Paul C. Weiler, Howard H. Hiatt, Joseph P. Newhouse, William G. Johnson, Troyen A. Brennan, and Lucian L. Leape. 175 pages. Cambridge, Massachusetts: Harvard University Press; 1993. $29.95.

Undoubtedly this decade's most important book about medical negligence, A Measure of Malpractice provides a welcome antidote to the mythology and disinformation that has permeated most policy debate on the subject. This terse report, dense in data but not in style, summarizes in surprising detail the monumental work of the Harvard Medical Practice Study—distilling into one slender book the observations and analysis reported in fuller but more fragmentary form in various earlier publications. It should be required reading for every participant in the health care reform effort.

In 1986, the New York State legislature commissioned an interdisciplinary team—physicians, attorneys, economists, statisticians, and social research experts—to diagnose the "crisis" of runaway professional liability insurance premiums. Their mission was to replace the anecdotal evidence so often bandied about with systematic, empirical data that would support informed judgments about the viability of the present tort system and proposed alternatives. Beginning without preconceived preferences, the team quickly realized that, to paraphrase de Tocqueville on democracy, finding fault with the tort system is easy; what is difficult is identifying an alternative that, on balance, will do better.

To this end, the team engaged in a massive study of 30 000 randomly sampled records from 51 acute care nonpsychiatric hospitals, conducted 2500 patient interviews, surveyed 1000 physicians, and reviewed insurance company files for the almost 70 000 claims of medical negligence filed in New York over 14 years. Because sample bias is potentially a serious confounder in such work, I will note just one example that conveys the impressively representative quality of their sampling: Of the patients appropriate to interview about the financial consequences of their in-hospital injuries, the team managed to locate 90%, and of that group, 90% agreed to be interviewed.

A Measure of Malpractice recounts the painstakingly exact methodology, replete with cross-validation and verification techniques, through which the study team developed an epidemiology of medical injury, documented the extent of related patient loss, and examined the role of malpractice litigation in injury compensation and prevention. The report is loaded with so much intriguing information that it is difficult to select even a few items to highlight. Probably the most startling finding is the extraordinarily high incidence of medical injury that can fairly be characterized as due to clinical negligence, and the correspondingly low rate of malpractice claims filed. The study shows that about 1% of all hospitalized patients experience negligent medical injury. In 1 year in New York, more than 13 000 fatalities and more than 7000 severe or permanent disabilities could be attributed to in-hospital medical negligence. By extrapolation, medical negligence kills at least 75 000 Americans annually, eclipsing the carnage of both the American workplace (6000 to 10 000 deaths per year) and the American highway (50 000 deaths per year). Physicians surveyed by the study team underestimated the incidence of medically caused morbidity and mortality by a factor of 10.

Even more counterintuitive for most physicians, I suspect, is the study team's key finding that "while the legal system does in fact operate erratically, it hardly operates excessively ... . (W)e found several times as many seriously disabled patients who received no legal redress for their injury as innocent doctors who bore the burden of defending against unwarranted malpractice claims. Our data make clear, then, that the focus of legislative concern should be that the malpractice system is too inaccessible, rather than too accessible, to the victims of negligent medical treatment." Only a small fraction of patients who suffer disabling injury caused by a health care provider's negligence ever file a malpractice claim at all, let alone receive any payment. For every 7.5 patients who incurred negligent injury, 1 malpractice claim was filed; one claim was paid for every 15 negligent injuries actually inflicted in hospitals. The report states that "the underlying assumption that too many groundless malpractice suits are initiated is unfounded." This is all the more noteworthy because unlike those injured by other kinds of negligent conduct, malpractice claimants rarely obtain any compensation unless a lawsuit is filed; 90% of all money paid to malpractice victims is received after litigation, compared with only about one third of money received by automobile accident claimants.

However, the report heartily endorses the general view that the litigation process consumes far too much money relative to the amount that reaches deserving victims. It also concludes that a solid majority of the malpractice claims that are filed are not valid—"false positives"—albeit the result more of a lack of medical information and understanding on the part of plaintiffs and their lawyers than of meretricious motive. Although the team determined that the legal system ultimately does an efficient job of filtering out these unfounded claims, they decry the economic and emotional cost to health care providers inherent in that process.

Can the virtues of the tort liability system be preserved while its problems are solved? In the end, the study team thinks not. They recommend scrapping the tort system for a different approach to the twin challenges of providing fair compensation for past injury and encouraging prevention of future injury. Their candidate: a "no-fault" scheme, analogous to workers' compensation, that would pay solely out-of-pocket expenses and lost earnings—nothing for pain, fear, loss of enjoyment of life, or loss of function (except limited payments for a few specified impairments). The scheme would cover only patients who suffer longer-term injuries for costs not otherwise reimbursed by insurance. In a version of the "enterprise liability" now supposedly favored by the Presidential task force on health care reform, hospitals would cover the patients of any physician they admit to privileges, even for out-of-hospital adverse events.

As a matter of political feasibility, the researchers suggest gradual implementation of this scheme on an "elective" basis. On admission, hospital patients would be offered the option of the "administrative compensation system" in lieu of their rights under common-law tort liability. (The report does not address the critical question of whether hospitals could limit their patient populations to those who "elect" no-fault.)

A book review does not permit extensive critique of this proposal. It is evident that no one on the study team has direct experience with the workers' compensation system from a claimant's perspective. For its evaluation of the workers' compensation model, the team relied on the work of conservative, not critical, scholars. Their assertions about the positive impact of workers' compensation on danger in the workplace and the "ease" and "efficiency" of establishing the causal relation between job and injury are, sadly, laughable to those who represent injured and disabled workers. As with some of the other premises adopted by the study team—that there should be no compensation for shorter-term disabling injuries caused by medical negligence because "these relatively modest losses can and should be covered by the victim's personal resources," for example, or that state agencies charged with oversight of physician misconduct have been meaningfully "expanded and invigorated"—better reality testing is urgently required. Despite these problematic aspects of the study team's arguments for reform, however, it is hard to see how anyone can quarrel with the quality and utility of the data they have generated. For that, all who want to see a better health care system are in their debt.