Those numbers are about the same. A 2003 Public Citizen study found 7% of the doctors were responsible for 67% of the number of payments. The report found that those same doctors made 69% of the total amount of the payouts. See Figure 3:

New York’s Dangerous and Undisciplined Doctors

If there are twenty-five doctors, and only five have made malpractice payments, and one made a single payment of $4 million, and four others each made 10 settlements of $100,000, then 4% of the doctors are responsible for 50% of the dollars paid, but it’s clearly not where the problem resides.

And you think the Department of Health should focus their investigations on the 20 that made no payments instead of the five that did? That is certainly an interesting philosophy.

Remember this: The fight doesn’t always go to the strongest or the race to the swiftest. But that’s the way to bet.

You repeat the misrepresentation in the comments here by falsely implying that Public Citizen was counting the number of payments rather than the amount of payments. If there are twenty-five doctors, and only five have made malpractice payments, and one made a single payment of $4 million, and four others each made 10 settlements of $100,000, then 4% of the doctors are responsible for 50% of the dollars paid, but it’s clearly not where the problem resides.

Because I had already linked to Overlawyered for the piece and thought the debate in the comments (the type that would never make it into a daily newspaper op-ed but would be of interest to many readers) was worthy of a second link because it was so in depth.

I’m disappointed that he’s trying to minimize his error by calling this post a “minor correction” when it pointed out that the most persuasive and jarring statistic in his op-ed was essentially made up.

You’re no doubt alone in believing that, since:
1. The main focus of the op-ed was the Dept of Insurance screwing things up during the Pataki years; and
2. The error merely shortchanged my own argument, as I’ve already addressed in Comment 14:

Payments made are likely to be far more important than suits filed. That is, of course, because anyone can start a suit and a suit may have multiple defendants (particularly in a hospital setting with many treaters). But a significant payment really doesn’t track that way. It is targeted to the tortfeasor that the insurance company is worried about based on the merits (in settlement situations) or that the jury and appellate review courts deem to be responsible.

In essence, by saying suits instead of payments, I probably shortchanged my own argument on the 4%/50% issue. The fact that so many payments come from such a small crowd makes it easier for investigatory authorities than trying to scan the number of suits.

Because the same misleading statistic has been used in other states, and Public Citizen chose to mislead readers of its report by calculating it for New York.

Incidentally, Brennan, Studdert, Caplan, Caplan, Merenstein, Entman, and Sloan do add up to seven. (Studdert-Mello, whom anyone in the field should be familiar with given the huge burst of publicity it received when it was released with a misleading press release that is frequently cited by plaintiffs’ lawyers who don’t want people to read the actual study, is at 354 NEJM 2024 (May 11, 2006).)

I agree that readers should check out the studies for themselves, including the discussions of them on this blog and Point of Law. But Wright hasn’t cited anything that supports the idea that the legal system does a good job distinguishing good doctors from bad doctors, which is what the discussion was about.

I have only read the first study you cited, the Brennan et al NEJM article, and the Harvard study on which it is based, upon which you and others primarily rely on this issue. I hope to find time to locate and read the other studies you cited, sometime in the not too distant future, especially the Studdert-Mello study (for which you provided no citation). However, given your assertion that the Brennan et al studies support your claim, and your further citation to the Anderson critique of the Harvard study, I do not expect to find sound support for your claim in any of the other studies you cite.

I admit that the severe underclaiming of medical malpractice injuries that the Harvard study (and every other study of which I am aware) documents does indicate a shortcoming of the tort system, at least in terms of efficient deterrence, and to a lesser extent in terms of corrective justice. While it is important from the standpoint of efficient deterrence that all valid claims be brought and successfully prosecuted, from the standpoint of corrective justice it is important only that claimants be able to file and successfully prosecute valid claims if they want to, but they can waive that right, and studies such as the Rand studies show that — contrary to popular perceptions fed by the anti-tort industry — the great majority of claimants do waive that right, except when they suffer serious injuries not covered by insurance for which they believe others are at fault. Unfortunately, it is clear that many claimants do not file or pursue valid medical malpractice claims today, despite a desire to do so, because of the great difficulty and expense of doing so, which has been greatly exacerbated by “tort reforms.”

However, it is not the underclaiming of medical malpractice sufferers that bothers you or other anti-tort advocates, but rather the alleged overclaiming. Yet, rather than responding to my criticisms of the Brennan et al studies upon which you and others principally rely to support the overclaiming argument, you provide a link to Richard Anderson’s critique of the Harvard study and assert that it shows that the Harvard study overstates the amount of malpractice.

It does not. Anderson rather argues that the Harvard study overstates the number of (admitted) adverse events because, although they could be classified as such only if they resulted in a prolongation of hospital stay or a disability upon leaving the hospital, they were “transient injuries (full recovery within a month) or occurred during the final stages of life in which therapeutic misadventures may subtract only hours or days from the life of a dying patient.” These are real injuries, not overstated adverse events. Prolonged stays in hospitals incur very real and high costs for the patient and/or its insurer. It may be true that there is no need to worry, at least from the standpoint of corrective justice, if the great majority of patients decide not to pursue claims even if these supposedly “minor” adverse events were negligently caused; they nevertheless are real adverse events.

More to the point, Anderson claims that the frequency of medical error or negligence is overstated in the Harvard study. (He wrongly describes the study procedure when he asserts that the study’s determinations of negligence were based on an average of the ratings by the two initial reviewers. I described the actual study procedure, which required a concurrence by the two initial reviewers or by one of the initial reviewers and the tie-breaking senior physician, in my initial post above (#16).) Anderson is upset that the finding of negligence was based on the legal “more likely than not” standard rather than some stricter standard. Yet, as I noted above, a stricter standard than the relevant legal standard actually was imposed, by refusing to classify an incident as negligent, even if one physician rated is as such under the “more likely than not” standard, if neither the other initial reviewer or the senior tie-breaker physician agreed. If even just one physician reviewer was willing to rate it as negligent, that is certainly enough to rate it as at least a plausible claim of negligence, worthy of proceeding to a filed claim and even a trial. The resulting understatement of plausible claims of negligence that I noted, which was exacerbated by the exclusion of “close calls,” is confirmed rather than refuted (as Anderson assumes) by the fact that, while at least one of the initial physician reviewers found negligence was more likely than not in 58 percent of all the adverse events, only 28 percent ultimately met the study criteria for negligence.

In the last two paragraphs of my post I discussed or cited several studies directly on point with respect to the question posed, which found much higher rates of medical error than those found in the Harvard study and which also found that the great majority of malpractice claims studied were valid or at least plausible and legitmately contestable, rather than being, as asserted by tort opponents, “frivolous” (in the usually understood sense of “without any plausibility or merit,” which I believe is its “colloquial use in the political context”).

You do what you claim that I have done. Rather than attempting to refute my criticisms of the findings of the Harvard study upon which you rely, you instead cite Anderson’s flawed critique of the findings of that study that you and other tort opponents do not like. Rather than attempting to refute the findings of the other studies that I discuss or cite, which are direcly on point, you instead claim that they “hav[e] nothing to do with the question presented” and that you have addressed many of them elsewhere; the sole link you provide refers to none of them.

I don’t have the time or inclination to futher pursue this argument (or many of your similar posts on other issues), so you will, as usual, get the last word. I suggest readers check out the various studies for themselves.

I also recommend Richard Anderson’s lengthy critique of the Harvard study, showing that the Harvard study, contrary to Wright’s claim, overstates the amount of malpractice.

This result is confirmed by the Harvard study that is the foundation of the NEJM article by Brennan et al that you cite. However, the Harvard study is routinely cited by anti-tort commenters, such as Ted, not for its confirmation of the severe amount of underclaiming and undercompensation, but rather for its finding of supposedly random liability claims and payments. The study concluded that, while at best only one in eight patients with valid claims, based on the medical records, files a claim and only one in sixteen receive compensation from the tort system, only eight of the 47 claims actually filed (which could be matched with hospitalizations in the study sample) had been determined during the study’s medical record review to be negligently caused adverse outcomes. Harvard Medical Practice Study, Patients, Doctors, and Lawyers: Medical Injury, Malpractice Litigation, and Patient Compensation in New York 7-30 to 7-35, 11-4 to 11-5 (1990).

Apart from the fact that, as the study authors admitted, the sample size was too small to draw any general conclusions about the overall frequency or specific occurrence of invalid claims, the study method was flawed. It was based only on the hospitals’ medical records, and there are obvious reasons for believing – as confirmed by later studies (see below) – that doctors and hospitals often (generally?) will not record adverse events and possible negligence in their medical records. Moreover, the study’s definition of an adverse event employed a (too) high threshold. It was defined as an injury resulting from medical treatment, as opposed to the underlying disease process, that caused a prolonged hospital stay or a disabling injury upon leaving the hospital, which the study admitted “does not encompass all the valid grounds for a malpractice claim.” The study recorded an adverse event as being due to medical negligence only if each of two independent physician reviewers (why only physicians, rather than physician-lawyer pairs?) agreed that such was the case (based only on the hospital’s medical records), or, if the two disagreed, if a third “senior” physician broke the tie in favor of negligent causation. Thus, even though one of the physician reviewers evaluated the case as involving both negligence and causation, it would not be recorded as a negligently caused adverse outcome if the other initial reviewer and the third senior reviewer both found a lack of negligence or a lack of causation. Moreover, “close calls” on the less-likely-than-not (rather than more-likely-than-not) side were recorded as not being negligently caused adverse outcomes. Yet, in the context of medical litigation, there clearly is a plausible, non-frivolous liability claim if even one (independent) physician believes there was negligent causation, especially if he believes this is “more likely than not” to be true, but also even if it is a “close call.” The study also noted that some of the alleged adverse events, especially those involving alleged failure to diagnose properly, may well have not been picked up by its screening methodology. If the claims involving these sorts of issues were all treated as being plausible negligence claims, over half of the filed claims would have been recorded as being valid. See id. at 7-30 to 7-35.

At the time of the Harvard study, few of the filed claims had been closed, so it was not known how many of the claims had actually been pursued, settled, tried, dismissed, or found by a judge or jury to be proven or not proven. The NEJM article provides the followup information. Only one of the claims was resolved by a jury trial, in favor of the defendant (it is not stated how this claim was classified in the initial study). Of 24 closed claims initially classified as involving no adverse event, 10 resulted in a payment to the claimant. Of 13 classified as involving an adverse event but no negligence, 6 resulted in a payment to the claimant. Of 9 initially classified as involving an adverse event and negligence, 5 resulted in a payment to the claimant.

As the followup study notes, the fact that a claim was settled with a payment to the claimant does not establish that it was valid. However, especially for “captive” medical insurance companies operated by physician groups, it is unlikely that claims will be paid absent some minimally plausible argument for negligence, especially given the fact that studies report doctors’ winning with juries 70-90 percent of the time. Erik Moller, Trends in Civil Jury Verdicts Since 1985 at 15-19 (Rand Corporation Report MR-694-ICJ 1996); U.S. Department of Justice, Bureau of Justice Statistics, Special Report: Federal Tort Trials and Verdicts, 1994-95, at 1 (NCJ-165810, Dec. 1997); National Center for State Courts, Examining the Work of State Courts, 1993 at 25; National Center for State Courts, Examining the Work of State Courts, 1994 at 36; U.S. Dep’t of Justice, Bureau of Justice Statistics, Civil Jury Cases and Verdicts in Large Counties 5 (1995); Special Issue: Litigation Dimensions — Torts and Contracts in Large Urban Counties, 19(1) State Ct. J. 24, 32 (1995); Stephen Daniels, Verdicts in Medical Malpractice Cases, Trial, May 1989, at 23; cf. U.S. Department of Justice, Bureau of Justice Statistics, Special Report: Medical Malpractice Insurance Claims in Seven States, 2000-2004 (NCJ-216339, March 2007) (only about one-third of closed insurance claims resulted in a payout to the claimant in three states for which data was available; in the fourth state, Illinois, only 12 percent of closed claims had a payout to the claimant).

Conversely, the fact that a claim was not successfully pursued and legally proven does not imply that it was frivolous rather than a good faith, plausible claim, especially given the major expenses and hurdles imposed on the bringing of such claims by so-called “tort reform.” The authors of the followup study (actually, only one of them) reviewed the actual case files for only twelve of the 46 closed claims. Moreover, the review was once again one-sided: the reviewer looked only at the insurers’ files, not the claimants’ files. Four of these were the cases initially classified as involving negligent adverse events for which no payment was made; they remained so classified after the file review. The other cases were 8 of the 9 cases initially classified as not involving an adverse event or negligence in which a payment of over $25,000 was made; two of these were reclassified as involving a negligent adverse event after the (one-sided) case file review.

As previously noted, studies such as the Harvard study based solely on the medical records will inevitably misclassify many situations as not involving adverse events or negligence. A subsequent study (authorized after the chief surgeon experienced several instances of negligent treatment while a patient in his hospital) was undertaken in which trained observers contemporaneously recorded errors that were discussed by medical personnel during regularly scheduled day-shift weekday clinical meetings in a major urban hospital. Although the observers only attended regularly scheduled weekday (Monday through Friday) day-shift meetings, they found a much higher incidence of medical errors than has generally been assumed or reported: at least one medical error for 45.8% of the 1047 patients studied, and at least one serious adverse outcome (ranging from temporary physical disability to death) caused by medical error for at least 17.7% of the patients. 18.2% of the patients were subject to medical errors the seriousness of which was not discussed. Few of these medical errors were recorded in the medical records. In fact, some physicians admitted they did not include information about errors in the patient’s chart because they wanted to avoid litigation. Only 13 (1.2%) of the patients made liability claims. Eleven of these 13 had been identified by the study as having an adverse event due to medical error. Four years after the statute of limitations for filing suit on the adverse events had ended, only 3 of the 13 claiming patients had received compensation, 8 claims had been dropped, and 2 cases were still pending. Lori D. Andrews, An alternative strategy for studying adverse events in medical care, 349 The Lancet 309, 311-12 (Feb. 1, 1997); Lori D. Andrews, Medical Error and Patient Claiming in a Hospital Setting, American Bar Foundation Working Paper #9316 at 7, 10-11 (1995). The high rate of medical errors is confirmed in Institute of Medicine, To Err is Human (2000).

Studies of actual case files by doctors have found that there are very few instances of frivolous malpractice claims being brought. See, e.g., Mark Taragin et al., The Influence of Standard of Care and Severity of Injury on the Resolution of Medical Malpractice Claims, 117 Annals Int. Med. 780 (1992); Tonn & Associates, Medical and Hospital Professional Liability: A Report Prepared for the Texas Health Policy Task Force (July 1992) (funded by the Texas Hospital Ass’n, the Texas Medical Ass’n, and the Texas Trial Lawyers Ass’n); Neil Vidmar, Medical Malpractice and the American Jury: Confronting the Myths About Jury Incompetence, Deep Pockets, and Outrageous Damage Awards (1995).

Too, Sloan, Whetten-Goldstein, Githens, et al. (1995), found no relationship between birth outcomes and malpractice liability risk — suggesting that that risk can be reduced dramatically without any endangerment of safety in the obstetrical area.

Meanwhile, it’s now two days later, and Justinian Lane still hasn’t corrected his original misstatement.