Law & Contemporary Problems: Edward A. Dauer and Leonard J. Marcus, Adapting Mediation To Link Resolution Of Medical Malpractice Disputes With Health Care Quality Improvement, 60 Law & Contemp. Probs. 185 (Winter 1997)

Cited: 60 Law & Contemp. Probs. 185 (Winter 1997)
[*pg 185]

ADAPTING MEDIATION TO LINK RESOLUTION OF MEDICAL MALPRACTICE DISPUTES WITH HEALTH CARE QUALITY IMPROVEMENT

EDWARD A. DAUER^* AND LEONARD J. MARCUS^**

I. INTRODUCTION: THE HYPOTHESIS

II. MEDICAL INJURY
   A. Medical Errors
   B. Legal Liability Also Abounds
   C. Malpractice Liability Has Not Been Correlated With Malpractice Prevention
   D. The Liability Process and the Requisites of Prevention

III. THE HYPOTHESIS: CONFORMING THE RESOLUTION PROCESS WITH ERROR PREVENTION
   A. Injured Patients' Motivations Are Consistent With Error Prevention
   B. An Encouraging Pilot Project: The Massachusetts Voluntary Mediation Program

IV. SPECIFIC APPLICATIONS: NEXT STEPS
   A. What Would the Model Look Like?
   B. How Would the Results Be Implemented?
   C. How Would the Proposal Be Validated?

V. THE NEAR FUTURE: EXPANDING THE MASSACHUSETTS EXPERIMENT

VI. CONCLUSION AND FOREWORD

FOOTNOTES

I. INTRODUCTION: THE HYPOTHESIS

According to conventional theory, the tort liability system serves two objectives: compensating injured persons, and causing other persons to internalize the costs of their errors and thus to guard against them in the future. The system is driven by the energies of the claimant, whose self-interest in obtaining compensation is compatible with the larger social goal of reducing future risks. This coincidence of private and public objectives occurs both within the formal liability system and in the more frequent instances where the resolution of a claim is privately negotiated rather than publicly adjudicated -- so long as the bargaining takes place, to steal a phrase, in the shadow of the law.¹

There is good evidence that the tort liability system does not work that way in practice. And, to a considerable extent, those who are interested in satisfying either of the two objectives have gone their separate ways. Alternative dispute resolution procedures such as mediation have been developed and promoted as efficient techniques for resolving claims (the compensation objective), while quality improvement initiatives in health care have largely ignored the claims process as an avenue for effective error reduction. Ironically, devices such as the National Practitioner Data Bank that have attempted to facilitate the link between compensation and risk reduction are suspected by some of having exacerbated the difficulties in achieving more effective forms of claims resolution.

Recent empirical studies have demonstrated in a very sturdy way that the predominant motivation of iatrogenically-injured claimants is not the need or desire for economic compensation.² Rather, claimants are often simply at- [*pg 186] tempting to ensure that the error is not repeated. A pilot mediation program now being conducted under the sponsorship of the Massachusetts Board of Registration in Medicine confirms these findings, strongly suggesting that quality improvement can be an outcome of private dispute resolution through mediation. Almost all parties in the first cohort of cases in the Massachusetts program to which mediation was offered agreed to participate. Of the first ten cases concluded, and on which we report here, nine were resolved, with only four involving any payment of money. More commonly, outcomes tend to focus on future patient safety. Although the Massachusetts program's design does not allow these data to satisfy the requisites of statistical analysis, the results have been powerful enough to encourage a significant expansion of the initiative, and to lead us to pursue additional theoretical and empirical studies on the question.

"Mediation" means different things to different observers. To some working within the civil liability system, mediation connotes an alternative venue to more traditional settlement negotiations within which lawyers acting on behalf of clients pursue the settlement of pending lawsuits. Our conception of mediation presumes an elaboration of the objectives of the process, beyond the settlement of private disputes, bringing the private interests of the injured person into closer consonance with the public's interest in preventing injuries in the future. It also includes the potential to enhance the efficiencies of claim resolution, and offers direct and collaborative roles to patients and to doctors. Such a vision of mediation is not unusual in the world of dispute resolution, though we are not aware of its application until now to the problem of medical error.

As the predicate for further research, we hypothesize that mediation in either a fault-based or a no-fault environment can make claims resolution more efficient and simultaneously promote quality improvement in health care more effectively than does the litigation/settlement process. To be as clear about the limits of our hypothesis as we are about its importance, we do not suggest that mediating medical injury disputes is a cure-all for what many -- on every side of the debate -- see as a difficult and unsatisfactory system. Changes in the processing of disputes are seldom effective substitutes for substantive or institutional reforms. Neither do we suggest that mediation is the route to effective error prevention. The hypothesis we mean to explore is at once narrower and sharper -- in almost any foreseeable legal regime³ in which disputes about medical injury can arise, there are forms of mediation that are compatible with the goals of quality improvement and that can avoid the impediments to quality improvement that are structurally embedded in conventional tort litigation.

[*pg 187]

II. MEDICAL INJURY

A. Medical Errors

The investigators in the Harvard Medical Practice Study⁴ calculated through retrospective chart reviews that approximately four percent of all hospitalized patients suffer an iatrogenic injury,⁵ at least half of which are the result of preventable errors⁶ and one-quarter of which result from actionable negligence.⁷ The most frequent type of errors, medication errors, occur in two percent to fourteen percent of all hospitalizations, though few cause detectable [*pg 188] injury.⁸ Still, nearly fourteen percent of the 1.3 million iatrogenic injuries⁹ each year are fatal.¹⁰ As Lucian Leape has calculated, iatrogenic injury results in 180,000 deaths per year, or "the equivalent of three jumbo-jet crashes every [two] days."¹¹

There is no way to know the optimal number of medical errors. That calculation would require some consensus about the marginal value of incremental lives saved or lost and the acceptable social costs of preventing the losses. There does not seem to be any such consensus. Medical insurance programs, for example, will typically invest not more than $50,000 per year of high quality life for major interventions¹² such as coronary bypass grafts. Yet FDA regulations demand antigen testing¹³ of donated blood, to eliminate "window period" human immunodeficiency virus, at a cost of more than $5 million per marginal infection prevented.¹⁴

Such nice disquisitions aside, no serious observer has been found who believes the present rate of preventable errors is too low. Even those who are aware of the problems of measurement assume, if they do not explicitly say, that the rate is too high.¹⁵ In an absolute sense that seems reasonable; one [*pg 189] study of intensive care units found that on average 1.7 errors occur among the typical 178 activities per patient per day, an error rate of about one percent.¹⁶ Again Leape (following W.E. Deming) makes the comparison: Even a one percent error rate in aviation would result in twenty unsafe landings at O'Hare each day.¹⁷ At one percent, 320,000 checks would be deducted from the wrong account and 160,000 pieces of mail would be lost every hour.¹⁸ To put this in perspective, 1.3 million hospital injuries resulting in 180,000 deaths a year represents roughly four times the annual fatalities on all of America's highways.¹⁹ Medical injury abounds.²⁰

B. Legal Liability Also Abounds

The amount of liability incurred by the medical establishment is both certainly very large and probably much too small. Thirty-nine percent of all of America's physicians have at one time or another incurred at least one malpractice claim.²¹ The National Practitioner Data Bank ("NPDB"), which collects reports of malpractice payments made through either settlement or verdict, received 89,000 malpractice payment reports about 64,000 physicians in its first five years.²² In 1988, the medical liability insurance companies in the State of New York alone paid out about $850 million to pay and settle claims.²³

Despite the significant amount of liability incurred, it is probably too little. While some plaintiffs with non-meritorious claims may coerce a settlement from an insurer unwilling to pay even more to defend the claim than to settle it, those plaintiffs represent a small number when compared to the number of negligent injuries for which no claim, and a fortiori no payment, is ever made. Contrary to the cries of the tort reformers, there is no excess of malpractice liability.²⁴ For example, while jury awards in medical liability cases are not very much greater than jury awards for comparable injuries in tort actions gener- [*pg 190] ally,²⁵ in medical injury in particular, the proportion of all severely injured persons who bring any type of claim is in the neighborhood of only ten to twelve percent.²⁶

It is possible that the inaccurate imposition of medical liability is a sufficient explanation for the supra-optimal amount of medical error. The solution would then be to make liability investments more economically accurate. However, for a number of reasons it would still be appropriate to explore the possibility that the persistence of error can be attributed at least in part to process flaws. Although in a perfect world an ideal solution would be to enhance liability to the optimal degree, in the real world the goal of less error might be achieved more efficiently by adjusting some other part of the system, if the cost of doing that is less than the cost of achieving the first best solution.²⁷

Moreover, dollars may not measure what actually happens. Assume, for example, that physicians are highly averse to the processes of medical malpractice liability and claims, such that the process of adjudicating even a $10,000 loss results in intense distress and discomfort. To avoid that personal, nonmonetary loss, physicians might be willing to invest more than $100 (of someone else's money, to be sure) to avoid a one percent chance of incurring another $10,000 claim (also paid for with someone else's money).²⁸ There is very good evidence that physicians are enormously distressed by the prospect of [*pg 191] malpractice liability,²⁹ a distaste made all the worse by the fact that even a de minimis settlement results in a mandatory report to the National Practitioner Data Bank. The report is then distributed at least biennially to every facility at which that practitioner has practice privileges.³⁰

As the tort theorists would expect, one dollar of liability can have a large effect if its imposition is known to and becomes a concern of a number of physicians other than the parties to any particular suit.³¹ Unless physicians' responses to their perceptions of liability are themselves correlated positively with error prevention, however, increasing the frequency or the severity of liability will likely produce only dysfunctional or irrelevant responses.

C. Malpractice Liability Has Not Been Correlated With Malpractice Prevention

Liability that "concerns" health care providers and liability that "deters" them from injurious error are two very different things. Despite a significant investment in effective liability, medical error persists. Whether the ratio of liability to prevention is high or low or indeterminable,³² the linkage between the two seems far from perfect.

A physician who is motivated by the risk of a malpractice claim may adopt any one or more of several strategies.³³ Because some medical specialties are well-known to have higher malpractice claim rates than others,³⁴ those specialties might be avoided³⁵ in favor of "safer" practices. The practitioner may also be able to choose among techniques that may reduce the risk of incurring a claim, versus techniques that might reduce the possibility of error. Hewing closely to "practice parameters,"³⁶ for example, may reduce the risk of liability [*pg 192] while not necessarily changing the risk of iatrogenic injury. Other forms of "defensive medicine" likewise illustrate how liability can cause claim-reducing behavior that may or may not also be error-reducing.³⁷

The high frequency of claim avoidance relative to error avoidance is most likely related to physicians' perceptions about the accuracy of the tort system itself. Suffice it to say, it does not command their universal respect. Thus the Harvard Medical Practice Study found that "[m]ost physicians perceive their suits as arising from circumstances beyond their control ... . Physicians perceive that they will be sued for a bad outcome approximately [forty-five percent] of the time, irrespective of negligence. They perceive that the chance of being sued increases to only [sixty percent] if they act negligently."³⁸ "It would be better if the [tort] system were viewed by physicians as rational, for their re- [*pg 193] sponses and the deterrent effect then could be more rational than they are today."³⁹

Addressing the question directly, the Harvard Medical Practice Study concluded that the link between medical liability and deterrence of medical negligence was small and negative, though the correlation was weak.⁴⁰ One of that study's investigators, Paul Weiler, later elaborated the analysis in a separate work.⁴¹ Weiler is generally optimistic about the relationship between liability and deterrence (and generally convinced of the relationship between liability and defensive medicine), and argued that from a methodological point of view the correlation is devilishly difficult to establish:

One thing we certainly discovered was that however laborious a job it was to identify and tabulate the injuries and losses caused by medical treatment, the difficulty of that task paled by comparison with the statistical complexity of trying to isolate a causal effect of the prospect of tort suits on the quality of medical care.⁴²

To be fair, it is in this instance uncertain whether the absence of evidence is evidence of the absence. For example, there is the problem of endogeneity: A hospital with more negligent practice will have more liability, which, if the deterrence hypothesis is true, should result in less negligent practice. Over time the effects will cancel in the data. After applying highly sophisticated techniques that admittedly "posed considerable statistical risk of inflating the estimate of the tort impact," Weiler concluded that "only a fairly modest, though statistically significant, preventive effect of malpractice litigation is discernible in our data."⁴³ Among the empiricists, Weiler may be nearly alone. Most would agree with Sloan that evidence is almost completely lacking for the proposition that the tort liability system deters medical negligence.⁴⁴

D. The Liability Process and the Requisites of Prevention

The differences are large between the tort liability system⁴⁵ and medical [*pg 194] quality improvement processes.⁴⁶ That is not to say that even when considered from the quality improvement perspective, the tort system is flawed merely by reason of those differences. The two systems are meant to do different things. Deterrence is not the only function of civil liability. Compensation of injured people is another, along with the articulation of social mores, the nonviolent resolution of disputes, even punishment where that seems appropriate. Deterrence, however, is undoubtedly one of the explicit objectives of the tort system, and one without which many of the other parts of the tort apparatus would not make quite as much sense.⁴⁷ It would be a beneficial and practical step forward if it could be ascertained whether the deterrence function of the tort system is internally flawed, and if it is flawed, how it could be improved without compromising the tort system's other roles.

There are differences between torts and quality improvement that are structural but benign. There are other differences that represent opportunities unnecessarily lost. And there are differences that are both unnecessary and positively inimical to both medical quality improvement and the deterrence objectives of the tort system.

The tort system's narratives are historical. Its focus is on identifying the individuals who are to "blame" for having caused a plaintiff's loss. Quality improvement's narratives are the future. Its focus is on identifying the things that can be changed to alter the future. Torts asks the questions "who," "what," and sometimes "how." Quality improvement initiatives ask the question "why?" There is overlap, but there is also difference.

Consider an example in which a physician gives a patient no information about potentially harmful drug interactions for a medication newly prescribed. The patient follows the directions that accompany the new prescription and is injured by an interaction. The tort system asks, who is responsible for the injury? The doctor. How? She did not know about the specific interaction. Then torts asks how that physician's not knowing that fact compares with what other similarly situated physicians know and do not know. This physician did not know and did not check. Other physicians would have known or would have checked. This physician was therefore negligent, and harm resulted. The physician is ordered to pay the plaintiff an amount approximately equal to the resulting losses.

Quality improvement is by contrast a "backwards march of whys,"⁴⁸ inquir- [*pg 195] ing about the underlying causes of errors to determine where changes can be made for the future. Why did the doctor not know about the interaction risk? Because she was not told about it. Why was she not told? Because this health plan does not publish new drug bulletins for its admitted physicians. Why not ... ? Or, why did this doctor not stop to check? Because she was overwhelmed with a patient load that day. Why was she overwhelmed? Because the managed care utilization scheme provides incentives for heavy patient loads. Why ... ? Or, why did the pharmacy not spot the interaction risk? Because the pharmacy does not have at hand each patient's medication history. Why not? Because the nurses have an incentive to keep the charts in the medical departments. Why ... ?

Individuals do make errors and should be responsible for the quality of their work. Nevertheless, the "bad apple" approach⁴⁹ of the tort system focuses on outliers rather than on more pervasive influences. It looks at outliers as if they were significant when in fact they are most often highly unusual and sometimes random events. The strategy of quality improvement system design, by contrast, is to recognize that errors occur, to recognize that people work within systems, and to design the systems to do two things: (1) to make it difficult for individuals to make errors and (2) to make the whole system capable of "absorbing" individuals' errors when they occur by identifying and correcting errors before they can be harmful.⁵⁰ Even when a doctor has committed an error of judgment or skill, a systems approach demands to know how and why that infraction came about.

It is important to recognize that there are several kinds of differences between torts and quality improvement. For example, the tort and quality improvement systems have different objectives. It is thus not surprising that the tort system is not as good at quality improvement as are explicit quality improvement efforts. That fact, if it is so, would be more tolerable if the tort and quality improvement domains were independent of each other -- allowing the tort system to take care of blame and compensation while medicine's own quality devices (for example, peer review and the "M&M"⁵¹ conference) attend to fixing the future. The difficulty is that they are not independent domains. The conventional tort system is not only ineffective at deterrence, but it may also be positively interfering with the proper working of medical quality improvement, the sole purpose of which is the error prevention that the tort system apparently does not itself achieve. An outline of that analysis calls for a comparison of the requisites of techniques for error prevention with the features of the civil liability system.

There is a never-ending debate on the virtues of the civil liability system and the private bargaining that it shapes. Juries are criticized as people of av- [*pg 196] erage ignorance and manipulable passions, yet praised as bulwarks of the rights of individuals against the pressures of power and capital. Negligence law is either disdained as inefficient and dysfunctional, or canonized as the emblem of the norms of individual freedom and responsibility. The adversary process is either a wasteful food fight in a foreign language, or the crucible of rigorous challenge and testing from which emerges if not the truth, then at least the right answer for the nonce. That the civil liability system has endured is surely some evidence that it satisfies widely or strongly held values, or both. That it has in many ways and places been eroded is equally good evidence for the contrary.

The values and functions invoked in the lawyers' debates, however, are of most interest within the confines of a legal system that charts its terrain in terms of "rights." From the perspective of those in the medical field for whom medical injuries and the means to prevent them are of utmost importance, the values in which the lawyer-debaters keep rhetorical score do not mean much. With the exception of an occasional editorial from the American Medical Association, little discussion is had of "rights" in the literature of health care. On the present subject, malpractice liability, that literature includes no real debates. The medical literature is uniformly critical of the effects of the law's adversarial procedures on the prospects for medical error prevention. We are aware of no quantitative studies that measure the distortions conventional liability procedures may impose on health care quality improvement. There is, however, some evidence to suggest that the present regimes do not effectively deter even negligent error.⁵² That fact runs counter to ordinary economic theory and by itself suggests that some inhibitions or distortions are occurring. A summary comparison of the requisites of quality improvement with the related features of liability litigation, as offered in the following table, presents a highly plausible if not compelling case.⁵³

[*pg 197]



What the Tort System Does	What Error Prevention Needs

Identifies and focuses on the acts of individuals.	Exploration of underlying causes and comprehensive analysis of the systems in which individuals work.

Attaches serious consequences to typically infrequent events; reaches binary win-or-lose outcomes from a process of more-probable-than-not.	Acceptance of errors as windows into "process upsets" or "process inadequacies," which when taken in a comprehensive understanding of outcomes and deviations can illuminate opportunities for improvement.

Through its negligence standard, "blames" individuals for falling below the minimal standard of performance in their profession.	Perception of errors not as failures, but as opportunities to understand and improve every part of the system.

Occurs long after the events giving rise to the claim.	Investigation of errors as close as possible to the "time of event."

Resolves disputes in a highly public manner, causes humiliation to the defendant, and creates the perception that it is punitive.	Departure from a punitive frame of mind and movement toward the belief that errors are inevitable and manageable.

Sets co-defendants in opposition to each other and possesses no intrinsic organizational follow-up.	Cessation of physicians feeling they are targets of outside scrutiny, and involvement of physicians as "active participants in quality assurance."

Creates a defensive orientation in the defendant, which results in the hiding of errors.	Health care Total Quality Management ("TQM") and Continuous Quality Improvement ("CQI") demand the fullest access to information about errors and deviations, and full and free communication about incidents and their causes.

The last comparison in the list is the most often cited, as in a recent letter to the editor of the Journal of the American Medical Association, in which a reader commented that "the wasteful loss of information that could [otherwise] be derived from systematic study of adverse outcomes is the most pernicious [*pg 198] effect of malpractice litigation."⁵⁴ "Total Quality Management," or TQM, is said to require three things:⁵⁵ (1) a culture in which errors and deviations are regarded not as individuals' failures but as opportunities to improve the system, (2) a "grassroots" participation in identifying errors and their sources, and the ways to system modification,⁵⁶ and (3) a commitment to TQM from organizational leadership. These requirements are inhibited, if not prevented, by the traditional attributes of the adversarial civil liability process.

III. THE HYPOTHESIS: CONFORMING THE RESOLUTION PROCESS WITH ERROR PREVENTION

We thus arrive at a watershed. First, the tort system seeks two major goals in its role as the forum for managing claims of medical injury: compensating today's injured persons, and deterring injurious behavior for tomorrow. Second, the deterrence function does not appear to work. Third, as a compensation system, the procedures of civil liability are not independent from but rather appear to run counter to many of the principal requisites of medical quality improvement and, therefore, medical error prevention. And fourth, though for now we take this provisionally,⁵⁷ the tort system is both less accurate and less efficient than some alternatives can be at fulfilling the compensation function.

Looking outward from this watershed, we are not led to propose, as many others have, that the substantive legal aspects of the civil liability system be abandoned in favor of something else, whether it be no-fault⁵⁸ or, contrarily, [*pg 199] the restriction of damage recoveries or access to juries.⁵⁹ Our hypothesis deals with the process by which these claims are addressed, not by the formal rules that would apply to them if they were to be heard in a court. We suggest that a system of voluntary, mediation-based claims management can simultaneously achieve the compensation (and other) goals of the claimants, can be more satisfactory to the defendant physicians, and can add to, rather than interfere with, the larger orbit of medical error prevention.

In the following section, we explore why we believe this hypothesis to be plausible. Here we only suggest that the error prevention literature would incline those who are familiar with mediation toward our hypothesis.

In the words of Dr. Leape, "the most important reason physicians and nurses have not developed more effective methods of error prevention is that they have a great deal of difficulty in dealing with human error when it does occur .... [P]hysicians, not unlike test pilots, come to view an error as a failure of character ...."⁶⁰ Even more to the point, the emotional impact on a physician of an error that caused patient harm is often profound. Nonetheless, under the existing systems of external liability and internal peer review, "physicians are typically isolated by their emotional responses; seldom is there a process to evaluate the circumstances of a mistake and to provide support and emotional healing for the fallible physician."⁶¹

For its part, mediation, when properly employed, can be private, integrative, safe, nonjudgmental, and flexible in scope, process, and outcome. It can be a safe harbor with therapeutic potential, and can offer its participants the opportunity to address the source as well as the consequence of the immediate problem.⁶² Mediation may, in short, offer a process whose traditional attributes are consistent with, rather than antithetical to, the requisites of quality improvement.

As another observer put it, "[t]he paradox of modern quality improvement is that only by admitting and forgiving error can its rate be minimized."⁶³ This is strong medicine; we need not resort to anything quite that stark to make the [*pg 200] point. If mediation can offer the physician nothing more than a less punitive and isolating environment for dealing with the patient's injury and claim, it will offer something very different from the conventional tort system.

Mediation can offer more. The essence of mediation is direct communication between the parties most directly affected, in these cases the physician and the patient. Addressing medical error prevention (and not claims or liability or compensation), one reply to Leape's earlier work suggested that "patients have a priceless and unique perspective both on error prevention and on error resolution."⁶⁴

Spurred by their accrediting agencies⁶⁵ and surrounded by an environment of quality concerns among management theorists generally,⁶⁶ within the past decade hospital management has embraced a "revolution" in the approach to health care quality improvement: "Quality is what the customer says it is."⁶⁷ Quality improvement and medical error prevention are not the same thing, of course. Patients' perceptions of quality -- how cheerfully, for example, a nurse responds to a buzzer in the post-op recovery room -- may affect a patient's assessment of their hospital experience yet may have little or no effect on objective clinical outcomes. While that is certainly true, error prevention bridges the ground between quality improvement and risk management⁶⁸ and is a concern of both:

Even in clinical interactions, customers' perceptions are important indicators of qual- [*pg 201] ity. Consider the physician who explains a procedure in terms that are medically sound and thorough. That physician's explanation might be professional and accurate, but it may still fail to meet the patient's need, which is understandability. How can the physician possibly judge understandability as well as the patient?⁶⁹

The potential benefits from patient involvement in quality-of-care initiatives are widely believed.⁷⁰ There is also strong, though not yet broad, interest in the potential for "putting the [injured] patient into peer review"⁷¹ after an iatrogenic injury has occurred. No link has yet been devised, however, to utilize the occasion and the energy of an injured person's claim as a route for realizing that involvement. Our hypothesis is that applying a form of mediation to the claims process may be a way to accomplish a portion of that goal.

A. Injured Patients' Motivations Are Consistent With Error Prevention

Some of the people injured by what they perceive as a physician's medical error are motivated to take action. They address the physician or the hospital involved, and a few either find satisfaction that way or for other reasons let the matter end there. Others take their complaint to a state board of medical registration or licensing. Those procedures may end, or not, with action by the board. Still others voice a "claim," seemingly demanding compensation for the economic consequences of the error and the injury. The claim may begin with a personal confrontation or with service of process, or can go through several stages as it emerges unresolved from the earlier ones.⁷² We are concerned here with "claims," by which for this purpose we mean express demands for compensation accompanied by allegations of error and injury, and which have the potential to develop into civil liability actions in the courts.

In principle, every claim is a window through which an incident can be examined to determine its underlying cause and thereby contribute to preventive actions for the future. That is only true if someone is looking through the window and if the parties are prepared comprehensively to examine the cause. A malpractice claim is also an event marked by considerable energy invested by at least one of the parties, the claimant. The claimant wants something, and seems highly motivated to achieve it. If the claimant seeks only money through the conventional resolution process, then neither the physician nor the patient has any incentive to look through the window with an interest in the future.⁷³ [*pg 202] The cause will be considered only as a replication in miniature of the negligence regime, namely an individually focused assessment of that one practitioner's competence at that one instant in time. Most important, the remedy will be limited to a benefit for that one claimant.

Taking the present legal environment as a given, these claims will continue to be brought in substantial numbers. If these claims are lemons to the health care system, can we make lemonade? Could the energy causing the claims be redirected to a purpose more consonant with the public interest in medical error prevention?

The economic model of claiming behavior postulates an injured person facing an investment decision, motivated by the chance to gain or regain economic value through a lawsuit. The cost to bring the suit, the chance of success, the distribution of likely outcomes, and other economic variables are used to explain the claimants' behavior. In fact, however, the economic model does not explain very much.

Interest in claiming behavior has a long history, though empirical investigations have been numerous within only the past decade. Happily, for our purpose, most of the recent studies have examined claiming in cases of alleged medical injury.⁷⁴ In most of the studies, populations of patients were analyzed to identify the differences between those persons similarly injured by objective measures who brought claims against their physicians and those who did not. Some studies examined the characteristics of the people⁷⁵ and of the event, finding important correlations between claiming behavior and the affective qualities of the doctor-physician relationship.⁷⁶ Others examined economic⁷⁷ [*pg 203] and cultural⁷⁸ variables, as well as the process by which the potential defendant's own behavior can affect the attribution of blame and hence the propensity of the claimant to sue.⁷⁹ The factor with the greatest value in predicting lawsuits, for example, is how well the provider (physician or hospital) responds to the initial post-injury confrontation by the claimant.⁸⁰ "Stonewalling," which is what many defense attorneys instinctively and routinely advise, nearly guarantees a suit that might otherwise be avoided.⁸¹

Of most direct interest to this investigation is the more precise question not of what triggered the claim, but what the claimants wanted. In one investigation, parents of children who had been neurologically impaired at birth were studied to determine the differences between those families who claimed and those who did not, with special attention to their underlying motivations.⁸² Only a minority were motivated by the need for money.⁸³ The leading factors were these: ⁸⁴

33% Advised to do so by a third person;⁸⁵

[*pg 204]

24% The doctor was not completely honest, or lied;⁸⁶
24% Needed money for the child's future care;⁸⁷
20% Impulse reaction; realized the child had no future;
20% Could not get anyone to tell them what happened; and
19% Anger, revenge, or to assure it did not happen again.

In many cases like these perinatal injuries, a legal claim is pursued as a way to achieve something else (for example, an explanation), and a demand for money is most often the instrument by which some other motivation can be satisfied. That other motivation is most often noneconomic; it is often a search for understanding, communication, and correction. Correction is what error prevention requires. Corrective remedies may thus serve public as well as private goals.

Just as our observations about the requisites of quality improvement brought us to consider mediation as a process for addressing those needs,⁸⁸ these other motivations also lead us to our hypothesis about mediation's potential. Litigation translates plaintiffs' motivations into claims for money, even though the motivation may have originally been non-monetary. Money is the only remedy available to litigants in the civil liability system. Mediation, in contrast, is characterized by unlimited flexibility of remedy, by the effort to retranslate articulated demands back into recognized needs, and by an explicit strategy of creating rather than inhibiting effective communication between the participants.⁸⁹ Thus what we know about the patients combines with what we know about the doctors⁹⁰ and the theories of quality improvement to suggest [*pg 205] that a properly-crafted form of mediation⁹¹ applied to medical malpractice claims may contribute simultaneously to satisfaction of the parties' needs and to enhancement of the link between the claims process and the processes of improving patient safety.

B. An Encouraging Pilot Project: The Massachusetts Voluntary Mediation Program

The Voluntary Mediation Program ("VMP") is a pilot project conducted by the Massachusetts Board of Registration in Medicine in conjunction with the Program for Health Care Negotiation and Conflict Resolution at Harvard.

The Board of Registration is responsible for licensing and regulating physicians in the commonwealth. Among its tasks is receipt and investigation of complaints lodged by patients regarding their medical care. For the VMP, patient complaints are divided into three broad categories. The most egregious problems, in which there has been severe physician misconduct or dangerous substandard care, are handled immediately, and are not eligible for mediation. The Board will usually suspend a physician's license or impose limits on his or her practice in these cases. On the opposite end of the spectrum of complaints are minor issues, such as small billing disagreements, that are generally dismissed by the Board. Between these two extremes are a myriad of dissatisfied patients with legitimate concerns about conduct, outcomes, or communication. These cases are reviewed by the Board's investigatory committee and when appropriate are referred to the VMP.⁹²

The scope and reach of the VMP were constrained in large measure by the political climate in which the Board of Registration in Medicine operates. While considerable attention was devoted to the design of the project by its supporters at the Board, there were concerns that it might spark public criticism before it could demonstrate its results. Some at the Board, for example, feared that mediation would be understood as another instance of the Board "going easy on doctors," a political liability for a regulatory body that already has a lower rate of disciplinary action than occurs in most other states. These concerns have resulted in a useful project though one with very limited funding [*pg 206] and staff and, therefore, few cases. Similar, though more universal concerns resulted in the narrowing of the range of cases that could be submitted to mediation.⁹³

Among the major design questions facing the planners was the motive to mediate: Why would both physician and patient want to meet to resolve their issues? Since the process is voluntary, there must be an implicit impetus both to bring parties to the table and to help them reach resolution. For the physician, the motivation to mediate is the dismissal of the case by the Board. That is, if the case is resolved to the satisfaction of the patient, it will be removed from the permanent and public record of the physician.⁹⁴ For the patient, the mediation offers a low-cost opportunity to meet with the physician, to vent frustration with the medical experience, to make a specific request in the name of settlement, and, as a number of participants have said, to ensure that the same mishap does not befall another patient. The cost of mediation is covered by the state, so there are no financial obstacles to bringing the parties to the table, and when neither has yet retained an attorney to represent them, there are no legal costs. To date, every physician presented with the option has accepted mediation. Only a few patients have rejected it. Once they agree to mediation, the disputants meet,⁹⁵ usually in the offices of the Board of Registration in Medicine, with one or two mediators to review the complaint, discuss the issues, and review and assess the options for settlement.

For all cases deemed appropriate for mediation, the Board of Registration staff who initially contacted patients and physicians regarding their interest in the program heard no unwillingness to try the process. While several of the physicians wanted to contact their attorney or insurer before agreeing to mediation, all eventually seemed to believe that there was little to lose in meeting with the patient or family member. The allure of settlement was heightened by the possibility of removing the complaint from the files of the Board at a time when those files are becoming increasingly open to public scrutiny. While con- [*pg 207] sumers were somewhat intimidated by the prospect of meeting face-to-face with the physician, there was again a nearly uniform acceptance of the process.⁹⁶

Of the first ten cases mediated in the VMP, nine have been successfully resolved. The case that did not settle involved an elderly woman whose emotional instabilities rendered her unable to assess the settlement offer (this in spite of the fact that her son, who accompanied her to the mediation, tried to convince her of the reasonableness of the physician's perspective).⁹⁷

Of the nine cases that were resolved, only four involved any monetary transfer. Brief synopses of some of these cases follow.

In the first a patient experienced an unsatisfactory surgical outcome:

Medical Outcomes Dispute, May 1993
A well-educated male patient underwent groin surgery and complained of severe post-operative pain. The surgeon belittled and eventually ignored the complaints. The patient was examined by another surgeon who conducted a corrective procedure that required further hospitalization. The patient filed a complaint with the Board of Registration in Medicine, noting the poor outcome of surgery, the intransigence of the physician to discuss the problem, and the discovery through later investigation that this procedure has a ten percent failure rate. The surgeon had not disclosed this high risk prior to the first procedure.
The case was assigned to a senior mediator (the general counsel for a university hospital) and a mediator trainee. Both the patient and physician were gripped with anger about the case, though they both agreed to meet one another in mediation.
In his opening statement, the patient mentioned the ten percent failure rate. The physician was shocked to hear this, and corrected that figure: the risk was in fact one-tenth of one percent. The physician explained that he would never have done so risky a procedure without informing the patient. He went on to express his regret that the patient had experienced so much pain after the operation. He further explained that his attorney suggested he not return calls to the patient, as they were perceived as harassing.
The patient was receptive to the physician, and asked about the additional costs of hospitalization. The physician agreed to pay him for one-half of his out-of-pocket costs for the second hospitalization. Four hours had been scheduled for the media- [*pg 208] tion. After forty-five minutes, the patient and physician were shaking hands and asking of each others' welfare.

In another case, a cosmetic surgeon agreed to return the fee paid by a patient for elective surgery that was not covered by insurance:

Cosmetic Surgery Dispute, March 1995
A young woman who aspired to a singing career underwent cosmetic surgery to improve the appearance of her nose. The operation was not covered by her insurance, and she paid for it out of pocket. Unfortunately, the surgery failed, and her nose was severely malformed after the operation. She had a very antagonistic discussion with the physician in which each allegedly issued vindictive charges against the other. The physician offered to repair the nose, and the patient refused. She solicited the opinion of several other plastic surgeons, and chose one to correct the surgery. This second procedure, which she again paid for on her own, was a success. She filed a complaint with the Board of Registration in Medicine against the first surgeon. The disciplinary committee of the Board determined that this case did not reflect serious substandard care, though it certainly appeared to be a communication problem.
Based on a reading of the correspondence of the patient and physician, the case seemed an unlikely candidate for mediation. The hostility was such that it seemed neither could view the other rationally. Nonetheless, it was decided to engage in preliminary interviews. The case was assigned to a senior mediator and a mediator trainee (an attorney and social worker by training, now vice-president of a hospital). The patient was looking for a large sum to compensate her for her out of pocket expenses as well as for her pain and suffering. The physician would not hear of it: He felt his care was appropriate and he offered to fix what was problematic. With a great deal of prodding, it became evident that the physician was willing to repay her what she paid him, and that she would be willing to accept the amount. Because of the intense animosity, it was decided that it would be counterproductive for them to meet. The mediation was conducted entirely by telephone.
To the very last moment, this appeared to be a case that could not settle. Nonetheless, the physician did deliver a personal check for the amount paid by the patient. The patient agreed to have the complaint removed from the open files of the Board. Because the repayment came from the physician personally, and not through an insurer, it was not reported to the [*pg 209] National Practitioners Data Bank.

A third physician, who inappropriately hit a child during an examination, agreed to establish a college fund in the name of the child. And in the fourth, a daughter whose mother died of cancer while under the care of a urologist asked that the physician make a contribution to the American Cancer Society and enroll in continuing medical education:

Wrongful Death Dispute, December 1994
The daughter filed a complaint with the Board of Registration in Medicine stating that the urologist caring for her elderly mother had not properly diagnosed and treated what eventually became cancer of the urinary tract. Over the telephone and without a physical exam, the physician assumed that the reported blood in the patient's urine was an infection, and prescribed antibiotics for the mother. After the bleeding continued, the mother was hospitalized and diagnosed with cancer. At this point her condition was untreatable and she died shortly thereafter. The Board's disciplinary committee investigated and determined that in light of the physician's overall exemplary record, disciplinary action was not warranted. The dispute was referred for mediation.
The case was assigned to a senior mediator and a mediator trainee (an orthopedic surgeon). The daughter stated that she did not want to sue for damages, as no amount of money could return her mother. Rather, she wanted to be sure that this tragedy not be repeated. The daughter vented her feelings to the physician, and he patiently listened. He expressed his own sorrow for the death of a patient he had seen for many years and of whom he was personally fond.
The physician apologized to the daughter and agreed to enroll in a continuing medical education course on urinary cancer. He also agreed to make a contribution to the American Cancer Society in the amount of $1,000.

The remaining cases involved no monetary transfer. In one case, the physician and patient agreed to work together to change hospital research protocols so there would be no reoccurrence of a miscommunication between two research projects that left the patient in a potentially life threatening situation as a result of drug interactions. In two cases, the patient was satisfied with an apology and a face-to-face discussion with the physician and an explanation of what contributed to the problem. In two other cases, the complainants agreed to resolutions in which the physicians changed the practices of their offices, either by providing additional staff training or by revising the information given [*pg 210] to patients to reduce the likelihood of misunderstandings with serious quality-of-care risks. One of these cases typifies what may be the most significant outcomes of a mediation facility such as the VMP:

Child Poked by Unprotected Used Syringe, November 1996
While a mother was waiting in the examining room of her allergist, her three and a half year old toddler found an exposed used syringe and poked herself with it. Though upon subsequent testing the young girl was not found to have contracted HIV, the mother was emotionally distraught about the incident. To add to the emotional intensity, this child had been born only after extensive infertility treatment for the mother. The mother charged that following the event the physician and his office staff were rude and not appropriately responsive. She felt they created unsafe conditions in the office and were not behaving in a responsible manner.
The case was assigned to a senior mediator (an orthopedic surgeon) and a mediator trainee (attorney and senior vice-president of a health care system). The mother came with her husband. The physician was accompanied by an attorney who represented his malpractice insurance carrier.
The mother explained what happened and the emotional consequences it created for her family. The physician apologized for the careless conditions in his office as well as for the immediate response to the incident. He further explained the new office procedures, training, and other measures that had been instituted since the incident to ensure that it could not be repeated. The apology and report of these changes satisfied the parents, and they agreed to resolve the case. The defense attorney had come prepared to write them a check. Because their costs for counseling were covered by their health insurance plan, and because they were motivated primarily by the desire that the incident not be repeated, they did not demand a cash settlement. The defense attorney was amazed (but went along).

These above cases demonstrate one important finding of the project, namely the eagerness and desire of patients to ensure that whatever happened to them would not be repeated for others in the future.⁹⁸ This motivation for bringing the case to the Board is identical to one part of the complex of motivations that causes injured patients and their families to come to the malpractice liability system as well. Patient participation is a source of interest and energy that could provide considerable insight and information to the health care sys- [*pg 211] tem as it strives to better understand and address system-related deficiencies. By conducting these talks in the relative confidentiality of the Board, the physicians recognized that a sincere and diplomatic explanation (and occasional apology) along with an agreement to modify practices, procedures, or capabilities is a reasonable price for clearing his or her record with the state licensing agency.

Through their licensing, complaint review, and regulatory procedures, state boards are able to contribute to the overall quality of medical care in their jurisdictions. Through disciplinary action, boards are able to respond to the most egregious practice infractions. However, there is little they can offer consumers who have legitimate complaints that do not require disciplinary action. Following investigation, these cases often are disposed of with little or no state action, little or no settlement from the perspective of the patient, and considerable antipathy on the part of the physician. Moreover, an adverse event involving a patient's care often causes a defensive response by the physician, the insurer, and the implicated health care system. This defensive response also obscures what could be a valuable resource for the health system: the motive that the occurrence not be repeated.

Given the consequences of unsafe conditions in the health care system, efforts to build constructive avenues of correction should be encouraged. Voluntary mediation creates opportunities for constructive physician education. Physicians often go through a transition in the course of mediation, first being resistant to the complaint and allegations, and then as the process develops, recognizing the legitimacy of the concerns raised. If nothing else, it alerts them not to repeat the offense, so that they do not have a repetition of the complaint. At best, it helps them to reassess the fundamental reasons for the problem and to adopt a course of remedial action. In the VMP, we have seen both. Similarly, when consumers face barriers to addressing problems they experience in the course of their medical care, such as high legal costs, important information is lost to the medical care system. Mediation offers an accessible mechanism to identify these deficiencies.

The results of the VMP suggest that voluntary mediation offers a protected avenue to collect useful information and feed it back into the system. It does so while offering the patient or family member an outlet for their rage over the incident as well as the comfort that some good emerged from what was otherwise likely a tragedy or near tragedy. It further offers the physician an opportunity to do what he or she likely desires: the chance to apologize to the patient and, while expressing regret, to offer assurance that steps were taken to make a repetition unlikely. Mediation thus seems to create opportunities for resolutions that may be more satisfying for the consumer and the physician. The private motivations are highly consonant with the needs of future patient safety. Interestingly, in the VMP, we have found this to be especially true for those cases brought by a family member of a deceased patient. These complainants are reluctant to profit financially from the incident, and often state that all they really want is the impossible return of the deceased. They file the complaint [*pg 212] motivated by the desire to ensure that the same circumstances they experienced are not suffered by other patients or families.

IV. SPECIFIC APPLICATIONS: NEXT STEPS

The VMP was not originally designed to test the propositions developed in this paper, namely that a program of mediation could achieve error-correcting outcomes even when operating within a system of negligence-based liability. It was created as a practical program to offer mediation as an alternative to the formal regulatory complaint process. Under accepted methodological criteria, therefore, its results, while encouraging, do not alone validate our hypothesis, nor necessarily serve as a model for its application in the field. However, the observed results are entirely consistent with what mediation theory would have predicted.

The VMP was in some senses serendipitous. While motivated by interests other than those of a research study, it was getting underway just as we were exploring the theoretical and political dimensions of medical injury, malpractice, and future risk reduction. Moreover, the VMP has thus far addressed only "complaints" brought to a regulatory board, not "claims" brought to the civil liability system. We do not know much about the potential differences between the two. Thus a significant part of the challenge is to see whether the outcomes of the regulatory mediations, the VMP, can be replicated for at least some claims brought directly to the liability insurers or the courts. Our observations about the VMP nonetheless confirm the hypothesis sufficiently to warrant further, more rigorously structured investigation.⁹⁹ (The results have also caused additional activities in the field, a matter to which we turn below.) Prior to launching any such studies, however, a number of structural questions need to be addressed.

A. What Would the Model Look Like?

We can suggest with some degree of confidence that there is at least one reason why health care interests may wish to consider mediation for managing malpractice claims: the potential for prevention of future errors. Suppose, [*pg 213] however, that results like those of the VMP could be achieved regularly, and for civil liability claims in addition to those brought to a regulatory body. If one were thus emboldened to be more prescriptive, what more would one do? Would it be appropriate, for example, to develop a mediation program that had error-risk-reduction as an explicit remedial goal? If so, who would bring that perspective to the table? Would the mediator have a portfolio, so to speak, that encouraged him or her to steer the disputing parties toward a corrective outcome in any case where they seemed amenable to it, even if it was not originally one of their own ideas?¹⁰⁰ Suppose the parties would both be satisfied with a few hundred dollars changing hands and being done with it. Should the mediator overcome the presumption of party autonomy and insist that the conversation be more comprehensive?¹⁰¹

Perhaps the process could be built as a black box, as is much of alternative dispute resolution. Mediators would be trained to understand the wide range of motivations injured claimants have in medical malpractice cases. They would be trained to understand the requisites of effective error prevention and, perhaps, the relationship between the two. They might then conduct mediations without an explicit portfolio of their own, but would in some uninstructed and as-yet-unforeseen way increase the percentage of all mediated outcomes that meet an error correction objective.

If a case came from a regulatory body, would that body have its own representative at the table, and would that help to multiply the proportion of forward-looking outcomes? A somewhat less troublesome possibility would be to have the defendant's insurer or the implicated health care facility (or other sponsor) invite the claimant to participate in a process wherein the patient would both articulate a claim and have a forum to address the cause and cures [*pg 214] of the error. A patient who just wanted to try for the settlement money and leave could do so. Those who wanted a procedure through which to pursue other objectives would have that offered, and the monetary claim might or might not be resolved at the same time.

B. How Would the Results Be Implemented?

Some outcomes would be individualized to the practitioner. The participating doctor, for example, might agree to undertake a period of ongoing training in the relevant area. Other outcomes could have more systemic implications. For example, it might be agreed that a contributing cause of the error was the failure of the surrounding system to do or refrain from doing some particular thing. In either case the information should simultaneously be confidential vis-à-vis the individual, yet integrated into the ongoing system in which the error happened. How is that to be done? Can the necessary confidentiality be assured in states that have anti-secret settlement rules? ¹⁰² If a matter comes to the mediation program through a regulatory body, can that body delegate its public responsibility by agreeing not to pursue administrative remedies if the parties reach a mutually-acceptable solution?

C. How Would the Proposal Be Validated?

The hypothesis to be tested reads something like this: A program of voluntary mediation, designed with an explicit objective of encouraging resolutions of patient complaints and claims in ways that have the potential to reduce future errors, can remove barriers from and add useful information to efforts at medical error prevention, and at the same time achieve degrees of party satisfaction at least equal to those that result from the conventional civil liability process.

How could such a proposition be tested? A population of incidents (or voiced complaints and claims) could be randomized between experimental and control groups. The experimental group would be encouraged to participate in the mediation program; the control group would resolve their claims in whatever manner is conventional for that time and place. The outcomes of the mediations and of the control group settlements would be graded by a blinded panel (double-blinded to the extent possible) made up of medical quality improvement and risk prevention experts. The grading standards would be established in advance, and in ignorance of the structure of the mediation process and of the training received by the mediators. They would be crafted to meet the most significant requisites of quality assurance and risk prevention. Simultaneously, outcome and process satisfaction measures would be taken through coded surveys of the parties.

Would this be enough? How would we know whether the remedy or resolution chosen was appropriate to the matter? Would it be necessary to do in- [*pg 215] dependent file reviews of every case?¹⁰³ Could that information ethically be collected for purposes of the research yet be withheld from the mediating claimant?

Would it be necessary to measure actual future error rates, assuming the experimental and control groups could be identified in their ongoing practice settings? This seems implausible. But would the panel grading be measuring anything other than the conformity of equally theoretical ideas about medical quality assurance?

These are difficult questions. Only some of them are methodological.

V. THE NEAR FUTURE: EXPANDING THE MASSACHUSETTS EXPERIMENT

This has been a report of a work in process. Even as we continue to develop its theoretical and institutional predicates and wrestle with ways to verify our notions with empirical rigor, events have offered an opportunity to test in a pragmatic setting and to think more concretely about several of the issues we have raised.

In early 1997, the Governor of Massachusetts appointed a new administrator of the Commonwealth's Office of Consumer Affairs and Business Regulation. Among the many responsibilities of the Office is oversight of the Board of Registration in Medicine as well as that of eight other licensing and consumer agencies. Prior to his appointment to this position, the new administrator, Mr. Michael Duffy, had been director of the Massachusetts Commission Against Discrimination, where he had developed a highly visible and well-respected mediation system for resolving complaints.

Shortly after his appointment, the administrator met with the Executive Director of the Board of Registration, Mr. Alexander Fleming, and with one of us¹⁰⁴ to learn more about the VMP. Over the next several months, a series of meetings, reports, and agreements were developed that stem directly from the findings of the VMP. In the spring of 1997, and at the administrator's request, the Pioneer Institute awarded a development grant for a new mediation project. With this support, in July of 1997, a comprehensive analysis of Board procedures was completed which analyzed the full implications of expanding the VMP.

In its introduction, the report¹⁰⁵ frames the problem:

[*pg 216]

The Board of Registration in Medicine ... is one of nine agencies that are overseen by the Massachusetts Office of Consumer Affairs and Business Regulation ... . The Board licenses all physicians in the Commonwealth, and when complaints against physicians arise, the Board investigates and attempts to resolve them. If the Board finds merit to the complaint, it may take sanctions against the physician up to and including the revocation of the physician's license... .

There are several problems with this paradigm of administrative law: it is costly to administer, it is slow to resolve disputes, it is frustrating to the parties involved, it almost always involves the aggravation and expense of lawyers, and the results it provides are often not satisfying to the consumer and licensee alike, regardless of who prevails.¹⁰⁶

The report then discusses the use of alternative dispute resolution and its applications to the nine agencies administered by the Office of Consumer Affairs and Business Regulation, carrying the theme specifically to disputes presented to the Board of Registration in Medicine and proposing that the VMP shift to mediating the most serious cases that are presented to it.

The main goal of mediation [through the existing VMP] is to explore options for mutually agreeable resolutions through the help of a facilitator. It is used to maintain and enhance physician-patient relationships that are tainted by incidents of misunderstanding. Still, only a handful of cases have been sent to mediation since the birth of the Program. Their success, however, should be the impetus for exploring the benefits of mediation.¹⁰⁷

It is recommended that the Board's Complaint Committee review all cases, and arrange for mediation for all but the most serious of complaints,¹⁰⁸ such as those that would demand the immediate revocation of the physician's license.

In order to protect the public's interest in the mediation of complaints, the report recommends direct participation by a member of the Board of Registration's staff, thus implicating a number of the opportunities and issues already discussed.

This proposal also brings directly into play the problem of the public's right to know. Mediation is a confidential process. The Board is a public agency. How are the two mandates appropriately balanced? Although it may not be acceptable in every situation or in every jurisdiction, the Governor and sponsors of this initiative prefer generally to protect confidentiality, weighing the [*pg 217] loss of public disclosure as a small price to pay for what may be improved public protection through enhanced medical care.¹⁰⁹

Another of the structural issues discussed above is the problem of translating a goal-directed mediation outcome into the world of quality improvement. Discussions with the Administrator and with the Board thus led to the development of the "Patient Safety Statement" ("PSS").

Rather than describe the outcome of a mediation as a settlement, the product of this new program will be an agreement focused specifically on future patient safety. A complaint regarding a matter that occurred in the past, and that is the basis of a present mediation, becomes explicitly the basis for improved patient care in the future. A PSS could be directed to the individual physician, to the health care system, or to specific organizations or departments whose changed behavior or procedures could serve to enhance future error prevention. An individual response by a physician could include a continuing education course, a pledge of corrective actions, or even an apology. On the system level, the statement could cover office procedures, patient care protocols, or organizational communications. The agreement could include a monetary payment by the physician to the patient.

The PSS might include aspects that are confidential as well as aspects that are in the public domain. For example, if the physician agreed to take personal corrective action, that could be held in confidence. If, however, restrictions were imposed on his or her practice privileges, that information would be made available to the Board and to the National Practitioner Data Bank as required by law, and possibly to hospitals, colleagues, or others who would need to be aware of the agreed-upon sanctions.

VI. CONCLUSION AND FOREWORD

This paper has described an assembly of ideas, some well established and some new: (1) that the litigation process has demonstrated no significant effects on future error prevention in the area of medical malpractice; (2) that endemic features of the tort liability system are inconsistent with the requisites of quality improvement; (3) that malpractice claimants bring an energy to the compensation and complaint procedures that is motivated by, and may be satisfied by, resolutions other than money alone; and (4) that mediation of medical malpractice complaints offers a system that may harness the energy and the resources now devoted to litigation and direct them at least in part toward quality improvement. Additionally, we have described the results of a field trial, the VMP, which, while not originally established to test these ideas, tends to confirm them, albeit in a regulatory and otherwise somewhat specialized setting. We have attempted to outline some of the major questions, both theoretical [*pg 218] and experimental, that need to be addressed before any more rigorous and comprehensive empirical analysis is undertaken. Finally, we have described a new initiative that may permit practical and politically-feasible solutions to several of these problems.

This new model, which has the support of both Massachusetts's Board of Registration in Medicine and of its Office of Consumer Affairs and Business Regulation, views the patient-consumer experience as an asset to improved quality of care. The Patient Safety Statement and the new-found emphasis that it implies move the compensation and complaint function from an adversarial adventure into a context that has constructive problem solving as an explicit objective. The model offers a unique opportunity to link the advantages of mediation to the purposes of improved patient safety. These advantages include a less adversarial process, more open communication afforded by confidentiality, and facilitated and mutually educational discussion. Together, these advantages allow the patient to realize what has been the motivation for many consumers who have been injured by their medical care, namely the hope that what happened to them will not be repeated for someone else.

We conclude with the objectives expressed by the Director of the Massachusetts Office of Consumer Affairs and Business Regulation:

Using mediation to resolve medical disputes shows a willingness to refrain from establishing blame and move on to try and prevent future problems from occurring. Thus far, issues of malpractice, negligent care, and other patient-physician problems [have been] handled by trying to find the guilty party. In many cases, however, the guilty party is not simply the physician, and trying to distribute responsibility and blame may result in frustration and further agony. Shifting to preventive dispute resolution, that is, attempting to prevent past problems from recurring, is a much needed change ... . [M]ediation can be the catalyst that would bring about systemic change. The long run benefit is invaluable. ¹¹⁰

FOOTNOTES

*	Dean Emeritus and Professor of Law, University of Denver; President, National Center for Preventive Law.
**	Director, Program for Health Care Negotiation and Conflict Resolution, Harvard School of Public Health.
1.	Robert H. Mnookin & Lewis Kornhauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 YALE L.J. 950 (1979).
2.	Iatrogenic injuries are those induced in the patient by action of the physician or other medical personnel.
3.	In a number of states, public agencies and private groups continue to explore the feasibility of "no-fault" or "avoidable-harm" medical malpractice systems. These initiatives have not yet determined how to link the compensation process most effectively with the processes of quality improvement. Our theoretical investigations and these early results from the Massachusetts pilot may also be of assistance to statutory programs such as these.
4.	PATIENTS, DOCTORS, AND LAWYERS: MEDICAL INJURY, MALPRACTICE LITIGATION, AND PATIENT COMPENSATION IN NEW YORK (1990) [hereinafter HARVARD MEDICAL PRACTICE STUDY]. The error and negligence rates discovered for the State of New York conform reasonably well to the rates found for Colorado in a later study by some of the same authors. See Studdert et al., Can the United States Afford a "No-Fault" System of Compensation for Medical Injury?, 60 LAW & CONTEMP. PROBS. 1 (Spring 1997).
5.	See Howard H. Hiatt, Harvard Medical Practice (NY) Study, in ISSUES IN MEDICAL LIABILITY: A WORKING CONFERENCE 13, 13 (Mary L. Grady & Randie A. Siegel eds., 1991).
6.	See Studdert et al., supra note 4, at 18-29 (grading adverse events from record review of samples of cases in Colorado and Utah into those that would meet the "Swedish criterion" of "avoidability," and those that would meet the American criterion of "negligence"). Of all adverse medical events in their samples, 28-32% were caused by negligence and 54-55% were avoidable. See id. A "preventable" error is not necessarily a "negligent" error, though the concepts are similar, and confusion between them is always possible. A negligent error is one that could have been avoided by the exercise of due care, where "due care" is a construct of the risk of harm, the magnitude of the potential harm, and the burden or difficulty of avoidance. In medical malpractice, the professional standard of care short-cuts the calculation, inquiring whether the degree of care exercised was consistent with the standard of care generally found in the profession at the time and in a similar place. A preventable error could be the result of a non-negligent mistake. Implicitly, it too includes some contemplation of cost versus benefit. Any medication error could be avoided, for example, if a team of 100 nurses and pharmacists sat continuously at every hospital bedside. An error that could be prevented only in that way is not "preventable" in any useful sense. The concept of preventable error in the medical literature thus explicitly measures one half of the negligence equation-whether the error could have been avoided-but only implicitly and inexactly recognizes the other half, namely. the relationship between marginal cost and marginal gain. So far as we can tell, there is no standard definition for "preventable." A related and sometimes interchangeable concept is the "adverse event," defined as an injury that prolongs a hospital stay or results in disability at the time of discharge and "is caused by inappropriate medical management instead of the disease process." Anne C. O'Neil et al., Physician Reporting Compared With Medical-Record Review To Identify Adverse Medical Events, 119 ANNALS INTERNAL MED. 370, 371 (1993). Reasonable prevention strategies should decrease the number of adverse events. See Troyen A. Brennan et al., Hospital Characteristics Associated with Adverse Events and Substandard Care, 265 JAMA 3265, 3268 (1991). All negligent errors are preventable; but not all preventable errors are the result of negligence. In this article, we use "preventable" to mean both negligent and non-negligent errors, recognizing that while the civil liability system is concerned with the well-defined former, almost everyone else is concerned with the ill-defined latter.
7.	The remaining one-quarter of injuries result from unpreventable errors. See Hiatt, supra note 5, at 13; Brennan et al., supra note 6, at 3266. Brennan et al. note that adverse event rates do not vary with negligence rates, since different hospitals have different kinds of patient populations. In addition, negligence rates are highly nonuniform among hospitals, suggesting that "substandard care is not randomly distributed." Id. at 3268.
8.	See Lucian L. Leape, Error in Medicine, 272 JAMA 1851, 1851 (1994).
9.	That number is very likely lower than the true count. Errors that occur in hospitals can be identified and counted. Errors that occur in out-patient settings often do not generate a chart or other evidence.
10.	See Leape, supra note 8, at 1851.
11.	Id.
12.	See HIV-1 Antigen Test Unjustified for Blood Donor Screening, FDA Panel Finds, CCBC NEWSL. (Council of Community Blood Ctrs., Wash., D.C.), June 23, 1995, at 1 [hereinafter HIV-1 Antigen Test]. A quality adjusted life year ("QALY"), one year of high quality life, is the product of years alive times the proportion of high quality. Thus, for example, four years of 75% quality is equal to three QALYs.
13.	The standard human immunodeficiency virus ("HIV") tests detect antibodies to the virus. There is a "window" period of two to four weeks during which a blood donor could have been infected-and therefore be infectious-yet not have developed detectable antibodies. See id. The p-24 antigen test detects the virus more directly and earlier, thus closing the already small window a bit more. See id.
14.	Conversation with James MacPherson, Executive Director of America's Blood Centers (Nov. 1997). When p-24 antigen testing of donated blood was debated in 1995, it was estimated that it could save 24 incremental cases per year of post-transfusion HIV, "at an added cost of ... $2.3 million per [QALY]," see HIV-1 Antigen Test, supra note 12, at 1, while others believed the number of cases would be half that and thus the cost would be over $5 million per QALY, see Red Cross Reports First HIV Antigen Positive, Antibody Negative Donor, CCBC NEWSL. (Council of Community Blood Ctrs., Wash., D.C.), Sept. 27, 1996, at 1 (noting that some earlier predictions of the value of antigen testing had been in the range of four to six cases saved per year). In fact, as of September of 1996, only one infected donation had been identified out of eight million units tested. See id. And, as of this date, in all of Canada, p-24 antigen testing has identified no incremental infected units at all. See Conversation with James MacPherson, supra. The cost of antigen testing per QALY is therefore somewhere between $5 million and infinity. Moreover, nearly one-quarter of the persons receiving transfusions will die within one year in any case, from the illness or accident that called for the transfusion. See Eleftherios Vamvakas & Howard F. Taswell, Mortality After Blood Transfusion, 8 TRANSFUSION MED. REVS. 267, 270 (1994).
15.	See, e.g., PATRICIA DANZON, MEDICAL MALPRACTICE: THEORY, EVIDENCE, AND PUBLIC POLICY 29 (1985) ("[I]t is almost certainly still true that the cost of injuries due to malpractice far exceeds the cost of claims."); Lucian L. Leape et al., Systems Analysis of Adverse Drug Events, 274 JAMA 35, 35 (1995) (noting that of the 1.3 million iatrogenic injuries annually, "as many as two-thirds may be the result of errors in medical management").
16.	See Leape, supra note 8, at 1851.
17.	See id.
18.	See id.
19.	See Nat'l Highway Traffic Safety Admin., U.S. Dep't of Transp., Overview: Traffic Safety Facts 1996 (visited Nov. 5, 1997) <http://www.nhtsa.dot.gov/people/ncsa/overvu96.html>.
20.	This is not, of course, the measure of medical negligence. It is the consequence of errors in medical management, of which a substantial portion are preventable and a minority of which result from negligence. America is not awash with negligent practitioners. Errors nevertheless persist, in some abundance. Our concern is the prevention of medical error. We are agnostic on the question of how well identifying negligence contributes to that goal. A distribution of error rates classified by severity and avoidability (negligence versus adverse outcome) has been described in DANZON, supra note 15, at 22 tbl.2.2.
21.	Correspondence with Lucian L. Leape (June 1997) (on file with Edward Dauer). The rate among some high-risk specialties such as OB-GYN and a few surgical specialties is 70%. See FRANK A. SLOAN ET AL., SUING FOR MEDICAL MALPRACTICE 15 n.14 (1993).
22.	See BUREAU OF HEALTH PROF., U.S. DEP'T OF HEALTH & HUMAN SERVS., NATIONAL PRACTITIONER DATA BANK-1995 ANNUAL REPORT tbls.3, 5 (1995).
23.	See HARVARD MEDICAL PRACTICE STUDY, supra note 4, at 9.
24.	This is not the equivalent of saying that there is not a problem with the malpractice liability system. In fact, for that problem, some tort reforms-collateral payment offsets, attorneys' contingency fee rates, speculative non-economic damages, and so on, are an appropriate response. But that is another story.
25.	Holding injury and loss constant, medical malpractice jury awards are typically 150% of other tort awards. See Frank A. Sloan & Randall R. Bovbjerg, Medical Malpractice: Crises, Response and Effects, HEALTH INS. ASS'N AM. RES. BULL. (Wash., D.C.), May 1989, at 10. Nevertheless, even in medical malpractice, the compensation received by most injured claimants is less than their actual economic loss; and the degree of under-compensation is higher for those with greater losses. See SLOAN ET AL., supra note 21, at 220. For a comprehensive treatment of medical malpractice jury verdicts, see NEIL VIDMAR, MEDICAL MALPRACTICE AND THE AMERICAN JURY (1995).
26.	Andrews et al. reported serious injuries at a rate of 17.7%, with claims for compensation at only 1.2%. See Lori B. Andrews et al., An Alternative Strategy for Studying Adverse Events in Medical Care, 349 LANCET 309, 312 (1997). The Harvard study found that one out of seven patients injured through actionable negligence made claims (assuming that all claims made came from the pool of negligent events), and that one out of five cases where negligence caused death or at least six months of disability resulted in a paid claim. See Hiatt, supra note 5, at 14. Claiming rates for other kinds of torts tend to be higher. The most comprehensive study, though now a bit dated, is Richard E. Miller & Austin Sarat, Grievances, Claims and Disputes: Assessing the Adversary Culture, 15 L. & SOC'Y REV. 525 (1981). More recent though more limited studies include Sally Lloyd-Bostock, Propensity to Sue in England and the United States of America: The Role of Attribution Processes, 18 J.L. & SOC'Y 428 (1991), and Herbert M. Kritzer, Propensity to Sue in England and the United States of America: Blaming and Claiming in Tort Cases, 18 J.L. & SOC'Y 400 (1991). Miller and Sarat found tort claiming rates as high as 86%. See Miller & Sarat, supra, at 544-46.
27.	The following analogy has been suggested. Assume the perfect way to take a particular corner in your car is to drive at 30 mph and to turn the wheel 20 degrees. In the real world, you are approaching the corner at 85 mph. When it is not possible to achieve the perfect speed in time, it is folly nevertheless to turn the wheel the perfect 20 degrees. This example was given to an audience at a panel discussion on law and economics in the late 1970s by Professor Richard Markovits of the University of Texas.
28.	About 2% of physicians sued for malpractice pay any portion of the awards or settlements themselves; and about 6% pay for any part of their own defense costs. See HARVARD MEDICAL PRACTICE STUDY, supra note 4, at 9-1. Medical malpractice insurance provides even less incentives for deterrence than other liability insurance does, since the premiums are virtually never experience-rated, and co-payments are extremely rare. See DANZON, supra note 15, at 128-30.
29.	See HARVARD MEDICAL PRACTICE STUDY, supra note 4, at 9-58. Physicians' "great emotional distress" over tort suits is seen by them as a "psychological distraction" rather than an influence on their practices. See id.
30.	The report is distributed every time the practitioner applies for new privileges at another facility, regardless of how often that occurs. There is no sunset for these reports. An NPDB record is forever. See Health Care Quality Improvement Act of 1986, 42 U.S.C. §§ 11101-11152 (1994) (requiring the reporting of malpractice payments and maintenance of those records but not authorizing the deletion of records).
31.	Thus the cost of defensive medicine can easily exceed the value it contributes to patient safety. Defensive medicine is a reaction to the perception of the risk of being sued, which is on average seen as being three times higher than it actually is. See Alice Gosfield et al., Report of the Institutional Transformations Panel, in HEALTH CARE DELIVERY AND TORT: SYSTEMS ON A COLLISION COURSE? 17, 18 (Elizabeth Rolph ed., 1991).
32.	Sloan and Bovbjerg, writing in 1989, concluded that "there is no credible information available on the benefit of the [medical] malpractice liability system." Sloan & Bovbjerg, supra note 25, at 36; see also SLOAN ET AL., supra note 21, at 2 ("No-one presently knows with any degree of confidence whether tort law over- or under-deters.").
33.	See Sloan & Bovbjerg, supra note 25, at 34.
34.	Two examples of specialties with high claim rates are general surgery and obstetrics-gynecology.
35.	Some geographic areas might be avoided as well, or at least some practices in rural areas where specialized medical resources are less readily available to help deal with the unusual case or complication.
36.	Practice parameters are standard protocols for the treatment of designated conditions. In Maine, physicians from certain specialties who choose to participate in the initiative are immunized by statute from malpractice liability if they can demonstrate that their treatment of a case conformed exactly to the pertinent practice parameter. It is a matter of some debate whether these parameters improve safety by reining in rogue treatments or reduce safety by inhibiting necessary departures in individual cases. The purpose of enacting practice parameters was expressly to decrease the cost of defensive medicine and the cost of liability claims; their adoption was not motivated by a concern for quality of care improvement. See Edward David, Maine Medical Demonstration Project, in ISSUES IN MEDICAL LIABILITY: A WORKING CONFERENCE, supra note 5, at 11, 12.
37.	It is not uncommon to find a blend of the two strategies in medical risk-management systems. Thus, for example, the opening paragraph of an American Medical Association publication on risk management advises in its first paragraph: Many physicians find themselves involved in professional liability claims that are not prompted by actual medical negligence, but by deficiencies in administrative systems or communications in the practice setting. Lack of attention to details like recordkeeping, or inadequate communication between members of the health care team responsible for the patient, can place the physician in an indefensible position in a court of law. SPECIALTY SOC'Y MED. LIABILITY PROJECT, AM. MED. ASS'N, RISK MANAGEMENT PRINCIPLES AND COMMENTARIES FOR THE MEDICAL OFFICE 1 (1990). Certainly the defensive measures may be consistent with good practice, too, though the distinctiveness of efforts to avoid claims is obvious in the mix. Thus one example from the same source: There is a growing number of large medical liability judgments and awards arising from claims where defendant physicians failed to produce documented evidence of appropriate discharge or post-treatment continuing care instructions. Although most patients are provided with specific instructions before they leave the physician's office, documentation of such advice or warnings is seldom incorporated into the patient's record or, more importantly, provided to patients in writing. Id. at 4-5. One report from a state medical society roundtable reflects a unanimous conclusion that defensive medicine is practiced "not for the benefit of the patient but to protect [the physicians] from liability claims." David, supra note 36, at 12. Torts scholars may object that "defensive medicine" initiatives that are not in fact useful to error reduction will not succeed in reducing liability under the law, because they do not add to the measurable exercise of "due care." And if they do not reduce liability but are nevertheless costly, they will be abandoned by economically rational physicians. So why worry? Perhaps that is so in some theory of tort law; it is not so in the field, where many practitioners believe that incidents of malpractice liability are not in fact linked with negligent error in the first place, but instead arise from gross imperfections in the legal system. Those practitioners often take the view that steps taken to combat liability will have a better payoff than steps taken to improve a practice that (some believe) is not so riddled with negligence as the frequency of claims suggests. See Gosfield, supra note 31, at 29; VIDMAR, supra note 25, at 3-8. Interviewed directly about the practice-and the value-of defensive medicine, "many or most physicians clearly feel that any improvement achieved is not worth the expense." See Sloan & Bovbjerg, supra note 25, at 10.
38.	HARVARD MEDICAL PRACTICE STUDY, supra note 4, at 9-2, 9-34.
39.	Kirk B. Johnson, AMA/Specialty Society Medical Liability Project, in ISSUES IN MEDICAL LIABILITY: A WORKING CONFERENCE, supra note 5, at 18, 19.
40.	See HARVARD MEDICAL PRACTICE STUDY, supra note 4, at 10-2, 10-3. The investigators found no evidence of a deterrent effect-i.e., no evidence that higher claims rates reduce adverse events. While the data were far from compelling, the Study's conclusion was that "our findings are at best weak evidence of no deterrence." Id. at 10-44.
41.	See PAUL WEILER, MEDICAL MALPRACTICE ON TRIAL (1991).
42.	Id. at 88.
43.	Id. at 90. The fact that the result is "statistically significant" does not mean that the effect is significant. It means only that the measure of the dependent variable (negligence reduction), however large or small it may be, is more likely to be the result of the independent variable (liability) than to be a case of random non-causal association. Thus Weiler can be interpreted as saying that there is evidence to believe that some effect occurs, though the size of the effect itself is modest.
44.	See SLOAN ET AL., supra note 21, at 2. Another of the Harvard study's investigators (Hiatt) has nevertheless estimated that if the threat of tort liability were completely removed, the negligent error rate would rise from 1% to 1.7% of all hospitalizations. See Hiatt, supra note 5, at 15.
45.	As will become clear below, we mean to include in the tort liability system not principally its network of substantive rules, but much more so the procedures those rules and related rules employ. We mean, of course, civil litigation, and the private settlement negotiation that occurs in the shadow of the law. To a significant degree, the process of insurance claims management is also a part of the tort liability system.
46.	We use quality improvement as a generic term. In some contexts, it refers to specific forms of quality-enhancement activities; our use of the term is broader. We mean by it the entire range of error-reduction and quality-improvement activities in which health care organizations and practitioners do or reasonably could engage with those purposes in mind.
47.	This is obviously a larger proposition than we need to make at this point, and in any event one that is too large to be adequately addressed here. However, it should be kept in mind that "deterrence" is an answer often given when the question is asked why two people equally badly injured in the same hospital do not both recover when only one was injured by someone else's negligence.
48.	See Rebecca Voelker, "Treat Systems, Not Errors," Experts Say, 276 JAMA 1537, 1538 (1996) (describing the "sentinel event" process of the Joint Commission for the Accreditation of Health Care Organizations). Iatrogenic death is a sentinel event.
49.	See Heather Palmer, Professional Review Sanctions, in ISSUES IN MEDICAL LIABILITY: A WORKING CONFERENCE, supra note 5, at 51, 54.
50.	See Leape et al., supra note 15, at 35-36.
51.	Morbidity and mortality.
52.	See supra text accompanying notes 32-44. The same is true in other areas of personal injury. See generally George Eads & Peter Reuter, Designing Safer Products: Corporate Responses to Product Liability Law and Regulation, 7 J. PROD. LIAB. 263 (1984). Eads and Reutter believe that "[i]t is not possible to measure the improvement, if any, ... that has resulted from changes in regulation and law." Id. at 263. The RAND study on which the article reports recounts, however, numerous ways by which the imposition of liability distorts rather than enhances efforts at product safety. See id. at 272-76.
53.	The following tabulations combine information from several sources. See LYNNE CUNNINGHAM, THE QUALITY CONNECTION IN HEALTH CARE (1991); Leape, supra note 8; O'Neil et al., supra note 6; Voelker, supra note 48; The Quality March, HOSPITALS, Dec. 20, 1993, at 40. Also incorporated are comments made by participants at two discussion groups held at the Harvard Law School and at the Harvard School of Public Health during the Fall of 1996.
54.	David D. Grove (Letter to the Editor), 274 JAMA 458, 459 (1995).
55.	The related concept of Continuous Quality Improvement ("CQI") has similar requirements: (1) recognition that health care is delivered by systems rather than by collections of individuals; (2) a focus on service to the customers, recognizing patient definitions of quality as well as clinical definitions; (3) identification of key organizational processes that have impacts on quality; (4) teamwork; (5) an information-rich environment; and (6) procedures to assure continuous improvement. See, e.g., Gosfield et al., supra note 31, at 39. See generally CUNNINGHAM, supra note 53. A survey conducted in 1993 showed that 69% of the 3,303 hospitals queried had explicitly adopted some form of CQI. See The Quality March, supra note 53, at 40.
56.	An interesting corroboration can be found in the work of Morlock and Malitz, who report that "the only risk management program activity related to lower numbers of claims experienced was in-hospital educational programs regarding physician and nurse responsibilities in [quality assurance and risk management]." Laura Morlock & Faye Malitz, Do Hospital Risk Management Programs Make a Difference? Relationships Between Risk Management Program Activities and Hospital Malpractice Claims Experience, 54 LAW & CONTEMP. PROBS. 1, 21 (Spring 1991).
57.	The evidence is very substantial, though it is thus far indirect, that a system of voluntary, post-dispute mediation can achieve the standard compensation goals in medical injury cases with greater efficiency and no less accuracy. That is the hypothesis of one of the companion studies to this one, which seeks to validate or refute such claims through a controlled experiment. See Edward A. Dauer, Leonard J. Marcus & Susan M.C. Payne (forthcoming 1998) (reporting on a Johnson Foundation-funded research project).
58.	We should disclose that one of us has proposed a variant of no-fault-a "designated compensation event" system-in one narrowly-defined context, namely liability resulting from the use of blood or blood products. See Edward A. Dauer, Administrative and No-Fault Systems for Compensating Medically-Related Injuries, in BLOOD AND BLOOD PRODUCTS: SAFETY AND RISK 157, 168 (Henrik Bendixen et al. eds., 1996).
59.	The AMA, for example, has made such a proposal. See generally SPECIALTY SOC'Y MED. LIABILITY PROJECT, AM. MED. ASS'N, A PROPOSED ALTERNATIVE TO THE CIVIL JUSTICE SYSTEM FOR RESOLVING MEDICAL LIABILITY DISPUTES: A FAULT-BASED, ADMINISTRATIVE SYSTEM (1988).
60.	Leape, supra note 8, at 1851.
61.	Id. at 1852.
62.	See, e.g., LEONARD J. MARCUS, RENEGOTIATING HEALTH CARE (1995). We do not neglect the arguments that have been made against mediation, such as the argument that because the neutral mediator has no authority, mediated outcomes can replicate the parties' pre-existing inequalities of wealth and power. Our point here is not to argue for the use of mediation, but to offer a testable hypothesis and an explanation for why that hypothesis has at least initial plausibility. Eventually the criticisms must be addressed, though we are much less concerned with them in a voluntary post-incident process than we would be if the mediation was mandatory either by rule of law or by pre-dispute contract.
63.	David Blumenthal, Making Medical Errors Into "Medical Treasures," 272 JAMA 1867, 1868 (1994); see also Grove, supra note 54, at 459 ("[F]orgiveness will be needed for systematic improvement of medical practice to occur. Such forgiveness will be in short supply under a tort-based system.").
64.	Michael D. Fetters (Letter to the Editor), 274 JAMA 458, 458 (1995).
65.	One such agency is the JCAH, or Joint Commission for the Accreditation of Hospitals, now the JCAHO, or Joint Commission for the Accreditation of Health Care Organizations. The JCAH is described as follows: The JCAH was formed in the early 1950's by the American College of Surgeons, the American College of Physicians, the American Hospital Association, and the American Medical Association. Its purpose was to establish minimum hospital standards for patient care. For details for the program, see Dornette, The Legal Impact on Voluntary Standards in Civil Actions Against the Health Care Provider, N.Y.L. Sch. L. Rev. 925, 925-28 (1977); Holbrook & Dunn, Medical Malpractice Litigation: The Discoverability and Use of Hospitals' Quality Assurance Committee Records, 16 Washburn L.J. 54, 57 (1976). Jackson v. Power, 743 P.2d 1376, 1383 n.13 (Alaska 1987). Hospitals voluntarily seek accreditation for financial and professional prestige reasons. First, accreditation by the JCAH means the hospital qualifies to participate in the federal Medicare and Medicaid programs. Accreditation by JCAH is deemed substantial compliance with the Medicare conditions of participation. 42 U.S.C. § 1395bb (1982); 42 C.F.R. § 405.1901(d) (1986). Second, JCAH accreditation is often a prerequisite to obtaining approval of internship and residency programs. Finally, the institution's reputation and standing in the community is affected by whether it received JCAH accreditation. Id. at 1383 n.14 (citations omitted).
66.	See, e.g., HEALTH LAW SEC., AM. BAR ASS'N, ACHIEVING QUALITY IN MANAGED CARE 1 (John D. Blum ed., 1997) ("Quality medical care is a phrase that has become a mantra of health care providers, regulators, purchasers and consumer in the 1990s.").
67.	MARA M. MELUM & MARIE K. SINIORIS, TOTAL QUALITY MANAGEMENT: THE HEALTH CARE PIONEERS 57 (1992).
68.	Quality improvement, in the words of one of the Golden Oldies, is designed to "accentuate the positive," while risk management tries to "eliminate the negative." There is all too frequently a separation between the two, based on perceived distinctions approximating those of the old refrain. It is possible that while quality improvement has figured out how to incorporate patients' views, risk management has not.
69.	WENDY LEEBOV & CLARA J. ERSOZ, THE HEALTH CARE MANAGER'S GUIDE TO CONTINUOUS QUALITY IMPROVEMENT 34 (1991).
70.	See, e.g., CUNNINGHAM, supra note 53, at 107.
71.	Speaking of the potential for mediation in a proposed "avoidable-injury" compensation program, Dr. Eric Thomas of the Harvard School of Public Health has written, "[One] advantage mediation brings ... is the possibility of being a catalyst for CQI. The information gained from the mediation process [could] be fed back into the system to prevent similar injuries from happening again. ... [w]e [would thus ] have a way of using the increased number of reported events to improve quality." Letter from Dr. Eric Thomas, Harvard School of Public Health, to the authors (Dec. 13, 1996) (on file with Edward Dauer).
72.	It may, in addition, first be heard by the physician, or by the hospital, or by an insurer.
73.	There are, of course, exceptions. Some liability insurance carriers have well-developed risk-management programs that do use their claims experience effectively as a source of quality information, and that do define risk-management to include both error prevention and loss prevention. Still, the insurer's job is to pay for the defense of the doctor in an adversarial system. When the positions are that (1) the doctor did nothing negligent and (2) the focal point is the doctor, it is difficult both organizationally and strategically to pursue a comprehensive and objective investigation of the incident. We suspect this happens less frequently than it should and plan to measure it in future research. The National Practitioner Data Bank might also seem to be a positive link between claims and prevention. This is a controversial question and one too large to be addressed here. As a signaling device, the NPDB is not always taken seriously as a preventive tool. See OFFICE OF INSPECTOR GEN., U.S. DEP'T OF HEALTH & HUMAN SERVS., NATIONAL PRACTITIONER DATA BANK: USEFULNESS AND IMPACT OF REPORTS TO HOSPITALS ii (Feb. 1993) ("Data Bank reports rarely led hospitals to make privileging decisions they would not have made without the reports, even when the reports provided information that hospitals did not already know."); OFFICE OF INSPECTOR GEN., U.S. DEP'T OF HEALTH & HUMAN SERVS., NATIONAL PRACTITIONER DATA BANK: USEFULNESS AND IMPACT OF REPORTS TO STATE LICENSING BOARDS iii (March 1993) ("Most licensing Boards have not queried the Data Bank.").
74.	Perhaps this is so because the underlying records are more readily accessible? Or because medical malpractice claims have been a matter of intense political as well as academic interest?
75.	See, e.g., Marlynn May & Daniel Stengel, Who Sues Their Doctors? How Patients Handle Medical Grievances, 24 L. & SOC'Y REV. 105, 106-07 (1990); see generally Mark I. Taragin et al., Physician Demographics and the Risk of Medical Malpractice, 93 AM. J. MED. 537 (1992) (finding male doctors three times more likely to be in a high claims group than female doctors).
76.	See, e.g., Howard B. Beckman et al., The Doctor-Patient Relationship and Malpractice, 154 ARCHIVES INTERNAL MED. 1365, 1368-69 (1994) (identifying doctor-patient relationship issues as motivations for patient complaints, including complaints of desertion, devaluation of patient or family views, lack of information, and failure to understand the patient or family perspective); Gerald B. Hickson et al., Obstetricians' Prior Malpractice Experience and Patients' Satisfaction With Care, 272 JAMA 1583, 1586 (1994) (finding that frequency with which physicians are sued is related to patients' satisfaction with interpersonal aspects of medical care); Wendy Levinson, Physician-Patient Communication: A Key to Malpractice Prevention, 272 JAMA 1619, 1619 (1994) ("Malpractice attorneys asked to cite the primary reason the patient pursued a malpractice suit report that more than 80% are due to communication issues ... . 35% were due to physician attitudes (in a hurry, air of superiority), 7% were due to physician disparagement of previous care, and 5% were due to unrealistic patient expectation.").
77.	See, e.g., LaRae I. Huycke & Mark M. Huycke, Characteristics of Potential Plaintiffs in Malpractice Litigation, 120 ANNALS INTERNAL MED. 792 (1994) (citing lack of health insurance and outstanding medical bills); Molly McNulty, Are Poor Patients More Likely to Sue for Malpractice?, 262 JAMA 1391 (1989).
78.	See, e.g., Herbert M. Kritzer et al., Aftermath of Injury: Cultural Factors in Compensation Seeking in Canada and the United States, 25 L. & SOC'Y REV. 499 (1991).
79.	A number of studies by Sally Lloyd-Bostock and others are collected in 2 EDWARD A. DAUER, MANUAL OF DISPUTE RESOLUTION: ADR LAW AND PRACTICE § 17 (1994).
80.	The several studies have been collected and integrated at least twice, once by one of the present authors, see id., and once by Professors Penchansky and Macnee. Dauer attempted to fit a single model to the collection of findings, by using two axes of three variables each-Facts about the people, the injury and the setting; and Values including economic needs, personal comfort, and perceptions of fairness. He and his collaborator Peper concluded that "Fact" variables are of much less effect on the actual claiming decision than the "Values" are, and that within values perceptions of fairness seemed to be the most salient. Of all of the variables they examined, Penchansky and Macnee found the quality of the relationship and the quality of the communication to be the largest in reducing willingness to sue, exceeding every other variable relating to the doctor, the patient, or the injury. See Roy Penchansky & Carol Macnee, Initiation of Medical Malpractice Suits, 32 MED. CARE 813, 828 (1994).
81.	For further discussion, see DAUER, supra note 79, § 17.04.
82.	See Gerald B. Hickson et al., Factors that Prompted Families to File Medical Malpractice Claims Following Perinatal Injuries, 267 JAMA 1359 (1992); see also SLOAN ET AL., supra note 21, at 64-65 (discussing another study of families of birth-injured infants).
83.	See generally Hickson, supra note 82.
84.	See id. at 1361.
85.	Often, this third person was another member of the medical profession. See id. It is not possible to say with certainty how this variable interacts with the others. Observers who are inclined toward the economic view might see in it an attorney supporting the assessment that bringing a lawsuit will result in greater gains than costs. Others, following the attribution theories of Lloyd-Bostock and other empiricists, would see in the advice of others a psychological support for action, whatever the operative underlying motivation might be. A perceived injury does not often lead to a claim unless, as Lloyd-Bostock has observed, the injured person first "attributes" the injury to a wrongful act by the potential defendant. See Lloyd-Bostock, supra note 26, at 429. Thus for example the revenge motive could be unleashed only after the injured person attributes the blame to someone else. That attribution is made possible by the advice or support of third persons. See generally id.; May & Stengel, supra note 75, at 118-19. We are indebted to Professor Celia Taylor for having first raised this point. Celia Taylor, Remarks at a Colloquium of the Faculty of the University of Denver College of Law (April 1997).
86.	One observer suggests that there is an "implicit contract between the physician and patient for the physician to be truthful and caring and to be present when needed." See SLOAN ET AL., supra note 21, at 217.
87.	The relative size of this factor is not entirely surprising. The medical costs of iatrogenic injuries are covered by most health insurance plans, and the disability costs by most disability insurers, both public and private. Thus, for example, 85% of the U.S. population has medical insurance for at least 80% of the actual cost. See Edward A. Dauer & William Johnson, Study Under the Auspices of the Arizona Health and Hospitals Association (1995) (on file with Edward Dauer).
88.	See supra text accompanying note 62.
89.	See generally ROBERT A. BARUCH BUSH & JOSEPH P. FOLGER, THE PROMISE OF MEDIATION (1994); JAY FOLBERG & ALISON TAYLOR, MEDIATION: A COMPREHENSIVE GUIDE TO RESOLVING CONFLICTS WITHOUT LITIGATION (1984); CHRISTOPHER W. MOORE, THE MEDIATION PROCESS (2d ed. 1996); NEW DIRECTIONS IN MEDIATION (Joseph P. Folger & Tricia S. Jones eds., 1994); LAWRENCE SUSSKIND & JEFFREY L. CRUIKSHANK, BREAKING THE IMPASSE (1987); WILLIAM URY ET AL., GETTING DISPUTES RESOLVED (1988).
90.	For example, the need for a confidential and therefore safe place in which to confront and discuss errors and adverse outcomes. Our emphasis on claimants' motivations is not meant to elide the equally important interests of medical practitioners. Just as the purely economic view of tort claimants does not accurately describe the much richer field within which their actions take place, so too the stereotype of the purely defensive physician is not faithful to physicians' own interests in furthering quality of care. While both corrected portraits help to make the case for mediation, we emphasize the former here only because it is, we suspect, the more surprising and less well known.
91.	The details of the mediation procedure would go beyond the scope of this article. We elaborate them in a forthcoming paper with Dr. Susan M.C. Payne of the Case Western Reserve University School of Medicine. See Dauer, Marcus, & Payne, supra note 57. In that paper, we address such logistical questions as how mediation begins, who attends and what their responsibilities are, who pays, how it works when there are multiple defendants, and how a physician can participate if he or she is willing to do so but the insurer is not (or vice versa). See id.
92.	The claims have all been brought to a Board of Medical Registration, rather than to a court in the first instance. That fact reduces the value of the findings from this program so far as mediation of claims in court is concerned, though it does not make them irrelevant. We are aware of that limitation in these preliminary data.
93.	Concerned that physicians charged with significant substandard care or misconduct would have the cases settled through mediation, the Board limited eligibility to cases involving communication misunderstandings or relatively minor care issues.
94.	It is for this reason that cases of serious injury or abuse are ineligible. In these cases, the Board might be unable, in effect, to delegate its regulatory functions to the complaining party.
95.	The Board itself has no presence at the mediation, an important limitation that we discuss below. For the VMP itself, the limitation was politically necessary at the time. This limitation was not present for a similar mediation project at the College of Physicians and Surgeons in Toronto, Ontario. Leadership of the Massachusetts Board as well as its designers (Marcus and others) developed close contact with the Toronto project. The Toronto agency had a similar mediation project that brought together physicians and patients for face-to-face meetings to resolve complaints. The primary difference between the two projects was the Toronto agency's willingness to include cases of severe misconduct or malpractice. In such cases, a member of the College staff would participate in the mediation to represent the public interest. In other words, they would not agree to settle the complaint until concerns of public safety were resolved. The settlement thus could include disciplinary action against the physician. Though there was great interest in Massachusetts in extending the model to include such cases, there was consensus that the publicity such a change could generate would jeopardize the project. As a result, the Massachusetts VMP involved mediations only between complaining patients and family members and the specific physician named in the complaint.
96.	Since the Board covered the expenses of mediation, there were no direct costs for either the patient or physician.
97.	The case was accepted into the VMP in part as a test of whether mediation is a useful method for resolving disputes brought by psychiatric patients or patients with emotional limitations.
98.	This finding is not new. It does confirm in a real-time setting the empirical investigations of others, some of which have, of necessity, been conducted through retrospective questionnaires.
99.	Further investigation is also warranted by the fact that we do know about the population of claimants in medical malpractice cases. For example, because the Board of Registration in Medicine has no authority to award damages to an injured complainant, it may be that the population of complainants includes people who have already determined that they wish a non-monetary remedy, and that therefore the results of the VMP are attributable to that self-selection. While that effect is very likely operating, it is also notable that litigation is presently the route of choice for a majority of those who seek something other than monetary compensation. As Sloan and colleagues have shown, in a field (perinatal neurological injuries) in which as a practical matter there are available to the injured patients' families both a no-fault administrative system and the conventional tort system, families whose motivation for bringing a claim is "retribution" are ten times as likely to choose the more difficult tort-negligence pathway over the easier and less time-consuming no-fault pathway. See Frank A. Sloan et al., The Road From Medical Injury to Claims Resolution: How No-Fault and Tort Differ, 60 LAW & CONTEMP. PROBS. 35 (Spring 1997).
100.	A conundrum immediately arises, especially for a complainant who has seen a lawyer: How is anyone to know what the claimant's authentic motivations may have been? It is well established in the field of dyadic counseling that injured persons often (though not always) approach professionals (like lawyers) with an uncrystallized cloud of mixed and strong but inchoate feelings. Without explicitly meaning to do so, lawyers tend to filter out those motivations and possibilities that fall outside the scope of the remedies they can offer. Riskin refers to this as the tacit imposition of the "lawyers' philosophical map." See Leonard L. Riskin, Mediation and Lawyers, 43 OHIO ST. L.J. 29, 43-44 (1982). Redmount describes it as "pre-emptive" rather than "empathic" counseling, resulting in the joint adoption of the legal frame of reference. See Robert S. Redmount, Attorney Personalities and Some Psychological Aspect of Legal Consultation, 109 U. PA. L. REV. 972, 982-89 (1961). This tendency is well-stated in a quotation generally attributed to Abraham Maslow: "When all you have is a hammer, everything looks like a nail." The point is, that a claimant may emerge from a lawyer-client consultation with a clearly-formed set of objectives that, most likely, are agreeable to the lawyer and at least not obviously inconsistent with whatever the more authentic yet unformed objectives and motives of the client may have been. Some mediators, recognizing this lawyer filtering, attempt to deal with the parties directly to get at their pre-filtered interests. But how can one know what those were?
101.	The issue of a mediator with an agenda is a tautly debated matter among mediators and mediation theorists. Some believe that a mediator can never have an agenda, though he or she can refuse to participate in an outcome that is unlawful or discernibly harmful to unrepresented third parties. To enter the mediation without advising of that penchant is to breach the implied warranty of neutrality; to disclose it may be deleterious to ever getting started. One proponent of the opposing view is Lawrence Susskind. See generally Lawrence Susskind, Environmental Mediation and the Accountability Problem, 6 VT. L. REV. 1 (1981).
102.	See, e.g., FLA. STAT. ch. 69.081 (1997); TEX. R. CIV. P. 76a.
103.	One of us (Dauer) confesses to having some crabbiness about this subject. At one point, a critic of a proposed medical injury mediation program asked, "How could one know that the settlement accurately reflected what might have happened in court?" Dauer became sullen. Why should the shadow litigation result be the touchstone? If an injured patient might have collected $10,000 but settled instead for an explanation, is that not the injured person's choice? There is nothing about litigation results that commend them as the gold standard. Well, yes, but suppose that the mediator has a public agenda. Could the patient's economic needs ever be sacrificed in a three-way trade?
104.	Dr. Marcus, who had along with Mr. Fleming developed and implemented the VMP.
105.	See OFFICE OF CONSUMER AFF. & BUS. REG. (MASSACHUSETTS), ALTERNATIVE DISPUTE RESOLUTION: A PROPOSAL FOR THE PIONEER INSTITUTE (draft, July 1997) (on file with Leonard Marcus).
106.	Id. at 1-2. It is also expensive. The report contrasts the costs of mediation with the costs of a consent order and hearing through the Division of Administrative Law Appeals. A complaint managed by the Board through consent order requires an average 33.4 hours of Board of Registration in Medicine staff time and costs the agency $5,435. See id. at 28. The process demands 30 hours of a physician's time with a cost of $3,750. See id. at 29. Thus the total cost is $9,185. See id. The cost of resolving a complaint through the Division is considerably higher. With an average 245 hours devoted by the Board, the agency spends $27,554 per case. See id. at 30. The estimated cost to a charged physician is estimated at $15,000. See id. The total Division cost is $42,554. See id. By contrast (based on the experience of the VMP), the average mediated case is projected to require only 19 hours of Board time, or $2,940 of expenses. See id. at 31. Assuming that the physician would assume the professional fee of the mediator, the total cost to the physician would be $1,525. See id. The total cost for mediated cases would be $4,465. See id. at 32.
107.	Id. at 22-23.
108.	Another recommendation is to add the requirement that "all parties be reasonable and of stable mental condition. Mediation is a rational process and requires all parties to behave rationally as well." Id. at 23-24.
109.	See id. at 26 ("Rather than pointing a finger at the guilty party, mediation allows all parties to clear up misunderstandings or to identify the necessary steps to be taken in order to correct an error. At its core, mediation of medical disputes is a preventive measure.").
110.	Id. at 32-33.