine separate policies amounting to several times more than the actual cost of treatment, and they had "been hospitalized for more than four months [over the prior three years] and [had] claimed in some nine lawsuits against these companies during that time, over $15,000." Id. Nevertheless, the court ordered coverage because the insurance policy did not explicitly require medical necessity and because the wife had several small children at home who prevented her from resting. See id. at 471.

(27) "Medical necessity" is not intended to mean life-or-death necessity but merely medically appropriate or medically beneficial. The intent is to exclude coverage for care that is harmful, of no benefit, or nonstandard. See generally Dallis v. Aetna Life Ins. Co., 574 F. Supp. 547 (N.D. Ga. 1983) (collecting cases that discuss the meaning of the term), aff'd, 768 F.2d 1303 (11th Cir. 1985).

(28) See, e.g., Taylor v. Prudential Ins. Co., 775 F.2d 1457 (11th Cir.) (reversing summary judgment on bad faith liability for insurer's refusal to pay $11,500 for an eight-month hospitalization following stroke, even though Medicare had determined that hospitalization was not medically necessary), reh'g en banc denied, 781 F.2d 905 (11th Cir. 1985); Ex Parte Blue Cross-Blue Shield, 401 So. 2d 783, 783 (Ala. 1981) (holding that two-week hospitalization following fall and fracture should be covered, even though personal physician testified as the only expert witness that "he did not consider such hospitalization 'medically necessary'"); Abernathy v. Prudential Ins. Co., 264 S.E.2d 836 (S.C. 1980) (holding that facial hair removal was medically necessary within the meaning of the policy at issue).

Insurers were sometimes able to prevail, however. See, e.g., Margolis v. Prudential Ins. Co., 629 F. Supp. 195 (D.D.C. 1985) (holding that hospitalization of comatose patient was not medically necessary); Franks v. Louisiana Health Serv. & Indemn. Co., 382 So. 2d 1064 (La. Ct. App. 1980) (holding that insurer may deny coverage for hospitalizations to treat sore throat and colds); Kinzie v. Physician's Liab. Ins. Co., 750 P.2d 1140 (Okla. Ct. App. 1987) (holding that trial court was authorized to rule as matter of law that in vitro fertilization is not necessary to treat an illness; child birth is elective).

(29) See, e.g., Dallis v. Aetna Life Ins. Co., 768 F.2d 1303, 1304 (11th Cir. 1985) (holding that coverage exists for "immuno-augmentative therapy" for cancer, costing $11,000 and provided in a Bahamanian clinic, even though it has "never been approved by any of the various agencies of the United States Government, nor has it ever been proven to be effective"); Shumake v. Travelers Ins. Co., 383 N.W.2d 259 (Mich. Ct. App. 1985) (holding that insurer must pay $17,478 for laetrile and nutritional therapy for lung cancer, even though FDA disapproval meant that it was illegal to ship laetrile in interstate commerce), appeal denied, 425 Mich. 859 (1986); Tudor v. Metropolitian Life Ins. Co., 539 N.Y.S.2d 690, 692 (N.Y. Sup. Ct. 1989) (finding coverage for mercury vapor testing by a physician who "treats the whole person, with biochemical methods and emphasis on elemental deficiencies and food allergies"); Taulbee v. Travelers Co., 537 N.E.2d 670 (Ohio Ct. App. 1987) (finding coverage for immuno-augmentative therapy for cancer); Wilson v. Travelers Ins. Co., 605 P.2d 1327 (Okla. 1980) (holding that insurer must pay for laetrile). But see Free v. Travelers Ins. Co., 551 F. Supp. 554 (D. Md. 1982) (finding no coverage for laetrile); Bruno v. Security Gen. Life Ins. Co., 522 So. 2d 1242 (La. Ct. App. 1988) (finding no coverage for nutritional supplements prescribed by "holistic" physician); Jacob v. Blue Cross & Blue Shield, 758 P.2d 382 (Or. Ct. App. 1988) (affirming summary judgment for insurer who denied coverage for Gerson therapy (for cancer) administered in Tijuana, Mexico and immuno-augmentative therapy received in the Bahamas).

(30) Almost identical language was adopted by the government to define the coverage of Medicare and Medicaid. The Medicare statute precludes payment for treatments that are not "reasonable and necessary for the diagnosis or treatment of illness or injury." 42 U.S.C. section 1395y(a)(1) (1988). The Medicaid statute has been construed similarly to require states to cover all "medically necessary services." See Beal v. Doc, 432 U.S. 438, 444 (1977) (stating that "serious statutory questions might be presented if a state Medicaid plan excluded necessary medical treatment from its coverage," citing 42 U.S.C. subsection 1396, 1396a (10)(c) (1988)). Interpretive guidelines issued under Medicare state further that, "in making [a coverage] decision, a basic consideration is whether the service has come to be generally accepted by the professional medical community as an effective and proven treatment" and that payment may not be made "for any experimental, investigational, or unproven treatment." Experimental Investigational Items or Services, Medicare & Medicaid Guide (CCH) 28,152 (1976); Goodman v. Sullivan, 891 F.2d 449, 450 (2d Cir. 1989) (paraphrasing Medicare Part B Carrier's Manual); see also 54 Fed. Reg. 4302 (1989) (codifying these standards).

Additional congruity between the terms of coverage under private and public insurance results from the fact that Medicare contracts with private insurance companies to handle routine claims administration. See Timothy P. Blanchard, "Medical Necessity" Denials as a Medicare Part B Cost-Containment Strategy: Two Wrongs Don't Make It Right or Rational, 34 ST. LOUIS U. L.J. 939, 955 (1990).

(31) A typical provision excludes coverage for treatment that is "'not medically necessary--i.e., when in the judgment of the Carrier the medical services did not require the acute hospital bedpatient (overnight) setting'" and states that "'[t] he fact that a physician may prescribe, order, recommend, or approve a service or supply does not, of itself, make it medically necessary.'" Franks v. Louisiana Health Serv. & Indemn. Co., 382 So. 2d 1064, 1066 (La. Ct. App. 1980) (quoting the policy at issue in the case); see also Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp.586, 589 n.5 (E.D. Va. 1990)("'[B]enefits will be denied if the Plan determines, in its sole discretion, that care is not Medically Necessary.'" (quoting the policy at issue in the case)); Blue Cross & Blue Shield v. Smither, 573 S.W.2d 363, 364 (Ky. Ct. App. 1978) (involving language to the same effect). For a survey of typical contract provisions, see Miller, supra note 16, at 334-35.

(32) Typical language reads: To be considered medically necessary a service or supply must meet all of these tests. a. It is ordered by a doctor. b. It is commonly and customarily recognized throughout the doctor's profession as appropriate in the treatment of the sickness or injury. c. It is neither educational nor experimental in nature nor provided primarily for research purposes. Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131, 134 (D.N.J. 1989) (quoting the policy at issue in the case).

The precise meaning and relationship between the two terms "experimental" and "investigational" has never been clear. See, e.g., Amy Green, HMO Won't Fund Dying Patient's Experimental' Drug Therapy, AM. MED. NEWS, Jan. 13, 1989, at 1, 17 ("When [one subscriber] first read the [HMO] handbook, she . . . did not know what an experimental procedure was. It wasn't defined or referred to further. She knew she did not want to be a subject in a medical experiment, though.") It is generally supposed, however, that experimental procedures are those for which basic safetyor effectiveness are still in doubt, whereas investigational procedures are those thought to have medical benefit in some circumstances, but the precise clinical indications and protocols for treatment still have not been sufficiently standardized.

Because physicians constantly strive to improve safety and effectiveness by fine tuning all treatment methods, insurers have not been able to define when a new treatment passes out of the investigational stage and into the accepted-with-ongoing-improvement stage. See Pirozzi, 741 F. Supp. at 593 ("Use of a protocol does not, by itself, indicate that a procedure is experimental. Were this not so, much of medicine might be swept within the ambit of the experimental exclusion provision as there is ongoing investigation across the whole range of medicine, including many well-established procedures." (footnote omitted)).

(33) See infra note 42.

(34) See, e.g., Pirozzi, 741 F. Supp. at 594 ("Of course, a different experimental exclusion, or different expert testimony, or a plan that conferred broad discretion on the administrator might well require a different result."); Dozsa, 716 F. Supp. at 136, 139 (holding that insurer must cover ABMT for multiple myeloma cancer despite explicit testimony that such treatment was "investigational," because insurer excluded "experimental" care and care not "commonly and customarily recognized" in its contract, but did not exclude "investigational" treatment); Shumake v. Travelers Ins. Co., 383 N.W.2d 259, 265 (Mich. Ct. App. 1985) (approving laetrile only for thisone patient), appeal denied, 425 Mich. 859 (1986).

(35) For example, insurers have been unable to exclude coverage for temporomandibular joint syndrome (TMJ), a condition of chronic facial pain caused by misalignment of the jaw. TMJ is thought by many medical professionals to be over-diagnosed and over-treated. Some insurers have relied on a generic exclusion of "dental services" to refuse coverage for TMJ treatment by dentists and oral surgeons, but courts have struck down these denials reasoning that "dentistry" is too ambiguous a term to rely on in these circumstances. See Masella v. Blue Cross & Blue Shield, 936 F.2d 98 (2d Cir. 1991); Goss v. Medical Serv., 462 A.2d 442 (D.C. 1983); McFadden v. American United Life Ins. Co., 658 S.W.2d 147 (Tex. 1983). Insurers responded by excluding TMJ explicitly, in one case with a clause that goes on for 18 lines. See Alice G. Gosfield, Value Purchasing and Effectiveness: Legal Implications, in 1991 HEALTH LAW HANDBOOK 185, 203 (Alice G. Gosfield ed., 1991). When confronted with these more precisely worded exclusions, however, some courts have continued to balk, reasoning that the language is too technical and complex for lay policyholders to understand. See Ponder v. Blue Cross, 193 Cal. Rptr. 632 (Cal. Ct. App. 1983).

(36) See Duncan, 388 So. 2d at 470.

(37) See supra note 28 and accompanying text.

(38) See McLaughlin v. Connecticut Gen. Life Ins. Co., 565 F. Supp. 434, 449 (N.D. Cal. 1983); see also Shumake, 383 N.W.2d at 263 n.2 ("[The insurer] could limit coverage to drugs and medicines which have not been renounced by the American Medical Association or exclude drugs and treatment which are controversial.").

(39) See supra note 12.

(40) The courts' desire to rule in favor of the patient can result in reasoning that borders on the absurd. Consider, for example, Ex Parte Blue Cross-Blue Shield, 401 So.2d 783 (Ala. 1981). The court refused to allow Blue Cross to use to its advantage the trial court's ruling that the attending physician has sole authority to determine medical necessity. Blue Cross introduced the doctor's admission that hospitalization was not medically necessary, but the court noted that the policy did not bind Blue Cross to the doctor's opinion on the matter, because it allowed the carrier to make its own independent judgment of medical necessity. See id. at 785-86. Then, the court decided that since Blue Cross was not bound by the doctor's opinion, the issue of medical necessity should have been submitted to the jury. See id. at 786.

(41) See Hughes v. Blue Cross, 263 Cal. Rptr. 850 (Cal. Ct. App. 1989). The Hughes court stated:

If the insurer employs a standard of medical necessity significantly at variance with the medical standards of the community, the insured will accept the advice of his treating physician at a risk of incurring liability not likely foreseen at the time of entering the insurance contract. . . . [G]ood faith demands a construction of medical necessity consistent with community medical standards that will minimize the patient's uncertainty of coverage in accepting his physician's recommended treatment. Id. at 857. In Mount Sinai Hosp. v. Zorek, 271 N.Y.S.2d 1012 (Civ. Ct. 1966), the court stated:

Only the treating physician can determine what the appropriate treatment should be for any given condition. Any other standard would involve intolerable second-guessing, with every case calling for a crotchety Doctor Gillespie to peer over the shoulders of a supposedly unseasoned Doctor Kildare. The diagnosis and treatment of a patient are matters peculiarly within the competence of the treating physician. The diagnosis may be insightful and brilliant, or it may be wide of the mark, but right or wrong, the patient under his doctor's guidance proceeds upon his theories and sustains expenses therefor. Id. at 1016.

(42) See Marker v. Union Fidelity Life Ins. Co., 125 F.R.D. 121 (M.D.N.C. 1989), aff'd per curiam, 907 F.2d 1138 (4th Cir. 1990) (awarding summary judgment to the patient for his appendectomy done for purely preventative, exploratory reasons and asserting that "the determination of what was reasonable and necessary was for the licensed, treating physician"); Van Vactor v. Blue Cross Ass'n, 365 N.E.2d 638, 643 (Ill. App. Ct. 1977) (stating that where patient was hospitalized to remove impacted teeth, "[t]here was sufficient evidence . . . that the insured was justified in relying on the good faith judgment of his treating physician as to the medical necessity of services prescribed"); Carrao v. Health Care Serv. Corp., 454 N.E.2d 781, 788 (Ill. App. Ct. 1983) (stating that Van Vactor remains good law in the absence of an explicit contractual provision to the contrary); Little v. Blue Cross, Inc., 424 N.Y.S.2d553 (N.Y. App. Div. 1980) (insurer bound by the good-faith certification of the insured's doctor that private duty nurse was needed during recovery from heart attack, even while in hospital). See generally Aetna Life Ins. Co. v. Sanders, 193 S.E.2d 173, 176 (Ga. Ct. App. 1972) (giving great weight to physician's recommendation); Siegal v. Health Care Serv. Corp., 401 N.E.2d 1037, 1043 (Ill. App. Ct. 1980) ("[T]he jury could have found that [the attending physician's] recommendation, absent substantial reasons for the insurer's rejection, was sufficient to demonstrate compliance with the requirements of [the] policy."); Shumake, 383 N.W.2d at 263-64 (confronting a coverage dispute involving laetrile and nutritional therapy for lung cancer, thecourt stated that "a physician's judgment should be accorded deference" and that "[a] physician is generally better equipped than lawyers and judges to discern what is medically necessary," although the court did not require absolute deference to such judgments).

Similar rulings have been made under Medicare and Medicaid. See Weaver v. Reagen, 886 F.2d 194, 200 (8th Cir. 1989) ("The Medicaid statute and regulatory scheme create a presumption in favor of the medical judgment of the attending physician in determining the medical necessity of treatment."); Pinneke v. Preisser, 623 F.2d 546, 550 (8th Cir. 1980) ("The decision of whether or not certain treatment or a particular type of surgery . . . is 'medically necessary' rests with the individual recipient's physician and not with clerical personnel or government officials.").

(43) A substantial and increasing minority of courts, though, have held that the sole province to determine coverage does not lie with the attending physician and that insurers have a proper role in determining medical necessity. See infra notes 115 & 143.

(44) In the rare instances where this position is considered as an explicit public policy prohibition, its striking weaknesses are readily revealed. Consider the cogent critique by one court of an argument by a group of psychiatrists that they should not be bound by a health plan's requirement of prior approval for coverage of mental health services:

[T]hese psychiatrists have joined this Plan. Having sought to be members and contractually agreed to be members, how can they now be heard to challenge the provisions to which they agreed? How can they voluntarily enter a contract and then challenge its terms? And if challenged, why would the contract not simply be void? How can plaintiffs challenge only the provisions they do not like and ask the Court, in effect, to modify the contract to their liking? Plaintiffs do not address this.

But the most persuasive argument against the plaintiffs, it seems to me, is ... [that the] [p]laintiffs say, in effect, "Irrespective of any obligation I have to my patients and to my profession, my judgment as to what is in the best interests of my patients will not be determined by the exercise of my medical judgment, but by how much I will be paid for my services." Plaintiffs are saying in effect, "Since I am weak in my resolve to afford proper treatment, [the insurer's] preauthorization program would induce me to breach my ethical and legal duties, and the Court must protect me from my own weakness." In other words, protect me from my own misconduct. This is strange stuff indeed from which to fashion a legal argument.

Varol v. Blue Cross & Blue Shield, 708 F. Supp. 826, 833 (E.D. Mich. 1989).

(45) See Harvey M. Sapolsky et al., Corporate Attitudes Toward Health Care Costs, 59 MILBANK MEMORIAL FUND Q. 561, 561-85 (1981) (describing the lack of concern by companies over the cost of health benefit programs.) The federal government was more concerned about the cost of its health care programs, but most of the government's emphasis was on controlling the prices paid for specific services and not questioning the individual decisions made by physicians. See KAREN DAVIS ET AL., HEALTH CARE COST CONTAINMENT 104-29 (1990).

(46) See, e.g., Rolla Edward Park et al., Physician Ratings of Appropriate Indications for Three Procedures: Theoretical Indications vs. Indications Used in Practice, 79 A.M. J. PUB. HEALTH 445 (1989). In this study, a panel of physicians judged that 17.3% of coronary angiographies, 10.5% of endoscopies, and 28.5% of carotid endarteretomies actually performed were clearly inappropriate. See id. at 446-47.

(47) See, e.g., John E. Wennberg et al., Are Hospital Services Rationed in New Haven or Over-Utilised in Boston?, LANCET, May 23, 1987, at 1185, 1185-88. A number of studies have found that hospitalization rates vary across geographic areas even when those areas have similar demographic characteristics. Physicians in Boston were found more likely to admit patients for "medical and minor surgical ills" than were their counterparts in New Haven. See id. at 1187. In the case of major surgery the variation was not as consistent; for some procedures (carotid endarterectomy and hip and knee replacement) rates were higher in Boston while for others (coronary bypass, thyroidectomy, excision of intervertebral disc, cholecystectomy, hysterectomy, extended simple/radical mastectomy, and splenectomy) the rates were higher in New Haven. See id. at 1187-88. Procedures with low discretion in decisions to admit are more closely correlated to incidence rates in an area than are high discretion procedures. See id. at 1185. The fact that Boston hospitals are more likely to admit patients for more discretionary procedures does not seem to be reflected in the occupancy rates of the hospitals, which were approximately equal in the two areas. See id. at 1186.

Because of the nature of mortality statistics, it is not possible to come to a definitive conclusion as to which area's practice patterns are more effective. The one concrete result of the differences is that Medicare reimbursements in Boston are almost twice that in New Haven. See John E. Wennberg et al., Hospital Use and Mortality Among Medicare Beneficiaries in Boston and New Haven, 321 NEW ENG. J. MED. 1168, 1168-73 (1989).

(48) See ROLLA E. PARK ET AL., PHYSICIAN RATINGS OF APPROPRIATE INDICATIONS FOR SIX MEDICAL AND SURGICAL PROCEDURES 24 (1986) (tentatively concluding that "there really is disagreement about the value of these procedures for many indications, and that the disagreement reflects the lack of detailed evidence about the circumstances in which these procedures are efficacious").

(49) See DAVIS ET AL., supra note 45, at 50; see also Gerard F. Anderson & Earl P. Steinberg, Hospital Readmissions in the Medicare Population, 311 NEW ENG. J. MED. 1349, 1353 (1984) (suggesting that PROs monitor readmissions as well as general hospital programs).

(50) The number of employees working for self-insured firms doubled from 21% to 42% between 1981 and 1985. See Gail A. Jensen & Jon R. Gabel, The Erosion of Purchased Health Insurance, 25 INQUIRY 328, 333 (1988).

(51) See Evan J. Ellman, Monitor Mania: Physicians Regulation Run Amok!, 20 LOY. U. CHI. L.J. 721 (1989).

(52) As with prior insurer innovations, this change was urged by the courts. See, e.g., Van Vactor v. Blue Cross Ass'n, 365 N.E.2d 638, 644 (Ill. App. Ct. 1977) (stating that insurer may not determine medical necessity because "there is no mechanism by which the insured can ascertain in advance how such a determination will be made and whether or not he will be covered for a specific treatment or hospitalization"); Myerson v. Associated Hosp. Serv., 314 N.Y.S.2d 834, 836 (N.Y. Sup. Ct. 1968) ("It would be a different situation if [the patient] were aware from the very beginning that he was primarily [receiving excluded services].").

(53) Prospective review is not possible, for example, in emergency cases.

(54) These techniques include pre-admission certification, concurrent review of hospitalization length of stay, second opinion requirements for surgical procedures, and high-cost case management.

(55) See CHARLES EBY ET AL., A BRIEF HISTORY OF HEALTH CARE COST CONTAINMENT 5 (Health Ins. Ass'n of Am. ed., 1991) (survey of HIAA members).

(56) See INSTITUTE OF MEDICINE, CONTROLLING COSTS AND CHANGING PATIENT CARE? THE ROLE OF UTILIZATION MANAGEMENT 3, 17-18, 66 (Bradford H. Gray & Marilyn J. Field eds., 1989) [hereinafter INSTITUTE OF MEDICINE, CONTROLLING COSTS].

(57) See id. at 71.

(58) See id. at 71-73.

(59) See id. at 73-77.

(60) See id. at 4, 77.

(61) See supra notes 41-44 and accompanying text. For instance, sharp criticism has been leveled against the Medicare rule that coverage decisions may bnot be challenged until the treatment has been rendered and a bill submitted for payment. The Supreme Court acknowledged the rule in Heckler v. Ringer, 466 U.S. 602, 621-22 (1984), over a dissenting opinion criticizing its harshness.

(62) Wickline v. State, 239 Cal. Rptr. 810, 812 (Cal. Ct. App. 1986).

(63) Cf. Zuckerberg v. Blue Cross & Blue Shield, 464 N.Y.S.2d 678 (N.Y. Sup.Ct.), rev'd on other grounds, 487 N.Y.S.2d 595 (N.Y. App. Div. 1985), aff'd, 490 N.E.2d 839 (N.Y. 1986). In Zuckerberg, the court ordered the insurer to pay for cancer treatment received in Mexico, which consisted of organically grown fruits and vegetables and certain medications and vitamins. The court quoted John Greenleaf Whittier's poem Maud Miller when it wrote: "A possible path was opened, had it not been taken, what then? At least, it was tried. 'For of all sad words of tongue or pen, The saddest are these: "It might have been!"'" Id. at 683.

(64) See, e.g., Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131, 140 (D.N.J. 1989) ("Weighing the various equities presents no difficulties....Failure to provide treatment will probably result in death in a matter of months."). Interestingly, the courts rarely appear to consider placing some or all of the risk of monetary loss on the providers that insist on the procedure's efficacy.

(65) 562 N.Y.S.2d 908 (N.Y. Sup. Ct. 1990).

(66) In addition to Bradley, the court in Dozsa, 716 F.Supp. 131, issued a preliminary injunction preventing the insurer from denying coverage for an ABMT treatment. See id. at 140. The court found that the procedure was not experimental, despite the following exhaustive consideration that revealed many explicit references to its "investigational" status:

Dr. Plocher [the insurer's medical director] and his associates analyzed peer review medical literature on the subject; he reviewed the procedure with outstanding experts in the pertinent specialty field; he went to independent organizations of technology assessment; he submitted a proposal for coverage to 90 or more Prudential medical directors and other persons in the field for their comments. Thereafter the question of coverage of ABMT procedures was presented for a final decision to Prudential's headquarters vice-presidents for claims, marketing, law, underwriting and contract. Id. at 135. When the particular claim came in for this patient, Dr. Plocher also "sought the advice of an outside expert," who wrote: "'I do not think ABMT for relapsing multiple myeloma can be considered anything other than investigationalat the present time.'" Id. at 136 (quoting the expert). The court was unpersuaded by this because it was phrased as "investigational" rather than "experimental." See id. at 136-38. Two doctors, however, testified that the procedure was commonly recognized as an appropriate treatment for multiple myeloma. See id. at 138-39.

In contrast, the basis for decision is more substantial in Rollo v. Blue Cross/Blue Shield, No. 90-597, 1990 U.S. Dist. LEXIS 5376 (D.N.J. Mar. 22, 1990). See supra text accompanying notes 1-3.

(67) Bradley, 562 N.Y.S.2d at 910. The court denigrated the consent form as "the bar's contribution to the defense of potential medical malpractice litigation," concluded that it did not accurately describe the nature of the treatment, and gave it little weight. Id. It is interesting that the court attributed the consent form to a hypersensitive bar rather than characterizing it as an honest response to the judiciary's own development of the informed consent doctrine. See also Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 593-94 (E.D. Va. 1990) (finding ABMT for breast cancer covered despite explicit research protocol and the fact that it is still being subjected to randomized clinical trials elsewhere).

(68) See Benjamin Freedman, Equipoise and the Ethics of Clinical Research, 317 NEW ENG. J. MED. 141, 141 (1987) ("The ethics of clinical research requires equipoise--a state of genuine uncertainty on the part of the clinical investigator regarding the comparative therapeutic merits of each arm in a trial."); Eric Kodish et al., Ethical Considerations in Randomized Controlled Clinical Trials, 65 CANCER 2400, 2400-01 (1990) ("The IRB[s] evaluate ... whether proposed research satisfies equipoise conditions, i.e., whether the scientific data show that there is real uncertainty about which arm of the protocol is likely to result in better patient outcomes ....").

(69) In clinical research jargon, the procedure had been subjected to only Phase I studies. Randomized controlled trials constitute Phase II.

After the Bradley decision, the National Cancer Institute (NCI) announced that it would conduct randomized clinical trials for women with metastatic breast cancer undergoing ABMT. See Suplee, supra note 13, at 1. This event verifies the experimental status of ABMT even for more conventional uses; the NCI will saction randomized clinical trials only if the procedure is experimental and the medical efficacy of the procedure has not been demonstrated. Such trials still have not been approved for ABMT's application to AIDS patients.

(70) See Kenneth S. Abraham, Judge-Made Law and Judge-Made Insurance: Honoring the Reasonable Expectations of the Insured, 67 VA. L. REV. 1151 passim (1981).

(71) See Jon R. Gabel & Gail A. Jensen, The Price of State Mandated Benefits, 26 INQUIRY 419 (1989) (analyzing the effects of state-mandated insurance on the cost of insurance, the offering of insurance by small employers, and self-insurance by large employers).

(72) See Abraham, supra note 70, at 1188 ("Any increase in a policy's package of insurance protection will often increase its price....[S]ome people will choose not to buy it at all--the increase in cost ... will have priced them out of the market.").

(73) Thus, the most thorough analysis to date of the cost of statutorily mandated coverage concluded that "nearly one of every six small firms that do not offer health insurance would in an essentially [legislatively] mandate-free environment." See Gabel & Jensen, supra note 71, at 428. It is reasonable to infer that judicially mandated insurance imposes at least an equal cost.

(74) Cf. Abraham, supra note 70, at 1192 ("Balancing the equities between parties to an insurance contract is an acceptable judicial function. That function, however, cannot be pursued liberally without sacrificing the certainty and predictability afforded by written policies of insurance.").

(75) For a more detailed discussion of contracting alternatives, see infra text accompanying notes 171-212.

(76) Peter Huber writes: "Take away contract and there is nothing left to insurance; either bargains of this sort are enforceable on their own mutually understood and accepted terms, or they will not be written at all." PETER W. HUBER, LIABILITY: THE LEGAL REVOLUTION AND ITS CONSEQUENCES 149 (1988).

(77) See supra text accompanying note 20; infra text accompanying notes 96 & 112.

(78) See supra text accompanying note 60.

(79) See PEPPER COMMISSION REPORT, supra note 5, at 40-41; supra notes 46-48.

(80) This evaluation is usually performed through randomized clinical trials. See David A. Kessler et al., The Federal Regulation of Medical Devices, 317 NEW ENG. J. MED. 357, 359 (1987). The Food and Drug Administration does not review medical procedures such as new surgical techniques and does not require clinical trials for all possible uses of a drug or device, only the uses indicated in the sales literature. See id. at 359 (stating that "[s]afety and effectiveness are assessed with special reference to the uses for which the device is intended, as set forth in the labeling on the device").

(81) See COMMITTEE FOR EVALUATING MEDICAL TECHNOLOGIES, INST. OF MEDICINE, ASSESSING MEDICAL TECHNOLOGIES 136-40 (1985) [hereinafter ASSESSING MEDICAL TECHNOLOGIES]; Maxwell J. Mehlman, Health Care Cost Containment and Medical Technology: A Critique of Waste Theory, 36 CASE W. RES. L. REV. 778, 784-94 (1986) (comparing three different concepts of "waste," the first considering only the effect of the technology on the patient, the second incorporating the cost of the technology, and the third converting patient benefits into dollars so that benefits can be compared with the cost of the treatment). (82) See Zuckerberg v. Blue Cross & Blue Shield, 487 N.Y.S.2d 595, 600 (N.Y. App. Div. 1985), aff'd, 490 N.E.2d 839 (N.Y. 1986) (approving insurer's denial of coverage for nutritional therapy treatment of cancer and stating that the denial would have "the desirable effect of affording greater protection to ... patients who are especially vulnerable to unfounded claims of miraculous cures"). Perhaps the patronizing nature of this justification explains why courts have rejected it so consistently. Except in the most obvious cases of quackery, it stands to reason that insurers should defer to patients' and doctors' own decisions about whether a treatment is safe. Therefore, insurers might fare better in court if they used a less disingenuous justification, that is, if they defended primarily on the basis of cost effectiveness, or even granting cost effectiveness, on the basis of uninsurability. See infra text accompanying notes 156-58.

(83) For instance, the five criteria that the Blue Cross and Blue Shield Association use in making its technology assessment recommendations to its constituent plans ask only whether the technology improves net health outcomes equal to the alternatives. They contain no consideration of whether equally effective alternatives are less costly or whether the incremental increase in net health outcomes is justified by the additional costs. See Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 590-91 (E.D. Va. 1990); see also Kalb, supra note 15, at 1115-16 ("[P]rivate insurers virtually never consider the costs of a technology when determining whether it is reasonable or necessary....Specifically, private insurers have not attempted to contain costs by systematically excluding wasteful technologies from coverage.").

(84) See Miller, supra note 16, at 335-36 ("Third-party payors are attempting to focus more on the cost effectiveness of new procedures and reviewing general cost/benefit principles on which to evaluate the effect of one procedure as an alternative to other surgical or medical procedures."); cf. Glenn Kramon, Medical Second-Guessing-in Advance, N.Y. TIMES, Feb. 24, 1991, at D12 ("Insurers say the day is coming when denial of payment will be far more common."). On the Medicare proposal to incorporate cost considerations into coverage decisions, see Medicare Program; Criteria and Procedures for Making Medical Services Coverage Decisions that Relate to Health Care Technology, 54 Fed. Reg. 4302, 4308-10 (1989) (codified at 42 C.F.R. pts. 400, 405). See also Robert Pear, Medicare to Weigh Cost as a Factor in Reimbursement: Fundamental U.S. Shift, N.Y. TIMES, Apr. 21, 1991, at A1 (reporting on a confidential draft of the new Medicare rules awaiting clearance).

(85) An Institute of Medicine committee stated:

At this time, the committee does not see utilization management moving toward intentional rationing of clinically necessary medical services.... Nurse and physician reviewers do not make explicit case-by-case assessments of whether the expected clinical benefits of a hospital admission or other proposed service for a specific patient are, in some way, worth not only the clinical risks but also the economic costs.

INSTITUTE OF MEDICINE, CONTROLLING COSTS, supra note 56, at 148. Whether insurers, or anyone else, should make health care rationing decisions is a very controversial topic. See Maxwell J. Mehlman, Rationing Expensive Lifesaving Medical Treatments, 1985 WIS. L. REV. 239, 301-03 (arguing against the use of cost/benefit criteria). It suffices for the present discussion to observe: (1) someone must make rationing decisions for health care dollars to be spent wisely, see Clark C. Havighurst & James F. Blumstein, Coping with Quality/Cost Trade-Offs in Medical Care: The Role of PSROs, 70 NW. U.L. REV. 6, 6-20 (1975); William B. Schwartz, The Inevitable Failure of Current Cost-Containment Strategies, 257 JAMA 220, 223 (1987); and (2) some rational consumers might decide that their insurers are in the best position to make these decisions, as HMOs are thought to do. Thus, we are not entertaining the strong argument that indemnity insurers should or must be rationing agents. We suggest only that the law should not foreclose this option if a significant number of purchasers desire it. We repeat, however, that at present the issue is two steps removed from this more controversial stage because insurers are bogged down in the first level of assessment, namely, basic safety and effectiveness.

(86) See Joanne B. Stern, Bad Faith Suits: Are They Applicable to Health Maintenance Organizations?, 85 W. VA. L. REV. 911, 913-20 (1983) (reviewing the law on bad faith claims).

(87) For example, in Taylor v. Prudential Ins. Co., 775 F.2d 1457, 1459 (11th Cir. 1985), the court reversed a summary judgement granted to an insurer on a bad faith claim even though it relied on Medicare's determination that no continued hospitalization was medically necessary. In Hughes v. Blue Cross, 263 Cal. Rptr.850, 852, 855-58 (Cal. Ct. App. 1989), the court affirmed an award of $700,000 in punitive damages for the insurer's use of only the physician's orders, progress notes, and nursing notes in denying coverage, despite the insurer's repeated requests to the patient and physician to forward all relevant documents. In Aetna Life Insurance Company v. Lavoie, 505 So. 2d 1050, 1051 & n.1 (Ala. 1987) (per curiam), the jury awarded $3,500,000 in punitive damages (reduced to $500,000 on appeal) on a claim of only $1650, even though the insurer based its decisions on the type of documents that were missing in Hughes--the admitting sheet, admission history and physical, discharge summary and, lab reports. The insurer initially failed to consider the nursing notes and progress notes. It later corrected this mistake, but the court ruled that "the insurance company cannot later seek to justify its denial by gathering information which it should have had in the first place. '[A]n insured purchases insurance and not an unjustified court battle when he enters into the insurance contract.'" Id. at 1053 (alteration in original) (quoting Gulf Atlantic Life Ins. Co. v. Barnes, 405 So. 2d 916, 925 (Ala. 1981)).

(88) See Wilson v. Blue Cross, 271 Cal. Rptr. 876, 883-85 (Cal. Ct. App. 1990); Hershey, supra note 19, at 62-63.

(89) 29 U.S.C. sub section 1001-1461 (1988).

(90) See Blum, supra note 19, at 210. It was once taken as established that punitive damages were unavailable under ERISA. The Supreme Court's recent decision in Ingersoll-Rand Co. v. McClendon, 111 S. Ct. 478, 486 (1990), suggesting that federal courts retain inherent authority to award such damages, casts the issue in some doubt, however. Compare Haywood v. Russell Corp., 584 So. 2d 1291 (Ala. 1991) (holding punitive damages are available under ERISA) with Gaskell v. Harvard Coop. Soc'y, 762 F. Supp. 1539 (D. Mass. 1991) (holding to the contrary, despite Ingersoll-Rand).

(91) Cf. Wilson, 271 Cal. Rptr. at 883 (alleging that contract breach caused patient's suicide).

(92) See David G. Barnum, The Supreme Court and Public Opinion: Judicial Decision Making in the Post-New Deal Period, 47 J. POL. 652 (1985).

(93) In her book, Medicare: The Politics of Federal Hospital Insurance, Judith M. Feder details how the federal government was forced to adopt the same insurance system provided in the private sector in order to avoid a boycott by the hospital industry and the medical profession. See JUDITH M. FEDER, MEDICARE: THE POLITICS OF FEDERAL HOSPITAL INSURANCE 143-56 (1977).

(94) See DAVIS ET AL., supra note 45, at 10-32, 108-28 (detailing the various public and private cost containment initiatives that have been used in the United States).

(95) See id.

(96) The leading critic is Professor Havighurst, who has characterized insurers' cost containment record as a "tacit conspiracy...policed by organized medicine." Clark C. Havighurst & Glenn M. Hackbarth, Private Cost Containment, 300 NEW ENG. J. MED. 1298, 1299 (1979); see also Clark C. Havighurst, The Questionable Cost-Containment Record of Commercial Health Insurers, in HEALTH CARE IN AMERICA: THE POLITICAL ECONOMY OF HOSPITALS AND HEALTH INSURANCE 221, 254 (H.E. Frech ed., 1988) (noting that "commercial carriers appeared for a long time to be nearly useless in combating the influence of moral hazard on the cost of insured health services"); Havighurst, supra note 20, at 338 (stating that "[f]or a long time, the insurers made a positive virtue of noninterference in professional decisions").

(97) See Robert Pear, Medicare Prognosis: Unwieldy Growth Fueled by More Fees and Beneficiaries, N.Y. TIMES, March 10, 1991, section 4, at 4 (noting that Medicare is "the fastest-growing major program in the Federal budget").

(98) See Uwe E. Reinhardt, Health Care Spending and American Competitiveness, HEALTH AFF., Winter 1989, at 5, 5-21 (describing the general perception among business and union executives that health care costs are driving up the cost of doing business in the United States, but disputing the factual basis for this impression).

(99) See Louis Uchitelle, Insurance Linked to Jobs: System Showing its Age, N.Y. TIMES, May 1, 1991, at A1, D23 ("The A.F.L.-C.I.O. estimates that three-fourths of the days lost to strikes in 1989 involved disputes over health care [benefits].").

(100) See supra text accompanying notes 46-48.

(101) See Stuart Altman & Robert Blendon, Introduction to MEDICAL TECHNOLOGIES: THE CULPRIT BEHIND HEALTH CARE COSTS? 1, 1-2 (1979) (Proceedings of the Sun Valley Forum on National Health, available through the U.S. Government Printing Office); Gerard Anderson & Earl Steinberg, To Buy or Not to Buy: Technology Acquisition Under Prospective Payment, 311 NEW ENG. J. MED. 182, 182-185 (1984); Schwartz, supra note 85, at 222-23.

(102) See David Salkever & David Bice, The Impact of Certificate-of-Need Controls on Hospital Investment, HEALTH & SOC'Y, Spring 1976, at 185, 185-88.

(103) See Anderson & Steinberg, supra note 101, at 182.

(104) See William L. Roper et al., Effectiveness in Health Care: An Initiative to Evaluate and Improve Medical Practice, 319 NEW ENG. J. MED. 1197, 1198-99 (1988).

(105) See 42 U.S.C. section 299 (1988).

(106) See Linda J. White & John Ball, Integrating Practice Guidelines with Financial Incentives, 16 QUALITY REV. BULL. 50, 51-52 (1990).

(107) See Rhonda L. Rundle, AMGEN Cleared to Sell Kidney Patient Drug, Still Faces Big Hurdles, WALL ST. J., June 2, 1989, at A1.

(108) See Denis Cournoyer & C. Thomas Caskey, Gene Transfer into Humans: A First Step, 323 NEW ENG. J. MED. 601, 601 (1990).

(109) See Anderson & Steinberg, supra note 101, at 184.

(110) See supra text accompanying note 80.

(111) Cf. Susan Foote, Assessing Medical Technology Assessment: Past, Present, and Future, 65 MILBANK Q. 59, 62-72 (1987) (describing the difficulty the federal government faces in conducting technology assessments).

(112) See Kalb, supra note 15, at 1119 ("Limiting insurance coverage to those technologies that have been demonstrated to be safe, effective, and cost-effective would enhance social welfare by making adequate health insurance less expensive and more accessible."); see also ASSESSING MEDICAL TECHNOLOGIES, supra note 81, at 11 ("Decisions about payment for medical care should be based on more than safety, efficacy, and research status of the care."); id. at 213 ("The authority to apply reimbursement sanctions to implement the findings of assessment, even if quality is at stake, must be clearly spelled out in the law."); id. at 223 ("Reimbursement offered, or withheld, is a prime tool for the enforcement of socially necessary decisions.").

(113) See generally Clark C. Havighurst, Decentralizing Decision Making: Private Contract Versus Professional Norms, in MARKET REFORMS IN HEALTH CARE 22 (Jack A. Meyer ed., 1983) (advocating consumer choice over professionally defined norms).

(114) The major exception comes from those who stand to lose as a result. See Gerald W. Grumet, Health Care Rationing Through Inconvenience: The Third Party's Secret Weapon, 321 NEW ENG. J. MED. 607 (1989). Grumet, a physician, rails against "managed care's arsenal of cost-control weaponry" that supersede the physician's autonomy by a managerial-review process in which armies of claims clerks, administrators, auditors, form processors, peer reviewers, functionaries, and technocrats of every description insinuate themselves into a complex system that authorizes, delivers, and pays for medical service.

Id. at 608. It is noteworthy that the prime focus of this polemic is Medicare and Medicaid, which are not subject to the forces of market discipline that we argue serve to keep private utilization review in proper balance. See infra text accompanying notes 168-70. Also, insurer review is obviously opposed by patients at the time they are sick. As we develop below, however, the proper perspective on this policy question is the perspective of subscribers at the time they purchase their insurance. See infra text accompanying notes 150-53. Thus, a Louis Harris poll found that 57% of the public viewed prospective utilization review to be acceptable, as opposed to 41% of doctors. See ANNE STOLINE & JONATHAN P. WEINER, THE NEW MEDICAL MARKETPLACE 167 (1988).

(115) Accord Blue Cross & Blue Shield v. Smither, 573 S. W.2d 363 (Ky. Ct. App. 1978). The court wrote:

We do not believe a treating physician should be placed in this unassilable position. One need only look to the Medicare and Medicaid System for alleged evidence of fraud which may occur on the part of doctors .... Since a large part of today's rising medical costs are borne by organizations which offer medical benefits plans, such as Blue Cross and Blue Shield, we believe these organizations should be entitled to some measure of protection and should be allowed to challenge decisions made by doctors.

Id. at 365; see also Free v. Travelers Ins. Co., 551 F. Supp. 554 (D. Md. 1982). The Free court held:

[T]he plaintiff's unfettered right to select a physician and follow his advice does not create a corresponding responsibility in the defendant to pay for every treatment so chosen.... [T]o require insurers to pay for every remedy, proven or unproven, prescribed by a physician, could invalidate the actuarial basis of current premium rates. Id. at 560.

(116) In a classic fable depicting medical inflation in terms of the eating habits of the citizens of the mythical country "Gourmand," the story concludes:

Large numbers of people spent all of their time ordering incredibly elaborate meals. Kitchens became marvels of new, expensive equipment. All those who were not consuming restaurant food were in the kitchen preparing it. Since no one in Gourmand did anything except prepare or eat meals, the country collapsed.

Judith R. Lave & Lester B. Lave, Medical Care and Its Delivery: An Economic Appraisal 35 LAW & CONTEMP. PROBS. 252, 253 (1970), quoted in Memorial Hosp. v. Maricopa County, 415 U.S. 250, 274-76 (1974).

(117) See, e.g., Bruce J. Hillman et al., Frequency and Costs of Diagnostic Imaging in Office Practice: Comparison of Self-Referring and Radiologist-Referring Physicians, 323 NEW ENG. J. MED. 1604, 1604-1608 (1990) (finding that physicians with a financial investment in diagnostic imaging equipment ordered four times more imaging examinations than physicians without a financial investment in such equipment).

(118) See Reilly v. Blue Cross & Blue Shield United, 846 F.2d 416, 426-27 (7th Cir.) (Posner, J., concurring and dissenting), cert. denied, 488 U.S. 856 (1988). Concerning the physicians who testified in favor of reimbursement for in vitro fertilization, Judge Posner observed that "[a]s it happens these physicians are specialists in the treatment of fertility and naturally want to encourage the use of exciting and promising treatment." Id. at 427.

(119) See John P. Bunker et al., Evaluation of Medical-Technology Strategies: Effects of Coverage and Reimbursement, 306 NEW ENG. J. MED. 620, 622-23 (1982) (documenting the "obfuscation," "miscoding," and "deliberate effort to conceal" that commonly occurs when physicians submit new experimental procedures for reimbursement); Dennis H. Novack et al., Physicians' Attitudes Toward Using Deception to Resolve Difficult Ethical Problems, 261 JAMA 2980, 2980 (1989) (finding that the majority of physicians surveyed were willing to "misrepresent a screening test as a diagnostic test to secure an insurance payment").

(120) See infra text accompanying notes 239-42.

(121) See Steven DiCarlo & Jon Gabel, Conventional Health Insurance: A Decade Later, HEALTH CARE FINANCING REV., Spring 1989, at 77, 82.

(122) See Reilly, 846 F.2d at 424 ("In the long run, if Blue Cross were to grant too many claims, ... it might be replaced as the plan's administrator.").

(123) See Nazay v. Miller, 949 F.2d 1323, 1335 (3d Cir. 1991) ("The employer... had incentives to avoid the loss of morale and higher wage demands that could result from denials of benefits."); see also INSTITUTE OF MEDICINE, CONTROLLING COSTS, supra note 56, at 109 ("[I]n a roundtable discussion with benefits managers from several large companies that the IOM held in Dec. 1988, some firms were portrayed as slow to adopt prior review out of concern that they would antagonize their white-collar employees."); Havighurst & Hackbarth, supra note 96, at 1301 ("Because, in a competitive market a plan would have to satisfy both consumers and providers about the fairness of the mechanism employed in making these decisions, hardship should be kept to a minimum.").

(124) See William L. Scogland, Fiduciary Duty: What Does it Mean?, 24 TORT & INS. L.J. 803, 829 (1989) ("Indeed, liability has been imposed in a situation in which an administrator has paid persons who were not entitled to benefits under the fund ... because of the fiduciary's duty to defend the fund."); see also Brown v. Blue Cross & Blue Shield, 898 F.2d 1556 (11th Cir. 1990), cert. denied, 111 S. Ct. 712 (1991). In Brown, the court observed:

Decisions on behalf of a plan in the form of a trust lend themselves less readily to the accusation of conflicting interests and are more easily justified.... Fiduciaries are obligated to act not only in the best interest of beneficiaries, but with due regard for the preservation of trust assets,... bearing in mind the interests of all participants and beneficiaries. Id. at 1567-68 (citations omitted). The Brown court errs, however, by erecting impossibly complex burdens of proof that fiduciaries must meet in order for their group-based decisions to be sustained. See infra text accompanying notes 214-20.

(125) See generally Daniel Fischel & John H. Langbein, ERISA's Fundamental Contradiction: The Exclusive Benefit Rule, 55 U. CHI. L. REV. 1105, 1126-38 (1988) (discussing the desirability of aligning the plan administrators' interest with the employer's interest).

(126) See infra text accompanying notes 149-54.

(127) The Institute of Medicine's report on utilization review observes that there are three categories of reviewers: (1) freestanding, (2) insurer-based without provider contracts (i.e., traditional indemnity insurance), and (3) insurer-based with participation contracts (i.e., Blue Cross plans and preferred provider organizations (PPOs)). The report states that "[i]n the first category...the party at risk may be an employer, an insurer, or even an HMO or PPO that contracts with the organization for utilization management services." See INSTITUTE OF MEDICINE, CONTROLLING COSTS, supra note 56, at 62-63. Frequently, though, these utilization review firms have been purchased as subsidiaries of insurers and thus are not wholly independent.

(128) Value Health Science is one firm that sells utilization review criteria. Many of the senior members of Value Health Science were originally employees of Rand Corporation and were funded by the federal government to develop appropriateness criteria. See Kramon, supra note 84, at 12.

Even when utilization review is done by the insurer in-house, the corporate organizational structure provides some assurance that the individual people who actually make medical assessment decisions are sufficiently insulated from undue economic or managerial conflicts if a departmental separation exists between underwriting and claims administration.

(129) See INSTITUTE OF MEDICINE, CONTROLLING COSTS, supra note 56, at 85 ("[Reviewing doctors] make little direct use of formal screening criteria. They are expected to make decisions based upon their own clinical judgment.").

(130) None of the review organizations surveyed by the Institute of Medicine paid their physician reviewers on a basis that rewarded the number of denials. See id. at 278-79. Likewise, most review organizations themselves are compensated by their clients on a per-person or per-review basis, not on the size of their savings. See id. at 169.

(131) The insurance industry has argued that a fourth avenue of protection is available: the "experience rating" pricing system that prevails for most health insurance sold in the United States. "Experience rating" means that the annual recalibration of premiums will reflect each insured group's actual claims experience for the prior period. Therefore, it is argued, an insurer is at little risk and has little temptation to refuse payment, because all amounts it pays are recouped in next years' premium increases. See Brief of Amici Curiae ACLI and HIAA in Support of the Petition for Certiorari, Blue Cross & Blue Shield v. Brown, cert. denied, 111 S. Ct. 712 (1991) (No. 90-494). This argument, however, does not accurately reflect the true market dynamic. Although some insurance is written to create an actual stop-loss provision for the insurer that allows it to recoup from the employer claims paid in excess of a stated maximum, conventional, open-ended insurance leaves the insurer exposed to the risk for the entire term of the contract. The ability to soften any losses by next year's increases is determined by how much competing insurers will bid for the employer's business. The fact that they too will look to past claims experience in pricing their bids does not mean that the final price will create a windfall. Instead, the experience-rated price will be the market's best estimate of the following year's actual costs. Therefore, the risk borne in prior years goes unmitigated. This argument boils down to the observation that the risk is limited to a year (or whatever the stated renewal date is). But during that year, the risk is fully borne by the insurer, as it will be during the following year, and the year after, as long as the policy is renewed.

Nevertheless, the insurance industry's argument is well taken to the extent that many experience-rated contracts are written in a manner that severely dampens the risk actually borne by the insurer. The preceding discussion assumes prospective experience rating, where past experience is used to fix the next year's rate in advance. Many contracts, however, use various forms of retrospective experience rating, which adjusts the total premium, either up or down, at the end of the term to reflect the group's actual experience during the term of the contract. See CHARLES W. WRIGHTSON, HMO RATE SETTING AND FINANCIAL STRATEGY 216-17 (1990). These contracts are actually hybrids between self-insured and fully-insured--the insurer and employer share the risk. To th extent that these arguments shift the risk to the employer, the insurer's potential conflict of interest is indeed mitigated.

(132) "There are more than 800 commercial insurers, 90 BC/BS [Blue Cross/Blue Shield] plans, and over 600 HMOs doing business in the United Stated." Henry T. Greely, AIDS and the American Health Care Financing System, 51 U. PITT. L. REV. 73, 104 (1989).

(133) The fluidity and competitiveness of the health insurance market is manifested by the rapid erosion of the share of business that has been lost to self-insured employers over the past decade. See DiCarlo & Gabel, supra note 121, at 77 (noting that from 1977 and 1987, the percentage of privately insured employees working for self-insured employers rose from 9% to 36%).

(134) See Havighurst & Hackbarth, supra note 96, at 1300 ("Numerous conflicting interests must be addressed, and the plans that are most successful in accommodating these conflicting interests will be most successful in the marketplace. Plans that are too cumbersome will not and should not survive in a competitive market.").

(135) "The A.F.L.-C.I.O. estimates that three-fourths of the days lost to strikes in 1989 involved disputes over health care [benefits]." Uchitelle, supra note 99, at D23.

(136) See ASSESSING MEDICAL TECHNOLOGIES, supra note 81, at 53 (stating that the insurance industry delegates decisions to the profession by consulting medical literature and medical experts, and does not evaluate efficacy per se but rather "providers' acceptance of a technology as standard practice").

(137) See Havighurst & Hackbarth, supra note 96, at 299.

(138) The fact that they have chosen established practice itself as the judge of appropriateness does not necessarily mean that accepted practice is ideal. This has occurred only because there is no other external, objective source for assessing medical appropriateness. See Clark C. Havighurst, Practice Guidelines for Medical Care: The Policy Rationale, 34 ST. LOUIS U. L.J. 777, 796-97 (1990). If an alternative objective standard for assessment should emerge, such as the findings of the external technology assessment organization that we advocate, see infra text accompanying notes 190-92, then the parties to the insurance contract should not be foreclosed from choosing this new standard to replace the present professional standard.

(139) Moreover, the fact that this medical standard has been inflated by the very financial incentives that the insurance system creates (as well as by defensive medicine concerns) means that it contains a large margin for error. Therefore, even if an insurer were to stray below the perceived professional norm, subscribers still have some confidence that coverage decisions will remain within the broad range of acceptable practice patterns.

(140) See Reilly v. Blue Cross & Blue Shield United, 846 F.2d 416, 423 (7th Cir.) (stating that "ERISA's provisions do not permit such potential abuses; decisions and their rationales are reviewable" and that therefore the insurer is not the sole authority), cert. denied, 488 U.S. 856 (1988); Nalezenec v. Blue Cross, 569 N.Y.S.2d 264, 265 (N.Y.App.Div. 1991) (holding that "sole judgment" language "does not give [the insurer] an unfettered right arbitrarily to reject claims for skilled nursing care; ...a denial of coverage is subject to judicial review"); cf. RESTATEMENT (SECOND) OF TRUSTS Section 187 cmt.k (1959) ("It is against public policy to permit the settlor to relieve the trustee of all accountability....[T]he trustee should be answerable to the courts, so far at least as the honesty of his conduct is concerned.").

(141) See infra text accompanying notes 220-70. How stringently courts should review insurers' decisions under this standard is an issue we defer to Section III.

(142) See, e.g., Franks v. Louisiana Health Serv. & Indem., 382 So.2d 1064, 1068 (La. Ct. App. 1980) (finding that allowing the insurer discretion in defining medical necessity does not constitute invalid "potestative condition" since the insurer is "under an obligation to make an honest, sincere effort to determine whether hospitalizations...were medically necessary" and "was required to make a good faith effort by use of medical experts to determine the necessity of hospitalization"); Jacob v. Blue Cross & Blue Shield, 758 P.2d 382, 384 n.1 (Or. Ct. App. 1988) (reasoning that insurer review is not unconscionable because "Blue Cross does not have unlimited discretion to decide what is and is not covered" and "must apply the objective standards set forth in the exclusions and ... carry out its obligations under the contract in good faith").

(143) See supra note 115; see also Myerson v. Associated Hosp. Serv., 314 N.Y.S.2d 834, 838 (1968) (Hecht, J., dissenting) ("I am ... mindful of the fact that documented studies ... have established that unnecessary hospital utilization is a major contributing factor in ever-rising subscriber rates ...."); Lockshin v. Blue Cross, 434 N.E.2d 754, 756 (Ohio Ct. App. 1980) ("[A] function, basic to the insurer, is the right '[to determine] whether a claim should be allowed or rejected.' ... Without such a right, an orderly establishment, administration and dispensation of insurance benefits would be virtually impossible." (quoting Stuhlbarg v. Metropolitan Life Ins. Co., 55 N.E.2d 640, 642 (Ohio 1966))); Jacob v. Blue Cross & Blue Shield, 758 P.2d 382, 384 n.1 (Or. Ct. App. 1988) (rejecting unconscionability of insurer deciding medical necessity).

For decisions under Medicare and Medicaid, see Rush v. Parham, 625 F.2d 1150, 1154 (5th Cir. 1980) ("[A] state may adopt a definition of medical necessity that places reasonable limits on a physician's discretion."); Cowan v. Myers, 232 Cal. Rptr. 299, 303, 306 (Cal. Ct. App. 1986) ("[P]laintiffs are in error when they assert the physician is the sole arbiter of what constitutes a medical necessity.... Not only would such a rule result in inconsistent and unfair applications based on the variation between physicians, but the State's requirement of reimbursement would be limited only by the imagination of physicians."), cert. denied, 486 U.S. 846 (1987).

(144) See, e.g., Ex parte Blue Cross-Blue Shield, 401 So. 2d 783, 785-86 (Ala. 1981) (requiring insurer to pay for a two-week hospitalization following a fall because it was a medical necessity, even though the contract stated necessity was to be determined in the judgment of the Carrier); Blue Cross & Blue Shield v. Smither, 573 S.W.2d 363, 364-65 (Ky. Ct. App. 1978) (remanding case for a full trial despite Blue Cross's judgment that a 77 day hospital stay was not medically necessary).

(145) See supra notes 70-91 and accompanying text.

(146) This topic has received much consideration in other areas. See generally MILTON R. WESSEL, SCIENCE AND CONSCIENCE 32-37, 42-55 (1980) (evaluating the judicial resolution of socio-scientific controversies); Milton R. Wessel, Adversary Science and the Adversary Scientist: Threats to Responsible Dispute Resolution, 28 JURIMETRICS J. 379 (1988) (same); Lon L. Fuller, The Forms and Limits of Adjudication, 92 HARV. L. REV. 353, 394-404 (1978) (discussing lack of appropriate training); Nathan Glazer, Should Judges Administer Social Services?, 50 PUB. INTEREST 64, 66-68 (1978) (criticizing the judiciary's reliance on theoretical rather that practical or clinical knowledge in their decision making); Daniel P. Moynihan, Social Sciences and the Courts, 54 PUB. INTEREST 12, 12-15 (1979) (same). In the context of insurance disputes, see Costich, supra note 12 (detailing difficulties that courts have applying de novo review to disputes over medical appropriateness).

(147) See Anthony Champagne et al., An Empirical Examination of the Use of Expert Witnesses in American Courts, 31 JURIMETRICS J. 375, 388 (1991) ("The willingness of experts to reach firm conclusions was viewed [by jurors responding to a survey] as more important in determining believability of experts than were impressive educational credentials ... or a reputation as a leading expert in the field ....").

(148) Eminent jurist Judge Jack Weinstein observed that "an expert can be found to testify to the truth of almost any factual theory, no matter how frivolous, thus validating the case sufficiently to avoid summary judgment and force the matter to trial....Juries and judges can be, and sometimes are, misled by the expert-for-hire." Jack B. Weinstein, Improving Expert Testimony, 20 U. RICH. L. REV. 473, 482 (1986); see also Michael H. Graham, Expert Witness Testimony and the Federal Rules of Evidence: Insuring Adequate Assurance of Trustworthiness, 1986 U. ILL. L. REV. 43, 45 ("Today practicing lawyers can locate quickly and easily an expert witness to advocate nearly anything the lawyers desire."); Peter Huber, Safety and the Second Best: The Hazards of Public Risk Management in the Courts, 85 COLUM. L. REV. 277, 333 (1985) ("The scientific community is large and heterogeneous, and a Ph.D can be found to swear to almost any 'expert' proposition, no matter how false or foolish.").

(149) Cf. HUBER, supra note 76, at 185-86 ("The only human reaction to the individual tragedy, viewed close up, is unbounded generosity, which any large corporation or insurer can surely afford to underwrite....The jury's focus is always on private harm after the accident, not on public benefit beforehand.").

(150) See Mehlman, supra note 81, at 876-77 (espousing the individual-in-need perspective for determining Medicare coverage).

(151) See Goodman v. Sullivan, 891 F.2d 449 (2d Cir. 1989) (stating that the Secretary's regulation denying payment for MRI scans does not presume to supervise or control the practice of medicine, but simply refuses subsequent Medicare reimbursement for certain kinds of services; while this may influence some medical decisions, such tangential influence is inherent in Medicare itself); Association of Am. Physicians & Surgeons v. Weinberger, 395 F. Supp. 125, 134 (N.D. Ill.) (stating that the medicare utilization review mechanism "does not prohibit a physician from performing any surgical operations he deems necessary....It merely provides that if a practitioner wishes to be compensated for his services by the federal government, he is required to comply with certain guidelines and procedures enumerated in the statute"), aff'd mem. sub nom. Association of Am. Physicians & Surgeons v. Matthews, 423 U.S. 975 (1975). In Wickline v. State, 239 Cal. Rptr. 810 (Cal. Ct. App. 1986), a case concerning a patient's premature discharge caused by the California Medicaid program's (Medi-Cal) denial of reimbursement, the court wrote:

The decision to discharge is ... the responsibility of the patient's own treating doctor....[W]hile Medi-Cal played a part in the scenario before us in that it was the resource for the funds to pay for the treatment sought,...Medi-Cal did not override the medical judgement of Wickline's treating physicians. Id. at 819; see also Mark A. Hall, Institutional Control of Physician Behavior: Legal Barriers to Health Care Cost Containment, 137 U. PA. L. REV. 431, 463 ("When insurance companies refuse to pay for medical treatment they consider unnecessary, we do not think of them as dictating to physicians how to practice medicine. These third-party payors are merely setting limits on what treatment they are willing to reimburse."). Havighurst and Hackbarth observe:

It is important to recognize that a coverage limitation in an insurance policy is nothing more than that. An exclusion from coverage does not necessarily imply anything about treatment or procedures, or the competence or efficiency of any providers whose services are not covered. All that is involved is a voluntary contractual limitation on the right of the insured to draw on the common fund.... Therefore, any service exclusion that plans might adopt should not become the subject of a debate over what is or is not good medical practice or over the medical 'need' for the excluded service in particular circumstances. Havighurst & Hackbarth, supra note 96, at 1300.

(152) JOHN RAWLS, A THEORY OF JUSTICE 12 (1971).

(153) The recent decision of Congress to repeal catastrophic health insurance may be instructive. This legislation, which eliminated upper limits on Medicare benefits, would have benefitted most those individuals with very expensive illnesses. After being confronted with the full cost of this benefit, groups representing the elderly protested and were able to force the legislation's repeal. See Thomas Rice et al., The Medicare Catastrophic Coverage Act: A Post-Mortem, HEALTH AFF., Fall 1990, at 75, 80.

(154) Cf. Rollo v. Blue Cross/Blue Shield, No. 50-597, 1990 U.S. Dist. LEXIS 5376, at *22 (D.N.J. Mar. 22, 1990) ("[S]ubscriber premiums should not have to pay for procedures which are purely experimental or investigative or subsidize every scientist stirring magic potion in some laboratory at the top of a mountain with lightning flashing about.").

(155) Thus, the many courts that have ordered payment for ABMT treatment for cancer patients adopt as their mantra the observation that ABMT, proven or not, is the patient's "only hope," changing the issue to whether there is any possibility of success rather than whether the treatment is proven effective. See Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 588 (E.D. Va. 1990) (examining testimony of doctor who determined that ABMT for breast cancer was the patient's "best chance for any type of meaningful survival," despite the fact that the only evidence to support his contention was that it temporarily shrinks the size of the tumor); Dozsa v. Crum & Forster Ins. Co., 716 F.Supp. 131, 138 (D.N.J. 1989) (finding that claim for ABMT of multiple myeloma patient with six months to live was not experimental because "[i]t was simply the only appropriate treatment available to treat plaintiff's condition"); see also Bradley v. Empire Blue Cross & Blue Shield, 562 N.Y.S.2d 908 (N.Y. Sup. Ct. 1990) (finding ABMT for AIDS patient not experimental, even though never performed by any other doctor).

(156) See Bunker et al., supra note 119, at 621-22 (stating that fees are initially set to reflect the technologies' development costs and that due to the lack of competitive pressures and insurer oversight, they rarely drop after the technology becomes more commonplace).

(157) It might be objected that this cost is slight in group insurance because rates are set each year based on the prior year's experience; therefore, there is only a small gap for unanticipated procedures to arise. On the other hand, this same observation means that subscribers would be giving up very little by not insuring new procedures. Therefore, at the margin, they may find it cost beneficial to sacrifice this increment of coverage.

It also can be argued in response that experimental therapies for terminally ill patients are the very type of risks that are the most insurable, in the sense that insurance is most appropriate for high cost, low incidence events with catastrophic consequences. This is true, but the proper balance of insurable and uninsurable elements of experimental care can only be struck in the marketplace. For instance, the market is perfectly capable of producing an insurance product that excludes experimental therapy except in cases of life-threatening illness. Indeed, this appears to be the construction that a number of insurers give to their experimental exclusion, when it is said that they apply it on a "case-by-case basis." Interview with Joel Miller, Health Insurance Association of America, in Washington, D.C. (April 16, 1991).

(158) There is obviously a Catch-22 at play here since the refusal to reimburse at the outset will greatly slow the dispersion of the new technology. The parties to the contract may nevertheless desire this effect. As detailed above, many analysts in the health policy community believe that new technological innovations are diffused much too rapidly due to the absence of the market test that is usually imposed on innovations. See supra text accompanying note 96. If new technologies must prove themselves in an adverse reimbursement environment, only those that are cost saving or that provide a clear cost beneficial improvement in outcomes will be selected.

(159) See, e.g., John H. Goddeeris, Medical Insurance, Technological Change, and Welfare, ECON. INQUIRY 56, 60 (1984) ("[T]he individual acts ex post as though his expenditure decisions have no effect on the insurance premium."). Peter Huber argues:

Wise policy can only be based on the broader perspective. Efficient deterrence looks at risk in general for the population at large; harsh though it sounds, the circumstances of the single plaintiff now in court are all but irrelevant.... A workable compensation system always looks beyond the individual tragedy to the solvency of the system as a whole. HUBER, supra note 76, at 60. David Eddy, a leading researcher and policy analyst on medical effectiveness, discusses this point with respect to ABMT treatment for breast cancer in particular, but uses terminology that contrasts a societal perspective with an individual patient perspective: "[M]any activities in medicine that make great sense from the point of view of an individual patient might not make sense when the perspective is widened to encompass other activities for other people--what we call society." David M. Eddy, The Individual vs. Society: Is There a Conflict?, 265 JAMA 1446, 1449 (1991). For a rare decision recognizing the critical nature of this distinction, see Nazay v. Miller, 949 F.2d 1323, 1338 (3d Cir. 1991) (observing that the district court erroneously "focus[ed] on this case rather than the overall administration of the Plan").

(160) See PAUL MENZEL, MEDICAL COSTS, MORAL CHOICES: A PHILOSOPHY OF HEALTH CARE ECONOMICS IN AMERICA 48-51 (1983) (distinguishing between the high-risk and low-risk perspectives on valuing the benefits of treatment).

(161) This inherent judicial myopia is demonstrated by the results in Medicaid coverage disputes, where the group perspective should be much more evident due to extremely limited funding for the poor. Despite the obvious group impact of ordering coverage for expensive and questionable services, some courts (notably, the Eighth Circuit) nevertheless consistently reach extreme results in favor of treatment. See Meusberger v. Palmer, 900 F.2d 1280 (8th Cir. 1990) (holding that pancreas transplant must be covered by Medicaid even though considered experimental by Medicare); Weaver v. Reagen, 886 F.2d 194 (8th Cir. 1989) (finding it arbitrary and capricious to deny Medicaid coverage for unapproved use of AZT, an expensive AIDS drug); Pinneke v. Preisser, 623 F.2d 546 (8th Cir. 1980) (holding that Medicaid must fund sex-change operations); cf. American Soc'y of Cataract and Refractive Surgery v. Sullivan, 772 F. Supp. 666 (D.D.C. 1991) (requiring Medicare to pay for investigational intraocular lens implants that have not yet received FDA approval).

(162) See HUBER, supra note 76, at 192; see also Robert E. Scott, The Case for Market Damages: Revisiting the Lost Profits Puzzle, 57 U. CHI. L. REV. 1155, 1201 (1990) (observing "the tendency of legal analysts to tell the story of contract breach ex post, and thus to define [contract terms] by reference to what has happened in a particular case...[rather than] at the time of contract, as though we did not know how the story would end").

(163) See BERNARD SCHWARTZ, ADMINISTRATIVE LAW 235-37 (3d ed. 1991).

(164) Fuller, supra note 146.

(165) He referred not merely to judicial litigation but to the institution of adjudication in its broadest possible sense, as contrasted with the two other fundamental forms of social ordering: contracting and voting. See id. at 363; see also supra note 146 and accompanying text.

(166) Fuller, supra note 146, at 403.

(167) See Eddy, supra note 159. The inherent unsuitability of medical appropriateness questions for judicial determination is underscored by the wide variation in practice patterns, documented by medical epidemiologists such as Wennberg. See supra note 47. These researchers have demonstrated that the courts operate under a false impression when they assume the existence of a uniform, scientifically determined medical consensus. Because medical decisionmaking is highly variable and judgmental, there is no objective criterion to inform a nonexpert judge how to decide questions of medical appropriateness, or even how to choose among differing expert witnesses.

(168) See FTC v. Indiana Fed'n of Dentists, 476 U.S. 447 (1986):

Any recommendation that benefits be denied...is based on the professional judgment of a licensed dentist that the materials available to him...are sufficient to indicate that the treating dentist's recommendation is not necessary to the health of the patient. There is little basis for concluding that, where such a divergence of professional judgment exists, the treatment recommendation made by the patient's dentist should be assumed to be the one that in fact represents the best interests of the patient. Id. at 464 n.4. (169) Although HMO coverage decisions have prompted some litigation, there has been much less than for traditional indemnity insurers. In part, this diminished level of controversy results from the fact that the process by which an HMO denies coverage frequently fails to disclose to the subscriber the existence of any controversy or difference of opinion. See CLARK C. HAVIGHURST, HEALTH CARE LAW AND POLICY 1209 (1988); ASSESSING MEDICAL TECHNOLOGIES, supra note 81, at 129. Therefore, courts raise no fundamental public policy objection to HMOs' determinations of medical necessity even though those determinations are infrequently exposed to judicial or subscriber scrutiny and so, to some extent, carry fewer procedural safeguards than similar determinations made under indemnity plans.

(170) See The HMO Act of 1973, 42 U.S.C. [Section] [Section] 300e to 300e-17 (1988).

(171) See Lee N. Newcomer, Defining Experimental Therapy--A Third-Party Payer's Dilemma, 323 NEW ENG. J. MED. 1702 (1990).

(172) See supra note 35.

(173) See id. at 639-42; see also Cline & Rosten, supra note 15, at 130 ("[I]f contract language is expanded to deal with a number of specific situations, a court is apt to find the language unenforceable as being a 'sea of print' or not inclusive enough if sought to be applied generally." (footnote omitted)).

(174) Thus, one court has held that, as a matter of law, the term "mental illness" is ambiguous and therefore must be construed in favor or the insured as not encompassing autism. See Kunin v. Benefit Trust Life Ins. Co., 910 F.2d 534, 541 (9th Cir. 1990), cert. denied, 111 S. Ct. 581 (1990). But see Brewerv. Lincoln Nat'l Life Ins. Co., 921 F.2d 150, 154 & n.2 (8th Cir. 1990) (disagreeing with Kunin).

(175) See Interview with Gordon L. Tobias, Medical Director of SVP Health Services, in Concordville, Pa. (Feb. 23, 1991); see also In re Appeal of the Medical Doctor Provider Class Plan Determination Report, Mich. Dept. of Licensing and Regulation. Ins. Bureau, No. 90-11109-BC (Aug. 5, 1991) (requiring Blue Cross to cede authority to treating physicians on determination of medical necessity).

(176) See Interview with Joel Miller, Provider Affairs Dept., Health Insurance Association of America, in Washington, D.C. (April 16, 1991).

(177) Brown v. Blue Cross & Blue Shield, Inc., 898 F.2d 1556, 1570 n.16 (11th Cir. 1990), cert. denied, 111 S. Ct. 712 (1991); see also Van Vactor v. Blue Cross Ass'n, 365 N.E.2d 638, 644 (Ill. App. Ct. 1977) ("[S]ignificant policy exclusions contained in a master contract but omitted from the brochure distributed to policyholders should not be enforced."); Waldrip v. Connecticut Nat'l Life Ins. Co., 566 So. 2d 434, 436-37 (La. Ct. App. 1990) (finding that a state statute requiring that "insured ha[ve] in his possession the entire contract" prohibits any incorporation by reference). But see Johnson v. District 2 Marine Eng'rs Beneficial Ass'n, 857 F.2d 514, 516 (9th Cir. 1988) (finding no coverage for liver transplant because it was not contained in the "complete listing" of covered procedures referred to but not itself listed in the Plan). Thus, it does not appear the courts would consistently uphold Havighurst's approach, that policyholders subscribe to one of several competing sets of clinically detailed practice guidelines, see Havighurst, supra notes 138 & 177, at 796-800, unless it is feasible to distribute insurance contracts of this bulk.

(178) See Monaco & Burke, Gatekeeper Part Two, supra note 15, at 402 n.7 ("If only those treatments specifically excluded in the policy could legally be denied coverage, a policy might be as long as the Yellow Pages, and insurers would have to send out frequent riders so as to catch all the new unproven or questionable treatments that came down the pike.").

(179) See Hall, supra note 151, at 475-77. One of the most common symptoms which exemplifies this problem is chest pain. Many people come to the emergency room with such pain but because of the many variables involved there is no set way in which to treat a given case. In most cases the decision is left to the physician who must combine all factors affecting that case (such as the patient's history, EKG results, what the physician has been taught about this decision, etc.) to reach his decision. Even this does not lead to consistency. While the EKG can be used to decide who can be sent home and who must be admitted immediately, there remains a group of patients with abnormal results who are not necessarily having a myocardial infarction. Because symptoms and pain thresholds vary between patients and because one can suffer a heart attack and maintain a normal EKG reading, there is no clear cut way to determine the proper course of action. See Lee Goldman et al., A Computer-Derived Protocol to Aid in the Diagnosis of Emergency Room Patients with Acute Chest Pain, 307 NEW ENG. J. MED. 588 (1982); Michael W. Pozen et al., A Predictive Instrument to Improve Coronary-Care-Unit Admission Practices in Acute Ischemic Heart Disease, 310 NEW ENG. J. MED. 1273 (1984).

(180) Moreover, because of the considerable cost involved in developing these treatment protocols, they are necessarily marketed to insurance companies by proprietary firms that protect them as trade secrets. Therefore, doctors and patients may not even have access to the rules that govern their reimbursement. See Hershey, supra note 19, at 58 (noting that UR firms "jealously guard[] the entire set of medical criteria,...[which are] viewed as proprietary data and constitute an essential resource of the business"); see also Kramon, supra note 84 (describing Value Health Science's UR system, which is criticized by the AMA's office of quality assurance for "fail[ing] to share with physicians the criteria on which it [is] basing its judgments").

(181) See Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 590-91 (E.D. Va. 1990).

(182) Id. at 590; see also Bucci v. Blue Cross-Blue Shield, 764 F. Supp. 728, 733 (D. Conn. 1991) (criticizing Blue Cross for not specifying a "defined validated, justifiable standard" required before experimental status ceases).

(183) The Pepper Commission estimated that only 10-20% of medical procedures have been subjected to clinical trials. See PEPPER COMMISSION REPORT, supra note 5, at 41. Although drugs and devices do require a clinical trial for their primary intended use, these trials are not required for the diffusion of alternative uses. See supra note 80.

(184) For instance, in Reilly v. Blue Cross and Blue Shield, 846 F.2d 416 (7th Cir.), cert. denied, 488 U.S. 856 (1988), the Seventh Circuit noted that the plaintiffs' argument resisted the use of a set percent success rate as a litmus test because they contended that the insurer's discretion over setting the standard could result in the denial of coverage, "for example, for [all] treatments administered to terminally ill patients" for whom the success rate approached zero. Id. at 423.

(185) Newcomer goes so far as to suggest covering procedures only if they are proven effective through randomized controlled clinical trials. He recognizes, however, that this is infeasible for the vast bulk of medicine. See Newcomner, supra note 171, at 1702-03.

(186) See infra text accompanying notes 202-03.

(187) See Kalb, supra note 15, at 1121-24 (advocating the option to select in the contract among levels of wastefulness versus cost effectiveness of new technologies).

(188) See Peter Doubilet et al., Use and Misuse of the Term "Cost Effective" in Medicine, 314 NEW ENG. J. MED. 253 (1986) (detailing the confusion that surrounds the use of these terms).

(189) See Cline & Rosten, supra note 15, at 136 ("[P]olicy language alone cannot induce social change or force courts to recognize that certainty of coverage facilitates the cogent economic goal of providing a product at a reasonable cost to the consuming public.").

(190) See Kalb, supra note 15, at 1125; Monaco & Burke, Gatekeeper Part Two, supra note 15, at 409.

(191) Federal agencies performing technology assessments include the Agency for Health Care Policy and Research as well as the Office of Technology Assessment. Private organizations include physician organizations such as the American Medical Association and the American College of Physicians, hospital based groups such as the Johns Hopkins Program for Medical Practice and Technology Assessment, individual private insurance companies and their trade associations, and private technology assessment firms such as Battelle and Health Technology Associates. See Resources Supply Technology Information, HOSPITALS, Aug. 5, 1989, at 42.

(192) See, e.g., Waldrip v. Connecticut Nat'l Life Ins. Co., 566 So. 2d 434, 435 (La. Ct.App. 1990) (noting that the contract defines "experimental" as any treatment "'not approved or accepted as essential to the treatment of injury or sickness by any of the following: (1) the American Medical Association; (2) the United States Surgeon General; (3) the United States Department of Public Health; or (4) the National Institutes of Health'"); Bradley v. Empire Blue Cross & Blue Shield, 562 N.Y.S.2d 908, 909 (N.Y. Sup. Ct. 1990) (noting that the plan in question excluded experimental treatment "according to guidelines established jointly ... by the State of New York, Blue Cross and Metropolitan Life Insurance Company").

(193) See, e.g., Reilly v. Blue Cross & Blue Shield United, 846 F.2d 416, 423 (7th Cir.) (stating that defendant's reliance on Blue Cross and Blue Shield Association's determination that in vitro fertilization is experimental created "an inherent risk of abuse"), cert. denied, 488 U.S. 856 (1988); Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 590 (E.D. Va. 1990) (noting that the medical director "merely 'applied the policy' [memorandum from Blue Cross/Blue Shield] and rejected plaintiff's preauthorization for [ABMT], solely on the ground that the Association deemed the procedure experimental"). This illustrates the impossible bind that courts can impose on insurers; in other decisions they have criticized insurers for the conflict of interest that occurs when they fail to delegate the coverage decision.

(194) See Kunin v. Benefit Trust Life Ins. Co., 910 F.2d 534 (9th Cir.) (noting that the insurer's medical director consulted with three psychiatrists whose knowledge about autism was nowhere in the record and that the director failed to consult with plaintiff's physicians), cert. denied, 111 S. Ct. 581 (1990); McLaughlin v. Connecticut Gen. Life Ins. Co., 565 F. Supp. 434, 439 (N.D. Cal. 1983):

The only investigation [on immuno-augmentative therapy as a lung cancer treatment] done at all was by [the medical advisor]. He read a few articles ... and made a phone call to another doctor to discuss it. However he did not contact Mrs. McLaughlin's United States doctors, the Immunology Researching Centre where Mrs. McLaughlin received treatment, or any doctor who supports the therapy.

(195) See Waldrip, 566 So. 2d at 436-37 (holding contract terms invalid under Louisiana statute that requires the entire agreement to be contained in a single document).

(196) Thus, in the Bradley case, the court observed that the designated entities had issued no assessment decisions on the particular use of ABMT at issue in that case. See Bradley, 562 N.Y.S.2d at 909; see also Rollo v. Blue Cross/Blue Shield, No. 90-597, 1990 U.S. Dist. LEXIS 5376, at *13 (D.N.J. Mar. 22, 1990) (noting that Blue Cross and Blue Shield Association had considered ABMT for several forms of cancer, but not for the particular type of cancer in this case); PEPPER COMMISSION REPORT, supra note 5, at 41 ("[O]nly 10 percent to 20 percent of medical practices are supported by randomized controlled trials."); Newcomer, supra note 171, at 1703 (observing that the inability to perform randomized clinical trials will create "orphan diseases" for which there are no assessments).

(197) Wilson v. Blue Cross, 271 Cal. Rptr. 876 (Cal. Ct. App. 1990), which concerned the liability of utilization review organizations, dramatically illustrates the pitfalls of failing to specify in the insurance contract the review mechanism to be used in making medical necessity assessments. The doctor for the Wilson's son ordered a month-long hospitalization for major depression, drug dependency, and anorexia, but Blue Cross announced that it would discontinue coverage after two weeks, following the recommendation of Western Medical Review Organization, with whom Blue Cross had contracted for concurrent utilization review services. See id. at 877, 880-81. Two weeks after his discharge, the patient committed suicide. See id. at 887. The court held that Western Medical must stand trial in a suit by his parents seeking compensation, in part because it "used the concurrent utilization review process without knowledge as to whether the decedent's Alabama Blue Cross policy allowed for such review," which along with other evidence, was sufficient "to raise a triable issue of material fact as to whether Western Medical's conduct was a substantial factor in causing the decedent's death." Id. at 883.

(198) Delegating these review functions to industry trade associations themselves, as the Blue Cross plans have tended to do, is not likely to alleviate these conflict problems in the perception of the courts. Obviously, the precise composition of the external entity will determine how convincing a case insurers can make for its independence and authoritativeness. Two options, plus a hybrid and several variations, are feasible. One is to construct a new private entity, perhaps housed at a prestigious university, whose board of directors reflects various interest groups such as academic researchers, providers, labor, business, insurers, and the public. The second option is to provide funding to an existing public sector entity, such as the Agency for Health Care Policy and Research, within the Public Health Service. A hybrid would be to construct a new, quasi-public entity to serve the same purpose, perhaps modeled on organizations like the Institute of Medicine within the National Academy of Sciences.

The option ultimately chosen must satisfy two fundamental concerns. One is the business needs of the insurers that will fund the organization. They will require assurance that timely assessments will be issued in sufficient quantity on the procedures they deem most important to assess, that the methodology used is a reputable one, and, most critical of all, that the terminology in which the assessments are issued will be sufficiently definitive and precise to fit snugly the coverage terminology employed in their contracts. The second basic concern relates to antitrust risks. The potential effect of the entity's decisions on the financial well being of medical equipment firms, drug manufacturers, and providers means that the entity must be constructed in a manner that avoids any serious accusation of collusive denials of reimbursement by the insurance industry. This goal is best accomplished by removing insurers' control over the operation of the entity and by eliminating any uniformity in their response to the agency's decisions, but these measures also tend to undermine the business purpose of the entity. Nevertheless, if it is clear that insurers do not control the content of the entity's decisions (only its funding and structure), and if each insurance company acts unilaterally in implementing the entity's decisions, antitrust concerns could be held to a minimum.

(199) See John P. Bunker et al., Evaluation of Medical Technology Strategies: Proposal for an Institute for Health-Care Evaluation, 306 NEW ENG. J. MED. 687 (1982) (second of two parts); Arnold S. Relman, Assessment of Medical Practices: A Simple Proposal, 303 NEW ENG. J. MED. 153, 154 (1980): cf. Arnold S. Relman, Reforming the Health Care System, 323 NEW ENG. J. MED. 991, 992 (1990) (advocating that more money be invested in a "wide-ranging technology assessment program").

It is not essential to this proposal that there be either a new entity or a single entity. We note, however, that the existing assessment organizations are not presently situated to perform this function. Primarily, this is because the entity's tasks must be carefully specified in advance for it to solve the insurance contracting problems. Existing agencies are too varied in the processes by which they conduct their reviews and the terminology in which they express their conclusions for insurance contracts to attach cleanly to them. If, however, an existing agency were to modify itself along the lines outlined in the following text, it could meet the requirements of this proposal.

(200) See Telephone Interview with Dan Dragalin, M.D., former Medical Director of Prudential Insurance Company of America (Mar. 16, 1991).

(201) These categories are similar to those in Paul Kalb's proposal. See Kalb, supra note 15, at 1121-24.

(202) There is a vast amount of technology assessment literature that gives much more meaning to these various terms. For instance, effectiveness can be defined in terms of various measures of health outcome (lives saved, years of life, or quality-adjusted life years), or by more intermediate measures of benefit such as increased certainty of diagnosis. It is not necessary, however, to incorporate any further definitional precision into the insurance contract. It is only necessary to make certain that the terms used by the reviewing entity match the terms in the contract.

(203) One level might cover all care, or different levels might be chosen for preventative versus curative care, hospitalization versus outpatient, or surgery versus medicine, etc. Moreover, different levels might be chosen according to the cost of the service, requiring a more demanding demonstration of appropriateness for more expensive services. Alternatively, lower levels might be chosen for treatments that address life-threatening conditions. These decisions will be shaped in part by the ability of contractual language to distinguish precisely and predictably among these various categories of treatment.

(204) See Pirozzi v. Blue Cross-Blue Shield, 741 F. Supp. 586, 590-94 (E.D. Va. 1990) (relying on testimony that ABMT has been routinely prescribed for breast cancer since 1988 to reject the validity of a Blue Cross technology assessment adopted in 1988).

(205) The contract might specify that a determination of unproven (experimental) status is binding for two years, accepted (safe and effective) status for five years, and unsafe or ineffective status for ten years. Or the parties might specify that an earlier decision need not be revisited unless brought into question by a certain quality and quantity of evidence, for instance, two or more subsequent articles published in the peer-reviewed scientific literature or at least one subsequent randomized clinical trial. See also White v. Caterpillar, Inc., 765 F. Supp. 1418, 1421-22 (W.D. Mo. 1991) (criticizing Blue Cross for failing to update a 1985 report on ABMT for breast cancer). See generally H. David Banta & Stephen B. Thacker, The Case for Reassessment of Health Care Technology: Once is Not Enough, 264 JAMA 235 (1990) (examining four clinical practice areas to demonstrate the need for continuously reassessing existing technologies).

(206) For example, under typical contracts, the insurer must determine whether the condition treated is an illness and whether the services are excluded as cosmetic or purely elective.

(207) The Permanente Medical Group, for instance, has a "Bone Marrow Transplant Advisory Board" composed of seven physicians from its medical staff that determines when and under what particular circumstances the procedure is no longer experimental for a particular disease. It relies principally on data published in peer-reviewed journals rather than on anecdotal case reports or unreviewed abstracts. For rare diseases with no published data, however, the Board relies on the consensus derived from the combined clinical knowledge of its members. See Interview with D. Blair Beebe, Chief of Staff of the Permanente Medical Group, in Washington, D.C. (Feb. 23, 1991).

(208) See supra note 198.

(209) See infra text accompanying notes 271-75 (describing alternative dispute es olution models); see also EZEKIEL J. EMANUEL, THE ENDS OF HUMAN LIFE: MEDICAL ETHICS IN A LIBERAL POLITY 204-11 (1991); IRVING LADIMER, DEMOCRATIC PROCESSES FOR MODERN HEALTH AGENCIES 149-69 (1979) (detailing mechanics of designing and contracting for arbitration within HMOs).

(210) See, e.g., Jones v. Laborers Health & Welfare Trust Fund, 906 F.2d 480, 481 (9th Cir. 1990) ("Because the Board of Trustees consists of both management and union employees, there is no conflict of interest to justify less deferential review."); Facchina v. NECA-IBEW Local 176 Health & Welfare Fund, 702 F. Supp. 641, 643-44 (N.D. Ill. 1988) (noting that Board of Trustees consisting of other workers and union members met to deny claim). See generally AMA ET AL., GUIDELINES FOR HEALTH BENEFITS ADMINISTRATION 10 (undated, issued in 1989) [hereinafter HEALTH BENEFITS GUIDELINES] (endorsing alternative dispute resolution, including "union representatives, plan benefit managers, and established plan mediators"); LADIMER, supra note 209, at Foreword and Introduction (advocating use of greater subscriber representation on HMO committees that resolve coverage disputes and other matters).

(211) See HEALTH BENEFITS GUIDELINES, supra note 210, at 2, 9 (recommending that utilization reviewers should respond within two business days, allow the attending doctor to talk with the reviewing doctor, and give the patient an opportunity to appeal a denial including the right to a review by another medical consultant chosen by the insurance company).

(212) Additionally, there are mounting efforts spearheaded by medical practitioner groups to regulate utilization review functions through legislation. See Milt Freudenheim, Doctors Press States to Curb Reviews of Procedures' Costs, N.Y. TIMES, Feb. 13, 1991, at A1, D3. To the extent that contracts comply with these explicit regulatory requirements, they should be deemed sufficient, as a matter of law, for purposes of judicial review of public policy and reasonableness considerations. Alternatively, courts might choose to defer to voluntary accreditation standards being developed internally by the industry, in consultation with provider and business groups, as setting a consensus standard of care for utilization review procedures.

(213) See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 110-11 (1989) ("ERISA abounds with the language and terminology of trust law....In determining the appropriate standard of review..., we are guided by principles of trust law.").

(214) Professors Fischel and Langbein provide a trenchant critique of ERISA case law premised on these and similar observations. See Fischel & Langbein, supra note 125, at 1157 ("These [welfare benefit] plans are complex multiparty arrangements, and it was unwise for ERISA to attempt to capture the complex responsibilities of plan fiduciaries by analogy to the simpler world of the private gratuitous trust.").

(215) See RESTATEMENT (SECOND) OF TRUSTS, Subsection 170 (1959) ("The trustee is under a duty to the beneficiary to administer the trust solely in the interest of the beneficiary." (emphasis added)). The leading treatises on trust law are virtually silent on multiple beneficiary problems. Fischel and Langbein suggest the trust law concept of the "duty of impartiality" as a solution but they give little guidance on how this concept applies to resolving conflicts between one beneficiary and the group, as opposed to conflicts among blocks of beneficiaries. See Fischel & Langbein, supra note 125, at 1159.

(216) In Brown v. Blue Cross & Blue Shield, 898 F.2d 1556 (11th Cir. 1990), cert. denied, 111 S. Ct. 712 (1991), the court stated:

"The beneficiary need only show that the fiduciary allowed himself to be placed in a position where his personal interest might conflict with the interest of the beneficiary. It is unnecessary to show that the fiduciary succumbed to this temptation....Indeed, the law presumes that the fiduciary acted disloyally, and inquiry into such matters is foreclosed.... [The] sole purpose and effect [of this rule] is prophylactic." Id. at 1565 (quoting Fulton National Bank v. Tate, 363 F.2d 562, 571-72 (5th Cir. 1966)); see also Fischel & Langbein, supra note 125, at 1114-15 ("The duty of loyalty is prophylactic....The idea is to prevent misbehavior by erecting an irrebuttable presumption of wrongdoing whenever the trustee engages in conflict tainted transactions.").

(217) Courts implicitly recognize this fact by stressing the trustee's fiduciary responsibility to the fund, see supra text accompanying note 124, but they fail to see that this is a bastardization of the fiduciary notion, which usually extends to an individual and allows no play for competing considerations.

(218) See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115 (1989).

(219) See Brown, 898 F.2d at 1566 ("[W]e hold that when a plan beneficiary demonstrates a substantial conflict of interest on the part of the fiduciary responsible for benefits determinations, the burden shifts to the fiduciary to prove that its interpretation of plan provisions committed to its discretion was not tainted by self-interest."). All of this must be decided before undertaking a review of the merits. A non-circular analysis is impossible, however, since the court indicates that part of the assessment of taint includes the quality of evidence and reasoning that the plan administrator used in reaching his decision. See id. at 1564, 1569. Therefore, the court's complex, ten-page discussion of burden-shifting presumptions boils down to the result-oriented rule that the court will use intensive scrutiny if it disagrees with the decision, but will defer if it agrees.

(220) See supra text accompanying notes 139 & 168.

(221) This presumption should not hold, however, in unusual cases where the insurer's decisions are not subject to market discipline, such as where the insurer is liquidating in the process of bankruptcy or has otherwise discontinued selling health insurance altogether. See Fischel & Langbein, supra note 125, at 1132.

(222) See supra note 46.

(223) For example, the Institute of Medicine describes one UR firm whose two physician reviewers held different opinions on which of two procedures is appropriate for a particular condition. Rather than forcing one of the doctors to compromise his clinical judgment, the firm assigned patients to these physicians at random. See INSTITUTE OF MEDICINE, CONTROLLING COSTS, supra note 56, at 85. A more demanding level of review might view this reasonable solution as irrational, but an arbitrary and capricious standard is capable of accommodating decision norms that produce divergent opinions, so long as each choice is itself reasonable.

(224) Many decisions have misapplied the arbitrary and capricious standard in this regard, but more informed analyses under ERISA are beginning to recognize this error. Compare Kunin v. Benefit Trust Life Ins. Co., 910 F.2d 534 (9th Cir.) (conducting a full-length evidentiary trial even though appellate-type arbitrary and capricious standard applied), cert. denied, 111 S. Ct. 581 (1990) with Jones v. Laborers Health & Welfare Trust Fund, 906 F.2d 480, 482 (9th Cir. 1990) ("In making this determination, we review only the evidence presented to the trustees.") and McMahan v. New Eng. Mut. Life Ins. Co., 888 F.2d 426, 431 n.1 (6th Cir. 1989) ("[T]he district court should consider only the evidence that was available to New England at the time of its final decision in this case.").

(225) This does not mean that arbitrary and capricious is a rubber stamp for insurers. They will continue to lose if there is insubstantial basis in the record for the substance of their decision and if the record is inadequate due to procedural deficiencies. Also, in the hurly-burly of everyday claims denial or preadmission certification, it is likely that many cases will exist where the "record" is sufficiently vague that a trial will be necessary to determine precisely what constituted the basis for the decision. Courts should, however, guard against expanding this last exception to summary judgment so that it swallows the rule. Uncertainties in what constituted the record below can always be hypothesized to create an apparent need for trial. See Reilly v. Blue Cross & Blue Shield United, 846 F.2d 416 (7th Cir.), cert. denied, 488 U.S. 856 (1988). The Reilly court found summary judgment inappropriate, even under an arbitrary and capricious standard limited to review of the record, and emphasized that a court must know:

(1) Who made the ultimate decision by Blue Cross that IVF is experimental?

(2) What are their qualifications and on what basis was that decision made?

(3) How many IVF procedures were analyzed to make this conclusion?

(4) What other evidence was reviewed by the decisionmakers which suggested that it was not experimental?

(5) How are the decisionmakers compensated by Blue Cross?

(6) How did this decision affect other Blue Cross health plans? Id. at 424. In most cases, the discovery process should provide sufficient opportunity to reveal these issues and remove the basis for speculation. As Judge Posner observed in Reilly, if the plaintiff has been unable to uncover any actual defects in the processes and qualifications of the insurers' decisionmaker, then there is no basis for a trial or a remand. See id. at 427 (Posner, J., concurring and dissenting).

(226) 906 F.2d 480 (9th Cir. 1990).

(227) The court observed that "[i]n making this determination we review only the evidence presented to the trustees.... We will not upset the review process the parties have bargained for...." See id. at 482.

(228) 857 F.2d 514 (9th Cir. 1988).

(229) Id. at 516. The court also observed that the Plan trustees had "a duty to keep the Fund financially stable ... [and] 'to provide benefits to as many intended beneficiaries as is economically possible.'" Id. at 517 (quoting Elser v. I.A.M. Nat'l Pension Fund, 684 F.2d 648, 656 (9th Cir. 1982)).

(230) 450 N.W.2d 470 (Wis. Ct. App. 1989).

(231) See id. at 475-76. The husband was afflicted with Parkinson's disease, ischemic heart disease, anemia, and carpal tunnel syndrome. The wife had cerebrovascular disease, chronic dementia, hypothyroidism, chronic depression, and hypertension. See id. at 471.

(232) See id. at 476

(233) See, e.g., Brown v. & Blue Cross & Blue Shield, 898 F.2d 1556, 1559 (11th Cir. 1990), cert. denied, 111 S. Ct. 712 (1991); Kunin v. Benefit Trust Life Ins. Co., 910 F.2d 534, 536-37 (9th Cir.), cert. denied, 111 S. Ct. 581 (1990).

(234) See 5 U.S.C. section 706 (1988).

(235) See Baxter v. Lynn, 886 F.2d 182, 187 (8th Cir. 1989); Travelers Ins. Co. v. Blakey, 342 S.E.2d 308, 309 (Ga. 1986) (per curiam).

(236) See Chevron, U.S.A., Inc. v. Natural Resources Def. Council, 467 U.S. 837, 844 (1984).

(237) See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 110-11 (1989). As to precisely how the contract needs to be phrased, see Cathey v. Dow Chem. Co. Medical Care Program, 907 F.2d 554, 559-60 (5th Cir. 1990), cert. denied, 111 S.Ct. 964 (1991) (distinguishing between language that merely authorizes or appoints insurer to act as claims administrator and language that makes the insurer's decision binding).

(238) See Cathey, 907 F.2d at 559-60; Martin v. Masco Indus. Employees' Benefit Plan, 747 F. Supp. 1150, 1153 (W.D. Pa. 1990).

(239) See Firestone Tire & Rubber Co., 489 U.S. at 115.

(240) See Brown v. Blue Cross & Blue Shield, 898 F.2d 1556, 1562-63 (11th Cir. 1990), cert. denied, 111 S. Ct. 712 (1991). This interpretation of the holding is actually far from clear. What the court said is much more complex. It held that, where a conflict exists, the insurer is presumed to have been affected by it, unless it demonstrates to the contrary. See id. at 1565-66. Only if it is successful in rebutting this presumption will the arbitary and capricious standard be maintained. See id. at 1566-67. This complex burden-shifting formulation, however, so confounds the determination of the standard of review with the determination of the merits that it essentially requires the court to review the merits de novo before deciding whether to review them under the arbitrary and capricious standard. See Anderson v. Blue Cross/Blue Shield, 907 F.2d 1072, 1076 (11th Cir. 1990) (describing the Brown standard as similar to de novo review). Not only do these mental gymnastics overestimate the abilities of rational thought, even a court capable of following all these twists and turns necessarily must conduct a factual investigation, which alone defeats a central purpose of having the more deferential standard.

(241) See WALTER GELLHORN ET AL., ADMINISTRATIVE LAW: CASES AND COMMENTS 475-85 (8th ed. 1987).

(242) Id. at 475. For recent examples of inappropriate exercises of "hard look review," see New York ex rel. Bodnar v. Secretary of Health and Human Servs., 903 F.2d 122, 126 (2d Cir. 1990) ("Given the Secretary's second-hand knowledge, we must necessarily demand that his review of the record be probing, precise, and accurate. Our review of the entire record reveals that the Secretary plainly missed items of pertinent weight in the hospital records."); Reilly v. Blue Cross & Blue Shield United, 846 F.2d 416, 427 (7th Cir.) (Posner, J., concurring and dissenting) (noting that the majority "conjure[d] up a host of unanswered questions concerning the qualifications of the members of the advisory committees" in discussing whether in vitro fertilization is "experimental"), cert. denied, 488 U.S. 856 (1988).

(243) Indeed, it may be considered a virtue of the arbitrary and capricious standard that it contains a mechanism well suited to solving the problems of conflict of interest that have plagued ERISA courts applying the less satisfactory trust-oriented model. See Van Boxel v. Journal Co. Employees' Pension Trust, 836 F.2d 1048, 1052-53 (7th Cir. 1987); supra note 240. Also, employed judiciously, this flexibility in the standard allows a necessary safety value for the few heart-rending cases like Rollo that otherwise might undermine public confidence in the judicial system.

(244) See supra text accompanying notes 20-69. Moreover, a stringent application of the arbitrary and capricious standard should not be used as a springboard to the routine award of punitive damages as well. The failure of courts explicitly to recognize that they are exercising hard look review leads them to the fallacy that any erroneous denial of coverage automatically justifies the imposition of punitive damages because the denial has now been labeled "capricious." See supra text accompanying notes 86-87.

(245) See GELLHORN ET AL., supra note 241, at 475-78.

(246) In Rollo v. Blue Cross/Blue Shield, No. 90-597, 1990 U.S. Dist. LEXIS 5376 (D.N.J. Mar. 22, 1990), for instance, the court defined the issue as whether ABMT is proven ineffective, not for cancer generally or for solid tumor cancers or even for cancer of the kidney, but for relapsed kidney cancer in a minor. See id. at *12. Thus, the court was able to conclude that the insurer's review process was "wholly inadequate to the task." Id. at *11-*12. Other courts have used semantics to vitiate the intent of the contract. In Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131 (D.N.J. 1989), although the court found the insurer's review process exemplary, the court was able to end-run the insurer by observing that it had addressed the issue in terms of whether ABMT is "investigational," whereas the contract used the term "experimental." See id. at 137-38.

(247) See Aetna Life Ins. Co. v. Lavoie, 505 So. 2d 1050, 1051 & n.1 (Ala. 1987) (per curiam) (jury awarded $3,500,000 on claim of $1,650; reduced to $500,000 on appeal); Hughes v. Blue Cross, 263 Cal. Rptr. 850, 853, 855-58 (Cal. Ct. App. 1989) ($700,000 punitive damages on a $17,000 claim).

(248) Lavoie, 505 So. 2d at 1052-53. But see infra note 264.

(249) See Hughes, 263 Cal. Rptr. at 857-58. The court objected that the requests did not specify what information the reviewer already had and what he sought. See id. On the other hand, the providers did not bother to call and inquire further. See id. at 855-57.

(250) See Wilson v. Blue Cross, 271 Cal. Rptr. 876, 881-82 (Cal. Ct. App. 1990). This specification does tend to exist, however, in health insurance plans negotiated by labor unions and in HMO contracts. Requirements for federal qualification of HMOs mandate a grievance procedure for resolving all subscriber disputes. See 42 C.F.R. section 417.107 (g) (1991). The same requirement is often imposed through state health department regulation of HMOs. For instance, one Michigan HMO is required to maintain a four-step grievance process culminating in review by the HMO member board. See Interview with Maureen Mangotich, Associate Medical Director of Health Alliance Plan, in Detroit, Mich. (Feb. 23, 1991). See generally LADIMER, supra note 209, at 86-89, 101-122 (detailing legal requirements and contractual provisions for HMO grievance procedures).

(251) This contractual void does not necessarily mean that the parties actually lack an understanding on the procedures to be followed, for these details may be spelled out in informal, extra-contractual documents. For instance, an insurer may retain an independent utilization review firm to issue preadmission certification under procedures specified in the insurer's contract with the UR firm, or the insurer may conduct this review itself according to internal, written operating procedures. These procedures may also simply be understood from prior practice. Nevertheless, since these understandings are not adequately preserved for judicial cognizance, they will likely do little to settle procedural disputes efficiently.

(252) See Varol v. Blue Cross & Blue Shield, 708 F. Supp. 826, 834 (E.D. Mich. 1989) (sustaining Blue Cross prior authorization procedures under substantive due process principles).

(253) See Mathews v. Eldridge, 424 U.S. 319, 343 (1976).

(254) See Richardson v. Perales, 402 U.S. 389, 410 (1971).

(255) See id. at 408.

(256) See Gray Panthers v. Schweiker, 716 F.2d 23, 37 (D.C. Cir 1983) (approving the use of the telephone line in disputes with less than $100 at issue).

(257) See id. at 26-27.

(258) See Schweiker v. McClure, 456 U.S. 188, 198-200 (1982).

(259) See, e.g., Bi-Metallic Inv. Co. v. State Bd. of Equalization, 239 U.S. 441, 445-46 (1915) (dealing with a dispute concerning property tax assessments).

(260) See, e.g., Heckler v. Ringer, 446 U.S. 602, 617 (1984) ("The Secretary's decision as to whether a particular medical service is 'reasonable and necessary' and the means by which she implements her decision, whether by promulgating a generally applicable rule or by allowing individual adjudication, are clearly discretionary decisions.").

(261) See Henry J. Friendly, "Some Kind of Hearing," 123 U. PA. L. REV. 1267, 1278 (1975).

(262) See, e.g., Rollo v. Blue Cross/Blue Shield, No. 90-597, 1990 U.S. Dist. LEXIS 5376, at *9 (D.N.J. Mar. 22, 1990) (denigrating insurer's medical expert as one "who received his M.D. degree 53 years ago,... who has not practiced in almost ten years. and who has no experience in bone marrow transplants or Wilms' tumor").

(263) See Stan N. Finkelstein et al., The Process of Evaluating Medical Technologies for Third-Party Coverage, 1 HEALTH CARE TECH. 89 (1984) (comparing technology assessment by the federal government and private insurers, and concluding that the latter was much less formal and thorough).

(264) See Mordecai v. Blue Cross-Blue Shield, 474 So. 2d 95, 98 (Ala. 1985) ("We reject [the] claim that [the insurer] was under a duty to do more than review the documents claimant submitted to it.").

(265) Jones v. Laborers Health & Welfare Trust Fund, 906 F.2d 480, 482 (9th Cir. 1990).

(266) The chance to correct procedural errors should work both ways in the contract. If a patient fails to apply for preadmission certification, he should still have the opportunity to justify the appropriateness of treatment after the fact. The patient, however, suffers the disadvantage of losing the insurer's advance ruling on coverage and thereby exposes himself to the risk of bearing the responsibility for payment in the event of retroactive denial by the insurer. Also, to encourage the orderly administration of claims, the insurer should be allowed to impose a reasonable penalty (most commonly accomplished via an increased copayment) for failure to obtain precertification, just as the contract might shift some of the patient's procedural costs to the insurer where a court finds that the insurer employed an inadequate process.

(267) See supra text accompanying notes 43 & 120.

(268) See generally SCHWARTZ, supra note 163, at 354-57.

(269) See id. at 357.

(270) See Schweiker v. McClure, 456 U.S. 188, 195-200 (1982) (finding that it was not unconstitutional for Medicare claims administrators to hire their own private hearing officers to resolve disputes, and that there was no evidence these officers had an incentive to deny claims or that their independent judgment was otherwise compromised).

These comments assume a quasi-adjudicative decision. A quasi-legislative decision to define in advance a general category of coverage or noncoverage does not require the same degree of insulation from bias. Because these class-based decisions affect the insured pool more broadly, they are more subject to competitive market penalties if the decision goes astray. Therefore, participation of management personnel should not be absolutely precluded, just as it is permissible for legislators to have a predisposition on the merits of questions they address.

(271) For a comprehensive taxonomy, see STEPHEN B. GOLDBERG ET AL., DISPUTE RESOLUTION 7-15 (1985). One worthy model that has received no previous attention might be called "Solomonic dispute resolution," for it would simply split the difference between the insurer and the subscriber where there is disagreement between the two physicians. Indeed, this is how many prospective review mechanisms work in practical effect, for the only penalty to the patient for ignoring a precertification requirement is the imposition of an additional, modest copayment. See, e.g., Nazay v. Miller, 949 F.2d 1323 (3rd Cir. 1991) (finding no ERISA violation in imposing a 30% penalty on failure to obtain precertification).

(272) Also, this option allows the contract to incorporate a host of complex procedural detail simply by referencing the standard set of rules maintained by the American Arbitration Association. Therefore, it is less cumbersome than a custom-built process. See Arnulfo P. Sulit, Inc. v. Dean Witter Reynolds, Inc., 847 F.2d 475, 478-79 (8th Cir. 1988) (upholding arbitration clause in ERISA pension benefit plan); LADIMER, supra note 209, at 149-50, 173-80 (discussing state laws authorizing arbitration generally, arbitration of malpractice claims, and arbitration by HMOs); cf. Madden v. Kaiser Found. Hosp., 552 P.2d 1178, 1179-80 (Cal. 1976) (en banc) (upholding HMO arbitration of malpractice complaint). But see Wheeler v. St. Joseph Hosp., 133 Cal. Rptr. 775, 783-91 (Cal. Ct. App. 1976) (finding arbitration clause invalid where clause was not brought to hospital admittee's attention when he was signing a "Conditions of Admission" form); Obstetrics and Gynecologists William G. Wixted, M.D., Patrick M. Flanagan, M.D., William F. Robinson, M.D. Ltd. v. Pepper, 693 P.2d 1259, 1260-61 (Nev. 1985) (finding arbitration-of-malpractice-claim clause invalid where patient is required to sign agreement in order to receive treatment). Although the validity of agreements to arbitrate tort-law malpractice disputes may continue to be in doubt, the case is much stronger for disputes arising out of contract, where commercial law doctrine has long recognized the greater freedom to vary contractual remedies by agreement.

(273) See supra note 210 and accompanying text (advocating use of other plan subscribers as arbitrators).

(274) See LADIMER, supra note 209, at 149 ("Arbitration is a complete legal alternative to litigation....[The award] is enforceable in court, but not subject to judicial review except on narrow technical grounds, including fraud, duress, and procedural flaw.").

(275) See supra text accompanying notes 151-67. This approach, however, is not a panacea. Convening a panel in each case is more cumbersome for the insurer or employer than simply vesting the authority with a single medical director or benefits manager, and airing disputes in this fashion also raises problems of patient confidentiality.

(276) Purchasers of insurance may yet determine that the optimal arrangement is for insurers to pay for whatever care within the benefit package doctors and patients decide is in the patient's best interests, limiting the insurer's role to policing for fraudulent billing and designing proper financial incentives to keep the decisions of patients and doctors in check.
COPYRIGHT 1992 University of Pennsylvania, Law School
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:The Law and Policy of Health Care Rationing: Models and Accountability
Author:Hall, Mark A.; Anderson, Gerard F.
Publication:University of Pennsylvania Law Review
Date:May 1, 1992
Words:36346
Previous Article:Just health care rationing: a democratic decisionmaking approach.
Next Article:Professional judgment and the rationing of medical care.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters