#2007-06 March 2007 No-Fault Medical Malpractice Compensation: Policy Considerations for Design of a Demonstration Project by Jackson Williams The AARP Public Policy Institute, formed in 1985, is part of the Policy and Strategy Group at AARP. One of the missions of the Institute is to foster research and analysis on public policy issues of importance to mid-life and older Americans. This publication represents part of that effort. The views expressed herein are for information, debate, and discussion, and do not necessarily represent official policies of AARP. © 2007, AARP. Reprinting with permission only. AARP, 601 E Street, NW, Washington, DC 20049 http://www.aarp.org/ppi Foreword The idea of a no-fault administrative system for adjudicating medical malpractice claims has been around for a long time, but efforts to bring such a system to fruition have never been as intense as they are right now. In 2004, Harvard researchers teamed with anti- lawsuit crusader Phillip Howard in a joint project funded by the Robert Wood Johnson Foundation to marry Harvard’s vision of no-fault with Howard’s notion of “health courts.” Early this year, Robert Wood Johnson added more funding to this collaboration, allowing Harvard and Howard’s organization, Common Good, to promote pilot projects in six states. Some no-fault advocates suggest that an administrative system that could be navigated without legal representation would, among other things, benefit older claimants who are disadvantaged under the current system. These advocates note that the current tort system provides higher damage awards for claimants who have lost large projected earnings and lower awards for those who have retired and that that personal-injury lawyers, who receive about one third of the recovery as their fee, may prefer to take the cases of younger claimants who are in their prime earning years over older claimants. In this analysis of issues relating to no-fault proposals, author Jackson Williams of AARP’s Public Policy Institute finds a lack of evidence to either support or contradict the contentions made by no-fault advocates about the advantages of the proposed system. Williams concludes that the proposals are still too amorphous to enable us to make conclusions about their impact. The fairness of a system, he observes, will ultimately depend on the details of its structure, and the paper explains exactly what elements are needed to ensure a patient-friendly design. The paper’s main caveat is that the fairness of a no-fault system could depend on whether its sponsorship is public or private. Currently, the Harvard/Common Good team envisions state sponsorship of its health court pilots, perhaps in partnership with the federal government. But the past decade has seen a trend toward private adjudication of claims, pursuant to mandatory arbitration clauses. Williams asserts that privately established health courts could be designed to subtly favor providers over patients, while the model of government sponsorship endorsed by Common Good would have public oversight. The current tort system has several shortcomings. Williams notes that a no-fault system would remedy some of these shortcomings but, in so doing, would trade off other values. Whether a no-fault system would result in increased compensation and greater perceived fairness, he says, will only be known after a demonstration project has been conducted and fully evaluated. David Gross, Ph.D. Interim Director, Policy Research and Development AARP Public Policy Institute i Table of Contents Foreword .............................................................................................................................. i Executive Summary ........................................................................................................... iii Background ......................................................................................................................... 1 1. Shifting Costs to Address the “Overhang” of Errors ...................................................... 2 2. The Role of Experts ........................................................................................................ 7 3. Equity Versus Equality in Damage Awards: “Predictability” and Its Effect on Deterrence ......................................................................................................................... 10 4. Boundary Between Tort and No-Fault.......................................................................... 11 5. Standard of Care and Evidence-Based Medicine.......................................................... 14 6. Standard of Care and Defensive Medicine ................................................................... 17 7. Evaluating the Success of a Demonstration.................................................................. 21 8. Public Oversight Versus Private Initiative.................................................................... 22 Conclusion ........................................................................................................................ 23 Table of Figures Figure 1A: Distribution and Sources of Compensation under Tort System ....................... 4 Figure 1B: Distribution and Sources of Compensation under No-Fault System................ 5 Figure 2: Continuum of Medical Evidence Basis for Procedure ...................................... 19 ii Executive Summary Background. A “no-fault” administrative compensation system for medical injuries offers the prospect of a simpler, fairer, more efficient, and less contentious means of adjudicating claims, while facilitating candid and self-critical discussion of the causes of medical errors. Some proponents of this approach suggest that it would also result in more injury compensation paid to elderly claimants. No-fault advocates are now beginning the difficult task of designing a demonstration project. Purpose. The purpose of this paper is to examine the quandaries posed by the collision of the no-fault approach’s ambitious goals with practical and cost constraints, uncertainty about medical science, and competing notions of fairness. The paper discusses how these dilemmas will manifest themselves in key design decisions to be made by demonstration project planners. Methods. The investigator reviewed scholarly literature discussing possible no-fault systems and the problems they are intended to ameliorate and documents reflecting no- fault proposals currently being propounded or formulated. Findings. Designers of a no-fault system face several key challenges: • A primary goal of no-fault is to compensate more of the injured patients for whom, under the tort system, claims are either not made or not paid; yet doing so could increase overall damages assessed against providers. A revised distribution of damage awards will result in “losers” as well as “gainers.” • Neutral medical experts will play a central role in a no-fault scheme, and exactly how and when their expertise is brought to bear will have tremendous impact on the number of injuries compensated. • No-fault’s elevation of predictability over “haphazardness” may force a choice between equity and equality. • A partial no-fault system that leaves residual cases in the tort system eases some problems but creates others. Separating cases for tort and no-fault treatment forces a choice between considerations of fairness in compensation and other system goals. • Proponents suggest that in an administrative system the traditional “standard of care” can be revisited, but it may prove impracticable to apply evidence-based medicine in adjudications or to reduce any “defensive medicine” that doctors may practice. • A rigorous evaluation of a no-fault demonstration will require considerable resources. • Policymakers must anticipate the possibility that a no-fault system could be tried, without public oversight and supervision, through the use of arbitration clauses. Summary and Conclusions. Replacing medical malpractice lawsuits with a no-fault system could offer broader compensation of injured patients while changing medical providers’ attitudes toward confronting medical errors. In the absence of more detailed iii information about how the system would be structured, it is impossible to determine whether the greater breadth of compensation awards would be perceived as fairer. Designers of an administrative system face a difficult task in reconciling the competing values and needs at stake. Striking a balance too far in the direction of relieving providers’ litigation burdens could result in patients losing not only the unique form of agency they wield in tort litigation but also some of their awards. In order to ensure that patients benefit from a no-fault project, basic consumer protections must be built into the system. iv No-Fault Medical Malpractice Compensation: Policy Considerations for Design of a Demonstration Project Background For many years there has been discussion in medico-legal academic circles of replacing tort-based compensation for medical errors with a no-fault system.∗ Advocates of a no- fault approach argue that the high transaction costs of the tort system—emotional as well as financial—might be avoided in a streamlined administrative system, freeing up money to compensate more injured patients. Such a system would make payment upon occurrence of a specified event rather than after proof of a provider’s negligence. Damages would be assessed according to a fixed schedule, rather than at the discretion of a jury. Administrative systems have replaced tort compensation for medical malpractice in New Zealand, Denmark, and Sweden.1 In America, the no-fault concept has moved from theoretical musings to the planning stage, with two research teams having been awarded grants to design demonstration projects. One, led by William Sage, would use the Medicare program as the vehicle for testing no-fault.2 The other, a collaboration of Harvard researchers long identified with the concept and Common Good, an advocacy organization known for its harsh critiques of the legal system, would test a “health court” system based at the hospital or health system level.3 In the meantime, bills are being considered by Congress that would potentially give federal authorization and funding to no-fault demonstrations.4 There are many values that can be promoted through an injury compensation system, several of which are in direct conflict. The traditional tort system, which governs all but a few medical injuries in the United States, and its alternative, the no-fault system, emphasize different sets of values. The tort system gives emphasis to granting full and complete compensation to those injured patients who prove negligence, incentives for professionals and institutions to avoid causing injuries (deterrence), and corrective justice. A no-fault system gives greater weight to granting essential compensation to the largest number of injured patients, ensuring predictability of compensation, facilitating communication within the health care system to avoid causing injuries, and minimizing transaction costs (efficiency).5 For the most part, each of these values is thought to be compatible with only one of the two systems. In other words, full compensation to a few individuals trades off against modest compensation to many injured patients; corrective justice trades off against minimal transaction costs; deterrence trades off against predictable compensation; and so on. Meanwhile, the different options for design of a no-fault system bring their own set of dilemmas, as is evident in the evolution of the Harvard proposal since its convergence ∗ The term “no-fault” will be used in this paper not in its literal sense, meaning payment made on the occurrence of an adverse event without regard to how avoidable the event was, but in its broader sense as shorthand for an administrative alternative to the tort system in which an individual provider’s fault or blameworthiness is not at issue. 1 with Common Good’s “health court” concept. Many of these dilemmas arise from the pragmatic need to keep the budget of a no-fault system within the boundaries of the current cost of tort compensation borne by medical providers. This Issue Paper discusses the policy considerations faced by designers of a no-fault system and how older Americans might be affected by some of the choices. It identifies seven sources of policy quandaries that arise when considering a switch from tort to no- fault compensation: the cost consequences of addressing the “overhang” of errors that go uncompensated under the tort system; the role to be played by medical experts; the trade- off between “predictable” awards and deterrence; the boundary between cases getting no- fault treatment and those remaining in the tort system; the adoption of a new legal standard of care; the mechanism for evaluating a demonstration; and the nature of the entity authorizing and administering the project, which could be either public or private. 1. Shifting Costs to Address the “Overhang” of Errors First and foremost among policy quandaries is the issue of costs. For three decades, providers have insisted that they are already overburdened by the costs of medical injuries and have persuaded many legislatures to shift some of those costs onto claimants or first-party payors (e.g., health and disability insurers and employers) through “tort reform” legislation. At the same time, empirical studies of medical malpractice have demonstrated that only a small percentage of medical errors ever give rise to tort claims.6 It is generally agreed that if a no-fault system were to compensate every medical injury, any savings in transaction costs would be more than offset by increased payments to injured patients—the so-called “overhang” of currently unbrought claims. This is because only a small portion of medical injuries, estimated to be between one in six and one in ten, currently result in claims. It is also generally agreed that for a switch to no-fault to be politically feasible, it could not impose greater costs on providers than they currently bear under the tort system.7 Thus, a primary practical imperative guiding designers of a no-fault system is to scale back the number of eligible claims and the size of awards so that provider-borne costs remain roughly the same. In their initial exploration of the cost of a no-fault system, Studdert et al. concluded that, if Sweden’s system of “compensable events” had been applied to Utah medical injuries in 1992, the total amount payable to patients would have been $82 million. However, Utah doctors paid only $22 million in malpractice premiums that year. This research team concluded that in order to make the no-fault system cost-neutral, it would be necessary to eliminate subrogation for any first-party insurance benefits; require the plaintiff to bear the first four weeks’ wage losses; replace wage losses beyond four weeks at 66 percent; eliminate recovery for any loss of household production (e.g., cleaning, food preparation, maintenance); and cap non-economic damages (e.g., pain, suffering, disfigurement) at $100,000. In their “Health Courts Proposal Skeleton,” the Harvard/Common Good project propose a scheme in which “economic damages would be compensated in full except…a deductible 2 period or out of pocket amount (we suggest that eligibility begin when patients reach 4–6 weeks lost work time or $3,000–4,000 in medical expenses).” This proposal also calls for eliminating first-party payors’ subrogation rights. The proposal would schedule non- economic damages according to “decision science research about utility losses” associated with different types of injury. It does not call for any arbitrary limitation, such as the $100,000 cap in the Utah proposal. Different combinations of damage elements to be included in compensation awards can benefit different constituencies at the expense of others. Under the traditional tort system, which is to say, as the tort system operates in states that do not place caps on non-economic damages or offset collateral source payments, non- pecuniary damages are rather prominent. One can imagine a no-fault system in which non-pecuniary damages are substantially lower, either because the utility loss calculations made by the plan designers concluded that current jury verdicts overstate the value of such losses, or because this element of damages has been capped in order to reduce insurance costs to providers. Non-pecuniary damages are often criticized because of their subjective nature. While wage losses and medical expenses leave paper trails (pay stubs and bills, respectively), concepts such as “pain and suffering” are not moored to invoices and juries are given no guidance in fixing damages. As such, awards for these damages can appear arbitrary and seem to some like a logical target for reduction if to sacrifice these damages means that many times more victims of medical errors will receive compensation. One consequence of award formulas that de-emphasize “paperless” damages such as pain and suffering or lost household production is that awards to retired claimants will be lower. To illustrate, imagine a 65-year-old female who is rendered paraplegic as a result of a medical error. According to Consumer Product Safety Commission estimates, by the time the average female reaches the age of 65, the present value of her future household work exceeds the value of her future wage earnings ($84,562 versus $28,311, in 1994 dollars). Limiting an award for pain and suffering to $100,000 would also likely be a drop in the value of that element of damages. Therefore, a regime in which damage elements retain the prominence they do under the tort system is more likely to represent an improvement in net compensation to the elderly than the one sketched out in the Utah proposal. The implications of consistently lower awards to the elderly are discussed in Section 3. Shifting responsibility for damage elements would also result in winners and losers among payors. Under the traditional tort system, the negligent provider is responsible for all the major elements of the claimant’s damages. Even when a claimant is fully insured for lost wages (sick leave, disability insurance) and medical expenses (health insurance), the collateral source rule holds the provider liable for damages covered by first-party insurance, and those insurers’ subrogation rights require the provider to reimburse them for their payouts to the claimant. 3 Under the Harvard model no-fault system, the provider’s responsibility is substantially smaller because the damage award wraps around first-party insurance. Subrogation rights are extinguished. Ordinarily, economists would frown on this because the deterrent effect of a compensation claim is thought to be most effective when liability lies with the party in the best position to prevent the injury. But to the extent that these insurance benefits are supplied by employers, abolishing subrogation may be beneficial because it increases the employer’s stake in patient safety. Of course, employers already have a stake in patient safety as injury to a worker causes a non-compensable loss to them also—this is why employers fund such patient safety initiatives as the Leapfrog Group. Abolishing subrogation would encourage employers to continue efforts to direct their employees to the safest providers. Figure 1A and B recapitulate the points discussed above to show how injury costs and compensation would be redistributed under a hypothetical no-fault system. The vertical axis represents four categories of damages sustained by an injured patient. The horizontal axis represents the many different patients who are injured by medical errors, with those having the “strongest” cases grouped on the right. For the sake of simplicity, if not accuracy, the diagrams assume all patients have health insurance and are employed by entities that provide paid sick or disability leave. Figure 1A: Distribution and Sources of Compensation under Tort System v u Pain & suffering Providers s t Household Providers Health insurers Employers a b Wage loss w x c d Medical y z e f 4 Figure 1B: Distribution and Sources of Compensation under No-Fault System s t r Providers Providers Health insurers Employers q a b w x c d y z e f The diagonal shading represents costs borne by negligent health care providers. In the tort diagram, the tall area at the far right represents tort compensation paid. This area is narrow, because only about one in eight injured patients makes a claim, and only about half of those patients receive an award. An additional area in diagonal shading, WXYZ, represents costs absorbed by capitated providers, who must bear the additional medical costs caused by their negligence regardless of whether there is a lawsuit. In the no-fault diagram, the diagonal shaded area at right, representing no-fault awards, is wider and shorter. It is wider because a greater number of patients receive awards. It is shorter for two reasons. First, the awards are expected to be less generous and exclude a deductible amount. Second, the award wraps around other benefits and extinguishes subrogation rights of third-party “collateral sources.” The doctrine of subrogation places a lien on a claimant’s recovery in the amount of the collateral source payment. For instance, ABCD represents “sick pay” earned by the tort claimant for which the employer is entitled to reimbursement. The area of CDEF represents medical expenses paid by a health insurer for which the insurer, in the traditional tort system, is entitled to reimbursement. In a no-fault system, these costs would be absorbed by the employer and insurer, and the shaded areas representing the costs they bear are larger in figure 1B. Recent wrangling over these reimbursements in Congress8 and the courts9 signals that employers and health insurers may be loath to give up their subrogation rights and thereby assume more injury costs. Finally, the diagonal shaded area is larger in the no-fault diagram, representing money freed up from reduced transaction costs to pay additional compensation. Overall, the shaded areas are larger in the no-fault diagram, representing increased payments to patients not only due to the transaction cost savings but also to restrictions on employers’ 5 and insurers’ subrogation rights. In both figures, unshaded areas represent costs borne by the injured patient. The redistribution of compensation among patients will result in net “gainers” and “losers” in terms of awards. Most at risk of losing are those claimants who could successfully obtain tort award in excess of what would be payable under a damage schedule (area STUV in figure 1A). Plaintiff lawyers can truthfully say that many of the people they represent would lose out; but this is true primarily because skilled lawyers choose clients they expect to have strong cases. It is more difficult to discern in advance categories of people who would gain or lose without knowing more precisely what damages are available. For this reason, the horizontal axes of the figures represent different types of patients, but do not break them down into categories. It is possible to speculate about characteristics of patients in area QRSA who presently go uncompensated but would receive awards in a no-fault system. Some of them would be patients who currently bring claims but are either forced to drop them or go to trial and lose. A recent Harvard study of closed medical malpractice claims estimated that about 27 percent of claimants who have valid cases for medical malpractice ultimately receive no payment;10 ideally, a no-fault system would compensate these patients. If the system works properly, many awards would also go to patients who, in a tort environment, are not even aware that they have sustained a medical injury because the error was not disclosed to them. Still others should be patients for whom, under the tort system, filing a lawsuit does not make economic sense. These would include patients whose damages are not sufficient to merit an award of a size that justifies the considerable investment of attorney resources required to build a tort case. It is thought that this would include those with less severe injuries as well as those who have severe injuries but for whom damages are not sufficiently quantified. Different people could sustain an identical catastrophic injury but those with higher income and more remaining years of expected employment would show a greater pecuniary loss and are thought to have more viable cases. Conventional wisdom holds that elderly patients have less compelling tort claims because their prime earning years are over. As such, they are thought to be net gainers from a switch to no-fault. However, as discussed above, some of the techniques proposed for reducing no-fault awards to compensate more claimants could result in minimizing awards to the elderly. Not depicted in figure 1A and B, but still very important, is the portion of compensation that is used to pay plaintiff’s attorney fees and litigation costs—about one-third of the gross tort compensation paid by providers. In a consumer-friendly no-fault system, the need to hire lawyers should be greatly reduced, if not obviated, leaving claimants with a higher share of net compensation. A final cost question is that of the time value of money. On average, about four years and nine months transpire between a malpractice incident and payment of tort 6 compensation.11 An administrative process that speeds payment of compensation to claimants would represent a substantial net gain for claimants at the expense of insurers, who reap investment income from their reserves. Faster resolution of claims would be an important consideration for elderly patients. 2. The Role of Experts The unique and defining characteristic of any claim for compensation of medical errors— whether through the tort system or a no-fault system—is its reliance on medical experts for guidance. Under the tort system, competing experts testify in adversarial proceedings, with lay jurors determining whether the standard of care has been violated. For some commentators, particularly those sympathetic to doctors, an ideal system would use neutral experts to guide presentation of evidence or determine the facts of a case. The Harvard/Common Good no-fault system would put this ideal into practice. For patients, the key consideration in revising the role of experts will be when and how neutral expertise is brought to bear. It is known from the various empirical studies of malpractice that independent experts reviewing medical charts will discover about eight times as many medical errors as there are claims brought into the tort system.12 Under the tort system, patients must first suspect that there has been negligence in their care, a suspicion that they may not be able to form without some knowledge of appropriate medical practice. They must then find a physician willing to testify that negligence took place. Access to these physician experts is controlled by the plaintiff’s attorney. The cost of retaining an expert is high, and attorneys will not invest their own money to consult on a client’s case unless, in the attorney’s opinion, there is likely to be a lucrative claim. A no-fault system in which claims were initiated by independent experts capable of finding every breach of the standard of care would be the most pro-patient permutation of the concept, leading to the highest number of adverse events being compensated. A system in which the patient is still required to discover the existence of a claim, but would have unimpeded access to an independent expert review, would still be more pro- patient than the current regime. A system in which the patient continues to be responsible for both discovering a negligent injury and marshaling expert evidence to prove it, but has expert judges making the ultimate determination in each case, probably would not be of much advantage to the injured patient, and could even be a detriment if a physician- dominated court were more sympathetic to providers than jurors are. Alternative models of expert input are suggested by a February 2005 policy report by Udell and Kendall13 and the Harvard/Common Good Health Courts Proposal Skeleton. Udell and Kendall cursorily describe a “health court review board” that would “investigate claims.” In this scenario, the onus is still on the patients to surmise that they may have been negligently injured, but the patients would be entitled to a seemingly independent investigation of the claim. The concept of investigating claims calls to mind the work of fair employment practices agencies that independently investigate charges of discrimination. Should the review board decide that there is not a “clear, uncontestable 7 case of malpractice” the case could go “to the health court for a full trial.” There, “expert witnesses would be hired by the health courts, not by plaintiffs and defendants.” The document does not state whether the expert witnesses would render opinions independent of that of the review board, i.e., that one or more its neutral experts would present a “minority report” to the health court. If so, the claimants would benefit from two opportunities to have an independent expert potentially provide the testimony needed to prevail in their claim. This could be of great benefit to patients who, under the tort system, not only must convince a lawyer that their claim is worth investigating but are also ultimately responsible for paying a high fee to the expert witnesses who testify on their behalf. The Harvard/Common Good Health Courts Proposal Skeleton envisions a process seemingly less advantageous to injured patients than the Udell/Kendall scheme. The Harvard document describes “a first level of review [that] would be an internal process at the involved hospital or insurer. This level of review is not intended to be a neutral adjudicatory process, but rather a formal mechanism for encouraging expeditious settlement of claims.” The document compares this process to a program currently operated by the COPIC medical liability insurer in Colorado. If an offer recommended after this review “did not lead to a resolution, then a health court hearing would be held on a prompt basis…an administrative law judge who specializes in health court claim adjudication would be assisted by medical experts with relevant expertise who come from a panel constituted through volunteers or selection by the court.” The Harvard draft indicates that the initial review panel would “operat[e] under state oversight and with discretion constrained by a legislative mandate to apply pre- established decision aids and damages schedule.” Presumably the constraints would include a requirement that settlement offers be made in good faith, i.e., that the insurer agree to settle all meritorious claims and that the dollar amount of the offers conform with the damages schedule, i.e., that the insurer refrain making discounted or “low-ball” offers. Both of these would represent new mandates on liability insurers that run counter to such insurers’ customary practices. Under the tort system, whether a settlement takes place and its dollar amount are influenced by the parties’ calculation of the likelihood that the plaintiff could prevail before a jury. A no-fault system should eliminate from this calculus the “haphazard” or normatively “inappropriate” factors, such as the appearance and manner of the parties, the skills of the lawyers and hired experts, or the predisposition of local jury pools. But the initial reviewers, even facing a good-faith mandate and potential sanctions, will, as agents of the defendant or insurer, be inclined to deny claims that would seem difficult to prove. It is hard to imagine that an insurer-appointed review panel would grant compensation as often as an independent panel would in close-call or fact-intensive cases. And it would certainly be the case that if the plaintiff needed to retain an attorney and a medical expert in order to prevail, the number of claims paid would not be much greater than the number paid under the tort system. 8 The specifics of the claim process will probably not matter in what might be called “classic” accelerated compensation events (ACE)—patient outcomes that are clear-cut cases of avoidable injury, apparent from the patient’s condition.14 An example would be a surgical error in which during reversal of colostomy, the patient’s bowel is connected to his bladder rather than distal portion of bowel, with the result that the patient is passing feces while urinating. In such a case, it is difficult to imagine either an independent or insurer-run review panel quarreling with the claimant over his eligibility for compensation. More problematic is the latent ACE, in which the overall context of the patient’s treatment and illness course must be examined. An example would be a “failure to diagnose lump in breast of hospitalized woman early enough to minimize the potential consequences of metastasis and ultimate death.” Whether a diagnostic test that was taken (or perhaps should have been taken) during a hospitalization a year or two years before the patient’s death, properly interpreted, would have revealed a tumor in time to save the patient’s life would not be immediately apparent to the decedent’s family; or even to the patient’s physician if it was not the same physician who had admitted the patient to the hospital previously. If the family members made the connection on their own and filed a claim, the merits of the claim might not necessarily be obvious to expert reviewers. For instance, a proper test may have been taken but with ambiguous or equivocal results, such that two experts could disagree as to whether the patient’s tumor was evident or death was avoidable. In this event, it would matter a great deal whether two opposing expert opinions reached by members of the reviewing panel would both be presented at a full hearing before the health court. If the reviewing panel decided to reject the claim, would the surviving spouse and children be required to retain a paid expert witness? Which side would have the burden of going forward with evidence and the ultimate burden of proof? In fact, similar situations and questions have arisen under the “no-fault” Virginia birth injury program.15 For the health court to fulfill its promise of compensating more patients in the injury “overhang,” structures and procedures will have to be crafted to ensure that avoidable injuries are disclosed or otherwise discovered, and that claimants have access to expert opinions that can make out a case for compensation not only in clear-cut cases but in fact- intensive or close-call cases as well. Claimants’ interests could be maximized by ensuring that there is an independent initial review; that reviewers are able to flag adverse events that were likely compensable for further investigation, even those that are not disclosed to or discovered by patients;∗ that initial review of each claim is provided by more than one expert, so that close-call cases will be flagged for further proceedings; and that the opinion of any initial reviewer in the claimant’s favor is allowed to satisfy the claimant’s burden of going forward with evidence, which is to say, that the claimant would not need to hire and pay an expert witness in order to move a denied claim forward for review by the health court. Another patient protection would be to impose some type of penalty on providers who do not initially disclose but are later found to have committed a medical ∗ For example, a review board could apply AHRQ’s Patient Safety Indicators computer algorithm to hospital discharge data or increase the number of autopsies of deceased patients. 9 error—one commentator, Paul O’Neill, suggests that patients receive treble damages if a cover-up is found.16 3. Equity Versus Equality in Damage Awards: “Predictability” and Its Effect on Deterrence One of the values embraced by no-fault advocates is “predictability” of awards, a concept that is sometimes stated in the negative sense, which is to say, the absence of “haphazardness” in awards. Some find haphazardness an objectionable trait of the tort system because similarly situated claimants do not receive similar awards. As noted earlier, differences can result from circumstances such as the “attractiveness” of the plaintiff and defendant, the local jury pool, or other random factors not related to the merit of a claim or the damages suffered by the claimant. Predictable awards have an additional benefit to providers and insurers: actuaries are better able to forecast payouts and set rates, reducing the possibility of large spikes in insurance premiums. The policy quandary arises from the proposition that the tort system’s haphazardness has virtues. This proposition holds that an element of randomness in the calculation of awards prevents those engaged in hazardous activities from making precise cost/benefit calculations in deciding whether to take a safety precaution, thus creating a bias in favor of greater precautions; and creates equality of opportunity for any individual claimant to receive a high-end damage award regardless of his or her demographic characteristics. How replacing “haphazard” awards with “predictable” awards could alter provider incentives is illustrated by the example of the current debate within health care provider organizations whether to purchase monitors that can detect “anesthesia awareness.” This terrifying experience of hearing and feeling invasive surgery while being paralyzed and unable to communicate can cause post-traumatic stress disorder, alcohol abuse, and fear of receiving medical care.17 In a no-fault system with scheduled damages, a hospital could use the incidence rate of intraoperative awareness and the number of surgeries it performs, multiplied by the award schedule, to determine whether it would be worth the cost to purchase a monitor. In the tort system, a precise calculation of damage awards is impossible, potentially increasing the incentive to buy the monitor. Further, because the size of no-fault awards under a damage schedule is in larger part dictated by the patient’s age and salary, a hospital could determine, by reference to age and occupation, whether the monitor should be used on an individual patient. The incentive structure would favor use of the device on younger, highly paid patients, but not on older, retired patients. By contrast, the possibility of a “haphazard” jury award with a large pain-and-suffering component discourages such gaming. It is assumed under both the tort and no-fault regimes that equity, rather than equality, should be the guiding principle for calculating damages; that is, that awards should vary in accordance with the severity of the injury and its costs to the claimant. Critiques of noneconomic damage caps that argue against their adoption because of their disparate impact on female and elderly claimants are, of course, arguing that equality of result should at least be a consideration in deciding damage award rules. But nobody has 10 seriously argued that equality should predominate over equity in the tort system because the availability of “random” punitive damage or non-pecuniary damage awards has obviated any need to choose between the two principles. A fundamental redesign of damage awards that elevates predictability over other considerations forces policymakers to think about this dilemma. Unless protections can be guaranteed that prevent discrimination, policymakers must at the very least give thought to whether damage awards should be based at least in part on injury severity, and not entirely on the claimants’ actual monetary losses. 4. Boundary Between Tort and No-Fault A key design consideration facing architects of a no-fault scheme is defining the limits of its jurisdiction vis-à-vis that of the tort system. Several commentators would explicitly limit no-fault to certain cases; others project an administrative system to eventually encompass 100 percent of all medical malpractice cases, but with a transition period during which there would be less than complete coverage. Thus far, proposals have placed boundaries on two dimensions of characteristics of a case arising within a hospital. The first of these “axes” on which intersecting lines could be drawn to demarcate boundaries between tort and no-fault jurisdiction might be called “individual characteristics”—case- or patient-specific attributes. Four possibilities are these: • Medicare patients would be covered by no-fault while other patients would remain in the tort regime, per Sage and Kinney.18 • Cases would be covered by no-fault with the exception of those involving “egregious, patently negligent errors,” which would be eligible for tort adjudication, per Robert Wachter.19 • Cases that fit the criteria for accelerated compensation events—“discrete events” that are “easily identifiable” as errors and “distinguishable from any similar outcomes that are less avoidable”—would be covered by no-fault, but “injuries that are not ACEs would remain in the tort system,” per Tancredi and Bovbjerg.20 • Cases involving “moderate injuries,” with damages below a $50,000–$80,000 threshold, would be covered by no-fault, with more serious injuries remaining in the tort system, per Tom Baker.21 The other axis on which intersecting lines of demarcation can be drawn is “clinical areas.” This type of distinction is the basis for the currently extant no-fault systems in the United States—the Florida and Virginia programs for neurologically impaired newborns, which were intended to carve out obstetrics from the tort system. The Harvard/Common Good draft proposes that no-fault be limited to “select clinical areas such as obstetrics and surgical/anesthesia,” perhaps because the groundwork for developing ACE descriptions in those areas has already been laid. How boundaries are drawn have two types of impact: the impact on patients, in terms of their eligibility for tort versus no-fault compensation processes; and the impact on 11 providers, in terms of changing their mind-set to address hypothesized perverse incentives created by the tort system. It is beyond the scope of this paper to discuss at length the relative merits to the claimant of tort versus no-fault compensation. It is generally agreed that plaintiffs’ goals in seeking compensation under the tort system include the following: • Meeting financial needs associated with the injury; • Finding out the true story of the cause of the injury; • Preventing future medical errors of the same type; • A desire to hold a perceived wrongdoer accountable, i.e., retribution. For the individual claimant with a strong case, the main advantages of the tort system in meeting those goals are its relative generosity of compensation, its broad array of fact- finding tools (e.g., discovery, subpoena power), and its dramatic culmination in a trial, where a verdict for the plaintiff constitutes a public rebuke of a negligent provider. The principal drawbacks to the tort system for potential claimants are the cost associated with the lengthy and complex proceedings and the high hurdle of proving negligence. For claimants, the optimal boundary is one that preserves the advantages of the tort system in cases where its retention is most warranted, but makes quick administrative compensation available in cases where transaction costs make tort claims impracticable. Three of the boundary proposals cited above had this aim in mind: the rationale for the Sage and Kinney proposal to move Medicare patients to no-fault was that their lack of wage earnings leaves potential damage awards too low to justify, economically, the cost of bringing a tort claim; Baker’s proposal for a dollar threshold recognizes the same dynamic; and Wachter’s proposed boundary recognizes the greater importance of accountability and retribution in cases involving a higher degree of culpability. From the provider point of view, the boundary question is wrapped up in the phenomenon of “legal fear” thought to pervade the medical profession as a result of its exposure to tort liability. A no-fault system is hypothesized to remove this burden, leading to the reduction of two perverse incentives. First, providers will be more open in discussing errors if they are not subject to tort liability. This should mean that errors will be disclosed to patients, and reporting of errors will create a rich trove of candid first- person accounts of the cause of errors, producing insights into patient safety that cannot be achieved from the second- and third-person accounts (“finger-pointing”) that are generated in the tort system. Under the tort system, the narratives of participating providers will likely be guarded, defensive, and less than candid as providers try to avoid the public rebuke that can result. In addition to discouraging candor in the investigation of errors that caused injuries, the tort system also discourages self-critical reports of near- misses or hazardous conditions, because these reports, if not acted upon, may touch on underlying causes of future injuries and form a paper trail that would help prove negligence as the cause of such injuries. 12 The second perverse incentive to be avoided is “defensive medicine.” The defensive medicine theory holds that doctors do not limit diagnostic testing or treatment to that deemed medically necessary, but rather tailor it to what they believe a lay jury thinks the treatment ought to be, resulting in the over-utilization of resources. If doctors knew that their medical judgment would be reviewed by experts rather than lay jurors, it is argued, they would order only the tests and treatments dictated by sound medical practices. Because both of these perverse incentives involve the subjective feelings of providers— fear of lawsuits, anxiety about public humiliation, lack of confidence in the legal system—the no-fault system must not only be objectively capable of remedying perceived deficiencies in the tort system, it must in fact remove those perceptions from the minds of providers. It would not suffice for respected hospital leaders to assure doctors and employees that the tort system no longer need be feared; the doctors and employees must believe it. The problem with boundaries is that a less-than-complete switch to no-fault would be expected to produce a less-than-complete belief among providers that the tort system need not be feared. Consider a third dimension of case types corresponding to support services departments within a hospital in which a medical error could take place. These departments, such as radiology, pharmacy, and laboratory, cut across clinical areas and error characteristics. For instance, a radiology department could be responsible for aspects of care of a patient who is a Medicare beneficiary or an under-65 patient and for an obstetrics patient as well as an angiography patient. Suppose that in a hospital radiology department, an employee realizes that there is a recurring problem with patients being sent from various other parts of the hospital without notification as to other drugs the patient may be taking. Ideally, we would want this employee to record instances of errors or near-misses, along with suggestions on how to resolve the problem, and communicate the information up the hierarchy. The employee would be more likely to take this step in a no-fault legal environment. However, if no- fault applies only to surgical or obstetrical patients, the memo could provide a “smoking gun” if such an error were to occur in a patient sent from the emergency department. In those circumstances, it is unclear whether hospital leaders would be any more likely to encourage the writing of such a memo than they are under the current legal regime. Even within the clinical areas covered by no-fault, if there are case-specific boundaries that blur practitioners’ understanding of its applicability, perverse behaviors may continue. Consider the case of a surgical error that the surgeon is aware of but could remain hidden in the absence of its disclosure to the patient. If the boundary between no- fault and tort were a bright line, such as Medicare versus non-Medicare, the surgeon could be fairly confident that an error during a Medicare beneficiary’s surgery could be disclosed without any tort consequences. But if the boundary would become apparent only after a threshold event is triggered—the patient suffers damages in excess of $50,000 or alleges “egregious negligence”—the doctor would remain reluctant to make any disclosure. Similarly, within the clinical area covered by no-fault, there would remain 13 reluctance to record near-misses or general self-critical observations if the tort system retained certain cases. The preceding discussion illuminates two major dilemmas in the demarcation of boundaries between tort and no-fault: The more pro-claimant the design of the program, the less effective the program will be in combating perverse incentives; and the more conservative the program is in phasing in no-fault incrementally across clinical areas or case types, the less effective the program will be in combating perverse incentives. A fourth dimension in which a no-fault system would operate is time. The possibility of future invalidation of a no-fault program also would hinder the change in attitudes that no-fault is meant to effect. It is not unusual for actuaries to wait until a state’s highest court has denied constitutional challenges to a “tort reform” law before adjusting their forecasts to account for the new law. Presumably doctors would exercise the same caution in relying on a no-fault program to govern their actions. This means that if a no- fault program is established by state statute and subject to challenge as violating the state constitution, it could take several years before doctors and hospital personnel feel confident that the program has in fact extinguished their tort liability. Another question is provoked by the Sage and Kinney proposal to bring Medicare beneficiaries under a no-fault system. Would this constitute a benefit to the Medicare enrollee or a ghettoization of the Medicare population? The answer would depend on the generosity of the compensation and the effect that separate systems would have on provider behavior. The rationale for beginning no-fault with Medicare beneficiaries is the belief that elderly injured patients are less likely than their younger counterparts to secure legal representation. A generous no-fault scheme could lead to Medicare patients in the aggregate receiving more compensation. On the other hand, if awards were low and their predictability relieved providers from any prospect of a “blockbuster” jury verdict, a slackened deterrent effect could lead to Medicare patients receiving second-rate care. Finally, if a no-fault regime covered only Medicare patients, and the other half of a provider’s clientele remained under the tort system, would the provider fully assimilate the mind-set of greater openness that no-fault is meant to inculcate? 5. Standard of Care and Evidence-Based Medicine One of the most intriguing possibilities raised by a change from the lay-dominated fact- finding of the tort system to expert-based fact-finding is that of raising the standard of care required of providers. The Harvard “Skeleton” draft envisions a standard of “avoidability,” which it describes as a more demanding requirement of providers than the current negligence standard. Specifically, Avoidable events are injuries that are caused by treatment (or omission of treatment) and that could have been avoided had care been provided according to best practice. In other words, an injury is deemed avoidable if it might have been prevented had a better system of care been in place. 14 Udell and Kendall wrote that in developing the health courts’ new “science-based common law,” “neutral experts would rely on scientific literature and consider evidence- based guidelines listed in the National Guidelines Clearinghouse.” One health policy expert cited by health court promoter Philip Howard is David Eddy. Eddy wrote at length about courts’ consideration of expert evidence in a 2001 article. In it, Eddy noted that, when judging the quality of medical care in a malpractice lawsuit, courts apply what might be called an “internal test.” Rather than try to determine the standard of care for itself, the court hears the testimony of experts from within health care who are expected to know the standard of care and compares their answers to what was done. What the court actually hears depends on what the lawyers want to present, but it generally involves the answers to any of three questions: What do the majority of practitioners do? (What is the “community standard”?) What do groups of experts say (e.g., the recommendations of specialty societies or national panels)? Or what does an individual expert believe?22 Eddy further argues that the premises upon which the “community standard” is based are fallacious, citing the hundreds of studies of variations in practice patterns and inappropriate care. When rates of procedures vary across communities by factors of two, five, ten and more, and when 10 percent, 20 percent, 50 percent, even 70 percent of practices are judged by peers and experts to be inappropriate or equivocal (providing no advantage of benefits over harms), it is impossible to believe that there is a single community practice out there that the majority of practitioners are following. Indeed, given the very high rates of inappropriate care that can prevail in communities, if we actually measured what practitioners were doing and used that to define the standard of care, we would run a high risk of installing an inappropriate practice as the standard of care. The well-documented overuses of hysterectomies, antibiotics, bypasses and C-sections are examples. Eddy posits an alternative in which the “community standard” is abandoned in favor of evidence-based medicine. If the Health Court abandoned the “community standard” and imposed a regime of evidence-based medicine, a plaintiff who received inferior treatment could bring suit, marshal the scientific evidence favoring effective but not yet customary treatment, and receive a ruling endorsing the new approach. If the ruling became binding precedent, this would have the effect, promised by Howard, of raising standards and improving quality. Presumably at the very least, this would have the effect of elevating practice guidelines that have received general approbation among experts and advocates in the health care quality movement to constitute the standard of care. For instance, consensus guidelines recommend preventative antibiotic administration prior to surgery.23 In practice, 15 however, according to Medicare’s HospitalCompare databank,24 this occurs in only 70 percent of surgeries nationwide; in some places, such as Texas, the percentage is lower still (60 percent). Imagine a case in which antibiotic prophylaxis does not occur as recommended, and the patient is injured by an infection. Under the tort system, the plaintiff would have to obtain an expert witness to testify that this violated the standard of care. If the case occurred in Texas, a defense expert could truthfully testify that in two out of five surgeries, doctors deem this precaution unnecessary. Under the health court rules, it appears that the claim would be decided in the patient’s favor unless the doctor had good cause, justified by findings in the medical literature, for not ordering the antibiotic. Another possible innovation suggested by a more exacting standard of care is the use of low rankings on performance measures as evidence or proof of an avoidable, compensable injury. As Atul Gawande speculates in a 2004 New Yorker article,25 a court could look at the distribution of average outcomes produced by comparable providers of a particular treatment, e.g., heart bypass or cystic fibrosis centers, with “being in the bottom half [of performers] be[ing] used against doctors in lawsuits.” This possibility was further explored by Kesselheim and colleagues,26 who concluded that performance measures would likely be inadmissible in tort trials. Long-standing rules of evidence prohibit jurors from considering a defendant’s conduct beyond the parameters of the individual plaintiff’s treatment, because of such evidence’s potential prejudicial effect. But if experts, rather than lay jurors, are evaluating the evidence, the rationale for the prohibition crumbles. An expert panel could carefully weigh the value of performance measures and determine whether a particular provider’s rating falls so low as to constitute unacceptable practice. A legal system that imposed higher quality standards would be good for consumers, but several potential pitfalls are apparent: • Because standards would be raised retrospectively in individual cases, doctors would not be free from the “fear of lawsuits” that is the primary rationale Howard cites for his proposal. While a legal regime of evidence-based medicine would “provide incentives for doctors to keep up with the latest developments in medicine,”27 they would still face dilemmas when confronted with conflicting guidelines. One commentator predicts that this approach would have the effect of “increasing the amount of uncertainty physicians face in determining what the law requires of them.”28 • Howard says that “a reliable medical court could provide incentives for improvements in health care—in systems or in the use of medical technology— without the heavy hand of regulation.”29 But would the piecemeal development of legal standards of care via litigation be less heavy-handed than regulation or market approaches (report cards, pay for performance)? 16 • Howard’s central critique of the legal system is that “doctors and other providers no longer feel comfortable making sensible judgments.”30 But as Eddy notes, if the “community standard” is replaced by an evidence-based medicine regime, the primary impact would be to reduce legal deference to providers’ “clinical judgment.” This deference is the reason that doctors win about three-fourths of all malpractice jury trials. If the Health Court were to make “binding rulings about what is good care and what is not” it would mimic the Washington Supreme Court’s controversial decision in Helling v. Carey,31 which overruled experts’ testimony as to customary practice and made a ruling as to what doctors should be doing. This would constitute a potential expansion of malpractice liability as it relates to quality of care, possibly making providers responsible for more costs. In adopting the “community standard” approach, the courts made a conscious decision to defer to medical providers in recognition of the fact that judges and juries lack scientific expertise. Special medical courts staffed by expert judges would not have that limitation, and the trade-off that the common law made in adopting a simplified standard—a lower quality threshold—could be reversed in a redesigned system. This is an intriguing proposal and a higher quality standard might be seen as a fair exchange for consumers’ loss of tort rights. But it seems unlikely that medical providers would endorse a proposal that reduces their autonomy and increases their liability exposure—the American Medical Association has said that no-fault must not “[lower] the burden of proof or the evidentiary requirements claimants must meet.”32 Certainly any hospital or health system opting for the arrangement would have to be quite certain that the professionals practicing within it could meet this quality standard. Assuming liability under these circumstances would be the harsher, opposite tack of pay-for-performance; it would constitute a penalty for nonperformance. 6. Standard of Care and Defensive Medicine While the Udell and Kendall passage cited in the previous section suggests that the health court standard of care would be higher, other passages in that document, as well as the writings of Phillip Howard, suggest that the standard of care could, at least in some circumstances, be lower. This is because of the health court’s promoters’ focus on “defensive medicine.” In the absence of clear legal standards, it is only natural that doctors order extra tests to limit their exposure to lawsuits—a practice that adds unnecessary services to the health care bill. Under what circumstances is a CAT scan excessive or unnecessary testing? Is it ever OK not to do another test? The current medical justice system, with its lack of precedents and perverse incentives, simply cannot support rational, science-based cost containment measures.33 The positive defensive medicine theory holds that the tort system promotes unnecessary diagnostic tests because doctors fear liability. The precise definition of defensive medicine is unclear, but it involves circumstances where “a patient present[s] with a probable minor condition but [there is] a small chance of a potentially very serious or fatal condition.” For its 1994 study, the Office of Technology Assessment (OTA) defined 17 defensive medicine as “tests, procedures, or visits [ordered]… primarily (but not necessarily solely) to reduce their exposure to malpractice liability.”34 Lewin-VHI’s study recognized, however, that defensive medicine might also encompass “interventions (e.g., sophisticated diagnostic imaging) … carried out when the expected potential benefit to the patient is extremely small.”35 The OTA definition seems to speak to the idea of the doctor’s professional judgment, which may in fact be informed by the Lewin-VHI notion, which is essentially one of cost-effectiveness. That is, as doctors think through their diagnosis they may be implicitly applying a cost/benefit analysis when deciding whether a particular test is warranted. How would having an expert health court bear on these issues? There are various ways in which its rules or effects could diverge from the tort system’s, corresponding with various different possible states of the medical evidence and expert consensus. Consider a hypothetical clinical scenario in which a doctor sees a patient in circumstance A and must decide whether to order procedure X. It could be that there is a surfeit of medical evidence, and a consensus on the best practice, both pointing to an affirmative decision. Antibiotic prophylaxis for surgery would be such an example; there is no question that antibiotic prophylaxis is not a “defensive” practice, and if the doctor fails to administer it and injury results, compensation should be paid. Or it could be that the medical evidence and expert consensus point to a negative decision. An example would be preoperative electrocardiograms (ECGs). By the 1980s it was evident that such tests were not warranted in surgical patients younger than 40 years of age who had no history of heart or circulatory system disease, and, encouraged by “published research, organizational endorsements and cost-containment measures,” by 1987 the rate of such tests had dropped below 10 percent.36 The researchers who tracked the decline in this testing, Macario and colleagues, speculated that the small amount of continued testing may have been motivated by liability fears. Presumably the benefit of replacing lay jurors with expert judges would be to free doctors from fear that a hired expert retained by a plaintiff’s lawyer could persuade jurors to ignore the expert consensus. If so, the incidence of these “unindicated electrocardiograms,” if they were motivated by liability fears rather than force of habit or disagreement with guidelines, would drop to zero. It could be that there is no evidence on the usefulness or cost-effectiveness of test X for this patient history, and doctors have relied upon their own judgment in deciding whether to order them. OTA’s 1994 study identified numerous clinical scenarios in which surveyed doctors split on the need to order certain tests. For instance, doctors were split 50/50 in their responses to a scenario involving a 42-year-old jogger complaining of chest pains: half would have ordered an exercise ECG and half would not. Of the half who would order the test, 17 percent cited fear of liability. The vast majority of doctors surveyed sincerely differed in their clinical judgment. 18 Figure 2 charts the three possible states of medical evidence on a given procedure. In area II are procedures like antibiotic prophylaxis for which consensus dictates that the intervention take place. Procedures in this area clearly do not constitute defensive medicine. In area I are those procedures like preoperative electrocardiograms for which consensus dictates that the intervention need not take place. If a doctor orders such a procedure out of fear of liability, the procedure unquestionably constitutes wasteful defensive medicine. Figure 1: Continuum of Medical Evidence Basis for Procedure II Necessary procedures O I Unnecessary procedures In area 0 are procedures for which there is no expert consensus. OTA estimated that about 8 percent of the time these procedures are motivated by fear of liability rather than a genuine belief in their usefulness. What is the best legal rule to be applied in this gray area when the test is not given and the patient dies? In the tort system, dueling experts can offer their opinion and jurors choose which expert to follow. This is said to lead to “haphazard” results, which, over time, could influence additional cautious doctors to order the test. Some passages in the Udell and Kendall document suggest that, in the absence of consensus, the health court’s decision rule would be to deny liability “if the provider’s actions were within the range of reasonable actions given the circumstances.” Would this constitute an improvement over the tort system’s “haphazard” results? At bottom, either system is still at the mercy of imperfect information. In the tort system, imperfect information results in the case going to the jury, with the outcomes “randomly” alternating between liability and no liability. The health court could take as its decision rule a blanket denial of liability until such time as evidence and consensus dictate otherwise, but because the health court has no better idea of what future medical research 19 will reveal than do jurors, the blanket denials are no more likely to lead to the correct outcomes than are jury verdicts. This would have the benefit of saving the transaction costs involved in jury trials. It would also save money in terms of utilization forgone in the absence of conclusive evidence. However, if doctors refrained from ordering procedures thought to be helpful but, as yet, unproven to be so, some patients would, in the end, turn out to have been incorrectly diagnosed or treated. Indeed, it would discourage tests such as the jogger’s exercise ECG that some 30 percent of doctors felt was warranted. Yet another quandary is that if unproven procedures were eliminated across the board, researchers would not be able to conduct retrospective comparisons of outcomes in patients given or not given test X. Another possibility would be to continue to apply what Eddy called the “internal test” of what “the majority of practitioners do,” but to have this determination made by experts rather than jurors. Instead of drawing the liability line at either the top or bottom of area 0, this approach would install a new line within area 0, adopting the majority rule until such time as medical evidence breaks one way or the other. The recognition of the health court as an expert decision maker capable of distinguishing justified care from unwarranted care could be sufficient to eliminate defensive medicine as described by Macario et al.; that is, testing that consensus holds is unnecessary and is motivated solely by liability fears. But there other factors at work that complicate the ability of a health court to eliminate unnecessary testing. The first, mentioned above, is imperfect information. As Macario et al. noted, it was not until “many assessments of medical technology… concluded that routine preoperative tests… yield an extremely low rate of true-positive test results” and that this research was “presented at national meetings and endorsed by national societies” that procedures such as the preoperative electrocardiogram were conclusively moved from the realm of category 0 to category I (“unwarranted”). It would seem unlikely that a health court could succeed in reducing unnecessary medical procedures in the absence of more comparative effectiveness data. The second is the role of fee-for-service payment as both an enabler and an accelerator of overutilization of health care resources. The theoretical model of defensive medicine assumes demanding patients who will sue if the cost-ineffective measure is not taken and reluctant doctors who would prefer not to order an unwarranted diagnostic test or procedure. But in settings where there is insurance, patients demand the test because they do not bear its full cost. Conversely, if patients must pay out-of-pocket for the test, doctors might very easily persuade them that the test is a waste of money. Meanwhile, in fee-for-service settings, doctors may profit by providing the test or procedure. Unless doctors treat patients on a capitated basis, doctors have no financial incentive not to order the test; forbearance must be motivated by something beyond their dismay that medical resources are being unnecessarily deployed. Finally, there is the inevitable discrepancy between individual and social risk preferences. Economists or other experts, after reviewing the medical evidence, may determine that it 20 is not cost-effective over a larger population to perform procedure X in circumstance A and relieve the doctor of liability for not performing it; but an individual patient who is risk-averse may still insist that it be given. Defensive medicine is widespread in the Netherlands, even though malpractice lawsuits are rare.37 In Dutch practice, which is typically capitated, the term is applied to tests ordered to reassure a worried patient who requests further testing; to avoid being reported to disciplinary authorities; or to avoid conflict in the doctor/patient relationship. In fact, the main factor pushing down the use of unnecessary preoperative testing during the 1980s was not any slackening of the tort system but, according to Macario et al., the fact that insurer “cost containment measures began to include threats of punishment.” Thus it would seem that the health court could not be expected to reduce unnecessary testing in the absence of more medical evidence on tests’ usefulness as well as some type of coercive cost containment pressure from insurers. Moreover, those doctors who see their role as advocates for patients may bristle at a health court’s suggestion that they ought not provide a treatment that has some net benefit for the patient, just as they have when managed care organizations impose such constraints. 7. Evaluating the Success of a Demonstration A no-fault demonstration should be evaluated on how well it achieves its twin goals of broadening compensation and increasing patient safety. A comprehensive evaluation would study change on both these dimensions by benchmarking compensation and safety performance under the tort system. This would require either taking measurements of compensation and error rates in the same unit (hospital, integrated delivery system, or other entity involved in the pilot) before and after implementation; or in comparison units; or, ideally, both. A proper research design would also include as a comparison unit providers using so-called “apology” or “voluntary resolution” programs.38 Those programs use some of the techniques of no-fault but operate within the tort system rather than under an alternative regime. Such programs may achieve some of the aims of no- fault, but they have not been studied so there is no publicly available data about their performance. If they do so, creation of formal no-fault programs may prove to be unnecessary. Studying the effects of no-fault on the distribution of malpractice insurance funds would be the most straightforward evaluation task. A successful pilot project would distribute a larger portion of the liability insurance dollar to patients and a smaller portion to both plaintiff and defense attorneys. At the same time, one would expect insurance funds to be distributed to a larger number of patients. But the broader distribution of insurance funds to patients would not necessarily in itself constitute a success. Ideally, all injured patients under each system should be surveyed to compare the overall quality of those patients’ recovery from the injury. Of greatest interest would be how all injured patients fare financially, i.e., what percentage are able to maintain their lifestyles without resort to bankruptcy, public assistance, or demands on family members or charities. It would be a hollow victory for no-fault if, despite a greater number of patients receiving medical injury awards, net hardships for the entire 21 population of injured patients remained the same because the greater breadth of compensation was offset by lower depth. Another possible negative outcome would be that if injured patients no longer self-select for compensation claiming as they do under the tort system, compensation might be redirected to patients less in need of awards (i.e., those who have other resources and might refrain from claiming in a tort environment). Also of interest would be whether no-fault eliminates the “compensation effect” thought to hinder recovery from illness in patients with pending tort claims.39 Evaluating improvements in patient safety will present several research issues. One avenue would be to compare overall error rates. This could be accomplished through any of several methodologies: chart reviews,40 observation,41 or surveys of patients.42 Another avenue would be to evaluate improvements in communication about errors, which is agreed to have instrumental value in improving patient safety. This could involve comparisons in error reports made within the unit or to a state patient safety agency; observation of morbidity and mortality conferences; or surveys of patients who report experiencing a medical error as to whether they were informed of the error by hospital personnel. A study of defensive medicine could also be undertaken as part of the demonstration. Here, the methodology of Macario et al.—reviewing charts to determine how often unwarranted tests were given—would reveal whether physicians’ greater confidence in an expert tribunal can reduce costly over-utilization. As the above discussion demonstrates, a thorough evaluation of a no-fault demonstration could be almost as arduous and labor-intensive as the design and implementation of the project. But anything less than a thorough evaluation would allow the many controversies over no-fault’s merits to linger as they have over the past thirty years. 8. Public Oversight Versus Private Initiative A final question to consider is, Under what auspices would a health court operate? Would it be a public court system or a private arbitration system? No-fault and health court proponents have devoted relatively little discussion to the structure, staffing, or oversight of the tribunals that would administer their proposed programs. The most detailed proposal for an administrative infrastructure is found in Senator Enzi’s proposed “Fair and Reliable Medical Justice Act” (MJA), a bill introduced during the 108th and 109th Congresses.43 MJA, which would authorize state demonstration projects, envisions two possible entities: • A state “Board” that would include representatives of medical licensing agencies, patient advocacy groups, medical providers, and attorneys. This Board would adopt the ACE classification scheme, set the compensation schedule, and establish procedures for claim hearings and appeals. 22 • A special court “presided over by judges with health care expertise who meet applicable state standards for judges,” which could also be advised or guided by the Board described in the first paragraph. In either event, under MJA, there would be a second level of oversight from a federal Review Panel that would include stakeholder representatives to be appointed by the Comptroller General. While many might take for granted the idea that a judicial system would operate under government auspices, there has in recent years been a trend toward private adjudication. The Federal Arbitration Act permits parties to a contract to opt out of the court system and refer disputes to private arbitrators. Such provisions are increasingly found in contract forms drafted by health care providers.44 Given this trend and the potential difficulty in passing no-fault legislation (as well as in fending off challenges to its constitutionality), health care providers and their malpractice insurers may find it expedient to impose a no-fault regime by use of an arbitration clause. Patients’ assent to a no-fault system through an arbitration clause would be similar to their assent to a government-sponsored demonstration: presumably, in either case, patients would be given notice and an opportunity to opt out. The exercise of the opt-out would most likely require patients to use a different health care provider, which may cause inconvenience or discontinuity in care. The key difference would be that establishing no-fault through an arbitration clause would allow a health care system and its insurer to design the system unilaterally, rather than through the give-and-take process involving stakeholders envisioned by the Enzi bill. A provider-designed system could be contrived to reduce the provider’s overall liability, which has been the goal of the provider community in both its legislative campaigns and arbitration efforts. This goal could be accomplished by establishing high procedural hurdles for plaintiffs, a stingy schedule of benefits, restrictive definitions of ACEs, biased tribunals, or any combination of the preceding. If the provider’s overall liability is lowered, the incentive to provide safe medical care is diminished. As such, no- fault-by-arbitration could pose considerable potential dangers to patients. Conclusion Replacing medical malpractice lawsuits with a no-fault system could offer broader compensation of injured patients while changing medical providers’ attitudes and behavior in confronting medical errors. Designers of such a system face a difficult task in reconciling the competing values and needs at stake. Striking a balance too far in the direction of relieving providers’ litigation burdens could result in patients losing not only the unique form of agency they wield in tort litigation but also some of their awards. In order to ensure that consumers benefit from a no-fault project, some basic protections must be built in: 23 • Initial review of claims should be made by an independent entity rather than an insurer-dominated entity and each claim should be reviewed separately by more than one expert. • If at least one expert reviewer is convinced that a patient’s injury was preventable, the claim should proceed to a full hearing with that expert’s opinion admitted for consideration. • If the demonstration project participant pays out less in compensation than it would have expected to pay under the tort system, the difference in funds should not be retained by the provider or insurer unless an evaluation of the project reveals a decrease in the error rate; in other words, there should be no incentive to manipulate the system to reduce net payouts. • Any pilot or demonstration project should be conducted under governmental auspices and public supervision; it should not be created solely by contract with patients or purchasers. A final question meriting consideration in discussion of a switch to no-fault malpractice compensation system is: What can we realistically expect in the way of improving the quality of medical care and dispute resolution? Some recently gathered data suggests that we should not count on vast improvements in either patient safety or the reliability of malpractice adjudication. A recent Commonwealth Fund survey of sicker adults in six countries found that in New Zealand, which uses a no-fault system, patients were somewhat more likely to have medical errors disclosed to them by providers: 61 percent of New Zealand patients reporting an error were not told by providers compared to 70 to 83 percent in the other five countries. But the rate of medical mistakes made in New Zealand was not significantly lower than in the other five countries. A number of commentators have noted that there is no evidence demonstrating a chilling effect of litigation on providers’ open discussion of errors and unsafe practices.45 Others believe that it is fear of disciplinary action, not litigation, that chills reporting,46 which might explain why even in New Zealand the majority of errors remain undisclosed. Then there is the question of the tort system’s reliability. Much of Common Good’s call for expert adjudication is premised on the beliefs that, as Udell and Kendall write, “the medical justice system… makes inconsistent and often incorrect judgments” and that “for every valid claim, four unfounded claims are filed.” While those are hyperbole, there is a large margin of error in tort adjudications and settlements. According to the most recent study of closed malpractice claims,47 about 37 percent of the claims brought by attorneys involve no medical error. In about 27 percent of claims, the tort system reaches the wrong result: either an award was paid to a patient without a medical error or a patient who should have been compensated was not. Critics have attributed these dysfunctions to characteristics unique to the tort system, such as lay juries or contingency-fee attorneys. But another possibility is that such dysfunctions are inherent in the process of human interaction with scientific knowledge. 24 Consider the finding of Dartmouth researchers that about one-third of Medicare expenditures are unnecessary—medical treatment ordered by physicians that provides no improvement in health outcomes.48 This finding indicates that doctors and lawyers overuse their respective systems at almost exactly the same rate. Consider also the finding of the Rand study of health care quality, which found that patients receive appropriate care, as judged by consensus guidelines, only about 55 percent of the time.49 By this benchmark, the tort system slightly outperforms the health care system in getting it right. Actors in the tort system—lawyers, experts, adjusters, jurors, and judges—and actors in the health care system, such as physicians and hospital administrators, face many of the same challenges. These include uncertainty about the medical evidence favoring a particular course of treatment, cognitive limitations, and the need to make money. When doctors in New York City hospitalize patients in their last six months of life for nearly double the number of days as their counterparts in Rochester, Minnesota, with no better result, there is obviously a great amount of uncertainty about proper care, and one could hardly expect lawyers and other tort system actors to achieve complete accuracy in judgments about appropriate care. And it goes without saying that when lawyers and doctors are paid on a fee-for-service basis, there is an incentive to practice their respective professions more aggressively. It is interesting to note that in Great Britain, where both doctors and plaintiff lawyers are paid on a capitated basis, there are lower health care costs and fewer lawsuits. If problems of tort system reliability are inevitable because of human factors and the nature of medical knowledge, it may be too much to expect great improvement in accuracy from a no-fault system. 1 David M. Studdert, et al., “Beyond Dead Reckoning: Measures of Medical Injury Burden, Malpractice Litigation, and Alternative Compensation Models from Utah and Colorado,” Ind. L. Rev., 33 (2000): 1643, 1675; David M. Studdert, et al., “Can the United States Afford a ‘No-Fault’ System of Compensation for Medical Injury?” Law & Contemp. Probs., 60 (1997): 1, 25-29. Marie Bismark and Ron Paterson, “No- Fault Compensation in New Zealand: Harmonizing Injury Compensation, Provider Accountability, and Patient Safety,” Health Aff (Millwood), Jan-Feb 25(1) (2006):278-83. 2 William M. Sage and Eleanor DeArman Kinney, “Medicare-Led Malpractice Reform,” Academy Health Issue Brief (January 2005). 3 Harvard School of Public Health and Common Good, “Design of a 'Health Courts' System Demonstration,” Working Draft, October 17, 2005. 4 Fair and Reliable Medical Justice Act, S 1337, 109th Cong., 1st sess. 5 Michael J. Saks, Daniel Strouse, and Nicholas Schweitzer, “A Multiattribute Utility Analysis of Legal System Responses to Medical Injuries,” DePaul L. Rev., 54 (2005): 277. 6 Harvard Medical Practice Study, “Patients, Doctors, and Lawyers: Medical Injury, Malpractice Litigation, and Patient Compensation In New York” (1990). 7 David M. Studdert, et al., “Can the United States Afford a ‘No-Fault’ System of Compensation for Medical Injury?” Law & Contemp. Probs., 60 (1997): 1, 25-29. 8 Theo Francis, “House Bill Provision May Help Insurers Recover Medical Costs,” Wall Street Journal, June 8, 2006. 9 Arkansas Department of Health & Human Services v. Ahlborn, U.S. Supreme Court, No. 04-1506 (2006). 25 10 D.M. Studdert , M.M. Mello, A. A. Gawande, T. K. Gandhi, A. Kachalia, C. Yoon, A.L. Puopolo, and T. A. Brennan. “Claims, Errors, and Compensation Payments in Medical Malpractice Litigation,” N Engl J Med., 354(19) (May 11, 2006): 2024-33. 11 2004 Annual Report, National Practitioner Data Bank. 12 Harvard Medical Practice Study, “Patients, Doctors, and Lawyers: Medical Injury, Malpractice Litigation, and Patient Compensation in New York” (1990). 13 Nancy Udell and David B. Kendall, “Health Courts: Fair and Reliable Justice for Injured Patients,” Progressive Policy Institute Policy Report (February 2005) 14 L.R. Tancredi and R.R. Bovbjerg, “Creating Outcomes-Based Systems for Quality and Malpractice Reform: Methodology of Accelerated Compensation Events (ACEs),” Milbank Q, 70(1) (1992):183-216. This article is the source for the two clinical scenarios discussed in the text accompanying this endnote. 15 Va. Birth-Related Neurological Injury Comp. Program v. Bakke, 46 Va. App. 508 (Va. Ct. App. 2005). 16 Paul H. O’Neill, Testimony before United States Senate Committee on Finance, March 8, 2006. 17 Preventing, and Managing the Impact of, Anesthesia Awareness, JCAHO Sentinel Event Alert, October 6, 2004. 18 William M. Sage and Eleanor DeArman Kinney, “Medicare-Led Malpractice Reform,” Academy Health Issue Brief (January 2005). 19 Robert Wachter and Kaveh Shojania, “Internal Bleeding : The Truth Behind America's Terrifying Epidemic of Medical Mistakes,” Rugged Land (2004). 20 L.R. Tancredi and R.R. Bovbjerg, “Creating Outcomes-Based Systems for Quality and Malpractice Reform: Methodology of Accelerated Compensation Events (ACEs),” Milbank Q, 70(1) (1992):183-216. 21 Tom Baker, The Medical Malpractice Myth, Chicago (2005). 22 David M. Eddy, “The Use of Evidence and Cost Effectiveness by the Courts: How Can It Help Improve Health Care?” J. Heath Pol. & L., 26 (2001): 396. 23 http://www.hospitalcompare.hhs.gov/Hospital/Static/About- HospQuality.asp?dest=NAV|Home|About|QualityMeasures#TabTop. 24 http://www.hospitalcompare.hhs.gov/. 25 Atul Gawande, “The Bell Curve,” The New Yorker (December 6, 2004). 26 Aaron S. Kesselheim, Timothy G. Ferris, and David M. Studdert, “Will Physician-Level Measures of Clinical Performance Be Used in Medical Malpractice Litigation?” JAMA, 295 (2006):1831-1834. 27 http://cgood.org/brochure-hcare-4.html . 28 Michelle M. Mello, “Of Swords and Shields: The Role of Clinical Practice Guidelines in Medical Malpractice Litigation,” 149 U. Pa. L. Rev., 645 (2001). 29 Wayne Guglielmo, “A Legal Crusader’s Solution to the Malpractice Mess,” Medical Economics (January 24, 2003). 30 “Legal Malpractice,” The Wall Street Journal (January 27, 2003). 31 519 P.2d 981 (Wash. 1974). 32 “Statement for the record of the American Medical Association to the Committee on Health, Education, Labor and Pensions Re: Medical Liability: New Ideas for Making the System Work Better for Patients,” June 22, 2006. 33 Nancy Udell and David B. Kendall, “Health Courts: Fair and Reliable Justice for Injured Patients,” Progressive Policy Institute Policy Report (February 2005). 34 Office of Technology Assessment, Defensive Medicine and Medical Malpractice (1994). 35 Robert J. Rubin and Daniel N. Mendelson, “Estimating the Costs of Defensive Medicine,” Lewin-VHI (January 27, 1993). 36 A. Macario, M.F. Roizen, R.A. Thisted, S. Kim, F.K. Orkin, and C. Phelps. “Reassessment of Preoperative Laboratory Testing Has Changed the Test-Ordering Patterns of Physicians,” Surg Gynecol Obstet., 175(6) (1992): 539-47. 37 Van Boven, et al., “Defensive Testing in Dutch Family Practice. Is the Grass Greener on the Other Side of the Ocean?” J Fam Pract. 44(5) (1997): 468-72. 38 Rachel Brand, “Doctors Taught to Admit Mistakes and Apologize,” Rocky Mountain News, March 10, 2004. 26 39 Edward Hickling, “Psychological Effects of Litigation: Compensation Neurosis, Symptom Exaggeration and Secondary Trauma,” Paper delivered at 11th Annual Clifford Symposium, DePaul University College of Law, April 14, 2005. 40 Harvard Medical Practice Study, “Patients, Doctors, and Lawyers: Medical Injury, Malpractice Litigation, and Patient Compensation in New York” (1990). 41 Lori Andrews, “Studying Medical Error In Situ: Implications for Malpractice Law and Policy” DePaul L. Rev., 54 (2005): 357. 42 C. Schoen, et al., “Taking the Pulse of Health Care Systems: Experiences of Patients with Health Problems in Six Countries,” Health Aff (Millwood) (Nov 28, 2005): [Epub ahead of print]. 43 Fair and Reliable Medical Justice Act, S. 1337, 109th Cong., 1st sess. 44 Linda J. Demaine and Deborah R. Hensler, “Volunteering to Arbitrate Through Predispute Arbitration Clauses: The Average Consumer's Experience,” Law & Contemp. Probs., 67 (2004): 55. 45 David A. Hyman and Charles M. Silver, "Speak Not of Error," Regulation, 28(1) (Spring 2005): 52-57. Available at SSRN: http://ssrn.com/abstract=807524; G.J. Annas, . The Patient's Right to Safety– Improving the Quality of Care Through Litigation Against Hospitals. N Engl J Med., 19 (2006): 2063-6. 46 At the recent Patient Safety and Health IT Conference (June 4–7, 2006) sponsored by the Agency for Healthcare Research and Quality, during one panel on “Innovative Applications of Reporting Systems,” panelists focused solely on the chilling effect of disciplinary processes and discounted the role of litigation in discouraging reporting. 47 D.M. Studdert , M.M. Mello, A. A. Gawande, T. K. Gandhi, A. Kachalia, C. Yoon, A.L. Puopolo, and T. A. Brennan. “Claims, Errors, and Compensation Payments in Medical Malpractice Litigation,” N Engl J Med., 354(19) (2006 May 11): 2024-33. 48 Center for the Evaluative Clinical Sciences, Dartmouth Medical School, The Care of Patients with Severe Chronic Illness: An Online Report on the Medicare Program by the Dartmouth Atlas Project. 49 E. A. McGlynn, S. M. Asch, J. Adams, J. Keesey, J. Hicks, A. DeCristofaro, and E.A. Kerr. The Quality of Health Care Delivered to Adults in the United States, New England Journal of Medicine, 348(26) (2003): 2635-2645. 27