Cato on Facebook
Cato on Twitter
Cato on YouTube
Cato RSS1000 Massachusetts Ave, NW
Washington, DC 20001-5403
Phone (202) 842 0200
Fax (202) 842 3490
Contact Us
The Cato Review of Business & Government
Reforming Medical Malpractice
and Insurance
Joseph P. Newhouse and Paul C. Weiler
Joseph P Newhouse is the John D. and Catherine T. MacA rthur
Professor of Health
Policy and
Management at Harvard University and director of the
Division of
Health Policy Research and Education.
Paul C. Weiler is a professor at Harvard Law School.
Total malpractice insurance costs for doctors and hospitals across the country soared from $60 million in 1960 to more than $7 billion by the late 1980s. Although premiums have now leveled off at about 1 percent of the nation's total medical care bill, the sense of a "crisis" in malpractice has been fueled by the erratic path followed by premiums: in the mid-1970s and again in the mid-1980s total premiums doubled in brief three-year time spans.
The financial dislocation from these eye-catching price increases, together with a strong sense of grievance felt by doctors at being personally embroiled in litigation with patients who challenge their professional competence, has put physicians' associations at the forefront of the movement for tort reform. Most of that effort has been directed at state legislatures, and some reform has occurred at the state level. The Bush administration's 1991 budgetary proposals contain a federal response in the form of financial inducements for the state to enact a number of the changes in liability and damage rules favored by doctors, and Sen. Orrin G. Hatch has introduced legislation to prod adoption of variety of malpractice screening devices and alternative dispute-resolution procedures.
The standard physician refrain in favor of tort reform is that most malpractice suits have little to do with actual malfeasance, but juries lack the capacity to filter out unwarranted claims brought by appealing plaintiffs with serious physical disabilities and financial needs that might be allayed by seeking redress from the deep pocket of the malpractice insurer. This creates an incentive for plaintiffs' counsel to file spurious claims in hopes of collecting an offer of settlement form a defendant seeking to avoid the expense of litigation. The result is a legal regime that does little or nothing to improve the safety of medical care, but that induces wasteful practices of defensive medicine that are even more costly than direct liability premiums.
Physicians' arguments are vigorously resisted by trial lawyers and consumer groups, who present a radically different view of the tort system. This alternative perspective argues that there are far more incidents of negligent medical injuries' (that is, actual torts') occurring in the medical care system than ever surface in the legal system. The litigation system does do a reasonably good job of filtering out the valid from the invalid cases, and it pays the vast bulk of settlement or verdict dollars to legitimate malpractice claimants. Even more important, the prospect of being sued and having to pay damages for injuries caused by substandard treatment operates as a powerful mechanism for quality assur ance in medical care. From this vantage point the true malpractice crisis is that there is an insufficient rather than an excessive number of tort claims.
Notwithstanding this counterargument, doctors have enjoyed considerable success in winning statutory revisions to common law tort doctrines and procedures. The empirical literature appraising the efficacy of the various reforms that have been enacted has determined, not surprisingly, that the largest effect comes from a cap on damages. For example, Zuckerman, Bovbjerg, and Sloan show that a cap on damages can reduce insurers' costs by more than 20 percent, depending on the type of cap. On the other hand, reforms focusing on liability rules (for example, informed consent) or procedural access (for example, reduced statutory limitations periods) have only a marginal impact.
To the extent that debates and analyses have focused on how best to minimize malpractice claims and premiums, however, their emphasis is misplaced. The object for public policy in this area should be to devise a medical liability regime that provides efficient insurance for those injuries that do occur and optimal incentives to avoid preventable injuries. The Harvard Medical Practice Study of the incidence and nature of medical injuries and malpractice litigation in New York State, in which we have participated, has been investigating how these latter policy objectives can best be realized. We shall summarize many of our findings in this article.
The Tort System in Theory
The appeal of the tort system is that it combines compensation of victims with incentives to prevent future injury Under certain assumptions the tort system operates as a socially optimal response to the medical injury problem. The question is whether those assumptions are valid.
It is assumed, first, that the physician personally bears the cost of negligent injuries. Therefore, when deciding what tests and resources to use in treating a patient, the physician will invest resources to the point where the marginal value of injuries avoided equals the marginal cost of avoiding the injury. Because the physician is better informed than the patient, it is appropriate to place liability on the physician to make the appropriate investment in injury prevention. It is also assumed that transactions costs are negligible, so that (assuming riskaverse patients) full compensation for past injuries is optimal insurance.
| The object for public policy in reforming medical malpractice and insurance should be to devise a medical liability regime that provides efficient insurance for those injuries that do occur and optimal incentives to avoid preventable injuries. |
In the real world of medical care none of these assumptions is satisfied. In the first place virtually all physicians have malpractice insurance that pays the cost of tort damages. Because experience rating of malpractice insurance is minimal, the cost of any negligent act is diffused over all premium payers, generally all physicians in a given specialty and area. Thus, the physician actually internalizes very little of the cost of injury to the patient. But the signal to the physician is distorted even further. Evidence indicates that only about 10 percent of cases involving negligence result in claims. Even in the cases involving more serious injuries-death or disability lasting more than six months among patients under seventy years of age-only about a third of the cases result in claims. Because the great bulk of injuries remain outside the scope of the tort system, the system falls far short of forcing doctors to fully internalize negligent injury costs.
In addition, the system obviously has high transactions costs. About 55 percent of malpractice premium dollars represent administrative costs, largely attorneys' fees and expert witnesses' fees. If the tort system had a sufficiently large preventive effect on medical injury and if there were no other way to obtain the effect, such costs might be justified. Viewed purely as a compensation mechanism, however, the tort system is much more expensive than other insurance programs. These high transactions costs have numerous consequences.
First, most small claims are barred from redress. Because any actual compensation system will have some administrative costs, sufficiently small claims will always be barred, but the threshold in the tort system is obviously fur higher than one would want for this particular purpose.
Second, the expense of the tort system means that its norm of full compensation is not necessarily appropriate. Even if all losses were monetary, full compensation is not optimal in a system with administrative costs.
Third, the nature of both medical injuries and medical records makes it difficult for the patient to determine from his own resources whether a claim is valid. The malpractice system must thus rely heavily on discovery following filing of a claim to determine whether there is evidence of actual negligence. Malpractice claims also generally name all parties involved in a patient's care (for example, the internist, surgeon, anesthetist, and hospital) in attempting to detect who the negligent party is. It is only through this discovery process that a plaintiff's attorney can judge whether it is worth investing resources to pursue the case further. The need for such a process implies, however, that there will be dropped rather quickly. Although dropped quickly, these cases inevitably impose some costs of defense, add to the administrative cost of the tort system, and lead to a perception of unfairness among doctors.
Furthermore, these costs of defense are not insured. They include costs of retrieving records and time away from a practice to give depositions and otherwise defend the claim, not to mention the stress associated with being a defendant. These uninsurable costs give physicians an incentive to minimize the likelihood of a claim. From this source springs an incentive for defensive medicine. Patricia Danzon has noted that health insurance already promotes the overutilization of preventive measures. Thus, some of what is labeled as defensive medicine would occur even without the tort system. Nonetheless, induced defensive medicine is a cost not considered in the simple model.
Modifying the Tort System
Over the past several years a number of state legislatures have introduced several modifications of the tort system. Some of these changes represent improvements, but none of them goes to the root of the problem. A fundamentally different system for dealing with medical injuries is needed.
The variety of measures that seek to make malpractice claims harder to lodge and to collect on-compulsory certificates of merit, screening panels, tighter statutes of limitation or repose, and elimination of joint and several liability-seem particularly misguided. The premise of these reforms is that doctors suffer from an excessive incidence of invalid claims (false positives), but research indicates that there is a far higher incidence of potentially valid claims that are not filed (false negatives). Erecting steep and expensive hurdles to pursuing malpractice claims in an attempt to address the smaller problem seems counterproductive when the evident effect is to increase the much bigger problem.
| In addition to advocating restraints on tort awards, we would encourage the adoption of a new damage category that would compensate the negligently injured plaintiff for reasonable attorney fees. |
A simple principle should govern efforts at tort reform: the primary focus of tort damages should be to reimburse the actual financial losses of the injured plaintiff. Adherence to this principle lends support to several reform proposals-for example, a collateral source offset rule (with an accompanying bar to insure subrogation or lien claims) to make better use of cheaper first-party loss insurance, a scale for pain and suffering damages running from a floor to a ceiling containing standardized injury profiles and specified damage amounts that will govern the parties' and the juries' appraisal of particular claims, tight constraints on punitive damages (an occasional but troubling phenomenon in recent malpractice litigation), and structured periodic awards that pay at least the larger damage amounts over the actual life span of the individual victim (a lifetime that may be longer instead of shorter than the actuarial projection). In addition to this typical panoply of restraints on tort awards, we would also encourage the adoption of a new damage category that would compensate another form of actual financial loss from the injury-the reasonable attorney fees the successful (the negligently injured) plaintiff must expend to secure the various forms of tort redress just mentioned. Again, this proposal for one-way fee-shifting reflects our position that the aim of a legitimate tort reform effort is not to cut back on claims, awards, and liability premiums as such, but rather to ensure that these claims, awards, and premiums are used for sensible compensation and effective prevention.
A Different Method for Addressing Medical Injury
Although there are decided improvements that can be made in the current malpractice system, there is also a powerful argument for a thoroughly recast system of tort damages to make this a more sensibly designed mode of accident insurance. We would make three key changes in the current system.
First, the focus of legal liability should be shifted from the individual physician to the hospital or health care organization connected with the incident. Those practitioners not affiliated with any hospital or organized delivery system would remain, at least initially, subject to the present tort system.
| We advocate shifting legal liability from the physician to the hospital or health care organization connected with the incident, expanding liability to include all medically caused injuries rather than only those caused by negligence, and compensating nonpecuniary damages in accordance with a comparatively modest schedule of benefits paid for permanently disabling injuries. |
Second, liability should be expanded to include all medically caused injuries rather than only those caused by negligence. Only damages incurred after the initial six months following the injury, however, would be eligible for compensation; that is, there would be a six-month deductible. Furthermore, the system would be the second payer after the patient's medical and disability insurance. Thus, there would be the equivalent of a collateral source offset.
Third, nonpecuniary damages should be compensated in accordance with a comparatively modest schedule of benefits paid for permanently disabling injuries.
We believe that such a system would be an improvement over the current system with respect to administrative cost, injury compensation, and incentives for prevention.
Cost. Above all, such a system would be cheaper to operate. Workers' compensation, an analogous system, spends about 20 percent of its costs on administration. The medical malpractice system we have proposed would not be so inexpensive to operate as workers' compensation, of course. For many work-related injuries the worker can be presumed to have been healthy, and therefore the question of causation (did the injury occur on the job?) is usually easy to resolve. In the case of medical injuries causation would be more difficult to determine because the patient is already sick. It would need to be determined whether the observed disability is attributable to some deficiency in the medical care rendered or to the underlying disease. In the latter case the disability would not be compensable.
In our New York study we asked physicians reviewing medical records to evaluate both causation (did medical care as opposed to the underlying disease cause the injury?) and negligence (if medical intervention caused the injury, did the care delivered fall below customary professional standards?). We further asked the physicians to rate their confidence in these evaluations. Confidence levels ranged from virtually certain to highly uncertain about the judgment, but we found much more uncertainty about the negligence call than about the causation call. Because uncertainty discourages settlement, a system that must make a judgment about negligence (as well as causation) will incur higher administrative expense (as a percentage of all expenses) than a system that must make a judgment only about causation.
Moreover, evolving standards of medical practice and efforts to introduce greater efficiency by not performing procedures with a small positive benefit (for example, managed care) both lead to litigation under the current system because one must determine whether practice fell short of the customary standard. Our system would largely avoid such problems. Our system should also produce fewer nonmeritonous claims because claims are directed at a single organization rather than at multiple defendants, only one of whom may have been negligent. There should be associated reductions in administrative costs and gains in perceived fairness. Furthermore, by covering all injuries and by placing liability at the level of the organization rather than the physician, the incentives for wasteful defensive medicine should be reduced. That is, the uninsurable costs to the physician of a claim should be less, in part because the system does not have to determine negligence and in part because the physician is not the defendant.
Partially because of the reduction in administrative expense, we estimate that a system such as we propose could be put in place in New York State for approximately what is now being spent on the malpractice system. It would, moreover, offer significant gains in both compensation and prevention.
Compensation. The principal gain in compensation would be that nonnegligently caused injuries could also receive some redress. According to our study in New York, such injuries comprise 72 percent of all medically caused injuries. Although they tend to be less severe on average than negligently caused injuries, a substantial number of the nonnegligent injuries have severe consequences. For example, half of the medically caused deaths and two-thirds of the cases of permanent impairment came from nonnegliently caused injuries.
| Not all patients would gain under the proposed reforms because to minimize administrative costs, our system would be a secondary payer to other insurance. In addition, we would include a six-month deductible. We would also reduce the awards some plaintiffs would receive for pain and suffering from what the current tort system might award. |
An example might help clarify the point. Consider a patient whose time away from work is prolonged by a severe reaction to penicillin. If the patient was known to be allergic to penicillin but the physician forgot to ascertain that and wrote orders for the patient to be given penicillin, or the nurse administered penicillin when it was not ordered, the patient was negligently injured and is entitled to full compensation. If the patient had never before been given an antibiotic and was therefore not known to be allergic, the patient is not entitled to any compensation. From the point of view of the patient's demand for insurance, however, the injury is equivalent, so it makes little sense to have full insurance in one case and none in the other.
Some might argue that the absence of a private market for insurance against non-negligently caused injuries suggests that patients may not be willing to pay to protect themselves against such risk. The lack of such insurance could stem from other causes, such as potential adverse selection, but even if the absence of such insurance is attributable to an unwillingness to pay, the argument proves too much. Because the loading for the compulsory insurance provided by the tort system is much higher than the loading for any first-party insurance, the argument implies that patients are also unwilling to pay for the compulsory insurance provided by the tort system.
Not all patients would gain under the proposed reforms. We have included three provisions that would leave some patients with less compensation than the present system.
First, to minimize administrative costs, this system would be a secondary payer to other insurance. To the extent possible, compensation should flow through existing first-party insurance (disability insurance and medical insurance) because the loadings on such insurance are much less (5 to 10 percent for group medical insurance). In states that do not currently have a collateral source offset rule, recoveries under the existing tort system are greater than they would be under our system.
Second, we would include a six-month deductible. Although transactions costs under the present system mean that small claims are effectively barred, it is possible that there are some large damages during the first few months that result in claims in the present system that would be prohibited in our system. We could remedy this by expressing the deductible in dollar terms rather than in time terms, but there is a substantial administrative gain from having the deductible in time terms. It becomes much easier to make the causation judgment once an appreciable period has elapsed from the original illness and its treatment. Making this deductible period somewhat shorter-for example, four months-would not, however, substantially increase financial or administrative costs.
Third, the schedule we envision for pain and suffering would reduce the awards some plaintiffs would receive for pain and suffering from what they might have received under the present system. By making awards less variable, however, a schedule should promote settlement and lower transactions costs as a percentage of awards. Also, as mentioned above, we would permit recovery of attorneys' fees in meritorious cases, whereas under the present system awards for pain and suffering may be a device for compensating the plaintiff for attorneys' fees.
In addition to providing redress for severe, non-negligent injuries, our system would improve on the present system for negligently caused injuries because compensation would be paid more promptly. Under the present "long tail" of malpractice disposition, payments are usually made several years after the accident. Delay may pose a hardship if the injured individual has little liquidity and may even mean a reduction in compensation if the ultimate award does not allow for prejudgment interest.
Prevention. Few would dispute that a no-fault plan provides more sensible and administratively economical compensation for past injuries than malpractice litigation. The major objection to replacing malpractice litigation is that it would eliminate an essential incentive to prevent future injuries. That concern is especially legitimate when one views the actual design of the broad medical injury compensation schemes in New Zealand and Sweden and the special programs for brain-damaged infants in Virginia and Florida. But unlike these programs, which basically function as targeted forms of social insurance, the program we envisage is a form of no-fault liability imposed on the health care providers who have an explicit connection with the iatrogenic injury. On the basis of research by Moore and Viscusi into the actual impact of the analogous model of workers' compensation, we believe that this liability model would enhance rather than dilute the legal incentives to reduce medical injuries.
First, virtually all serious injuries-those resulting in death or in disability lasting longer than six months-would be compensated under our proposal. Under the existing system claims are brought by only a third of patients under age seventy who are seriously injured through negligence (and even fewer by those over seventy). The signal to the medical community to prevent injury would therefore be strengthened.
Second, under the present system damage awards are generally paid for by all physicians in an area in a given specialty. Under our system the burden of liability would be borne by a smaller group, for example, the medical staff of a hospital. There is far less of a free-rider problem.
Third, the group responsible for paying the damages would have an internal organization that makes it possible to undertake actions to reduce the frequency of injury Under the present system, if a particular physician is frequently negligent, it is difficult for other physicians (who pay for those awards) to do anything to alter his behavior. Rather, the logic of the tort system is that the signal from the tort system to that one physician is sufficient to induce him to change his pattern of behavior. By contrast, the medical staff of a hospital or a health delivery system (for example, a health maintenance organization) is the logical group to investigate a physician's behavior and to determine whether the physician's staff privileges should be curtailed (for example, by being barred from undertaking certain operations) or denied altogether. Moreover, the medical staff would be free to allocate the liability it shared across its members. Thus, there could be a finer degree of discrimination in setting premiums than is observed under the present system (for example, adjusting for the frequency of operations performed).
| The program we envisage is a form of no-fault liability imposed on the health care providers who have an explicit connection with the iatrogenic injury. This liability model would enhance rather than dilute the legal incentives to reduce medical injuries. |
Fourth, a large class of accidents may not be attributable to the physician, or the means to prevent them (or ameliorate their consequences) may be outside the control of an individual physician. For example, it was discovered that approximately one-third of the injuries attributable to anesthesia occurred when the anesthesiologist's attention had been diverted from the monitor indicating that the anesthetized patient required more oxygen. To address this problem, a mechanism was devised so that a bell would ring when oxygen flow was impaired. This type of preventive measure is difficult for a single physician who may be adjudged negligent in one case to introduce. Such remedial action benefits all anesthesiologists at a particular hospital, and placing liability at the level of the organization would provide a stronger incentive to address this class of problem.
Finally our no-fault liability system would offer a legal and financial incentive to prevent or ameliorate the consequences of non-negligent injuries, that is, those injuries that are not avoidable by currently feasible techniques. Although there are market and professional rewards for the development of such injury prevention techniques, malpractice law offers no specifically legal reason for making such investments. A system that imposes liability irrespective of fault closes that major gap in responsibility.
Toward Implementation
We propose introducing our system on a voluntary basis at the level of a hospital or an organized delivery system. At the time of admission all patients at a hospital participating in the system would be informed about and would agree to be compensated for medical injury through our no-fault plan. They would give up their right to pursue a remedy in tort. If a patient preferred to retain his tort rights, he could simply use another hospital.
It is possible to introduce our changes at either a narrower or a broader level than the hospital (or health delivery system). One could, for example, offer our compensation program to individual patients. When the patient entered the hospital, he would either agree to be subject to that system or not. We do not recommend this limited focus at the patient level because of the substantial comparative advantages in prevention from acting at the institutional level.
One could also introduce our method geographically. A state, for example, could introduce such a compensation program. Although we would ultimately prefer to see this system adopted universally and therefore do not object to introducing it in an entire area, there are enough uncertainties and details to be worked out that it is better to begin on a smaller scale. (A small scale could be all providers in a metropolitan area of a million or so, however, or even all providers in a relatively small state.)
One considerable uncertainty about our proposal relates to a possible adverse selection problem in our method. Patients are at varying risk of an injury The New York study showed, for example, that the elderly were more likely to suffer an injury than the nonelderly. Risk that varied with measurable characteristics such as age could be adjusted for in setting premiums. Risks might, however, vary with patients' characteristics that would be observable by the hospital (or by the physician) but not by the risk adjuster. In that case hospitals would have an incentive to discriminate against ("dump") high-risk patients since they would bear the costs of any injuries that befell such patients. How well adjustments for varying risk can be made is an open question.
Introducing our proposed reform on a small scale could help determine the magnitude of administrative costs (particularly how costly it would be to determine causation), the ability to make adjustments for varying risk of injury (particularly whether selection behavior would be observed), and the degree to which an institution should be experience rated. Work by other researchers suggests that in the case of torts most hospitals are a large enough unit to be almost completely experience rate. This finding should be even more applicable to our system where cases are more numerous and less variable. Nonetheless, small hospitals may have a nontrivial community rating component.
Finally, although our estimates with New York data suggest that the cost of our proposed system would not be dissimilar to the cost of the present tort system, the present system has a potential for serious cost escalation that our proposed system does not have. One cause of the two previous malpractice "crises" was an increase in the number of claims being brought. Given that two-thirds of the relatively serious negligent injuries are still not generating claims under the present system, it is clear that there could be another steep jump in malpractice premiums at any time. The present lull in malpractice liability offers a favorable opportunity for experimentation with alternative methods to address medical injuries.
|
Selected Readings Danzon, PM. Medical Malpractice. Cambridge: Harvard University
Press, 1985. |
 |
 |  |
||
|
||||
|
|
Subscribe to Regulation |
|
|
 |
|
 |
||
Feeds for Cato events: 