A Flood of Litigation? Predicting the Consequences of Changing Legal Remedies Available to ERISA Beneficiaries
In November 1997, the President's Advisory Commission on Consumer Protection and Quality in the Health Care Industry published a report entitled Consumer Bill of Rights and Responsibilities. The report outlines the rights identified by the commission as consistent with high-quality health care, including health plan information availability, provider choice, access to emergency services, consumer participation in treatment decisions, respectful and nondiscriminatory care, confidentiality of medical information, and fair and efficient resolution of complaints and appeals. Despite a flurry of legislative activity regarding patients' rights in recent sessions of Congress, new legislation has not yet been enacted into law.
Rival patient protection bills vary in scope, but one particularly contentious element in many of them relates to the final right delineated in the Consumer Bill of Rights and Responsibilities--that of fair and efficient resolution of complaints and appeals. Specifically, leading proposals contain provisions that would increase the exposure of health plans to civil litigation. These provisions either (1) decrease the federal protection against lawsuits under state law enjoyed by entities administering employer-sponsored health benefit plans, (2) supplement the remedies available to beneficiaries of employer-sponsored health benefit plans under federal law, or (3) do both. Either provision requires amendments to the Employee Retirement Income Security Act (ERISA). ERISA has been interpreted by the courts to limit drastically the remedies available to individuals who want to pursue legal action against employer-sponsored health plans.
Proponents and opponents of expanding the exposure of ERISA health plans to liability have starkly different views about the effects of such changes. The fundamental assertions of each side are summarized in Table 1.
Our strategy is to examine assumptions about litigation and the legal system that underlie the fundamental assertions of proponents and opponents of changing ERISA liability provisions--that allowing ERISA beneficiaries legal remedies against health plans can change the behavior of plans for the better and improve enrollees' health outcomes (proponents) and that changing the remedies will result in a flood of litigation (opponents). We conclude that support for claims on both sides of the debate is weak. We then propose a model for estimating what litigation rates are likely to be in a legal environment that is less constrained by ERISA. The model reveals that most of the data necessary for accurate estimates are not available. Using the model, we show how litigation consequences might differ depending on what assumptions are made about the behavior of key actors.
What Is ERISA and How Does It Affect Patient Rights?
ERISA was enacted in 1974 to protect the pension and welfare benefits that employers provide their workers. It currently covers about 2.5 million health plans and 125 million workers, retirees, and dependents. With the exception of government and church employees and several other miscellaneous categories of employees, ERISA covers all individuals who receive health and/or pension benefits through an employer-sponsored plan. ERISA regulates health and pension plans at the federal level and preempts state regulation of these plans so that companies operating in multiple states face a uniform set of requirements and are free from the burden of complying with different regulations in each state of operation. ERISA specifically provides that a state law is preempted if it "relates to" an employee benefit plan and is not "saved" from preemption by falling within a traditional domain of state regulation, specifically insurance, banking, or securities.
One of the effects of the ERISA preemption of state laws relating to employee benefits has been the protection of employer-sponsored health plans from civil suits challenging benefits administration brought by ERISA-plan beneficiaries in state courts. The scope of the legal immunity has been defined by the federal courts in a series of decisions handed down over the last decade or so. In the key decision, Pilot Life Insurance Co. v. Dedeaux, the U.S. Supreme Court ruled that ERISA preempted an employee's state common law tort and contract claims against his employer-sponsored disability benefits insurer, which he alleged had improperly processed his workplace injury claim. The courts subsequently recognized the same form of preemption in the context of health insurance. The leading case, Corcoran v. United Healthcare, Inc., involved preauthorization review in obstetrics care. Mrs. Corcoran's obstetrician, who was concerned about her pregnancy, sought hospitalization three weeks before her delivery date. Early admission was denied in precertification review, the fetus died, and the Corcorans sued, alleging that an erroneous utilization review decision had resulted in the wrongful death of their child. The Corcorans were covered by employer-sponsored health insurance, and the Court of Appeals for the Fifth Circuit, invoking the ERISA preemption, did not allow the state court action to proceed.
Although barred from bringing state law causes of actions, ERISA beneficiaries may sue in federal court, where their remedies are defined by the act. ERISA provides that plaintiffs may recover benefits due to them, enforce or clarify their rights under their insurance policy, and enjoin any practice that violates either the act or terms of the plan, but the act does not provide for "extracontractual" damages for personal injuries--including compensation for income loss or pain and suffering or punitive damages. In effect, a party who believes him- or herself to be injured by wrongful denial of payment for care may, if he or she proves the case in court, recover the cost of the care itself or enjoin a plan from withholding care contrary to the terms of a policy, but may not recover any of the costs that may have followed from the denial.
Some inroads into expansive ERISA preemption have been made in recent years. In New York Conference of Blue Cross and Blue Shield Plans v. Travelers, the Supreme Court limited the reach of ERISA, suggesting that state laws that had "merely . . . an indirect economic effect" on employee-sponsored health benefit plans were not preempted. This holding has subsequently influenced several health insurance and general ERISA cases, apparently emboldening federal courts confronted by claims that state regulation should be preempted. In general, federal case law over the past three years has become increasingly mixed about ERISA preemption: For example, one court might interpret a state "any-willing-provider" law as preempted by ERISA, while another court rules that this type of law is not preempted.
The boundary of allowable civil actions against ERISA plans has also been tested in recent years. Courts have allowed ERISA plans to be held "vicariously" liable for the malpractice of affiliated physicians; actions have also proceeded against plans for negligent selection and monitoring of their medical personnel. Increasingly, the courts have relied on a distinction between disputes over quality of health care versus disputes over quantity (i.e., authorization of and payment for care), upholding the right of litigants to pursue claims over quality of care in state courts while limiting claims over quantity to federal court and the set of remedies circumscribed by ERISA.
But despite these developments in ERISA law, the act remains a potent force in narrowing opportunities for suits against managed care organizations that serve employer-sponsored health plans. Most of the managed care bills under consideration would change that situation.
In discussions about the consequences of relaxing ERISA restrictions against litigation, decisionmakers and their representatives are often vague about whether their assertions pertain to medical malpractice litigation or some other basis of suit, such as bad faith. It is important to clarify the types of litigation that health plans are likely to face if the ERISA provisions are amended to increase access to state court or expand remedies available in federal court. As currently framed, many of the proposed bills would pave the way for litigation against health plans for delays in or denials of benefits coverage. Such actions may be brought under a tort or contract theory, or, as in the case of "bad faith" claims, some hybrid of the two (Gergen, 1994). Plaintiffs can also be expected to lodge standard medical malpractice claims. There is ample precedent for such claims against managed care organizations (see, e.g., McClellan v. HMO of Pennsylvania, 604 A.2d 1053 (Pa. Sup. Ct. 1992); Harrell v. Total Health Care, Inc., 781 S.W.2d 58 (Mo. 1989)) although a larger volume of suits would no doubt draw into the fray organizational types, such as independent practice associations (IPAs), that have not hitherto been frequent targets of malpractice suits.
Recognition of the variety of claims that may follow amendments to ERISA is important because, while we know a good deal about claiming behavior and outcomes in the medical malpractice context, there has been no research about the dynamics of claiming against institutions for wrongful denial or delay of care authorization or payment. Hence, in predicting what litigation might follow from loosening the ERISA restrictions, it is not certain what previous data on claiming are relevant. Our review discusses what is known about litigation behavior in a variety of circumstances, none of which is exactly apposite to the situation that might evolve if the litigation environment for claims of wrongful denial and delay were to change significantly.
Leading Proposals Changing Legal Remedies Available Under ERISA
The proposals in the 106th Congress speed the process of change begun in courts' interpretations of the ERISA provisions and perhaps move it in more radical directions by directly amending ERISA. Leading proposals under consideration in Congress include the following: the "Patient Protection Act" (H.R. 4250) passed by House Republicans in July 1998 (Goldstein and Eilperin, 1998) and reintroduced in the 106th Congress (H.R. 448); a Senate Republican bill; two Democrat bills; two bipartisan bills, one sponsored by Rep. Norwood (R-Ga.) and the other by Senator Chafee (R-R.I.); and an independent bill sponsored by Rep. Ganske (R-Iowa).
The litigation provisions of the bills vary across three key dimensions. First, they differ in their approach to modifying ERISA preemption of state-law-based actions: Some augment the limited federal remedies allowed under ERISA, while others eliminate the preemption of civil suits against ERISA plans in state courts. The Patient Protection Act leaves the ERISA preemption intact but bolsters remedies under ERISA to include per diem penalties of up to $250,000 for benefits "not timely provided" and in breach of either ERISA or the plans' terms. The Democrat and Norwood bills, in contrast, seek to legislate away ERISA preemption of suits alleging delay or denial of benefits, thus opening the way for state courts to assume, more or less, unfettered jurisdiction.
Second, the bills differ in their prescriptions for the content of allowable remedies. The Democrat, Norwood, and Ganske bills would permit beneficiaries to seek standard common law remedies, including economic, non- economic, and punitive damages--although the Ganske bill prohibits punitive damages for benefits decisions in which the plan timely followed the recommendation of an external review panel. The Chafee bill, which opts for enhancing ERISA remedies rather than lifting the ERISA state litigation preemption, allows economic loss for beneficiaries injured by improper denials of care but draws the line at noneconomic and punitive damages.
Finally, in addition to or in place of creating new common law or statutory remedies, several bills grant the federal government power to impose civil monetary penalties on plan administrators for unreasonable delays or denials of covered benefits. The House Republican bill, for example, allows the Department of Labor to assess fines of up to $100,000 for a "pattern or practice of repeated adverse coverage decisions" in breach of ERISA or the plans' terms. Similarly, the Chafee bill empowers the Department of Health and Human Services to levy up to $250,000 for coverage practices that seriously jeopardize a patient's health.
Evidence on Contending Claims in the Debate over Liability Provisions
Underlying key assertions of proponents and opponents about the consequences of changing ERISA provisions regarding health plan liability are assumptions, often implicit rather than explicit, about the nature of litigation--including the behavior of consumers with respect to bringing claims, the behavior of lawyers whom they might seek to represent them, the outcomes of litigation, and the perceptions of risk associated with litigation. Advocates of expanding liability implicitly assume that private litigation is an effective deterrent mechanism. Opponents implicitly assume that were it not for ERISA, health plans would be awash in costly litigation. Here we examine the evidence for each of these assumptions, drawing primarily on the results of previous research on civil litigation in non-ERISA contexts.
How Effective Is the Legal System in Deterring Wrongful Practices?
The ability of the legal system to deter wrongful behavior rests critically on its effectiveness in screening claims, the quality of outcomes it delivers, and how these outcomes are perceived. Whether the legal system performs effectively as a deterrence mechanism is a matter of sharp dispute.
Does the Legal System Screen Claims Effectively? An efficient legal system would effectively screen claims so that individuals who have good reason to sue bring claims and, conversely, those who do not have good reason to sue are deflected from doing so or, at least, are not rewarded. Contrary to conventional wisdom, many people who have reason to sue do not. The Harvard Medical Practice study found that very few (approximately 3 percent) of the patients whom physician reviewers deemed to be victims of negligence in the course of their hospitalizations brought malpractice claims. To be sure, many of the patients who did not sue were either fully recovered from the injury within six months or were more than 70 years old when the injury occurred (Harvard, 1990). But a recent study of medical malpractice in Utah and Colorado found that only 33 percent of patients who were seriously injured by negligent care filed medical malpractice claims (Studdert et al., unpublished). Anecdotal evidence suggests that medical malpractice plaintiff attorneys are highly selective in deciding what claimants to represent because of the high costs of this specialized litigation and the risk of not obtaining compensation.
Individual behavior in other settings lends support to the empirical findings in the medical malpractice domain. RAND's Institute for Civil Justice (ICJ) conducted the first nationally representative study of claiming by accidental injury victims. Unlike the malpractice studies discussed above, the subset of these injuries that were directly ascribable to negligent conduct was not observed. However, virtually all of the accident victims were potential litigants in the sense that each individual suffered a loss as a result of injury, and there was a potential defendant other than the injured party or a family member in the vast majority of cases.
The ICJ study found that 80 percent of those injured did not even consider trying to obtain compensation for their losses from someone else, and only 10 percent took action to obtain compensation through the liability system. When individuals consulted lawyers about potential claims, about 20 percent were turned away, either because lawyers thought their claims were nonmeritorious or their damages were not worth the costs of litigation. Injured individuals were most likely to seek liability compensation in automobile accidents; they were quite unlikely to seek compensation in the vast number of injuries that occur outside of work and are unrelated to automobiles. These patterns held even when injuries were severe. The rate of legal representation varied dramatically by type of accidents: About 60 percent of individuals with moderate to severe injuries from automobile accidents hired lawyers to represent them. In contrast, when individuals were injured moderately to severely while using products other than automobiles outside the workplace--all of whom could have sued a product manufacturer--only 3 percent retained lawyers. These patterns reflect differences in how individuals attribute causation and blame, which vary dramatically according to circumstances, as well as differences in the costs and rewards of litigating different types of claims (Hensler et al., 1991). These patterns indicate the importance of understanding the context of claiming when trying to predict litigation rates.
Understanding the context of claiming is key when trying to predict litigation rates.
On the other hand, there is evidence that a substantial fraction of liability claims are either nonmeritorious or contain suspect components. That is, a substantial number of plaintiffs have no valid reason to sue or they seek more money than they deserve under the law. Several studies have reviewed medical malpractice claims files to determine the relative frequency of appropriate and inappropriate suits (Harvard, 1990; Cheney et al., 1989; Farber and White, 1991; McNulty, 1989) and found that between half and two-thirds of claims are brought with no apparent indication of negligence. Similarly, an ICJ study on automobile personal injury claims found that between 42 and 50 percent of claims for minor injuries were nonmeritorious or exaggerated (Carroll et al., 1995). By comparing claiming rates in traditional tort states with rates in states that have adopted no-fault regimes, the ICJ analysts further found that the incentive for filing such claims was the potential for securing noneconomic damages such as compensation for pain and suffering.
Are Legal Outcomes Appropriate? The ability of private litigation to discourage injurious behavior by health plans depends at least in part on the ability of courts to distinguish valid from nonmeritorious claims, to require payment only for the former, and to compensate the injured at levels appropriate to the harm incurred as well as the defendant's level of responsibility for those harms. Studies differ in their findings with respect to just how appropriately and consistently courts compensate claims. Some studies find that less serious claims are overcompensated by the courts and more serious claims are undercompensated with the most serious claims being the least well compensated (Carroll et al., 1991; Conard et al., 1964; King and Smith, 1988). Jury verdicts for cases with similar fact patterns vary dramatically across jurisdictions (Moller, 1996). Evidence on medical malpractice outcomes shows similarly mixed results. Once a claim is made, the presence of negligence increases both the probability and average level of payment (Taragin et al., 1992; Sloan et al., 1990; Farber and White, 1991). Severity of injury plays a strong role in determining the liability, however, and may even overshadow the effect of the presence or absence of negligence in a significant number of cases (Brennan et al., 1998).
Compensation levels are affected not only by the losses suffered by the claimant but also by claimant characteristics, attorney skills, and local legal culture. Defendants' payments are also affected by their responsibility for claimants' losses, as well as by various legal doctrines and statutes that may either increase or limit defendants' exposure without regard to their responsibility. Most claims are settled, not tried, for amounts that reflect not only their legal worth, but also the costs of litigation, the parties' aversion to the risk of taking a suit to verdict, and the availability of insurance. The divergence between outcomes, on the one hand, and claimant losses and defendant responsibility, on the other, interferes with the deterrent signals of the liability system.
How Accurately Is Litigation Risk Perceived? One important factor in the reaction of plans to litigation or to the threat of litigation is their expectations about that litigation. Without a history of claims on which to base a prediction, it is difficult to quantify the potential threat. Even insurers skilled at assessing risk may find it difficult to price policies in light of expanded liability exposure with little relevant empirical information to guide them. Over time, with observations on litigation rates, predictions about the frequency of future claims will be easier to calculate.
Even when sound statistical probabilities are available, however, decisionmakers may overestimate the risk and costs of litigation. Anecdotal data suggest that decisionmakers' estimates of litigation exposure are often derived from casual interactions with colleagues and others as well as from media reports of court cases. Empirical research comparing media reporting on litigation outcomes with actual outcomes shows that journalists are more likely to report plaintiff victories than defense victories, more likely to report cases that result in large dollar outcomes for plaintiffs, and more likely to report cases in which punitive damages are awarded (Garber and Bower, 1999; Bailis and MacCoun, 1996). Hence, individuals are likely to over- estimate the likelihood that plaintiffs will win cases that they bring to trial and that juries will award large damages. Indeed, the Harvard Medical Practice Study (Harvard, 1990) found that physicians overestimated the risk of being sued for malpractice by 200 percent, even despite the far greater incidence of negligence compared with claims.
Moreover, even when decisionmakers are informed about the distribution of litigation outcomes, decisionmakers may overreact to the small possibility of having to pay large penalties for certain behavior. According to organizational behavior experts, corporate decisionmakers often base decisions on worst-case scenarios, rather than on expected-value calculations (i.e., statistical averages) (Garber, 1998). So a small number of lawsuits over wrongful denial or delay of health care authorization or payment could lead health plan administrators to authorize unnecessary care even when such reactions are not justified by objective information.
In sum, the deterrent effect of new litigation against health plans would be imperfect.
Summary of Evidence on the Effectiveness of the Legal System in Deterring Wrongful Practices. Empirical evidence suggests that the legal system's capacity to deter wrongful behavior is weak. Proponents of relaxing ERISA restrictions against litigation argue that allowing ERISA beneficiaries legal remedies against health plans can improve plans' behavior and enrollees' health outcomes. But the results of previous research suggest that the behavior of claimants, lawyers, and court actors, combined with health plan decisionmakers' expectations regarding that behavior, may interfere with the deterrent effects of new litigation. Our findings are not at odds with other studies that have found moderate deterrent effects of the tort system (e.g., Schwartz, 1994). The threat of litigation may cause health plans to make better decisions, i.e., to approve more medically necessary care. But because some claimants bring nonmeritorious claims that are not always (or costlessly) weeded out of the court system while others with meritorious claims do not come forward, and because the costs of litigation for defendants do not always accurately reflect claimants' true losses and defendants' true degree of responsibility, health plans may react by changing appropriate behavior as well as by not changing inappropriate behavior. Health plans may deny some care that should be approved because they believe that a particular patient is unlikely to file a claim or that, if a claim is filed, it will be turned away or undercompensated by the legal system. Moreover, health plans--fearing that a valid health care decision will be called into question and possibly lead to large litigation costs, or not knowing if a health care decision will be seen as valid in a court of law--may be more likely to approve unnecessary medical care. The provision of more services than would have been provided in the absence of the change in liability, solely for the purpose of deterring litigation, is sometimes called "defensive" behavior. The idea that plan administrators would engage in approvals of unnecessary care to avoid litigation is reminiscent of assertions about defensive medicine practices by physicians concerned about medical malpractice litigation (U.S. Congress, OTA, 1994). Whether and how much defensive practice would arise is uncertain. In sum, the deterrent effect of new litigation would be imperfect: Decisions that cause injury may be deterred, but not all the time or in all cases; at the same time, health plans may react defensively by approving unnecessary care.
What Factors Influence the Volume of Litigation?
Those in favor of maintaining the legal immunities of ERISA health plans assume that ERISA is the main barrier standing between health plans and a flood of costly litigation. This assumption rests on beliefs about the size of the pool of potential claims for wrongful denial and delay, the responsiveness of laypersons and lawyers to the availability of new remedies, the likely character and outcomes of new litigation, the presence or absence of other barriers to litigation against health plans, and how willing courts and legislators are to curb runaway litigation.
The Pool of Potential Claimants. The absolute number of liability claims and lawsuits in any domain is a function of the rate of claiming and the size of the pool of potential claimants. However exaggerated automobile accident victims' claims may be, relatively few claims are filed without an accident having occurred (Carroll et al., 1995). Similarly, hundreds of thousands of women could not have come forward to claim compensation for alleged injuries associated with the Dalkon Shield or silicone gel breast implants had not millions of women been fitted with these devices (Hensler and Peterson, 1993).
The pool of potential claims for wrongful denial or delay of authorization of health care is defined, initially, by the number of instances in which payment or coverage is questioned by a health care administrator (e.g., in utilization review), a decision is made not to authorize that care or payment, and the patient who experienced a denial subsequently experiences a bad health outcome, which he or she attributes to that denial. Notwithstanding the concern expressed in the policy arena about likely litigation over wrongful delay or denial of health care authorization or payment, information about the magnitude of the pool is fragmentary. There is no available information about the relationships between denial of care or payment and health outcomes. Therefore, in our research, we sought information about
- the number of utilization review requests
- the percentage of requests denied
- the percentage of denials challenged through internal appeals processes
- the percentage of denials sustained after such challenge
- the percentage of denials challenged through an external review process (when that is in place)
- the percentage of external review challenges sustained by the external reviewers.
We construct the estimates of the rates at which individuals pass through the appeals and grievance system from two main sources: data from state agencies that require health plans to report on various aspects of their business and information from state agencies that assist consumers with complaints against their health plans. The information that states require managed care organizations to report varies widely. Some states have stringent reporting requirements, while others require only minimal data. In addition, the data that managed care organizations report are publicly available in some states but not in others. We did not contact every state to find out what data were collected and available, but rather collected data from a sample of state agencies reportedly more active in their monitoring of managed care organizations. Likewise, not all states have publicly available systems for health care consumers' complaints and not all states that have such systems report the results of the reviews. Selection issues may affect how representative the data are. For instance, it may be the case that states are more active in their monitoring because they experienced more problems with their managed care organizations compared with other states, or it may be that a state's close monitoring of managed care deters behavior by health plans that causes complaints.
Preauthorization Review. Utilization review (UR) generally occurs either prospectively (preauthorization review) or retrospectively for payment of medical services already received. Our data sources did not provide us with any information about the number of retrospective payment reviews conducted or the rate at which such reviews were approved. Thus, our figures reflect only preauthorization review requests. Among five plans reporting to the Texas Department of Insurance (TDI) and ranging in size from approximately 32,000 enrollees to over 230,000 enrollees, the median rate of preauthorization requests was approximately 1 for every 2 enrollees, and the median percentage of such requests denied was 3 percent, yielding a rate of initial denial of 1,500 per 100,000 enrollees (TDI, 1997).  Not inconsistent with the 3 percent denial rate reported by plans in Texas, a survey of about 2,000 physicians (Remler et al., 1997) revealed a first-round denial rate of recommended care of 2 to 6 percent across different services.
Internal (Within Plan) Appeals. The taxonomy used to describe individuals' disputes with plans is imprecise, and as a result the data collected by state agencies are imprecise as well. The Minnesota Department of Health (MDH) asks plans about their average "complaint" rate, which may include anything from an appeal of denied care to a problem with someone in a doctor's office being rude (considered by some a "grievance"). In contrast, in Texas, plans are asked specifically about appeals. The distinction is important because grievances carry a different potential for litigation than do appeals of plan decisions. In addition, different types of appeals carry different potential for litigation: Pre-service denials of care are more likely to result in a case with the potential for punitive or compensatory damages than are post-service payment refusals because of the greater likelihood of injury if care is delayed or not received. However, the data do not allow us to distinguish between these types of appeals. The median rate of written complaints received among four large plans reporting to the MDH was approximately 300 per 100,000 enrollees. About two-thirds of enrollees had their complaint settled "to their satisfaction," leaving 100 per 100,000 enrollees dissatisfied with the outcome of their written complaint (MDH, 1996). In Texas, among nine plans ranging in size from 32,000 to 330,000, the median rate of appeals filed was 320 per 100,000 enrollees and the median rate of denied appeals was 80 per 100,000 enrollees (TDI, 1997). 
External (State Agency) Appeals. About one-third of states, including Texas, Minnesota, and California, have a system in which managed care enrollees can seek assistance if they are unhappy with a decision made by their plan. The remedy available through the state system is typically limited--it does not include any punitive or compensatory damages for the individual. Rather, the individual might receive care that the state decides should have been authorized or receive reimbursement for care that was obtained. In addition, the decision of the state agency is not necessarily binding on the plan, which is the case, for example, in California--although most plans reportedly comply with the state's recommendations. The MDH reports a 1996 average annual rate of closed (resolved) medical and nonmedical (e.g., timeliness of claims processing) complaints of 40 closed complaints per 100,000 enrollees (MDH, 1996). The TDI reports a 1997 average of 30 complaints per 100,000 enrollees (TDI, 1998), and the California Department of Corporations' (DOC) data show a rate of request for state assistance of, on average, 10 per 100,000 enrollees (California DOC, 1997). Less information is available about the decisions made in external reviews compared with internal reviews. New Jersey instituted an independent utilization review organization (IURO) in 1997. Consumers were new to the system, so the rate of use by health care consumers was relatively low, but the decisions of the IURO are informative: In its first year, the IURO upheld the health plan's decision in about two-thirds of cases and, conversely, overturned the plan decision in favor of the consumer in about a third of cases (New Jersey Department of Health and Senior Services, 1998). In addition, the Center for Health Dispute Resolution (CHDR) reports that approximately two-thirds of their decisions uphold plans' decisions regarding care for Medicare enrollees (CHDR, 1998).
Estimates of the Potential Pool of Claimants. In theory, any or all individuals who are initially denied care in the UR process and who sustain an injury that they attribute to health care denial could potentially litigate. The pool of those who are denied care is about 1,500 per 100,000 enrollees, or about 1.9 million enrollees annually (based on the current population of 125 million ERISA enrollees).  However, many of these denials involve relatively routine decisions about coverage that are inconsequential in the sense that they do not result in adverse health outcomes; the possibility that such denials will support a lawsuit is remote. Almost no data are available about the profile of UR cases; however, a 1991 study of external appeals in the Medicare managed care system found that nearly 60 percent of cases reviewed at that level involved in-area emergency services and out-of-area urgent services. Twenty-seven percent of cases involved denials for inappropriate use of out-of-plan services (Richardson et al., 1993). Although the profile of internal plan disputes for the wider enrollee population may well differ, these findings suggest that many denials are related to contractual matters, such as services not covered, and explicit policy exclusions. In the main, such denials are unlikely to provide a foothold for litigation, either because they are not accompanied by sufficient patient injury, if at all, or because there is no reasonable basis for attributing fault to the plan.
Given these considerations, the figure of 1,500 per 100,000 enrollees who are denied care in initial UR almost certainly overestimates the fraction of enrollees who are realistic candidate litigants regarding a plan's coverage decisionmaking behavior. What fraction of population denied care at UR is an appropriate upper-bound estimate for potential litigants? As noted above, approximately 300 per 100,000 enrollees challenge the plan's initial decision through the internal appeals systems; the remaining enrollees do not. The Medicare data suggest that approximately 75 percent of disputed denials involve decisions unlikely to provide a strong basis for litigation. An assumption that one-third of all denials (100 per 100,000 enrollees) have such a basis is thus a conservative estimate.
While only a subset of enrollees who internally appeal in the current environment can be expected to sue in a new legal environment, it is also the case that some who litigate may come from the pool of patients denied care in UR who do not currently appeal (1,200 per 100,000 enrollees). It is logical to assume, at least for purposes of population estimates, that failure to appeal is correlated to some extent with a more limited basis for suit. Hence we estimate that 20 percent of nonappealers represent potential litigants (i.e., roughly one-half the proportion we estimated among enrollees who appeal). Again this estimate is conservative, but it is probably appropriate as a generous upper bound, given the limited information available. These additional assumptions about nonappealers yield an additional 240 per 100,000 (0.20 * 1,200) potential litigants. In sum, therefore, we calculate a high-end estimate of the pool of potential litigants at 340 per 100,000, or 425,000 potential litigants annually.
However, plans may stipulate in their contracts with enrollees that the internal appeal system be exhausted before a claim can be brought. At the same time, an "exhaustion" requirement could be legislated. If individuals are required to exhaust the internal appeal system, the potential pool of litigants is likely to be much smaller than the upper-bound estimate of 425,000 and to come from the set of individuals who appeal but whose denial is sustained, about 90 per 100,000 enrollees, or 112,500 enrollees annually.
In addition, because litigation is a costly process, it may behoove individuals to exhaust any available state review process before litigating although the damages available through the state process are often limited. Alternatively, legislation could require individuals to exhaust both internal and external review processes. At the same time, the set of individuals who continue to pursue recourse for denied or delayed treatment may be indicative of how many individuals will seriously pursue legal action. All of these factors suggest a somewhat smaller potential pool of litigants. Approximately 30 per 100,000 enrollees seek external review and about 20 per 100,000 have their denial sustained by the external review agency, resulting in a pool of potential litigants numbering between 25,000 and 37,500.
All of these estimates of the pool of potential claimants are based on data from health plans operating in the current environment in which legal remedies under ERISA are restricted. But were ERISA restrictions relaxed, health plans may change their behavior with respect to decisions about UR requests or appeals, and physicians and consumers may change their behavior with respect to requests for UR or appeals. In particular, plans may implement defensive practices such as lowering the threshold for approval of UR requests or appeals. Such behavior would be expected to decrease the potential pool of litigants. In response, however, once physicians and consumers realize that a request or appeal is less likely to be denied, they may be more likely to submit UR requests or to appeal denials of care, a sequence of events that will result in a larger pool of litigants. The magnitude and net effect of these changes in behavior is difficult to predict a priori, and we are not able to account for them in our estimates of the pool.
Another behavioral response--one that our estimates are sensitive to (see Table 2)--warrants mention. The pool of potential litigants would, in theory, be expected to decrease substantially if legal exposure prompts the introduction of rules that require enrollees to exhaust internal appeals processes before being eligible to bring a lawsuit. Plans themselves may introduce such exhaustion provisions, or legislators may be moved to do so. However, policymakers should be attuned to the one possible outcome that mandated use of internal appeal processes may stimulate: Exhaustion provisions could provide health plans with incentives to deny more care at the initial UR stage and then to act readily to overturn the denial when it is appealed.
Other examples of a litigation response to changes in doctrine include the increase in personal injury product liability cases in response to changes in product liability doctrine, the response of medical device users to the availability of tort damages under class action and other aggregative settlements, and the rise of sexual harassment claims in response to changes in legal doctrine regarding harassment and in the availability of punitive and compensatory damages. The dramatic growth in personal injury product liability suits from 1975-1987, in particular in the late 1980s, has been attributed to the product liability regime that emerged in the previous decades, including the product design defect doctrine and the application of defect tests to hazard warnings (i.e., the idea that inadequate hazard warnings are a type of design defect) (Viscusi, 1991).
Media attention to managed care issues may promote lawyers' and enrollees' interest in bringing claims against health plans.
Mass tort litigation has frequently resulted in the creation of large settlement funds, whose availability is widely advertised to potential claimants. When Dalkon Shield users were informed of the potential for compensation under a bankruptcy agreement, more than 200,000 women and their family members came forward; prior to seeking the protection of the bankruptcy court, the Dalkon Shield's manufacturer, A. H. Robins, had been sued by only about 20,000 people. Similarly, when women with silicone gel breast implants were invited to claim compensation under a proposed class action settlement, more than 400,000 filed intentions to claim; the parties to the settlement had expected a number one-quarter this size (Hensler, 1995).
In 1986, the courts recognized sexual harassment as a type of discrimination claim, and, in 1991, Congress allowed punitive and compensatory damages for employment discrimination cases brought under federal law. From 1992 to 1997, the annual number of sexual harassment claims increased from 10,500 to 15,600, and monetary settlements of claims, in the aggregate, increased from $12.7 million to nearly $49.5 million (U.S. EEOC, 1998). Not all of the increase in cases can be attributed to the changes in remedies available, however. For example, the increasing presence of women in the workforce is likely to have contributed to the rise.
The degree of responsiveness of laypersons and lawyers to changes in the availability of legal remedies depends on a number of factors. Media attention, lawyer advertising, and social networks facilitate dissemination of information about the availability of remedies to laypersons and also may convey the notion that suing is an appropriate option in certain circumstances (Hensler and Peterson, 1993). We would expect that the high level of media attention to managed care issues, coupled with consumer health care advocacy, would promote interest and willingness to bring legal claims against health plans.
Lawyers' calculus about whether to invest in learning a new line of litigation includes assessment of the financial value of the available remedies, litigation costs (including the development of intellectual capital to handle new types of cases), and the perceived risk associated with claims. The degree of responsiveness among lawyers is thus likely to be affected by how quickly information about remedies and strategies for bringing suits is disseminated through educational workshops, newsletters, and the Internet.
How lawyers would respond to loosening ERISA restrictions on suits over wrongful denial or delay is not clear. We interviewed a small, nonrandom sample of leading health law practitioners who represent plaintiffs in medical malpractice litigation and found that their interest in pursuing health plans was low. But recent publications aimed at trial lawyers evidence some interest in litigating against health plans (Perez, 1997; Liggio, 1997). If new state-law-based litigation produced high awards--for example, if punitive damages were available and juries evinced a willingness to award such damages--then lawyer interest would likely build more swiftly and spread further through the plaintiffs' trial bar. However, if claims were difficult to sustain in court, if punitive damages were not available or were sharply limited, and if juries proved unwilling to award large damages except in exceptional cases, then fewer lawyers might be interested in developing the expertise needed to litigate ERISA plans, and litigation would probably grow more slowly.
Character and Outcome of Suits. For the most part, the debate over loosening ERISA's restrictions against litigation has focused on individual suits. Opponents of amending ERISA are concerned that individual litigants would obtain high awards from juries faced with claims that health plan enrollees had been seriously injured by plans' administrative decisions. The specter of massive indemnity awards in suits against managed care organizations, health plans, and other prospective defendants is not without foundation. Several large, high-profile awards in recent years have created apprehension among health plans. For example, in 1993 a jury in Southern California awarded almost $90 million to the estate of Nelene Fox after her insurer, HealthNet, failed to approve autologous bone marrow transplantation and high-dose chemotherapy treatment for breast cancer (Eckholm, 1993). The case was subsequently settled for an undisclosed sum during the course of an appeal. More recently, a $120 million verdict was awarded against Aetna-U.S. Healthcare to the widow of David Goodrich (Johnston, 1999). Both cases involved non-ERISA enrollees and have been touted by opponents of loosening ERISA as representative of the kind of retribution that managed care organizations will face en masse in a less constrained liability environment.
The bulk of the verdicts in Fox and Goodrich--90 and 97 percent, respectively--are attributable to punitive damages. Punitive damages have occurred relatively infrequently in medical malpractice litigation (Rustad, 1998). Leading commentators have surmised that juries find physicians to be fairly sympathetic defendants (Rustad and Koenig, 1995), and physicians do not have great personal wealth compared with corporations even despite the quite comprehensive insurance policies most have. Managed care organizations, on the other hand, more readily fit the profile of classic punitive damages defendant (Polinsky and Shavell, 1998). Cases involving intentional torts, business decisions, contractual breaches, and financial injury are the most frequent contexts for punitive damages awards (Daniels and Martin, 1991). A RAND study found that about one in seven jury verdicts in cases alleging financial injury included punitive damages, averaging more than $5 million (Moller, et al., 1997). New opportunities for suits against health insurers raise the possibility that types of claims more closely associated with punitive damages awards, such as bad faith claims, will emerge as preferred ways to frame an action.
Defendants are particularly fearful of punitive damage exposure: Because punitive damage awards are unlimited in some jurisdictions, the risk of punitive damage exposure may drive defendants to settle cases that they would otherwise vigorously contest and--when they settle--to pay larger amounts in compensation than they would otherwise agree to. Large punitive damage awards--and defendant willingness to pay a "premium" in cases where they fear punitive damages--are likely also to stimulate more litigation (McGovern, 1989). Hence, the cost of punitive damages is observed both in the price to settle claims where punitive damages are at issue and in the frequency of litigation that ensues. As a result of these concerns, some of the bills proposing ERISA amendment place a cap on or prohibit punitive damages. 
Less attention has been paid to the question of how loosening ERISA restrictions against litigation might affect the potential for class action litigation. Would individuals band together to bring consumer class actions against health plans? The most likely occasions for class actions would be claims that enrollees had suffered a financial injury, for example, as a result of delays in reimbursement for care. Consumer class actions alleging wrongful practices in financial transactions appear to have increased dramatically in the last few years. Such suits may yield multimillion-dollar settlements and substantial litigation costs for defendants even when the underlying claims for individual losses are small (Hensler et al., forthcoming). However, several bills under consideration preclude claims based on financial injury and, thus, nullify the possibility of financial injury class actions. Whether enrollees could succeed in certifying class actions over personal injuries is highly uncertain. Where injuries are diverse and occur in vastly different circumstances, the mere fact of multiple claimants alleging that the cause of injury was wrongful denial would seem an unlikely basis for class certification. But if enrollees alleged a pattern of practice in breach of contract or violation of general statutory or regulatory codes that increased injury rates, they might be able to secure class status in some jurisdictions, absent a prohibition on such suits.
Current state law offers a variety of protections to health plans.
Barriers to Litigation. Current state law offers a variety of protections to health plans against being drawn into medical malpractice litigation. For example, state "corporate practice of medicine" laws have been interpreted to protect managed care organizations against medical malpractice litigation (Havighurst et al., 1998). In addition, many states allow "hold harmless" or indemnification clauses that prohibit suits between providers and plans. These may also provide protection to health plans against being drawn into malpractice litigation although a number of states have recently invalidated these clauses (Rothouse, 1998).
A substantial number of states have adopted general "tort reform" measures, including limitations on joint and several liability, caps on noneconomic damages, and caps on punitive damages. State-law-based suits against health plans would be subject to such limitations in jurisdictions that have adopted them. Moreover, in every jurisdiction, punitive damages are subject to reduction by trial and appellate court action; previous research indicates that, on average, only half of the dollars awarded in punitive damages are ultimately paid to plaintiffs (Moller et al., 1997). While not absolute barriers to litigation, such legal rules make litigation less attractive by reducing the benefits that accrue from litigation.
In addition, most states permit the use of binding arbitration in contracts between health plans and consumers, and many health plans include such clauses in their contracts. The ERISA preemption of state-law-based claims appears to have limited the use of arbitration in disputes between plans and consumers, but if exposure were to increase, plans might activate such clauses. Generally, arbitration--which relies on party-selected neutrals, rather than juries to decide disputes--is thought to eliminate extreme awards, hence diminishing somewhat the attractiveness of claiming (Rolph et al., 1997). In some other contexts--for example, banking--arbitration clauses have been drafted to include a waiver of rights to bring class actions.
Judicial and Legislative Responses to Runaway Litigation. Fears that litigation will surge in the wake of any change in ERISA should be tempered by previous responses to sharp changes in claiming rates that were made after shifts in legal doctrine. For example, in 1988 the California Supreme Court took action to stem insurance bad faith litigation,  which had burgeoned following a 1979 decision that allowed claimants to bring state-law private causes of action for bad faith torts under the state Unfair Trade Practices Act. Similarly, in 1988 the California Supreme Court sharply restricted bad faith liability for wrongful termination claims after wrongful termination lawsuits surged in the wake of an earlier court decision.
In addition to court-driven responses, legislators have also responded to contain litigation. Many states passed medical malpractice tort reform packages in the mid-1970s in response to dramatic increases in the frequency and severity of medical malpractice claims at that time--claim costs were increasing by up to 40 percent a year in some states. In response, many states enacted limits on the total award or some component of it, imposed collateral source offsets in the assignment of tort awards, provided for periodic rather than lump-sum payment of awards for future damages, and revoked a plaintiff's right to name a specific dollar amount in his or her complaint (Danzon, 1985).
The impetus for general tort reform measures was an increase in the frequency and value of certain types of lawsuits in the 1980s and early 1990s. Since 1995, more than 30 states have passed tort reform laws, and 23 states now have caps or prohibitions on punitive damages (Glaberson, 1999; Firestone, 1999).
Although past experience suggests that courts and legislators do act to curb runaway litigation, such responses may not occur for a number of years. In the meantime, defendants incur litigation costs, which they typically pass on to others, and may change their behavior so as to avoid litigation. As we have discussed, such changes may or may not be socially beneficial.
Summary of Evidence on Factors Influencing the Volume of Litigation. Based on empirical results of previous research on civil litigation, it is difficult to predict what the magnitude of litigation against health plans would be without the ERISA restrictions. Utilization review touches many health plan enrollees, more so when care is tightly managed. But the available evidence--while fragmentary--suggests that only a small fraction of requests for care are denied and that a significant fraction of those requests are ultimately authorized. The fraction of enrollees who suffer a significant harm when benefits are denied is unknown. If health care enrollees behave as Americans appear to in personal injury and medical malpractice domains when they have been injured, then litigation rates could be very modest. But to the extent that plans are the subject of widespread criticism, that the increased availability of remedies is widely publicized, and that lawyers perceive that suits against plans are likely to yield substantial financial rewards, litigation will grow. To win lawsuits, lawyers will have to overcome legal barriers other than the current ERISA restrictions, and enrollees who wish to claim may find that they are restricted to arbitration. If, nonetheless, there is a surge of costly litigation, courts and legislatures may respond by imposing restrictions on claims or damages. But such responses would probably take some years to emerge, and in the meantime health plans would respond to their increased exposure to litigation in ways that might or might not benefit health care consumers.
Estimating the Magnitude of Litigation
Despite uncertainty about the reactions of lawyers, consumers, health plan administrators, and other players to changes in ERISA, prior studies have sought information useful for understanding the potential for litigation if ERISA restrictions on litigation were loosened. Coopers and Lybrand (1998b) analyze litigation rates among individuals not covered by ERISA, focusing on lawsuits arising from benefit denial or delay. The study describes litigation rates among individuals covered by the California Public Employees Retirement System (CalPERS), employees of the State of Colorado, and employees of the Los Angeles Unified School District (LAUSD). The authors find annual litigation rates ranging from .3 to 1.4 cases per 100,000 enrollees across the different groups. If these litigation rates prevailed across the entire ERISA-covered population, then, in the wake of loosening ERISA restrictions on litigation, we might expect about 375 to 1,750 new lawsuits annually. However, the grievance and appeal systems available to the employees and retirees in the three groups in the study may be different from those available to the majority of individuals covered by ERISA. CalPERS, for example, has an extensive external appeals process, and individuals in CalPERS are required to exhaust the internal and external appeals systems before turning to civil litigation. Many plans may choose to stipulate use of the appeals system as part of their health care contracts with ERISA beneficiaries if ERISA were changed or exhaustion could be legislated, but it is possible that individuals will not have to exhaust internal remedies before litigating. Differences such as these are likely to translate into differences in litigation rates.
Coopers and Lybrand provide further evidence on likely litigation rates based on a single health care plan's experience with litigation before the Pilot Life decision (see the earlier discussion of Pilot Life Insurance Co. v. Dedeaux) and with recent litigation among individuals who purchase their own health insurance. Prior to Pilot Life, the health care plan had litigation rates of 3.2 per 100,000 enrollees annually, while the plan reported 9.3 lawsuits annually per 100,000 members enrolled in self-bought plans in the mid-1990s. Generalizability is an issue for both of these estimates because they are based on only one managed care organization, and managed care organizations vary widely in their use of utilization management, in the accessibil ity of their internal appeals process, in their handling of appeals, and in other factors likely to affect claiming rates. In addition, those covered by individual insurance may not be comparable to the population of individuals covered by ERISA. For example, limits on coverage in individual plans are likely to be lower than in group plans, and those covered by individual insurance are not likely to have access to an employer-sponsored external appeals process. Finally, the pre-Pilot Life litigation rates reflect claiming behavior in a different era. With the passage of time, individuals' litigation behavior may well have changed in ways unrelated to the court decision.
A Behavioral Model for Estimating Litigation Rates
To provide additional perspective on what litigation rates might be in an environment in which ERISA litigation restrictions were relaxed, we developed and estimated a behavioral model of claiming. Our model pertains to litigation over wrongful delays or denials of health care authorization or payment. We delineated the steps through which a suit arises. Using data discussed in the previous sections, we adopted "best estimates" of the number of health plan enrollees who would pass through each of the steps.
The six stages of the process out of which a suit may arise are outlined below.
- The health plan conducts utilization review or retrospective review.
- The benefit or coverage is denied.
- The doctor or patient challenges the plan decision.
- The health plan upholds the denial.
- The patient believes him or herself injured and seeks legal advice.
- A lawyer agrees to take the case.
In estimating litigation rates, we make different assumptions about the stages at which enrollees would decide to seek legal assistance. Consistent with our earlier discussion of how the pool of potential litigants may be constituted, we assume in our high-end scenario that some fraction of individuals who appeal initial denials in the current environment would bring a claim, and that litigants will also emerge from among the subpopulation of patients who do not currently contest adverse utilization review decisions through plan appeals mechanisms.
However, we have also discussed the possibility that regulations or plans themselves, in their contracts with enrollees, may stipulate that the internal system be exhausted before a claim can be brought. Thus, we also provide a set of estimates for litigation rates assuming that enrollees exhaust the internal appeals system. In a third litigation scenario, we maintain an exhaustion assumption and assume that the potential pool of litigants is approximated by the set of individuals denied care after pursuing the external appeal process. This set of estimates is considered for several reasons: first, because litigation is a costly process and individuals may seek it as a last resort; second, because legislation could require individuals to exhaust both internal and external review processes; and third, because the size of the group willing to pursue recourse for denied or delayed treatment to the level of external review may in fact be indicative of the number of individuals willing to invest the time and energy necessary to pursue legal action in a changed environment.
Our best estimates about individuals' behavior in Stages 1-4 are derived from the data, described earlier, from state agencies that require health plans to report on various aspects of their business to the agency and information from state agencies that assist consumers with complaints against their health plans. Based on these data, there is about one preauthorization request for every two enrollees per year. Nearly all of those UR requests are approved (97 percent), but the 3 percent that are denied result in about 1,500 denials for every 100,000 enrollees. Some managed care enrollees who are denied care go on to the plan's internal appeal process. In addition, some individuals who are retrospectively denied payment for care may appeal the managed care plan's decision in the internal appeal system. All told, about 300 appeals occur for every 100,000 enrollees. In just over two-thirds of internal reviews, care or payment is approved so that about 90 internal denials per 100,000 enrollees are sustained. About 30 out of every 100,000 managed care enrollees seek state assistance, and in about two-thirds of the cases (20 per 100,000 enrollees), the external review upholds the plan decision.
We used information from RAND and Harvard studies of claiming to estimate behavior in Stages 5-6. In estimating behavior at different stages, two critical assumptions were necessary: (1) the fraction of appeals or denials in which enrollees would perceive that they had incurred an injury or loss deserving a remedy and seek legal assistanceand (2) the stage of the process at which enrollees would exit to seek legal assistance. We estimate litigation rates under nine different scenarios, assuming alternatively that 50, 10, and 5 percent of denials result in enrollees consulting lawyers and that enrollees seek legal representation when denial of care or payment is first challenged, when a denial is sustained after internal plan review, and when a denial is sustained after an external review. In each scenario, relying on RAND's study of claiming among accidental injury victims (Hensler et al., 1991), we assume that 80 percent of those seeking legal advice retain lawyers.
Table 2 shows estimated litigation rates for the nine scenarios.
The estimates of the number of new cases filed include litigation over wrongful delays or denials of health care authorization or payment but do not include medical malpractice cases in which health plans are likely to be brought in as defendants. However, managed care plans may be held increasingly liable in medical malpractice cases even without a change in the ERISA preemption given recent trends in some courts. Therefore, even after ERISA is changed, separating any increase in the number of cases attributable to the change in legislation from increases for other reasons will be difficult.
The Coopers and Lybrand (1998b) findings for state government employees and retirees are lower than all but one of our estimates (see Table 2). But five of our estimates are within the range of the litigation rate reported by one managed care organization for individuals who purchase their own insurance (after Pilot Life). If the rate of claiming that Coopers and Lybrand found among non-ERISA- covered state government employees were to prevail among the ERISA population, we might expect 375 to 1,750 new legal claims annually. In our most extreme scenario at the high end, our calculations yield 136 claims per 100,000 enrollees, or 170,000 new claims annually. In our most extreme scenario at the low end, in which 5 percent of those whose appeals are turned down by external reviewers seek legal advice and 80 percent of those hire lawyers, our calculations yield 1,000 new claims annually.
To place these estimates in some perspective, there are about 285,000 negligent adverse medical events each year (Thomas et al., unpublished) and 50,000 medical malpractice suits each year (Weiler, 1994). There were about 2.5 million lawsuits claiming money damages filed in U.S. trial courts of general jurisdiction in 1990; about 1 million of these were tort suits, of which about half were a result of automobile accidents (Hensler, 1993). Our estimates for litigation over benefit denial and delay arising from the absence of ERISA restrictions range from 1,000 to 170,000 cases per year--a range that spans from far fewer than the number of medical malpractice cases to more than three times the annual number of medical malpractice cases. Higher-end scenarios would more likely prevail if punitive damages were allowable and awarded frequently and if large compensatory damages were awarded and sustained. Lower-end scenarios would be more likely if enrollees were required to exhaust internal appeals before seeking a court remedy, if cases were subject to curbs on punitive damages, and--over the long run--if courts and legislators imposed additional curbs in response to runaway litigation.
Although we can speculate about the factors that might produce lower and higher rates of claiming, the wide range of estimates that we calculated indicates that we do not know enough about critical aspects of claiming in a new ERISA environment to project litigation rates with any degree of confidence. Nor do we believe that others can make such projections without knowing more about legal claiming in this context.
Some proponents of expanding legal remedies for ERISA plan enrollees reason that ERISA beneficiaries should not be precluded from seeking legal remedies for improper health care decisions that are open to other consumers (non-ERISA beneficiaries), regardless of the relative costs and benefits of the legal changes. But, to a large extent,
the debate over ERISA change has focused on how health care outcomes would be affected by changes to ERISA and what the volume of litigation would be.
Many proponents of changing ERISA believe that increasing plans' exposure to litigation will improve health outcomes. Our review of relevant empirical evidence on the civil justice system's capacity to deter undesirable behavior sounds a cautionary note in this regard. The legal system is at best an imperfect deterrence mechanism in the contexts that have been studied.
Many opponents of amending ERISA fear the result will be a flood of costly litigation. Available data on litigation rates against health plans by non-ERISA enrollees analyzed in a previous study are more consistent with a less dramatic scenario while our review of the relevant empirical evidence on this point is inconclusive. Under some conditions, litigation might surge, imposing significant costs on health plans that we would expect to be passed on to consumers. Judges and legislatures might act to curb any such runaway litigation, but health plan administrators might change their behavior to reduce litigation costs before such responses emerged. In addition, because there are other legal barriers to suits against health plans besides ERISA and because some of the barriers created by ERISA itself are eroding, the consequences of amending ERISA could be less dramatic.
Expanding access to legal remedies is certain to lead to some new litigation. Moreover, whatever the outcomes of the current policy debate, managed care organizations are probably already considering changes in practice and policy in response to signals from the policy arena that their organizations' exposure to litigation may increase in the future. A critical question for future analysis is what the shape of those changes will be.
§514(a), codified as 29 U.S.C.A. §1144(a); §514(b)(2)(A), codified as 29 U.S.C.A. §1144(b)(2)(A).
Pilot Life Insurance Co. v. Dedeaux, 481 U.S. 41 (1987).
Corcoran v. United Healthcare Inc., 965 F.2d 1321 (5th Cir. 1992), cert. denied, 113 S.Ct. 812 (1992).
Massachusetts Mutual Life Ins. Co. v. Russell, 473 U.S. 134 (1985); Mertens v. Hewitt Associates, 508 U.S. 248 (1993); Medina v. Anthem Life Ins. Co., 983 F.2d 29 (5th Cir.), cert. denied, 114 S.Ct. 66 (1993); Bast v. Prudential Life Ins. Co., 1998 U.S. App. LEXIS 11358 (9th Cir.).
514 U.S. 645 (1995).
DeBuono v. NYSA-ILA and Clinical Service Fund, 117 S.Ct. 1747 (1997).
California Division of Labor Standards Enforcement v. Dillingham Construction, 117 S.Ct. 832 (1997); Boggs v. Boggs, 117 S.Ct. 1754 (1997).
Prudential Insurance Company v. National Park Medical Center, 154 F.3d 812 (8th Cir. 1998).
American Drug Stores, Inc. v. Harvard Pilgrim Health Care, Inc., 973 F.Supp. 60 (D.C. Mass. 1997).
Dunn v. Praiss, 606 A.2d 862 (N.J. Super. Ct. App. Div. 1992); Pacificare of Oklahoma v. Burrage, 59 F.3d 151 (10th Cir 1995); Prihoda v. Shpritz, 914 F. Supp. 133 (D.Md 1996).
Dukes v. U.S. Healthcare, 57 F.3d 350 (3d Cir.), cert denied 116 S. Ct. 564 (1995).
Again, managed care plans may be held increasingly liable in medical malpractice cases even without a change in the ERISA preemption, given recent trends in some courts.
Section 1201; reintroduced as H.R. 448, "Patient Protection Act of 1999" (Rep. Michael Bilirakis, R-Fla.).
S. 300, "Patients' Bill of Rights Plus Act" (Sen. Trent Lott, R-Miss., reintroduction of S. 2330).
H.R. 358, "Patients' Bill of Rights of 1999" (Sen. John Dingell, D-Mich., reintroduction of H.R. 3605) and S. 6, "Patients' Bill of Rights Act of 1999" (Sen. Tom Daschle, D-S.D., reintroduction of S. 1890).
H.R. 216, "Access to Quality Care Act of 1999" (Rep. Charles Norwood, R-Ga., reintroduced version of "PARCA," H.R. 1415, the Patient Access to Responsible Care Act of 1997) and S. 374, "Promoting Responsible Managed Care Act of 1999" (Sen. John Chafee, R-R.I., reintroduction of S. 2416).
H.R. 719, "Managed Care Reform Act of 1999" (Rep. Greg Ganske, R-Iowa).
There was some variability around the medians: The preauthorization request rate ranged from 1 for every 3 enrollees to 2 for each enrollee across the plans, and the UR requests denied ranged from 0.9 to 3.4 percent of preauthorization review requests. The median rate of initial UR denial among the plans was 1,450 per 100,000 enrollees.
The rate of complaints filed ranged from 120 to 520 per 100,000 enrollees among the four large plans reporting to the MDH, and complaints not settled to an enrollee's satisfaction ranged from 40 to 180 per 100,000 enrollees. The rate of appeals filed among the Texas plans ranged from 30 to 1,190 per 100,000 enrollees, and denied appeals ranged from fewer than 10 to 420 per 100,000 enrollees.
The estimates are compiled from a number of different types of managed care organizations, and the rates are applied to the entire population (125 million) of ERISA enrollees. Presumably, the pool of potential litigants would be larger in organizations that manage care more (stricter use of utilization review, for instance), and lower in indemnity plans in which only claims based on failure to pay in retrospective review would be filed. Therefore, our estimates of the pool are essentially a "managed care" average and could be larger to the extent that ERISA enrollees are distributed heavily in tightly managed care, or smaller to the extent that ERISA enrollees are in loosely managed or unmanaged care.
H.R. 719 "Managed Care Reform Act of 1999" (Rep. Greg Ganske, R-Iowa); S. 374 "Promoting Responsible Managed Care Act of 1999" (Sen. John Chafee, R-R.I., reintroduction of S. 2416).
Moradi-Shalal v. Fireman's Fund Ins. Companies, 46 Cal.3d 287, 758 P.2d 58, 250 Cal.Rptr. 116, 57 USLW 2138 (Cal., 1988).
Royal Globe Ins. Co. v. Superior Court, 23 Cal.3d 880, 592 P.2d 329, 153 Cal.Rptr. 842 (Cal., 1979).
Foley v. Interactive Data Corp., 47 Cal.3d 654, 765 P.2d 373, 254 Cal.Rptr. 211, 57 USLW 2396, 110 Lab.Cas. P 55,978, 3 IER Cases 1729 (Cal., 1988).
Funded by the Kaiser Family Foundation.
CalPERS is a somewhat less informative comparison group than the LAUSD or Colorado employees because it consists of both retirees, whose claiming behavior may differ from younger and employed individuals, and active employees (Burstin et al., 1993). Other groups of individuals not covered by ERISA include federal employees and Medicare and Medicaid recipients. However, litigation rates among individuals in these groups are not likely to be representative of those among ERISA beneficiaries: Federal employees are subject to limitations on litigation similar to those governing individuals covered by ERISA, while Medicare and Medicaid enrollees are subject to a special appeals system and may have different propensities to sue, because of demographic differences, than individuals covered by ERISA.
There is evidence, for example, that automobile accident claiming rates have risen over time.
We recognize that claims need not necessarily materialize contingent upon either initial denial (Stage 2) or a plan's upholding such a denial (Stage 4). Situations may arise in which the services received by an enrollee are influenced by plan policies in ways other than through straightforward denial. The patient does not appeal but later suffers an adverse outcome and sues. In such cases, the process is better represented by the following four-step process:
- The health plan conducts a utilization review or retrospective review.
- The benefit or coverage is approved (either in initial UR or on appeal).
- The patient believes him or herself injured and seeks legal advice.
- A lawyer agrees to take the case.
For example, the California Department of Corporations handles "Requests for Assistance" from health care consumers, and the Minnesota Department of Health assists consumers with complaints about their health plans. Some states have a single agency that handles minor grievances and benefit denial or delay problems, while others have independent utilization review organizations or agencies that specifically handle appeals of denied payment or care. The latter type of agency is a relatively new development in states: Texas and New Jersey implemented such agencies in 1997.
We remind readers that estimates of the pool are based on the behavior of health plans and consumers in the current environment. Behavior may change in an environment of relaxed ERISA restrictions, with health plans possibly approving more care and physicians or consumers requesting more care. The net effect of these changes is uncertain.
We assume that health plan enrollees will not consult a lawyer unless they believe that they have suffered a loss as a result of care or claim denial or delay, and that lawyers will not agree to represent claimants unless they can document losses as a basis for claiming financial damages. We are not aware of any data on the extent of injury or losses with denials of health care authorization or payment.
In reality, different fractions of enrollees might seek legal assistance at different stages of the process. In our model, we apply a uniform rate of exit and examine the effect of assuming that exit takes place at earlier and later stages.
Corresponding to high, medium, and low percentages of legal claiming that we observe in RAND and Harvard Medical Practice Study data.
Bailis, D., and R. MacCoun, "Estimating Liability Risks with the Media as Your Guide: A Content Analysis of Media Coverage of Tort Litigation," Law and Human Behavior, Vol. 20, 1996, p. 419.
Barents Group, LLC, Health Economics Practice, Impacts of Four Legislative Provisions on Managed Care Consumers: 1999-2003, prepared for the American Association of Health Plans, Washington, D.C., April 22, 1998.
Bovbjerg, R. R., L. C. Dubay, G. M. Kenney, and S. A. Norton, "Defensive Medicine and Tort Reform: New Evidence in an Old Bottle," Journal of Health Politics, Policy and Law, Vol. 21, No. 2, Summer 1996.
Brennan, T. A, C. M. Sox, and H. R. Burstin, "Relation Between Negligent Adverse Events and the Outcomes of Medical Malpractice Litigation," New England Journal of Medicine, Vol. 335, No. 26, 1998, pp. 1963-1967.
Burstin, H. R, W. G., Johnson, and S. R. Lipsitz, "Do the Poor Sue More? A Case-Control Study of Malpractice Claims and Socioeconomic Status," JAMA, Vol. 270, No. 14, 1993.
California Department of Corporations, Health Plan Division, Health Care Service Plan Complaint Data: 1996 Requests for Assistance, Los Angeles, Calif., June 1997.
Carroll, S. J., A. Abrahamse, and M. Vaiana, The Costs of Excess Medical Claims for Automobile Personal Injuries, Santa Monica, Calif.: RAND, DB-139-ICJ, 1995.
Carroll, S. J., J. S. Kakalik, N. M. Pace, and J. L. Adams, No-Fault Approaches to Compensating People Injured in Automobile Accidents, Santa Monica, Calif.: RAND, R-4019-ICJ, 1991.
Center for Health Dispute Resolution (CHDR), Medicare HMO/CMP Reconsideration Data, 1996-1997, Pittsford, N.Y., 1998.
Cheney, F. W, K. Posne, R. A. Caplan, and R. J. Ward, "Standard of Care and Anesthesia Liability," JAMA, Vol. 261, 1989, pp. 1599-1603.
Conard, A., et al., Automobile Accident Costs and Payments--Studies in the Economics of Injury Reparations, Ann Arbor: University of Michigan Press, 1964.
Congressional Budget Office, H. R. 3605/S. 1890: Patients' Bill of Rights Act of 1998, July 16, 1998a.
Congressional Budget Office, H. R. 4250: Patient Protection Act of 1998, July 24, 1998b.
Coopers and Lybrand, LLP, Estimated Costs of Selected Consumer Protection Proposals: A Cost Analysis of the President's Advisory Commission's Patient Access to Responsible Care Act (PARCA), prepared for the Kaiser Family Foundation, Menlo Park, Calif., April 1998a.
Coopers and Lybrand, LLP, Impact of Potential Changes to ERISA: Litigation and Appeals Experience of CalPERS, Other Large Public Employers and a Large California Health Plan, prepared for the Kaiser Family Foundation, Menlo Park, Calif., June 1998b.
Danzon, P. M., Medical Malpractice: Theory, Evidence, and Public Policy, Cambridge, Mass.: Harvard University Press, 1985.
Danzon, P., and L. Lillard, The Resolution of Medical Malpractice Claims: Modeling the Bargaining Process, Santa Monica, Calif.: RAND, R-2793-ICJ, 1982.
Donohue, J. J., III, and P. Siegelman, "The Changing Nature of Employment Discrimination Litigation," Stanford Law Review, Vol. 43, May 1991, p. 983.
Eckholm E., "$89 Million Awarded to Family Who Sued HMO," The New York Times, December 30, 1993, p. A1.
Farber, H. S, and M. J. White, "Medical Malpractice: An Empirical Examination of the Litigation Process," RAND Journal of Economics, Vol. 22, 1991, pp. 199-217.
Fireston, D., "Legislature Passes Cap on Damage Awards," The New York Times, June 2, 1999.
Galanter, M., "Shadow Play: The Fabled Menace of Punitive Damages," University of Wisconsin Law Review, Vol. 1, 1998, pp. 1-14.
Garber, S. "Product Liability, Punitive Damages, Business Decisions and Economic Outcomes," University of Wisconsin Law Review, Vol. 1, 1998, pp. 237-295.
Garber, S., and A. G. Bower, "An Econometric Analysis of Newspaper Coverage of Automotive Product Liability Verdicts," Law and Society Review, Vol. 33, No. 1, 1999, pp. 93-122.
Gergen, M. P, "A Cautionary Tale About Contractual Good Faith in Texas," Texas Law Review, Vol. 72, 1994, pp. 1235-1250.
Glaberson, W., "Some Plaintiffs Losing Out in Texas' War on Lawsuits," The New York Times, June 7, 1999.
Goldstein, A., and J. Eilperin, "Partisan House Swiftly Passes GOP Patients' Rights Bill," The Washington Post, July 25, 1998.
Harvard Medical Practice Study, Patients, Doctors, and Lawyers: Medical Injury, Malpractice Litigation, and Patient Compensation in New York, Boston, Mass., 1990.
Havighurst, C. C., J. F. Blumstein, and T. A. Brennan, Health Care Law and Policy: Readings, Notes, and Questions, 2nd ed., New York, N.Y.: Foundation Press, 1998.
Health Plan Division, Department of Corporations, State of California, Health Care Service Plan Complaint Data: 1996 Requests for Assistance, Sacramento, Calif., June 1997.
Hensler, D., "A Glass Half Full, A Glass Half Empty: The Use of Alternative Dispute Resolution in Mass Personal Injury Litigation," University of Texas Law Review, Vol. 73, 1995, pp. 1587-1626.
Hensler, D., "Reading the Tort Litigation Tea Leaves: What's Going on in the Civil Liability System," Justice System Journal, Vol. 16, No. 139, 1993, pp. 139-154.
Hensler, D., and M. Peterson, "Understanding Mass Personal Injury Litigation: A Socio-Legal Analysis," Brook Law Review, Vol. 59, 1993, pp. 961-1063.
Hensler, D. R., et al., Class Action Dilemmas: Pursuing Public Goals for Private Gain, Santa Monica, Calif.: RAND, forthcoming.
Hensler, D. R., et al., Compensation for Accidental Injuries in the United States, Santa Monica, Calif.: RAND, R-3999-HHS/ICJ, 1991.
Johnston, D. C, "$116 Million Punitive Award Against Aetna," The New York Times, January 16, 1999.
Kessler, D. P., and M. McClellan, "Do Doctors Practice Defensive Medicine?" The Quarterly Journal of Economics, Vol. 111, February 1996.
King, E. M., and J. P. Smith, Economic Loss and Compensation in Aviation Accidents, Santa Monica, Calif.: RAND, R-3551-ICJ, 1988.
Lewin Group, Consumer Bill of Rights and Responsibilities Costs and Benefits: Information Disclosures and External Appeals, Fairfax, Va., November 15, 1997.
Liggio, J. M., "Preparing the HMO Case (Medical Negligence)," Trial, Vol. 33, May 1997, pp. 26-29.
Localio, A. R., A. G. Lawthers, J. M. Bengston, et al., "Relationship Between Malpractice Claims and Cesarean Delivery," JAMA, Vol. 269, No. 3, January 20, 1993, pp. 366-373.
Localio, A. R, A. G. Lawthers, T. A. Brennan, et al., "Relation Between Malpractice Claims and Adverse Events Due to Negligence: Results of the Harvard Medical Practice Study III," New England Journal of Medicine, No. 325, 1991, pp. 245-251.
McGovern, F. E., "Resolving Mass Tort Litigation," BU Law Review, Vol. 69, 1989, pp. 660-690.
McNulty, M., "Are Poor Patients Likely to Sue for Malpractice?" JAMA, Vol. 262, 1989, pp. 1391-1392.
Milliman and Robertson, Actuarial Analysis of the Patient Access to Responsible Care Act (PARCA), Brookfield, Wisc., November 7, 1997.
Minnesota Department of Health (MDH), Health Policy and Systems Compliance Division, tables of complaints received by health plans and complaints received by the department, St. Paul, Minn., 1996.
Moller, E., Trends in Civil Jury Verdicts Since 1985, Santa Monica, Calif.: RAND, MR-694-ICJ, 1996.
Moller, E., N. Pace, and S. Carroll, Punitive Damages in Financial Injury Jury Verdicts, Santa Monica, Calif.: RAND, MR-888-ICJ, 1997.
Multinational Business Services, Inc., Regulatory Requirements of H. R. 1415, The Patient Access to Responsible Care Act, March 10, 1998.
Muse and Associates, The Health Premium Impact of H. R. 1415/S. 644, the Patient Access to Responsible Care Act (PARCA), Washington, D.C., June 29, 1998.
National Health Policy Forum, "ERISA Health Plan Denials: Exploring Models for External Review," Issue Brief, No. 720, June 19, 1998.
New Jersey Department of Health and Senior Services, Consumers Use Right to Appeal HMO Decisions on Care, Trenton, New Jersey, April 22, 1998.
Perez, R. A., "ERISA Preemption: Denying Employees' Rights to Benefits," Trial, Vol. 33, May 1997, pp. 72-74.
Peterson, M. A., R. T. Reville, and R. K. Stern, Compensating Permanent Workplace Injuries: A Study of the California System, Santa Monica, Calif.: RAND, MR-920-ICJ, 1998.
Polinsky, A. M., and S. Shavell, "Punitive Damages: An Economic Analysis," Harvard Law Review, Vol. 111, 1998, pp. 870-896.
Price Waterhouse LLP, Benefits and Costs of Consumer Protection in California, Menlo Park, Calif., June 1998.
Price Waterhouse LLP, The Impact of Managed Care Legislation: An Analysis of Five Legislative Proposals in California, prepared for the Kaiser Family Foundation, Menlo Park, Calif., November 1997.
Remler, D., K. Donelan, R. J. Blendon, et al., "What Do Managed Care Plans Do to Affect Care? Results from a Survey of Physicians," Inquiry, Vol. 34, Fall 1997, pp. 196-204.
Reynolds, R. A., J. A. Rizzo, and M. L. Gonzalez, "The Cost of Medical Professional Liability," JAMA, Vol. 257, No. 20, May 22/29, 1987, pp. 2776-2781.
Richardson, David A., James Phillips, and Dean Conley, A Study of Coverage Denial Disputes Between Medicare Beneficiaries and HMOs, Pittsford, N.Y.: Network Design Group Inc., September 1993.
Rolph, E., E. Moller, and J. E. Rolph, "Arbitration Agreements in Health Care: Myth or Reality?" Law and Contemporary Problems, Vol. 60, 1997, pp. 153-184.
Rothouse, M., "Issue Paper: Insurer Liability," Health Policy Tracking Service, October 1, 1998.
Rubin, R. J., and D. N. Mendelson, Estimating the Costs of Defensive Medicine, Washington, D.C.: Lewin-ICF, October 21, 1992.
Rustad, M. L., "Unraveling Punitive Damages: Current Data and Further Inquiry," Wisconsin Law Review, No. 1, 1998, pp. 15-69.
Rustad, M. L., and T. Koenig, "Reconceptualizing Punitive Damages in Medical Malpractice: Targeting Amoral Corporations, Not 'Moral Monsters,' Rutgers Law Review, Vol. 47, Spring 1995, pp. 975-1083.
Schwartz, G. T., "Reality in the Economic Analysis of Tort Law: Does Tort Law Really Deter?" UCLA Law Review, Vol. 42, 1994, p. 377.
Sloan, F. A., and C. Ruey Hsieh, "Variability in Medical Malpractice Payments: Is the Compensation Fair?" Law and Society Review, Vol. 24, 1990, pp. 997-1039.
Studdert, D. M, H. R. Burstin, B. I. Zbar, E. J. Thomas, E. J. Orav, and T. A. Brennan, Suing for Medical Malpractice in Utah and Colorado in 1992, unpublished manuscript.
Taragin, M. I., L. R. Willett, A. P. Wilczek, et al. "The Influence of Standard of Care and Severity of Injury on the Resolution of Medical Malpractice Claims," Annals of Internal Medicine, Vol. 117, 1992, pp. 780-784.
Texas Department of Insurance (TDI), A Health Maintenance Organization Purchasing Guide for Texas Consumers, Austin, Tex., June 1997.
Texas Department of Insurance (TDI), HMO Complaints for Calendar Year 1997 (January-December 1997), Austin, Tex., 1998.
Thomas, E. J., D. M. Studdert, H. R. Burstin et al., Adverse Events and Negligent Care in Utah and Colorado, unpublished manuscript.
U.S. Congress, Office of Technology Assessment, Defensive Medicine and Medical Malpractice, Washington, D.C.: U.S. Government Printing Office, OTA-H-602, July 1994.
U.S. Equal Employment Opportunity Commission (U.S. EEOC), Sexual Harassment Charges, EEOC and FEPAs Combined: 1992-FY1998, Washington, D.C., November 9, 1998.
U.S. General Accounting Office, Employer Based Managed Care Plans: ERISA's Effect on Remedies for Benefit Denials and Medical Malpractice, Washington, D.C.: U.S. Government Printing Office, GAO/HESS-98-154, July 1998.
U.S. General Accounting Office, HMO Complaints and Appeals: Most Key Procedures in Place, but Others Valued by Consumers Largely Absent, Washington, D.C.: U.S. Government Printing Office, GAO/HESS-98-119, May 1998.
U.S. General Accounting Office, Medical Liability: Impact on Hospital and Physician Costs Extends Beyond Insurance, Washington, D.C.: U.S. Government Printing Office, GAO/AMID-95-169, September 1995.
Viscusi, W. K, "The Dimensions of the Product Liability Crisis," Journal of Legal Studies, Vol. 20, January 1991, pp. 147-177.
Weiler, P. C., "Fixing the Tail: The Place of Malpractice in Health Care Reform," Pfizer Lecture, Rutgers University Law Review, Vol. 47, 1994, pp. 1157-1193.
Weiler, P. C., Medical Malpractice on Trial, Cambridge, Mass.: Harvard University Press, 1991.
Weiler, P. C., H. H. Hiatt, J. P. Newhouse, W. G. Johnson, T. A. Brennan, and L. L. Leape, A Measure of Malpractice, Cambridge, Mass.: Harvard University Press, 1993.
We thank Peter Jacobson and Paul Rheingold for their insights and research in an early phase of this project. We appreciate helpful reviews from our colleagues Steve Garber, Clark Havighurst, Robert Reville, Lloyd Dixon, M. Susan Marquis, and Elizabeth Rolph and valuable assistance from ICJ Director of Communications Beth Giddens.
RAND's Health Law program is a joint project of RAND Health and the RAND Institute for Civil Justice. The program represents a collaborative effort among researchers specializing in policy research on civil justice issues and health care to explore the evolving relationship between the legal and health care systems. Health Law research includes analyses of trends, outcomes, and policy options on a wide range of topics such as the role of the courts in shaping health policy, the benefits and costs of medical innovation, and the integration of managed care and liability. The program builds on a long tradition of RAND research characterized by an interdisciplinary, empirical approach to policy issues and rigorous standards of quality, objectivity, and independence.
For additional information about the Health Law program, contact David Studdert, RAND Health Law Coordinator at (310) 393-0411 ext. 6463, or write to: 1700 Main St., P.O. Box 2138, Santa Monica, CA 90407. E-mail: firstname.lastname@example.org.
Health Law materials are also available online from several sources. The RAND Health home page may be found at /organization/health/. The ICJ home page at /centers/icj/ provides information about the Institute for Civil Justice. Westlaw, the exclusive online distributor of ICJ materials, publishes the full text of many ICJ documents at http://www.westlaw.com/; Health Law publications will be among them.
RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND® is a registered trademark.
RAND issue papers explore topics of interest to the policymaking community. Although issue papers are formally reviewed, authors have substantial latitude to express provocative views without doing full justice to other perspectives. The views and conclusions expressed in issue papers are those of the authors and do not necessarily represent those of RAND or its research sponsors.
All rights reserved. Permission is given to duplicate this on-line document for personal use only, as long as it is unaltered and complete. Copies may not be duplicated for commercial purposes.Published 1999 by RAND
This Issue Paper is also available in hard copy.
RAND's Home Page