While the HIE findings are persuasive, two factors may limit the general applicability of the results. First, the study was conducted in the late 1970s; the comparison plan was the passive bill-payer prevalent in the insurance industry in that period, with no utilization control mechanisms. The more recent adoption by conventional plans of some of the cost-control measures once associated only with HMOs may mean that the difference in efficiency between the two types of plan has narrowed.

Second, the HMO used in the Health Insurance Experiment was a highly structured group-practice plan with many years of operating experience. Much of the growth in the industry in recent years has involved a different type of HMO, the individual practice association (IPA), which contracts with independent phyIsicians who see a mix of HMO enrollees and other kinds of patients. There is evidence that these more loosely structured HMOs have not achieved savings comparable to those observed in the HIE.48 Physicians may not modify their styles of practice in treating HMO enrollees if those enrollees constitute only a small share of their practice. In addition, some people believe that HMOs cannot impose cost-consciousness on practitioners who have not "signed on" to the concept of more efficient and less resource-intensive practice. Because so little is still known about the relative efficacy of different medical practices, external utilization controls may not be able to override individual physicians' judgment in many cases. The greater success of the "closed panel" plan, whose physicians treat HMO enrollees exclusively, has been attributed by some observers to the possibility that these plans attract physicians who are temperamentally more prone to conservative medical practice.

Because closed panel plans maintain their own medical facilities, they require greater start-up funding than IPAs. Federal funds were available to develop such plans in the 1970s, but new plans must now rely on private investment. Investors have favored IPAs, not only because they require less capital, but also because the wider selection of physicians makes them more attractive to consumers. This attraction may, however, be purchased at the price of reduced efficiency.

Finally, while some types of HMOs or similar organizations may be able to reduce costs relative to conventional plans, it is not clear that they have so far reduced growth in health care costs. Data from 1961 through 1981 suggest that HMOs may instead achieve

48 For the most recent findings, see Alan Hillman, Mark Pauly, and Joseph Kerstein, "How Do Financial Incentives Affect Physicians' Clinical Decisions and the Financial Performance of Health Maintenance Organizations?" New England Journal of Medicine 321 (2) (July 13, 1989): 86-92.

a one-time saving, after which costs rise at the same rate as those for other insurance programs. One expla nation that has been offered is that providers in HMOs are as likely as other providers to use new medical technologies.49 More recent data suggest that HMO premium increases have continued to resemble those of conventional insurance plans. The average HMO premium increase during 1988 was 17.2 percent, very close to the 19 percent increase for all employer coverage cited at the beginning of this report. 50

That HMO cost increases have paralleled those of other insurers does not necessarily mean that HMOs have reached the limit of their cost-saving potential. Because competition among health insurers was rela tively limited until recent years, many HMOs may not have faced the market pressures that could induce them to achieve greater savings. The next section reviews proposals to strengthen competition.

Competition and Consumer Choice-The competi tive strategy depends on the willingness of consumers to choose the most cost-effective plans. As was sug gested earlier, the consumers most likely to incur high costs may be least likely to choose the most efficient option. The problem of biased selection might persist even if conventional insurance plans were to disappear and consumers were able to choose only among managed care options. (Some industry analysts believe this will occur in the near future, chiefly because employers will refuse to offer conventional plans.) It is possible that the most costly patients, given a choice among competing managed care plans, would choose the plan that was least restrictive and potentially least able to achieve cost savings. The most efficient plans might continue to enroll the healthiest patients, for whom only limited savings are possible.

Some people believe that biased selection is largely attributable to the fact that consumers are economically sheltered from the cost of their choice of plan, because most of the premium is paid by the employer. Various schemes have been advanced to make the employee more cost-conscious. For example, the em ployer's contribution might be tied to the cost of the least expensive offering, with the employee bearing the full cost of the difference between that plan and other more expensive options.

49

Joseph P. Newhouse, et al., "Are Fee-for-Service Costs Increasing Faster Than HMO Costs?" Medical Care 23 (8) (August 1985): 960-966. 50 InterStudy, The Bottom Line: HMO Premiums and Profitability, 19881989 (Excelsior, Minn.: 1989). Staff and group model HMOs generally had lower increases, possibly confirming their greater efficiency. However, these HMOs also tend to be older than IPAS; age of the HMO was also a determi nant of the rate of increase.

However, selection bias can occur even when the choice of the more expensive plan has real financial consequences for the enrollee. Under the Federal Employees Health Benefits Program (FEHBP), the monthly employee share of premium costs in 1990 ranges from $20.54 in the least expensive high-option HMO to $234.07 in the most costly high-option conventional plan, a difference of $213.53 per month. 51 Under one possible fixed contribution scheme, the Federal share of both plans would be set equal to the full cost of the HMO ($82.16); the employee share would then be zero for the HMO and $265.29 for the conventional plan. If some Federal employees or annuitants are already willing to pay 11 times as much as others in order to obtain the conventional plan, it is not clear that even this change would cause all of them to shift to the HMO. For at least some subset of enrollees, the preference for unrestricted coverage is apparently sufficient to override even strong financial incentives.

One possible solution to the problem of enrollee self-selection is to abandon multiple choices and oblige all members of a covered group to enter a single plan, one selected by the employer or other buyer from among competing plans. Assuming that employers disregarded their own personal plan preferences and chose the least costly option, this approach would theoretically lead to competition among plans on the basis of efficiency. However, both employers and HMOs have been hesitant to enter into arrangements under which enrollees are unwillingly locked into a highly restrictive plan. For this reason, there have evolved arrangements even less restrictive than IPAs, known as open-ended or point-of-service plans.

The predecessor of these plans is the preferred provider organization (PPO). PPOs negotiate discounted rates with certain providers. Enrollees are given a financial incentive, in the form of reduced deductible or coinsurance requirements, to obtain care from providers participating in the PPO network. However, payment will be made under the plan for services furnished by any provider. PPOs thus differ from HMOs, which deny payment altogether for unauthorized non-emergent care provided by providers outside the HMO network. While some PPOs have adopted managed care techniques, such as the use of gatekeepers, most of the savings from a PPO are expected to result from encouraging enrollees to use the participating providers.

The newer, open-ended plans are hybrids, combining some features of HMOs and PPOs. Typically, the

$1 The conventional plan is national, while HMOs are offered only in specific locations. The comparison presented here applies only in one area (Tampa, Fla.) and represents the extreme of variation in the FEHBP system.

plan operates a structured health care system comparable to that of an IPA-model HMO. Enrollees are expected to access the system through a primary care gatekeeper and obtain services from other network providers upon referral by the gatekeeper. Like an HMO, the plan also imposes external utilization controls and negotiates price discounts with providers. As in a PPO, enrollees are free to use non-network providers for covered services, but must pay higher cost-sharing amounts if they choose to do so. Enrollees are also subject to higher cost-sharing if they use specialists within the network without the authorization of the gatekeeper.

Open-ended plans have been adopted by some employers as the single plan available to their workers, replacing systems in which the workers had a choice between conventional and HMO options. Their attraction has been that they overcome the possible selection bias in dual choice systems by enrolling all employees in an HMO-like program. At the same time, they can reduce the employee resistance that would probably greet a proposal for universal HMO enrollment, because they offer employees the safety valve of being able to choose non-plan providers.

Officials of some major insurers that have experimented with open-ended plans in multiple markets report that the plans appear to be reducing the rate of health care cost increases, relative to the increases for their conventional offerings in the same markets.52 Because these plans began operations only very recently, the data required for an objective evaluation are not yet available. Even PPOs, which have existed for a decade, have never been the subject of a controlled study. Some preliminary findings, however, suggest that the safety valve that makes PPOs attractive is potentially a serious weakness, one which may carry over to the newer hybrid plans.

One recent study of a PPO found that enrollees used the PPO's providers for preventive care and minor illnesses, but went outside the network about half the time for specialty care, major surgery, and hospitalization without surgery. 53 One study found a similar pattern among PPO enrollees who were actually employees of one of the providers in the PPO network.54 While these findings are not definitive, they suggest a dilemma that may be common to both PPOS and the newer types of managed care plans. If the price for going out of plan is not punitive, en

52 Personal communication with officials of Prudential and CIGNA. 53 Annemarie Wouters and James Hester, "Patient Choice of Providers in a Preferred Provider Organization," Medical Care 26 (3) (March 1988): 240255. The results may not be fully representative, because the PPO studied was somewhat skewed towards primary care providers.

54 Paula Diehr, et al., “Use of a Preferred Provider by Employees of the Preferred Provider," Health Services Research 23 (4) (October 1988): 537-554.

from .5% of health care dollars in 1960).5 True, $4 billion of these premiums charges were borne by physicians, as compared with $103 billion spent on physicians services. But while malpractice premiums do represent a somewhat higher share of physician services, four percent of gross physician revenues is not an extraordinarily high bill to pay for liability insurance.

These overall figures, however, tend to understate the acuteness of the problem in specific situations. There is a great deal of variation in premiums by specialty and by geographic location. For example, while general practitioners in Indiana or allergists in Arkansas paid less than $2000 in 1985,7 obstetricians or neurosurgeons in New York City were paying (or having paid on their behalf) from $150,000 to $200,000 annually. Even more disruptive is the fact noted earlier that premiums increased very quickly in the mid 1970s and in the mid 1980s. For instance, in many large states, surgical and obstetrical premiums increased three to four hundred percent in the period from 1980 to 1986.9 In addition to these financial expenses of insurance premiums, more and more physicians faced the personal stress of lawsuits brought by their own patients.

Beyond these fairly well documented concerns about insurance premiums and litigation stress are certain other costs of medical malpractice. First, premiums are not the only social costs attributable to malpractice litigation. Physicians and their representatives emphasize that professional liability premiums are smaller than the costs of defensive medicine. The American Medical Association (AMA), for instance, estimates that in 1985, the total cost of defensive medicine was $11.7 billion for physician services alone. 10 Others are less certain of the costs associated with defensive medicine, partly because the concept

"Levit and Freeland, "National Medical Care Expenditures," Health Affairs (Winter 1988): 124. See also Weiler, Medical Malpractice (ALI Background Paper, 1987).

• Premiums are based on claims history of the geographic location and the individual specialty. Since certain specialties are known to lead to many more claims than others, specialty designation has the greatest impact on one's insurance premium. For instance, St. Paul's rates specialties according to eight classes. Family practice, class four, is indexed at 1.0. Physicians who do no surgery, including allergists, dermatologists and psychiatrists, are in class 1A and indexed at .32. On the other hand, neurosurgical physicians are in class eight, and are indexed at 3.48. This means that in a given state, if family practitioners are charged $10,000 for malpractice premiums, psychiatrists are charged $3,200 and neurosurgical physicians are charged at least $35,000. In major metropolitan areas, malpractice premiums are quite a bit higher. For instance, St. Pauls charged class four physicians $43,900 in California in July of 1989. In Los Angeles, however, those physicians are charged $53,600. This compares with the $6,800 St. Pauls charges class four physicians in Arkansas.

"See General Accounting Office, Medical Malpractice: Insurance Costs (Washington, D.C. 1986).

See Florida Academic Task Force, 27.

• See Bovbjerg and Sloan, Medical Malpractice. 10 See AMA, Trends in Health Care.

itself is elusive.11 Defensive medicine is defined as care provided solely to decrease the potential for a lawsuit, in contrast to care that helps diagnose or treat a patient's ailment. What some term defensive practice, others might see as appropriate care elicited by the incentive effect of malpractice litigation. However, adding the direct costs of litigation to those of defensive medicine, some researchers calculated that in 1984 malpractice litigation cost $13.7 billion, 12 (just prior to the last big jump in premiums).

Another social cost is decreased access to care when malpractice litigation drives some physicians out of certain specialties and out of particular geographic locations. This problem is seen most sharply in studies of obstetrical services in low-income areas, where obstetricians, family practitioners or clinics are reducing this part of their practice, they say, due to the economic interaction of Medicaid cost containment and rising malpractice premiums. 13 Indeed, the Institute of Medicine's recent report on the delivery of obstetrical care focused on this particular issue, and emphasized that some relief from the burden of malpractice litigation must be given to individuals providing obstetric care for low-income women.14

The question remains, however, whether all these attendant costs of medical malpractice really do justify large-scale changes in the tort liability approach. To understand the most significant problems with the liability system, it is necessary to examine the apparatus in some detail and assess the strengths and weaknesses of its component institutions. In the following section, we will analyze the roles played by insurers, lawyers and health care providers in the overall increase in medical malpractice liability costs.

INSTITUTIONAL ANALYSIS

Insurers

Some have accused the insurance industry of creating the malpractice crisis in order to bring about increases in premiums. For instance the president of the Association of Trial Lawyers of America has stated, "the insurance industry itself has created the situation and now seeks to profit from it." 15 Consumer advo

11 See Tancredi and Barondess, "The Problem of Defensive Medicine," Science 200 (1978): 879.

12 See Reynolds, Rizzo, and Gonzalez, "The Cost of Medical Professional Liability," JAMA 257 (1987): 2776.

13 See Hughes, Rosenbaum, Smith and Fader, “Obstetrical Care for LowIncome Women: The Effects of Medical Malpractice on Community Health Centers, in Institute of Medicine," Medical Professional Liability in the Deliv ery of Obstetrical Care, vol. 2, 59-78.

14 See Institute of Medicine, Medical Professional Liability in the Delivery of Obstetrical Care, vol. 1, (1989): 1-14.

15 See Perlman, Presidents Page, Trial, January 1986, 5.

cates have made much the same charge. 16 That sentiment was reflected in the comment of J.B. Spence, one of the country's leading personal injury attorneys, who stated: "The insurance industry in this country has a gun to the head of the doctors [and t]he doctors in turn have a gun to the head of the legislatures." 17 This same diagnosis was made about the crisis atmosphere surrounding tort law generally in the mid 1980's, and the "reforms" this atmosphere provoked. 18

These accusations seem to have little merit, at least with regard to medical malpractice insurance. To our mind the largest share of the problem is the volatility of the liability insurance line. Insurance companies must rely on the predictability of the legal risk to be able to set their premiums, 19 but predictability is difficult to achieve in malpractice. To price their policies actuaries must try to estimate the frequency (under an "occurrence" policy) or at least the severity (under a "claims made" policy) of suits that will not finally be resolved for a decade or more. During this "long tail," there can be, and have been, substantial changes in patient propensities to sue and jury sentiments regarding awards. 20 Aggravating this problem is the insurance cycle-i.e., fluctuations in premiums attributable to changes in investment earnings on revenues, interest rates, the stock market, and the value of the dollar. 21

By contrast, the collusion hypothesis seems inherently implausible. 22 Collusion is rarely seen in an industry that has a comparatively low level of concentration in antitrust terms, as has the insurance industry. Moreover, any cartel formed in the insurance industry would be difficult to maintain, because cartel members can too easily attract business by improving

16 See Horwitz, "Nader Charges Insurers with Price Gouging," Washington Post, January 7, 1986, D1, col. 6.

17 See J.B. Spence, Testimony Before Academic Task Force for Review of the Insurance and Tort Systems, Vol. 2 at 136 (Miami, Fla., February 3, 1987), cited in Nye, Gifford, Webb, Dewar, "The Causes of the Medical Malpractice Crisis: An Analysis of Claims Data and Insurance Company Finances," Georgetown Law Journal, 76 (1988): 1495.

18 See Nader, The Assault on Injured Victims' Rights, Denver University Law Review 64 (1988): 625.

19 See generally, Abraham, Distributing Risk (1985).

20 George Priest has explored the problems posed for insurance by tort law generally. See Priest, "The Current Insurance Crisis and Modern Tort Law," Yale Law Journal 96 (1987): 1521 and Priest, "The Antitrust Suit and the Public Understanding of Insurance," Tulane Law Review 63 (1989): 999. Priest's concerns regarding eroding risk pools do not appear to be applicable to malpractice insurers.

21 See Abraham, The Causes of the Insurance Crises in New Directions in Liability Law, ed. W. Olson (New York: Academy of Political Science 1988). 22 See Priest, "The Current Insurance Crisis," and Priest, "The Public Understanding of Insurance," and most of the articles in the Symposium, Perspectives on the Insurance Crisis, Yale Journal of Regulation 5 (1988): 367– 516. The case for collusion is made in National Association of Attorneys General, An Analysis of the Causes of the Current Crisis of Unavailability and Unaffordability of Liability Insurance, May, 1986. A group of state attorneys general did file an antitrust suit against major casualty insurers: see In re Insurance Antitrust Litigation, 723 F. Supp. 464 (N.D. Cal. 1989). The suit was dismissed on McCarran-Ferguson grounds. See also Ayres and Siegelman, "The Economics of the Insurance Antitrust Suit: Towards an Exclusionary Theory," Tulane Law Review 63 (1989): 971.

the terms of coverage at the fixed price. In addition, during the malpractice crisis of the 1970s, and the product and municipality crisis of the 1980s, as big a feature of the problem was a lack of availability of insurance. It seems odd for a cartel to refuse to sell its products at even unduly high prices. Finally, and most importantly for medical malpractice, a substantial portion of this coverage is offered by "bedpan mutuals," insurance companies controlled by the doctors and hospitals themselves. 23 There is no reason to suppose that these carriers would be a part of a scheme to price gouge their own principals. For those several reasons the collusion hypothesis lacks theoretical economic support.

In addition, empirical analysis of long-term insurance premiums and costs supports quite a different di

23 The malpractice insurance industry has changed a great deal in the past decade. Physicians in private practice traditionally purchased insurance directly from malpractice insurers. Those physicians who worked for hospitals or Health Maintenance Organization (HMOs) had their malpractice insurance purchased by their employer. Since the late 1970s, however, more and more hospitals have chosen to self-insure rather than buying commercial insurance. Indeed a survey in 1980 indicated that there were already nearly 1,000 hospitals self-insured. See Needleman and Hackbarth, "The Malpractice Insurance System in Obstetrical Care: Recent Experience and Options for Change." Paper prepared for the Institute of Medicine. Washington, D.C. (1988). This self-insurance phenomenon undoubtedly accelerated during the malpractice crisis of the 1980s as all premiums increased dramatically.

Self-insurance is not the only change that has occurred in malpractice underwriting. As noted, before the insurance crisis of the mid-1970s, malpractice insurance for individual physicians was typically provided by commercial insurers. After this malpractice crisis, there was a great deal of consolidation in the market as many insurers withdrew. Medical societies themselves created their own insurance companies. The largest of these, the Medical Liability Mutual Insurance Company of New York, now possesses nearly 40 percent of the New York state market and nearly six percent of the overall national market. See Stern, "Medical Malpractice, Fidelity Insurance." Bests Review 89 (1988): 34. Hospital associations also developed insurance companies. The largest of these, the Pennsylvania Hospital Insurance Company, possesses 33 percent of the Pennsylvania hospital malpractice market. These so-called bedpan mutuals are, as one would expect, responsive to providers' input. This does not mean that they will be immune from market forces, or that they will behave differently than other insurers when faced with rising claims rates and increased severity of claims. One would expect, however, that they would be unwilling to act in collusion with other insurers in any effort to raise premiums artificially high.

In certain states, state legislatures had to step in to provide alternatives when the majority of medical malpractice insurers withdrew from the market. Most states created joint underwriting associations (JUAs), which are non-profit pooling arrangements funded initially by all entities providing insurance in a particular state. After the initial capital investment, the JUAs are intended to be self-supporting through physician premiums. At present, 13 states have functioning JUAs, with market shares ranging from three percent in Kansas to 83 percent in Rhode Island in 1986. See Kenney, Financial Condition of Medical Malpractice JUAs. (Shalmberg, Ill.: Alliance of American Insurers) (1987). Since insurance companies can face annual assessments if JUA deficits develop, the insurance industry keeps a close eye on the vitality of JUA programs.

The rest of medical malpractice insurance is offered by commercial insurers. Some insurers, such as the St. Paul's group, offers insurance in many

states.

Most malpractice policies are structured in a similar fashion. Medical malpractice insurance policies typically involve both monetary limits on individual claims or occurrences, and aggregate amounts that will be paid over a policy year. In some states, the limits are as low as $100,000 per occurrence and $300,000 in aggregate. More typical, is a minimum amount of $400,000 per occurrence and $1 million in aggregate, while many states have moved to policies involving $1 million per occurrence and $3 million in aggregate. There is very little experience rating in the medical malpractice insurance industry. Hospitals usually have much higher annual limits, ranging between five and thirty million dollars. See Institute of Medicine, supra note 15, Volume 1, Chapter 6, Obstetrical Malpractice Insurance.

agnosis. 24 In both New York and Florida, careful studies have documented the relationship between premiums and costs in the medical malpractice area. For example, the Florida Academic Task Force found that the primary cause in that state of increased malpractice premiums over the years 1980 to 1988 was the substantial increase in loss payments to claimants. The Task Force also demonstrated that insurance company profits are roughly equivalent to the profits made by other corporations in the United States. 25 There appears to be little support, then for the contention that the malpractice crisis was the result of insurance monopoly and excess profits.

Litigation System

The litigation system creates pressure for growth in malpractice premiums when there is either an increase in the number of claims being brought against insured individuals or institutions, or when the average payment on (i.e., the severity of) individual claims rises. Throughout the 1970s and much of the 1980s, both the frequency and severity of claims increased dramatically. 26

Overall, between 1975 and 1984, Danzon estimates that there was an average increase of 10 percent per year in claims brought against physicians, and that during the period from 1982 to 1986, the claim frequency per 100 physicians rose from 13.5 to 17.2 per year. 27 It appears that claims rates are now back

24 The New York figures are contained in the Report of New York State Insurance Department on Medical Malpractice, A Balanced Prescription for Change (New York: 1988).

25 See Nye et al., "The Causes of the Medical Malpractice Crisis," 1499. At the same time, the claims made by insurance companies that the malpractice line of insurance is unprofitable have been refuted by reports from the General Accounting Office (GAO). One study by the GAO concluded that profitability estimates for the medical malpractice line of insurance depends primarily on the adequacy of reserves for future payments of claims and the accounting principle used to estimate the value of the reserves. Using its own recommended accounting principles, the GAO determined that for the years 1975 to 1985 the malpractice line was quite profitable (estimated profitability being $2 billion): See General Accounting Office, Insurance-Profitability of the Medical Malpractice and General Liability Lines (Washington, D.C.: July, 1987). Another GAO study showed that of the property/casualty insurance lines that became insolvent from 1969 to 1978, none had medical malpractice as a primary line: see General Accounting Office, Insurer Failures-Property/ Casualty Insurer Insolvencies and State Guaranty Funds (Washington, D.C.: July, 1987).

26 The following empirical information is drawn from the same sources as were cited in footnote 2. We should note that claims data is notoriously unreliable. As Jacobson remarks "In evaluating these data, some important limitations must be considered. Aside from major gaps in available data, the reported data cannot be compared easily. Each data-reporting system used a different standard for defining the principal allegations of negligence, and each system uses different assumptions and numerical bases for data reporting. Some report data from all claims and expenses (ie, defense costs), including settlements and those in which no indemnity was paid. Others report averages only on a claims-paid basis or by looking at nonzero jury awards alone. Because many of the claims filed result in no indemnity payments, how those are incorporated into the figures makes a major difference in the apparent scope of the problem. One result of these different approaches is that estimates of average verdicts vary widely and statements about averages need to be scrutinized." Jacobson, "Medical Malpractice and the Tort System," JAMA 262 (1989): 3322.

27 Danzon, "Medical Malpractice" in Liability, 103.

down to 13/100 physicians per year. 28 As one might expect, claims vary a great deal by specialty.29 Overall, however, it appears that somewhat more than one in twenty physicians are sued successfully each year, 30

The mean amount paid for each claim is an even more important determinant of the legal cost pressures placed on insurers. There is little doubt that the size of malpractice claims rose a great deal during the 1970s and 1980s. Part of the increase in severity appears due to large jumps in jury verdicts. For example, the average award in jury verdicts jumped from $200,000 in 1974 to almost $650,000 in 1984 for malpractice cases in the California Superior Courts. 31 (To put this in some perspective, we should note that health care costs increased on the order of 60-70% over the same period.) These high jury verdicts induce, in turn, higher settlement amounts. The average malpractice settlement rose from under $12,000 in 1970 to $26,000 in 1975, to $45,000 by 1980, reached $80,000 by 1984 and topped $100,000 by 1986.32 As one might expect there is a great deal of variation from state to state in claims severity.33 For example, the Florida Academic Task Force has emphasized the

28 Physicians and Surgeons Update, 1.

29 The AMA estimates that before 1981, there were 3.2 claims per 100 physicians per year, and that by 1985, the incidence had tripled to 10.1. For obstetrician/gynecologists, the average annual incidence rose from 7.1 per 100 physicians in 1981, to 26.6 by 1985. For all surgeons, the increase was 4.1 in 1981 to 16.5. See Professional Liability Clearinghouse, American Medical Association Center for Health Policy Research, Professional Liability Update (Dec. 1986), cited in "Note, The Applicability of Experience Rating to Medical Malpractice Insurance," Case Western Reserve Law Review 38 (1987): 255. 30 The best estimates are that nearly 50 percent of all claims result in some sort of compensation for the plaintiff.

31 See American Medical Association, Professional Liability in the 80s, Reports 1, 2 and 3. (Chicago: AMA, October 1984, November 1984, March 1985.) While increases in California may have been particularly sharp, they are not unrepresentative. The Rand Institute for Civil Justice has developed information on jury verdicts in Chicago and San Francisco for the past twenty years. Their data indicates that awards for malpractice plaintiffs outstripped awards for other high profile tort actions, such as product liability. See Shanley and Peterson, Comparative Justice: Civil Jury Verdicts in San Francisco and Cook Counties, 1959-1980 (Santa Monica, Calif.: Rand Corporation, 1983), R-3006-ICJ. Other Rand data suggests that the average jury verdict in a malpractice case in these two cities increased from $50,000$100,000 in the early 1960s to $1.2 million in the early 1980s. See Peterson, Civil Juries in the 1980s (Santa Monica, Calif.: Rand Institute for Civil Justice, 1987). In the Miami area, the Florida Academic Task Force reports that jury verdicts on behalf of plaintiffs averaged nearly $900,000 from 1985 through 1987. See Florida Academic Task Force, 147-50.

32 See Weiler, Legal Policy for Medical Injuries: The Issues, the Options and the Evidence (January 1988). Danzon notes that severity increased at double the rate of the Consumer Price Index from 1975 to 1984. See Danzon, "Medical Malpractice" in Liability, 105. Bovbjerg and Sloan have estimated that the paid claim severity rose from approximately $45,000 per claim in 1980 to over $100,000 per claim in 1986. See Bovbjerg and Sloan, Medical Malpractice, 8.

33 In Arkansas, the average indemnity per paid claim was $31,000 in 1980 and $51,000 in 1984. See General Accounting Office, Six State Studies (1986). In Florida, on the other hand, the average indemnity was $80,000 in 1980 and over $140,000 in 1984. Unlike the frequency of claims, it appears that severity of claims has continued to increase. St. Pauls, for instance, estimates that the cost per claim increased 30 percent from 1982 to 1985. These increases have continued, so that the increase from 1982 to 1988 is nearly 80 percent. See St. Paul's Physician Update.