importance of large awards in driving up the overall average severity in Florida. 34

These huge claims have a much broader effect than simply dragging up the average cost per claim.35 As noted earlier, they induce a great deal of variation in insurance costs, making it difficult for actuaries to estimate total costs in an insurance market that is as segmented and thin as is malpractice. 36

The data regarding rising numbers of claims and severity of claims, combined with the effects of huge jury verdicts or settlements on risk pooling, provide strong evidence that the longer term increase in insurance premiums charged individual physicians and hospitals are a result of real changes in tort costs. The question remains, why have there been these increases in number of claims? Are physicians providing poorer quality care or are patients suing more frequently for other reasons? Research into iatrogenic injury and malpractice litigation can provide some answers to these questions.

Health Care System

Some have conjectured that the increase in claims and severity of claims is simply a matter of potential plaintiffs mining a mother lode of potential claims. Their assumption is that many potentially "litigable" and "compensable" events occur regularly in hospitals, and, in recent years patients are bringing a larger percentage of these cases into the tort process. This hypothesis is supported by the only published empirical study of iatrogenic injury and litigation, conducted in California in 1974.37 The results were

34 In 1975 in Florida, the largest paid claim was $500,000. In 1979, the largest paid claim was $600,000 but in 1981, there was a paid claim of $1.6 million. In 1984, the largest paid claim was over $5 million. In 1981, there were a total of 16 claims for more that $500,000 in Florida, while in 1983 there were 57 and in 1984 there were 105. Claims worth more than $1 million represented 4.9 percent of the total amount of paid claims in 1981, but in 1986 these claims accounted for 29.1 percent of the total compensation paid. See Nye et al., "The Causes of the Medical Malpractice Crisis," 1552. These are not the highest claims in the country. Nor has the huge verdict phenomenon abated. In 1988, the highest personal injury jury verdict in the country was rendered against a hospital in Houston for $52 million. See American Bar Association Journal, July 1989.

35 A few large awards lead to great discrepancies between the median amount paid per claim, the middle number obtained when the claims are arranged in ordinal fashion, and the mean amount paid per claim, the amount calculated by dividing the total indemnity by the number of claims. For instance, in 1984 the General Accounting Office noted that the median claim payments were $18,000 while the mean claim payments were $80,000. See General Accounting Office, Medical Malpractice: Characteristics of Claims Closed in 1984, 2.

38 See Priest, "The Current Insurance Crisis," 1545.

37 In 1974, the California Medical Association and the California Hospital Association commissioned a study by Don Harper Mills of the incidence of potentially compensable events in hospitals. California Medical Association, Medical Insurance Feasibility Study (Sutter Publishing, 1976). This study focused on 23 hospitals which had elected to participate. The researchers analyzed a sample of 20,864 records of inpatients who had been discharged in 1974. They defined a potentially compensable event as a disability that had been caused by health care management. The research team also evaluated

disturbing. The overall rate of potentially compensable events was 4.65 percent; 38 the rate of PCE's associated with negligence was .79%.39 Using this data, and combining it with information from the National Association of Insurance Commissioners for 1976 through 1978, Patricia Danzon estimated that in the mid 1970s, less than one out of 10 potentially compensable events associated with negligence gave rise to litigation. 40

The California study suggests that recent increases in rates of medical malpractice litigation are caused by the fact that more and more litigants are now actually bringing what always were potentially valid claims.41 Again, though, one must put these statistics about medical negligence in context. Modern health care is an inherently risky enterprise, employing a variety of intrusive procedures, equipment and drugs to combat diseases that would have been left to run their (usually fatal) course a generation ago. But however sophisticated its science and technology, medicine still must rely on the care and attention of doctors, nurses and technicians for the use and safety of these procedures. Unfortunately, like all human beings, medical personnel are prone to occasional oversights and mistakes. But when a usually conscientious surgeon makes a momentary inadvertent slip-up-which is negligence in the eyes of the law-this exposes the patient to severe and irretrievable harm for which a tort suit can be brought.42 Thus, it is revealing that

whether or not the potentially compensable event was caused by negligence, and graded each potentially compensable event according to severity.

38 Among these PCEs, 35.7 percent led to minor disability of 30 days or less and did not require surgery. Another 25.7 percent of the PCEs had minor disability that did not exceed 30 days but did require surgery. Major disability, lasting for more than 30 days but less than two years, occurred in 18.6 percent of patients. Another 6.5 percent of PCEs resulted in minor but permanent partial disabilities. 2.2 percent of the PCEs gave rise to major permanent partial disability and another 1.6 percent gave rise to grave permanent disability. Death occurred in 9.7 percent of PCES.

39 Forty-two percent of the PCEs that proved fatal were associated with negligence, whereas only 11.9 percent of the PCEs leading to temporary disabilities were caused by negligence. The overwhelming majority of PCEs occurred in the operating room (71.8 percent).

40 Her estimates suggest that if all cases that could potentially give rise to successful tort litigation did so, there would be 10 times the number of claims and 20 times the number of paid claims. See Danzon, Medical Malpractice (Cambridge, Mass.: Harvard University Press, 1985). Danzon has since estimated that at present greater than one in 10, perhaps as many as one in six or seven of potentially compensable events associated with negligence now lead to a claim, largely as a result of the increase in the number of claims over the period from 1975 to 1986. See Danzon, "The Frequency and Severity of Medical Malpractice Claims: New Evidence," Law and Contemporary Problems, 49 (1986): 57-84. 41 Of course, the available empirical information concerning the "mining" hypothesis is limited. First of all, the California study was a non-random sample, suggesting there may be considerable bias in its results. Moreover, little attention was paid during the conduct of the research to methodological issues such as reliability and validity of judgments concerning potentially compensable events or negligence. In addition, Danzon's work was based on a somewhat inadequate database, the National Association of Insurance Commissioners closed-claims data from 1975 through 1978. The Harvard Medical Practice Study Group, of which we are members, has just completed a study in New York, Patients, Doctors, and Laywers (1990), which addressed these concerns: we found that the present gap between negligent injuries and actual tort claims is even starker than the California estimates.

42 See generally Grady, "Why Are People Negligent? Technology, Nondurable Precautions and the Medical Malpractice Explosion," Northwestern University Law Review 82 (1988): 293.

since the mid 1970's there has been the same percentage increase in malpractice claims and costs in Canada and Great Britain (though from a much lower base level than the United States), despite their very different organization and financing of both legal and health care. 43

But whatever the explanation, the fact is there is no evidence we have been experiencing too many malpractice claims; if anything there appear to be too few. It would seem unfair, then, to encumber even further access to the tort system by those injured individuals with valid claims. On the other hand, large awards tend to have an unsettling effect on the insurance system. The erratic nature of awards, especially their pain and suffering component, produces unduly volatile premium changes during "crisis" periods and dislocation in the lives of individual practitioners, whether sued successfully or not.

RESPONSE-PUBLIC POLICY

In response to the pressure of rising malpractice litigation and premiums, all the state legislatures have sought to reform at least some of their law concerning malpractice liability. During the 1970's and 1980s, this effort took the form of "conventional" tort reform, under which the legislatures tinkered with different tort doctrines and procedures for malpractice claims. In the last few years, though, one finds growing interest in moving beyond tort law altogether, whether into administrative fault or no-fault compensation systems. We will first review conventional tort reform and then sketch some of the suggested alternatives to traditional tort litigation.

Conventional Tort Reform

A variety of types of conventional tort reform have been instituted by state governments in the last two decades, most fashioned in response to the first malpractice crisis of the mid 1970s. These statutory changes, designed to reduce the frequency and severity of malpractice claims and premiums, are aimed at one of three targets: initial access of patients to the courts, chances of success in establishing liability, and the potential amount of damages awarded. Since most of these reforms were instituted in at least some states more than a decade ago, sufficient time has elapsed to

4 See Dewees, Coyte and Trebilcock, Canadian Medical Malpractice Liability: An Empirical Analysis of Recent Trends (1989); Ham, Dingwall, Fenn and Harris, Medical Negligence: Compensation and Accountability, (1988): 11– 12 (on the United Kingdom).

allow empirical evaluation of the impact of such

measures.

We start with changes in liability rules, since this is perhaps the most distinctive feature of malpractice law. Unlike product liability law, courts have stuck closely to the fault principle, as reflected in the rule that only deviation from customary medical practice will suffice for liability. Few courts have strayed from this central assumption; ** nor have any legislatures relaxed it. However, a number of statutory enactments in the 1970s did modify the detailed rules embodying the fault principle, with a view to rolling back some of the common-law doctrines developed by judges in previous decades to make it easier for patients to establish physician fault. For instance, the judges had largely overturned the old locality doctrine under which practitioners were held only to the standard of practice in their same locale. Since 1970, approximately one dozen state legislatures have restated the standard of care by reference to local practice. 45

Another such effort sought to limit the operation of the doctrine of res ipsa loquitur. This judicial doctrine made it somewhat easier for patients to get a malpractice case to a jury in situations where an apparent mishap had led to a poor outcome. Yet another legislative attempt to overcome judge-made law concerned informed consent. In the 1960s and 1970s, some courts had defined the standard of informed consent in terms of what a reasonable patient needed to know, rather than what a reasonable physician felt it necessary to disclose. A number of state legislatures in turn explicitly enacted the reasonable practitioner standard.46

A more thoroughgoing method for changing liability standards would be to specify that practice guidelines developed by physicians should act as definitions of the standard of care for courts. Using outcome studies 47 and consensus panel determinations of practice standards, 48 health services researchers have begun to develop clinical parameters to determine appropriate care.49 For instance, the American College of Physicians and Blue Cross/Blue Shield have developed a series of guidelines for common diagnostic

** The most notable being Helling v. Carey, 519 P.2d 981 (Wash. 1974). 45 See Robinson, "The Medical Malpractice Crisis of the 1970s: A Retrospective," Law and Contemporary Problems 49 (Spring, 1986): 6.

46 See Meisel and Kabnick, "Informed Consent to Medical Treatment: An Analysis of Recent Legislation," University of Pittsburgh Law Review 41 (1980): 407.

47 See Wennberg, Mulley, Hanley et al., “An Assessment of Prostatectomy for a Benign Urinary Tract Obstruction: Geographic Variations in the Evaluation of Medical Outcomes,"JAMA 259 (1988): 3027.

48 See Kahn, Kosecoff, Chassin, et al., "The Use and Misuse of Upper Gastrointestinal Endoscopy," Annals of Internal Medicine 109 (1988): 664–70. 49 See Brook, "Practice Guidelines and Practicing Medicine: Are They Compatible?" JAMA 262 (1989): 3027.

tests. 50 While some may question the validity of such parameters 51 it seems likely that development of medical practice algorithms will accelerate. Indeed, recent legislation creates new opportunities for research on and development of, these methods, 52 and this Commission has recommended as part of a national system of quality assurance that national practice guidelines and standards of care be developed in a newly created agency for Health Care Policy and Research.

Similar practice guidelines could also be used to determine the standard of care for malpractice proceedings. However, their development is a very complicated and lengthy process. Therefore, while we now see the first uses of such guidelines in an effort to improve care and decrease litigation, 53 it is doubtful they will play a major role in the near future.

Another set of tort reforms placed a variety of restrictions on what was perceived as too easy access to common law courts. 54 One way to do this was to shrink the statutory limitation period. Using a judge fashioned discovery doctrine, many state courts interpreted their statute of limitations as not being tolled until a victim had, or should have, discovered that his injury was a result of doctor negligence. This innovation tended to give plaintiffs a much broader time span in which to file their suit. In response, a number of states repealed the discovery rule or required that claims be brought within a so-called statute of repose that set an outer bound to the discovery principle. 55 The rationale is that doctors and their insurers should enjoy some protection against litigation launched long after the original treatment, when there is considerable risk that the events will be appraised in light of quite different legal and medical norms than could fairly have been anticipated at the time.

Another such access rule was the limitation of attorney's fees. Plaintiff attorneys take malpractice cases on the basis of a contingent fee, meaning that if and only if the plaintiff wins, the lawyer takes his compensation as a percentage of the award. In response to physicians' concerns about the incentives this arrangement created for malpractice suits, many jurisdictions restricted the size of attorneys' fees, typically by im

30 See Sox, "Guidelines for Medical Practice: Necessary But Not Sufficient," Journal of General Internal Medicine 6 (1989): 551.

51 See Alper, "ACP Guidelines for Common Diagnostic Tests in the Practicing Internist," Journal of General Internal Medicine 6 (1989): 548.

52 See Omnibus Budget Reconciliation Act of 1989, P.L. 101-239, sec. 6103.

53 Berwick, "Continuous Improvement as an Ideal in Health Care." New England Journal of Medicine 320 (1989): 53.

84 See Bovbjerg, "Legislation on Medical Malpractice: Further Developments and a Preliminary Report Card," University of California at Davis Law Review 22 (1989): 499.

55 See McGovern, "The Variety, Policy and Constitutionality of Product Liability Statutes of Repose," American University Law Review 30 (1981): 579.

posing a flat percentage or by constructing a sliding scale based on the size of the award. For a time, Florida even experimented with the Anglo-Canadian rule of awarding fees to the prevailing party, but this law was replaced by a sliding scale based on timing of settlement and size of award. 56

Yet another way to limit access was to institute screening panels. Screening panels are composed of doctors and lawyers who review a case and evaluate its merits before a claim is filed. These screening panels are intended to decrease access to the legal system by influencing parties' evaluations of their probability of success in courts. While many states adopted screening panels in the mid-1970s, there has been less enthusiasm for them in the 1980s, largely because the panels have had little impact except to prolong the length of time and costs before resolution of a suit. 57

Finally, state legislatures have experimented with limits on damages, again in an effort to moderate the number and severity of claims. One such measure is periodic rather than lump sum payment of large damage awards: this allows the court to adjust the award to subsequent changes in health status of the victim. Another way to restrain damage awards is to reverse the collateral source rule, under which the courts traditionally held that damages payable by culpable defendants should not be reduced even if the patient's loss has been compensated by other sources, such as health or disability insurance. 58 Numerous states have responded to the arguments that this is a form of "double dipping," either by allowing defendants to introduce evidence concerning sources of compensation such as health and disability insurance, or even mandating that the courts deduct any such collateral insurance payments from tort awards. Of course, offsetting such collateral sources can reduce the deterrent threat of litigation, given that physicians would have a reduced responsibility for the social costs of medical injuries and personal injury lawyers whose fees are a percentage of the total damage award would have less inclination to pursue claims on behalf of a well insured individual, even when the care the individual received may have been grossly substandard.

56 See Hawkes, "The Second Reformation: Florida's Medical Malpractice Law," Florida State Law Review 13 (1985): 747.

57 See Danzon and Lillard, "Settlement out of Court: The Disposition of Medical Malpractice Claims," Journal of Legal Studies 12 (1983): 373-4; Shmanske and Stevens, "The Performance of Medical Malpractice Review Panels," Journal of Health Politics, Policy and Law 11 (1986): 525.

58 See generally Moceri and Messini, "The Collateral Source Rule in Personal Injury Litigation," Gonzaga Law Review 7 (1972): 310. Although it has attracted little attention from state legislatures, the concept of scheduling damages gained some scholarly support as a method for rationalizing the damage determination process: see Danzon, "Tort Reform and the Role of Government in Insurance Markets," Journal of Legal Studies 13 (1984): 527-30.

The most straightforward form of damages restriction is the damage cap. As mentioned above, a relatively small number of large awards comprise a substantial share of the total amount of money paid in tort claims. Since the mid 1970s, states have experimented with caps on damages. Some states place numerical limitations on pain and suffering. In the 1970s, most of these numerical restrictions applied to both economic and non-economic damages, 59 but in the 1980s almost all have focused simply on pain and suffering. To this point only a handful of states have indexed or adjusted their nominal dollar caps to subsequent changes in wage and price levels. 60

The efficacy of various tort reforms have been assessed by economists and health policy analysts using econometric methods. 61 Danzon, for example, traced the effect of mid-1970s legislation over an extended period of time and updated her research with claims information as late as 1984. She demonstrated that reductions in the statute of limitations result in a significant decrease in claims frequency. Moreover, the reversal of a collateral source rule reduced claims severity by somewhere between 11 and 19 percent, and claims frequency by around 14 percent. Finally, caps on damage awards reduced claims severity by 23 percent, but did not seem to have any effect on the propensity to initiate claims.

In a more recent study, Sloan and co-authors combined data on individual closed claims from the National Association of Insurance Commissioners' Study of Closed Claims From 1975-1978 with closed claims information from 1984 collected by the General Accounting Office of the United States Congress, and analyzed the effect of tort reforms on claims' severity.62 Just as did Danzon, Sloan et al. found few tort

59 See e.g., Indiana Code Annotated, Section 16-9.5-2-2.1 (Burn's Supplement 1986).

60 We have reviewed the major forms of tort reform in the 1970s and 1980s, but our list is not exhaustive. For instance, Bovbjerg's review of tort reforms touches on arbitration as an effort to limit the number of suits getting into court, restrictions of ad damnum clauses as a method for decreasing the size of recovery; expert witness standards as means of increasing the plaintiff's difficulty or cost of winning; and precalendar conference requirements and preferred scheduling for malpractice cases as means of decreasing the costs of the judicial process. See generally Bovbjerg, "Legislation on Medical Malpractice," 499.

of The most recent of these empirical analyses are Danzon, "The Frequency and Severity of Medical Malpractice Claims: New Evidence," Law and Contemporary Problems 49 (1986): 37-56; and Sloan, Mergenhagen and Bovbjerg, "Effects of Tort Reforms on the Value of Closed Medical Malpractice Claims: A Microanalysis," forthcoming, Journal of Health Policy, Politics, and Law. Unless otherwise stated we rely on these authors' analyses. This is not to overlook some of the important research which was done before 1985. See, in particular Reder, "An Economic Analysis of Medical Malpractice," Journal of Legal Studies 5 (1976): 267; Feldman, "The Determinants of Medical Malpractice Incidence: Theory of Contingency Fees and Empirical Evidence," Atlantic Economic Journal 7 (1979): 59; Danzon, "The Frequency and Severity of Medical Malpractice Claims," Journal of Law and Economics 27 (1984): 115; Adams and Zuckerman, "Variations in the Growth and Incidence of Medical Malpractice Claims," Journal of Health Politics, Policy and Law 9 (1984): 475; Sloan, "State Responses to the Malpractice Insurance Crisis of the 1970s: An Empirical Analysis," Journal of Health Politics, Policy and Law 9 (1985): 629.

62 See Sloan, Mergenhagen and Bovbjerg, "Effects of Tort Reform."

reforms that reduced the amount paid per claim. However, damage caps on the total awards, and even caps on just pain and suffering, substantially lowered the amounts paid. As well, changes in the collateral source rule produced sizable reductions in average payments. Laws that encouraged arbitration decreased the amount paid per claim, while changes in the statute of limitations reduced the likelihood that a claim would be closed with payment but did not affect the overall amount per paid claim. All the other reforms had little statistically significant impact.

Several conclusions can be drawn from these studies. First, it appears that changes in liability rules have little effect: malpractice litigation remains fault-based at its core and legislative tinkering changes this very little. Statutory obstacles to access also seem to have little effect, with the exception of changes in statutes of limitations. However, alterations in damage rules— i.e., caps or collateral source offsets-do have a significant effect on both the severity of claims or even the number of claims brought.

As one might have gathered from this review of the content of malpractice reform, most of these measures have been proposed by medical associations and championed by their legislative supporters. However, when a state legislature enacts a tort reform package, the legal dialogue does not end there. Patients and their attorneys can return to the state courts for decisions about how to treat the new statutory law-in particular, under the state constitution. After a decade of such constitutional challenges, it appears that patients have fared quite well in the judicial forum, with many of the state courts having struck down some the key reforms, especially in the damages

63 See in general, Vinson, “Constitutional Stumbling Blocks to Legislative Tort Reform," Florida State Law Review 15 (1987): 31; Turkington, "Constitutional Limitations on Tort Reform," Villanova Law Review 32 (1987): 1299; Smith, "Battling a Receding Tort Frontier: Constitutional Attacks on Medical Malpractice Laws," Oklahoma Law Review 38 (1985): 195; Learner, “Restrictive Medical Malpractice Compensation Schemes: A Constitutional quid pro quo Analysis to Safeguard Individual Liberties," Harvard Journal on Legislation 18 (1981): 143.

64 See e.g., Fein v. Permanente, 695 P. 2d 665, 38 Cal. 3d 137, 211 Cal. Rptr. 368 (1985), appeal dismissed, 106 S.Ct. 214 (1986).

65 See Brennan, "The Bill of Rights in the States: The Revival of State Constitutions as Guardians of Individual Rights," New York University Law Review 61 (1986): 535.

especially offended when they perceive physician. groups having secured statutory measures that substantially cut back on the tort rights of injured patients without the latter receiving some form of quid pro quo along the lines of no-fault workers' compensation. That judicial sentiment has been displayed in a number of constitutional guises; in the 1970s usually under equal protection or due process, but in the 1980s more often under the right of access to the court or to a jury trial.66

It is no accident that this judicial sentiment tends to surface in challenges to the constitutionality of restrictions on tort damages. While these are the measures that do offer the greatest relief to doctors from the burden of the malpractice regime, the cases where the caps are applied often involve the most seriously injured patients, and in situations where judges realize that some of the supposedly "excess" award is really going to pay the percentage contingent fee of the attorney whose services were necessary to secure any tort recovery for the injured party. In addition, as we have seen, the conclusion of more systematic research is that under the common law system, there are far fewer malpractice claims being made (let alone paid) in the legal system than there are malpractice events and injuries occurring in the health system.

One other alternative for overcoming the deficiencies of the present system might be to move from a tort to a contract regime. Some have argued that tort law smacks too much of "command and control" regulation, and thus retards the efficiencies and innovation that could arise out of a "free market" approach.67 They advocate much broader scope for voluntarily-negotiated contracts between patients and providers which would stipulate whether and how much compensation should be paid for injuries occurring in the course of therapy. In theory, these contractual alternatives would be more efficient in reflecting the mutual interests of the immediate parties while reducing the administrative costs associated with litigation. For example, providers and patients could stipulate that in the event of injury due to substandard care, certain scheduled damages would be available. The definition of substandard care in this case could be left up to binding arbitration which was faster and cheaper than jury trials.

66 Smith v. Department of Insurance, 507 So.2d 1080 (Fla. 1987); Kansas Malpractice Victims v. Bell, 757 P.2d 251 (Kan. 1988); Lucas v. United States, 757 S. W.2d 687 (Tex. 1988); Sofie v. Fibreboard Corp., 771 P. 2d 711 (Wash. 1989). In Boyd v. Bulala, 647 F.Supp. 781, 786 (W.D. Va. 1986) a federal district court held that legislation violated both the federal and Virginia's guarantee of a jury trial, but that state law interpretation was reversed in Etheridge v. Medical Center Hospital, 376 S.E. 2d 525 (Va. 1989), and in Boyd v. Bulala, 872 F.2d 119 (4th Cir. 1989), regarding the federal constitution. 67 See generally, Havighurst, "Private Reform of Tort-Law Dogma: Market Opportunities and Legal Obstacles," Law and Contemporary Problems 49 (1986): 143.

Of course, there are drawbacks associated with contractual arrangements of this sort. The contracts themselves could be the subject of a good deal of litigation, decreasing the potential gains in efficiency. More important, both state courts and state legislatures have exhibited great concern about the likely fairness of the terms of a contract struck between a provider and a patient, given the inequalities in bargaining power and knowledge that exist between these parties.68 Thus while contractual alternatives to malpractice litigation do represent another option that should be explored, they are clearly not a panacea for the problem of medical injury.

Beyond Tort Law

The reforms we have suggested for tort law might go some distance in terms of rationalizing the system. However, we are still left with the overwhelming empirical facts that few negligent adverse events now lead to claims, and conversely that many claims arise out of situations in which there is no evidence of negligence or even of injury. There are both too many unwarranted claims, and too many apppropriate claims that are never brought. Any comprehensive review of the medical malpractice crisis must at least consider what options lie beyond tort litigation.

The tort reforms discussed in the previous section modify the rights and obligations of patients and doctors in malpractice litigation. They do not, however, change the basic ground rules-that the patient must prove to a jury that he or she was injured due to the negligence of a physician or health care provider. There are, however, numerous alternatives to traditional tort litigation, some modest, some major. In this section we review several such methods through which one might provide compensation to patients injured as a result of their medical care, and also create incentives for better care by providers.

Alternative Dispute Resolution-Given the high administrative as well as emotional costs of litigation, many have sought alternative dispute resolution formats for medical malpractice claims. The term "alternative dispute resolution" is an umbrella concept. covering a variety of specific procedures designed to provide fair outcomes at lower costs. Included in ADR are such approaches as arbitration, or mediation.69 In the area of medical malpractice, the most prominent vehicle has been the pretreatment agreement through which patients accept binding arbitra

68 Havighurst, "Private Reform of Tort-Law Dogma," 163-4.

* See Goldberg, Green, and Sander, Dispute Resolution, (Boston: Little Brown and Company, 1985).