What we know about malpractice settlements

Iowa Law Review

Volumn 92, Issue 5, 2007, Pages 1783-1833

  Peters Jr , Philip G ^a  

^a UNIVERSITY OF MISSOURI   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 34948891685     PISSN: 00210552     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (292)

1. TOM BAKER, THE MEDICAL MALPRACTICE MYTH 12-13 (2005) (quoting President George W. Bush, Speech in Collinsville, Ill. (Jan. 2005)).
2. Troyen A. Brennan et al., Relation Between Negligent Adverse Events and the Outcomes of Medical-Malpractice Litigation, 335 NEW ENG. J. MED. 1963, 1963 (1996).
3. See generally HARVARD MED. PRACTICE STUDY, PATIENTS, DOCTORS, AND LAWYERS: MEDICAL INJURY, MALPRACTICE LITIGATION, AND PATIENT COMPENSATION IN NEW YORK (1990).
4. Brennan et al., supra note 2, at 1963, 1966-67.
5. Id. at 1967.
6. See, e.g., RICHARD A. EPSTEIN, MORTAL PERIL: OUR INALIENABLE RIGHT TO HEALTH CARE? 388 (1997) (stating, based on this study, that "[t]he system seems broken from both ends");
7. CAROLE R. GRESENZ ET AL., RAND CORPORATION, A FLOOD OF LITIGATION? PREDICTING THE CONSEQUENCES OF CHANGING LEGAL REMEDIES AVAILABLE TO ERISA BENEFICIARIES 5 (1999), available at http://www.rand.org/pubs/issue_papers/2006/IP184.PDF (citing the study for proof that "[s]everity of injury plays a strong role in determining the liability, however, and may even overshadow the effect of the presence or absence of negligence in a significant number of cases"); Common Good, Selected Malpractice Claim Data, http://cgood.org/ healthcare-reading-cgpubs- factsheets-14.html (last visited Apr. 1, 2007) (citing the study for the proposition that outcomes correlate with injuries, not with quality of care).
8. See infra Part I.
9. See infra Part I.A, p. 114 tbl. 1.
10. See supra note 5 (citing scholars' reactions to the Harvard study).
11. BAKER, supra note 1, at 1963;
12. Tom Baker, Reconsidering the Harvard Medical Practice Study Conclusions About the Validity of Medical Malpractice Claims, 33 J. L. MED. & ETHICS 501, 502, 509-11 (2005).
13. See infra Part II.
14. See infra Part III.
15. See infra Part III.C.
16. For this synthesis, the studies could have been organized in several different ways. First, the studies could have been classified according to the method used to determine whether or not the patient's malpractice claim had merit. Some studies have relied on the opinions provided to liability insurers by physicians retained specifically to evaluate the strength of the plaintiffs claim, while one study relied on the ratings provided by the hospital's risk managers. The best studies attempted to minimize bias by asking independent physicians to rate the quality of care given to the claimant. Second, the studies could be distinguished by the completeness of the information made available to the reviewers. Ratings that are given soon after the claim is made, especially those based entirely on the patient's medical records, are more likely to be inconsistent with the eventual resolution of the case than those rendered closer to the time of settlement. Although these different rating methods could be important in some circumstances, the studies using these methods have yielded surprisingly similar findings about the correlation between the medical negligence and settlement outcome. See infra p. 120 tbl.1, p. 121 tbl.2, p. 124 fig.4. As a result, I have not chosen to subdivide the studies on these bases.
17. See infra Part I.B.
18. Mark I. Taragin et al., The Influence of Standard of Care and Severity of Injury on the Resolution of Medical Malpractice Claims, 117 ANNALS INTERNAL MED. 780, 780 (1992). About 80% of these claims ultimately resulted in a lawsuit. See id. at 783 tbl.2 (showing that a total of 1,654 cases were resolved before suit was filed). Physician care was considered defensible in 62% of the claims, indefensible in 25% of the claims, and unclear in 13% of the claims. Id. at 781, 781 tbl.1.
19. See id. at 780-81.
20. At the outset, the defendant physician was contacted. Id. at 780.
21. The claim was deemed "indefensible" if the physician admitted error. Id. If the physician did not admit error, then a claims representative reviewed the claim. Id. If the claims representative concluded that the claim was "clearly defensible," then no further review was performed. Id. If he concluded otherwise, then external review was performed. Id. at 780-81.
22. Id. at 781.
23. Id.
24. Id. (finding a significance level of P<0.001).
25. Taragin et al., supra note 15, at 781. In addition, severity of injury had a small, but statistically significant, association with the likelihood of payment. Id. (revealing frequency of payment at 39% for low severity, 43% for medium severity, and 47% for high severity).
26. Id. at 781 (finding a significance level of P=0.16).
27. Id. at 780.
28. Id. at 780-82.
29. Id. at 782. In addition, roughly half of the cases in which the medical care was deemed defensible were reviewed only by a claims representative, while all initial determinations of poor quality were reviewed by other doctors. Id. at 781 tbl.1 (showing that 29% of the cases were deemed defensible with "no peer review"). This asymmetry in the evaluation procedure could bias the ratings in favor of defendants.
30. Henry S. Farber & Michelle J. White, Medical Malpractice: An Empirical Examination of the Litigation Process, 22 RAND J. ECON. 199, 203 (1991).
31. Id. at 204.
32. Id.
33. Id. at 204-05.
34. Id.
35. Farber & White, supra note 25, at 205 (finding a significance level of P<0.0001). Of the 252 cases, ninety-two were dropped or dismissed (36.5%), 147 cases were settled out of court (58.3%), and thirteen went to a verdict (5.2%). Id.
36. Id. at 204 tbl.1, 205.
37. Id. at 204 tbl.1.
38. Id.
39. Id. at 206 (stating that 40% of the variance was explained by severity and 8% by fault).
40. Farber & White, supra note 25, at 215 tbl.7.
41. Id. at 214, 215 tbl.7. The mean log settlement, using good care as the base, was 0.93 (.466) for cases with ambiguous care and 1.54 (.448) in cases with bad care. Id.at 206, 207 tbl.3.
42. Id. at 214.
43. Henry S. Farber & Michelle J. White, A Comparison of Formal and Informal Dispute Resolution in Medical Malpractice, 23 J. LEGAL STUD. 777, 777-79 (1994). Farber and White looked at the files of 355 complaints made to a single large hospital between 1976 and 1989 concerning the hospital or its providers (half of which were resolved without a lawsuit) and the files of 242 additional disputes initiated by the filing of a lawsuit. Id. at 786. The researchers had available to them the files of the hospital's patient-relations office and its legal-affairs office, including the opinions of the experts the hospital asked to assess the quality of medical care. Id. at 786-87. When informal complaints were received, the hospital got an evaluation from a supervisor or provider in the same specialty. Id. When lawsuits were filed, the hospital also retained outside experts. Id. at 787.
44. Id. at 798, 799 tbl.8. The coefficient, using good care as the constant, for ambiguous care was 1.69 (.245), and for bad care, the coefficient was 2.75 (.245). Id. at 799 tbl.8.
45. Id. at 799. The coefficient estimate of the log real settlement amount, using good care as the constant, was .579 (.273) for ambiguous care and 1.34 (.247) for bad care. Id. at 801 tbl.10.
46. Farber & White, supra note 25, at 215 tbl.7. In cases with similar severity, settlements with bad care were nearly four times larger than settlements in cases with good care. Farber & White, supra note 38, at 799. Likewise, settlements in cases with ambiguous care were nearly twice as large as settlements in cases with good care. Id. at 800 (showing that settlements in ambiguous-care cases were almost 80% larger than good-care cases). In a separate paper, White reexamined this data and found that, among settled cases, the average recovery in a case rated as negligent was five times higher than one in which the care was rated as good. Michelle J. White, The Value of Liability in Medical Malpractice, HEALTH AFF., Fall 1994, at 75, 80 exhibit 2 (finding the difference to be $205,000 in ambiguous-care cases as compared to $41,800 in goodcare cases).
47. This could cause poor-care cases to be misclassified as having good care; then settlement of these cases would show up as the settlement of a case involving good care.
48. Paul L. Ogburn, Jr. et al., Perinatal Medical Negligence Closed Claims from the St. Paul Company, 1980-1982, 33 J. REPROD. MED. 608, 609-10 (1988). The medical care was rated as "malpractice," "not malpractice," "or unsure." Id at 610 fig.1.
49. Id. at 608.
50. Id. at 609. The cases were reviewed a second time by a physician on the research team "to confirm prior judgments about medical negligence." Id.
51. Id. (finding a significance level of P<0.001).
52. Id. at 609-10.
53. Ogburn et al., supra note 43, at 609. The researchers also stated that "[t]he majority of cases judged to be medical negligence were not difficult to evaluate because they included gross physician neglect and/or mismanagement." Id at 610. This introduces the possibility that the physician reviewers were using an unduly demanding threshold for a determination of negligence. However, the reviewers had access to the insurance file that contained the most complete picture of case quality that is realistically available.
54. Frank A. Sloan & Chee R. Hsieh, Variability in Medical Malpractice Payments: Is The Compensation Fair?, 24 LAW & SOC'Y REV. 997, 1003-04 (1990). Florida requires insurance companies to report all closed claims to the Florida Department of Insurance, and the reports are publicly available. Id. at 1003.
55. Id. at 1010, 1014-17, 1018 tbl.3 (showing a regression coefficient of 0.69, with a significance level of P=0.01, in a two-tailed t-test).
56. Id. at 1014. These findings led the authors to conclude that "there is a relationship between the probability of payment and the degree of liability," Id.
57. Id. at 1003.
58. In addition, Sloan & Hsieh aggregated all claims made against separate doctors arising out of a single incident into a single case and used the information from the defendant, if any, who paid the highest indemnity. Id. at 1003. This methodology is likely to understate the number of unfounded claims and to overstate the average settlement paid.
59. FRANK A. SLOAN ET AL., SUING FOR MEDICAL MALPRACTICE 17-20 (1993). The study included some claims which never resulted in a lawsuit. Id. at 25 tbl.2.2. The authors rated a claim "liable" only if a majority of the four reviewers concluded that at least one of the defendants had caused the claimant's injuries through negligent treatment. Id. at 99-100. They rated a claim as "not liable" if a majority felt that none of the defendants had negligently caused the injuries. Id. The authors deemed all other cases "uncertain." Id. at 100, 107. The uncertain cases involved either a disagreement among the reviewers or an inability to give a rating due to the limited availability or quality of the medical records. Id. at 107-08. Each reviewer read and scored the claims individually. Id. at 98. The physicians evaluated them in three rounds. Id. at 98-99. The first round used only the information submitted to the state on the Florida closed claim form. Id. The form included the date of occurrence, age and gender of the patient, description of alleged actions that caused the claim to be filed, the nature of the procedures performed, and the principal injury giving rise to the claim. Id. In round two, reviewers were also sent information from interviews with the claimants, along with selected medical information taken from the medical records. Id. at 99. They were then asked whether they wished to change their ratings. Id. In round three, when the hospital charts could be supplied by the hospital (roughly one-third of the cases), the reviewers received abstracts of the hospital charts prepared by the research team. Id.
60. Id. at 6-8 (indicating the second set was limited to adults between the ages of twenty-five and fifty-four).
61. Id. at 98.
62. Id.
63. Id.
64. SLOAN ET AL., supra note 54, at 200, 206. The plaintiffs eventually dropped most of the no-liability cases. Id. In the cases that received a settlement payment, regardless of rating, compensation was less than economic losses in nearly 80% of the cases. Id. at 196-97, 197 tbl.9.4. Only one factor suggests any bias in the findings. Surprisingly, in eight of the twenty-eight cases given a "no liability" rating, the family had been told by a physician or nurse that the patient's injuries were caused by negligent medical care. Id. at 105. In one case, the defendant physician had encouraged the family to file a claim against him, stating that "in cases like this, we have insurance to cover it." Id. at 105, 107, 107 box 6.7. The classification of these claims as defensible may suggest reviewer bias.
65. Id. at 206.
66. Stephen J. Spurr & Sandra Howze, The Effect of Care Quality on Medical Malpractice Litigation, 41 Q. REV. ECON. & FIN. 491, 495 (2001). These claims represented all of the medical negligence claims brought against the hospital, its staff, and affiliated physicians that were closed between 1987 and 1995. Id. at 495. Of the 165 closed claim files, 91 were settled, 65 were abandoned, and 9 were tried to a verdict. Id.
67. -10). Id. at 505.
68. Id. at 496. The researchers' translations were based on the risk managers' assessments of the quality of care, not the claims' likelihoods of success at trial. Id.
69. Id. at 497-99, 499 tbl.2. The likelihood of settlement was not related to the severity of the patient's injuries. Id. at 499.
70. Settlements were more closely correlated with quality of care than were mediation awards. Id. at 505-06, 506 tbl.5 (showing a log settlement payment coefficient of 2.63 with a standard error of 0.57 and significance at the 5% level, and a log mediation award coefficient of 2.07 with a standard error of 0.79 and significance at the 5% level (P=.05)).
71. In addition, the risk managers presumably understood that their in-house evaluation of care quality could reach the ears of the affected doctors, thus potentially souring economically crucial hospital-physician relationships.
72. Ralph Peeples, Catherine T. Harris & Thomas Metzloff, The Process of Managing Medical Malpractice Cases: The Role of Standard of Care, 37 WAKE FOREST L. REV. 877, 877 (2002). The files were obtained from a major teaching hospital and a major North Carolina malpractice liability insurer. Id. at 881. The researchers looked only at lawsuits that had progressed to the entry of a court order directing mediation. Id. Thus, cases that were dropped, dismissed, or settled immediately after the answer were not included.
73. Id. at 881-82. The insurance files included expert- and physician-review summaries as well as witness-deposition summaries. Id. at 882. Because the contents of these files are so much richer than the bare medical record, insurer files are considered the "gold standard" for conducting medical malpractice research. Id.
74. Id. at 884. The reviewers were also usually from the same state. Id.
75. See id. (showing that in 80% of the cases that the insurers assessed as "probably liable," the medical experts also found liability likely, and that in 65.4% of the cases that the insurers assessed as "unlikely liable," the medical experts also found liability unlikely).
76. Id. at 899 tbl.6. When cases that went to trial are added to the total, payment was made in 93.1% of the cases where liability was rated as probable, 36.8% of the cases where liability was deemed to be uncertain, and 14.8% of the cases rated as defensible. Id. at 886 n.35, 888-89.
77. Peeples et al., supra note 67, at 884 (finding an average of 3.07 reviews if the insurer found no breach, 3.27 reviews if the insurer was uncertain, and 4.43 reviews if the insurer found a breach). Peeples and his colleagues concluded that insurers "proceed more carefully in those cases in which liability appears likely." Id.
78. Frederick W. Cheney et al., Standard of Care and Anesthesia Liability, 261 JAMA: J. AM. MED. ASS'N, 1599, 1599 (1989). The goal of the study was to define the impact of the "standard of care" on the likelihood and amount of financial recovery. Id. The reviewer instructions defined appropriate care as "that which met the standard of care for a prudent anesthesiologist practicing anywhere in the US at the time of the event." Id.
79. Id. at 1599-600. In each instance, a single volunteer anesthesiologist reviewed the lawsuit file and determined whether the care was appropriate or inappropriate. Id. For the 869 cases in which the appropriateness of care could be judged, the reviewer scored care as appropriate in 46% of cases and inappropriate or below standard in 54% of cases. Id. at 1601.
80. Id. Ten percent received no payment and payment data was missing for 8%. Id.
81. Id. at 1601 (finding a significance level of P< 0.01). Severity of injury did not influence the likelihood of payment. Id.
82. Id. (finding a significance level of P< 0.05).
83. Cheney et al., supra note 73, at 1599.
84. Roger A. Rosenblatt & Andy Hurst, An Analysis of Closed Obstetric Malpractice Claims, 74 OBSTETRICS & GYNECOLOGY 710, 710 (1989). The insurer covered the majority of the state's physicians. Id. It is also worth noting that one of the researchers - Rosenblatt - is a physician. Id.
85. Id. at 711.
86. See id. at 712 tbl.3 (showing that the 95% figure was calculated from data indicating that nineteen of thirty-three cases involved negligence and eighteen payments were made).
87. Id. at 712.
88. Id. at 713.
89. See supra note 5.
90. Brennan et al., supra note 2, at 1963-64; A. Russell Localio et al., Relation Between Malpractice Claims and Adverse Events Due to Negligence, 325 NEW ENG. J. MED. 245, 245 (1991).
91. Brennan et al., supra note 2, at 1964. The hospital charts for these patients had been reviewed as part of the larger study by nurses and then by physicians retained by the researchers. Id. at 1967
92. Id.
93. Id. The presence of a permanent disability was predictive of payment, however. Id.
94. Id. at 1967.
95. See supra notes 1, 5.
96. Baker, supra note 9, at 501-02.
97. Id. at 507-08. Baker also points out that the association that the researchers found between payment and injury severity "means something other than it first seems." Id. at 508. This is because the study only assigned disability scores to those injuries that the reviewers felt had been caused by the patient's medical care. Id. Thus, the study's regression analysis showed not, as is commonly believed, that every badly disabled patient is likely to get a settlement, but that a patient who is seriously disabled by her medical care is more likely to receive a payment than a patient whose care causes only temporary injuries or none at all. In other words, patients who can prove causation are more likely to get a settlement if their injuries are serious than if they are minor. Thus construed, the finding accords with both common sense and the economic realities of malpractice litigation. Larger claims are more likely to be worth the costs of preparing for trial and will have higher expected values than smaller claims with similar evidence of negligence.
98. See Brennan et al., supra note 2, at 1966. These are the numbers obtained after the ratings were revised to reflect information that had surfaced subsequent to the rating process that occurred ten years earlier. Id. Almost half of the settled cases which had initially been deemed undeserving by the Harvard team's methodology (a methodology which was intentionally designed to understate the frequency of credible claims) received less than $25,000. Id. at 1965. Most of these patients were simply given a write-off of the remaining fees due to the defendants for the contested medical care. Id.
99. See infra Part II.C.
100. Localio et al., supra note 85, at 249, The category of defensible cases were said to lack "positive" evidence of negligence. Id. at 250. Disagreements between the two physician reviewers on this issue were resolved by averaging the two scores and treating as negligent only those claims in which the average rating was at least "more likely than not." Id. at 246. Thus, any file in which the two reviewers were evenly split was classified as not caused by negligence. The authors conceded that their rating process was not like the resolution of a claim at trial. Id. at 249. ("In a lawsuit, a single expert opinion might be sufficient to support a finding of negligence; under our protocol it would not"). But they never explored how their calculations would have changed if the divided-opinion cases were treated as a separate category of claims.
101. See Michael J. Saks, Medical Malpractice: Facing Real Problems and Finding Real Solutions, 35 WM. & MARY L. REV. 693 (1994) (reviewing a 1993 book by the Harvard team). Of the twenty-one claims that were reviewed for the presence or absence of negligence, eight were considered by all reviewers to involve breach of duty and seven were considered by one reviewer to involve negligence. Localio, supra note 85, at 248 tbl.3. Only six claims were considered by both reviewers to lack evidence of breach of duty. Id. Thus, over two-thirds of the claims reviewed for breach of duty were determined by at least one of the two reviewers to involve negligent medical care.
102. David M. Studdert et al., Claims, Errors, and Compensation Payments in Medical Malpractice Litigation, 354 NEW ENG. J. MED. 2024, 2029 (2006) (finding that 40% of all malpractice claims are not meritorious and that most are resolved without any payment of money).
103. Id. Only 1,441 claims files were evaluated for merit. Id. at 2028.
104. Id. at 2025.
105. Id.
106. Id.
107. Studdert et al., supra note 97, at 2025. "Reviewers recorded their judgments using a 6-point confidence scale." Id. at 2026. The study classified the medical care as erroneous if the claim had received a score of four or above. Id. Only a single reviewer rated each file (though a small number of files were distributed twice, without the reviewers' knowledge); therefore, this study design cannot detect cases in which multiple reviewers would have reached divergent opinions. Id. at 2025. Cases that received a three or a four, both of which were described to reviewers as "[c]lose call" cases, were not sorted into a middle category of "unclear" cases. Id. at 2028. Errors were defined as "the failure of a planned action to be completed as intended (i.e., error of execution) or the use of a wrong plan to achieve an aim (i.e., error of planning)." Id. at 2026 (internal citations omitted).
108. Id. at 2025-26.
109. Id. at 2026, 2029 fig.2. For our purposes, it would have been ideal if the study had reported the settlement rate for all of the claims that did not go to a jury verdict. However, the authors excluded thirty-seven claims for which they could find no evidence of injury and nine with only dignitary injury, as well as two for which no rating of care quality was done. Id.
110. Id. at 2028 (P<0.001).
111. Id. (noting $313,205 as the mean amount paid to non-error claimants versus $521,560 as the mean amount awarded to meritorious claimants, P=0.004).
112. Studdert et al., supra note 97, at 2029 fig.2.
113. Id. at 2028. Only 6% of the cases in which payment was made involved plaintiff's verdicts. Id. at 2030 tbl.2.
114. Id. at 2029 fig.2 (noting that these numbers exclude claims with only dignitary injuries (nine), those with no injuries (thirty-seven), and those without judgments as to error (two)).
115. Id. at 2031.
116. This number excludes the thirty-seven claims that lacked evidence of causation.
117. This number includes fifty cases that went to a jury verdict, thirty-nine of which were won by defendants.
118. See supra Part I.B.3.
119. See supra Part I.A.
120. See supra Part I.B.
121. See infra p. 123 tbl.3, p. 131 fig.6.
122. See infra p. 123 tbl.3, p. 131 fig.6.
123. Although all of the three-category studies found this association, only three reported their data in form suitable for depiction in this format.
124. In the Rosenblatt and Hurst study, none of the fourteen cases without medical negligence settled. Rosenblatt & Hurst, supra note 79, at 712 tbl.3. Thus, no bar appears under "Good Care" in Figure 3, infra p. 123 and Figure 4, infra p. 124.
125. The simultaneous operation of forces driving the settlement rate up in good-care cases and down in poor-care cases explains why the authors found no statistical correlation between care quality and the settlement rate.
126. Shari S. Diamond, Order in the Court: Consistency in Criminal-Court Decisions, in 2 THE MASTER LECTURE SERIES: PSYCHOLOGY AND THE LAW 119, 125 (C. James Scheirer & Barbara L. Hammonds eds., 1983).
127. Shari S. Diamond & Hans Zeisel, Sentencing Councils: A Study of Sentence Disparity and Its Reduction, 43 U. CHI. L. REV. 109, 119 tbl.2 (1975) (finding a roughly 70% rate of agreement in two different cities among judges on whether an offender should be imprisoned).
128. See, e.g., Farber & White, supra note 25, at 204-05 (finding that in 30% of the cases, the experts disagreed or gave ambiguous evaluations); A. Russell Localio et al., Identifying Adverse Events Caused by Medical Care: Degree of Physician Agreement in a Retrospective Chart Review, 125 ANNALS INTERNAL MED. 457, 460-61 (1996) (finding a similar disagreement rate on both negligence and causation) ; Peeples et al., supra note 67, at 884 (finding that reviewers disagreed in 34.3% of the cases).
129. See David M. Eddy, Variations in Physician Practice: The Role of Uncertainty, HEALTH AFF., Summer 1984, at 74, 75 (outlining the sources of uncertainty);
130. Philip G. Peters, Jr., The Role of the Jury in Modern Malpractice Law, 87 IOWA L. REV. 909, 956-58 (2002) (reviewing the evidence).
131. There are two other interesting theoretical explanations for the payment of a settlement in low-odds cases. One is that these are cases in which a desire for vindication motivates the plaintiff. These patients may, for example, believe that they were lied to or treated with disrespect. Studies show that plaintiffs seeking vindication are less likely to drop cases in which the evidence of negligent treatment is weak. See SLOAN ET AL., supra note 54, at 161. This persistence may lead to a settlement offer. In addition, prospect theory predicts that low-odds cases will be settled for a premium. See infra notes 206-14 and accompanying text. This prediction is based on research finding that individuals facing a low-probability gain (such as a frivolous lawsuit) tend to be risk-seeking while those facing a high-probability gain tend to be risk averse. The low settlement rate in low-odds malpractice cases suggests that the predicted difference in risk tolerances either is not routinely present in malpractice litigation or that it is usually trumped by other sources of negotiating power that favor defendants. In either event, the group of low-odds cases that settle may be the set of cases in which the predicted asymmetry in risk tolerance is most powerful.
132. Another factor working against a perfect agreement between settlements and the ratings given by external reviewers is the differences in the factual records on which their evaluations are based.
133. PAUL WEILER ET AL., A MEASURE OF MALPRACTICE 125 (1993) ("We found marked variation among physicians in their willingness to label certain kinds of medical outcomes as iatrogenic, and an even more pronounced reluctance to label as negligent those treatment decisions that, ex post at least, were clearly erroneous.").
134. Id. at 172.
135. Ogburn et al., supra note 43, at 610.
136. See Thomas B. Metzloff, Resolving Malpractice Disputes: Imaging the Jury's Shadow, 54 LAW & CONTEMP. PROBS. 43, 74-75, 83 n.126 (1991).
137. Id. at 74 n.104.
138. Id.
139. Id.
140. Catherine T. Harris, Ralph Peeples & Thomas B. Metzloff, Placing "Standard of Care" in Context: The Impact of Witness Potential and Attorney Reputation in Medical Malpractice Litigation, 3 J. EMPIRICAL LEGAL STUD. 467, 480-82 (2006).
141. Id. at 482.
142. Id. at 479-80 (describing cases with probable liability); id. at 482 (describing cases with unlikely liability).
143. With the significant exception of attorney experience (which strongly favors malpractice defendants), we do not yet know whether strategic factors tend to favor one side more often than the other. See infra notes 237-60 and accompanying text.
144. Harris, Peeples & Metzloff, supra note 134, at 489.
145. Id. at 479-80 (describing cases with probable liability); id. at 482 (describing cases with unlikely liability).
146. Doctors commonly characterize a settlement payment, no matter how small, as victory for the plaintiff. See, e.g., Ralph Peeples, Catherine T. Harris & Thomas B. Metzloff, Settlement Has Many Faces: Physicians, Attorneys, and Medical Malpractice, 41 J. HEALTH & SOC. BEHAV. 333, 343-44 (2000) (noting the belief that "settlement means guilty" and concluding that "[s]ettlement means one thing to physicians and quite another to attorneys"). Thus, physicians are often dismayed when a weak claim is settled by their insurer, even for a very modest amount. In their eyes, these settlements belie the claim that liability is premised exclusively on proof of fault. See id. at 343. Physicians who feel this way implicitly believe that the truth is binary, i.e., a claim is either warranted or it is not. Just as a traffic light is red, green, or yellow, but never 40% red, a doctor's treatment either meets the standard of care or it does not. However, this binary ideal simply does not apply to subjective judgments of fault. See supra notes 116-17 and accompanying text. Reasonable people disagree on judgments of this sort and there often is no objective benchmark to resolve their disagreement. See id. Even as to factual determinations, the binary view ignores the limitations in our ability to reconstruct past events. Id.
147. See infra Part II.C.
148. Log coefficient of .931 (.466).
149. Log coefficient of 1.54 (.448).
150. Log coefficient of .579 (.273).
151. Log coefficient of 1.34 (.247).
152. Log settlement payment coefficient of 2.63 with a standard error of 0.57.
153. The $23,552 figure covers claims lacking both negligence and causation. The mean was $31,375 for claims lacking only negligence.
154. Brennan et al., supra note 2, at 1964-65, 1965 tbl.1 (showing many settlements under $25,000).
155. See SLOAN ET AL., supra note 54, at 220 (noting that settlements, unlike trials, will discount the damages to reflect the probability of success);
156. Spurr & Howze, supra note 61, at 502-04.
157. Spurr & Howze, supra note 61, at 499.
158. See Samuel R. Gross & Kent D. Syverud, Getting to No: A Study of Settlement Negotiations and the Selection of Cases for Trial, 90 MICH. L. REV. 319, 342 (1991) (stating that virtually every case has some probability, however low, that the plaintiff will prevail and every case has some settlement value). Gross and Syverud note that, under the basic economic model of settlement, zero-payment offers should be rare, even though they are not. Id. at 343 (reporting that no settlement offer was made in 25.2% of the personal injury cases in their sample).
159. Brennan et al., supra note 2, at 1963.
160. See supra p. 122 fig. 2.
161. See supra p. 114 tbl.1.
162. See supra Part II.B.
163. Id.
164. Just as the costs of litigation give defendants an incentive to make nominal payments to settle weak claims, the costs of litigation similarly provide plaintiffs who have strong evidence of negligence with an incentive to settle for an amount that is less than their actual damages. See RUSSELL KOROBKIN, NEGOTIATION THEORY AND STRATEGY 101 (2002) (noting that litigation costs lead plaintiffs to set lower settlement expectations).
165. Farber & White, supra note 25, at 207-08.
166. Id. at 208.
167. Id. at 207 (noting that the average settlement was $260,000 while the average recommendation was $446,000, and stating that only seventeen of the sixty-eight settled cases in their sample were promptly settled for the figure recommended by the panel). It seems reasonable to assume that the fifty-one cases rejecting the panel's recommendations would have settled for even less had a screening panel not reviewed them and made recommendations.
168. Stephen J. Spurr & Walter O. Simmons, Medical Malpractice in Michigan: An Economic Analysis, 21 J. HEALTH POL. POL'Y & L. 315, 337 (1996) (noting that 68.6% of the cases settled for less than the panel recommended and that the average settlement was $117,063 while the average recommendation was $125,934).
169. Id. at 337. Only 2.1% had a mediation award of zero. Id. at 336.
170. Walter Orlando Simmons, An Economic Analysis of Mandatory Mediation and the Disposition of Medical Malpractice Claims, J. LEGAL ECON., Fall 1996, at 41, 51 (finding that 15.6% of the settlements equaled the panel award).
171. Id.
172. Id. at 51 tbl.7.
173. Id. at 52.
174. See, e.g., Farber & White, supra note 25, at 208;
175. Spurr & Howze, supra note 61, at 495;
176. Spurr & Simmons, supra note 161, at 340.
177. See Peeples et al., supra note 67, at 884-85, 891-93.
178. See infra Part III.
179. See Peeples et al., supra note 67, at 892 (noting that physicians see the lack of peer review as a major flaw even though physicians currently determine the standard of care).
180. Id.
181. See supra Part II.C.
182. See supra Part II.C.
183. One could make the same point with a graph showing that average recovery included not only the cases in which a settlement payment was made, but also all of the zero-payment cases.
184. See George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13J. LEGAL STUD. 1, 26 (1984) (explaining that "tie relative [settlement] calculus . . . changes" when one party has more to lose from an adverse verdict);
185. Spurr & Simmons, supra note 161, at 338.
186. See Gross & Syverud, supra note 151, at 365-66 (suggesting that self-esteem is implicated more than reputation by the settlement decision).
187. See Peeples, Harris & Metzloff, supra note 140, at 343-44.
188. Id. at 342 (indicating that almost two-thirds of physicians who thought they were not liable expressed a desire for vindication).
189. See Gross & Syverud, supra note 151, at 366. For the same reason, plaintiffs who are strongly motivated by nonpecuniary objectives are less likely to drop their cases.
190. See Frank A. Sloan & Thomas J. Hoerger, Uncertainty, Information and Resolution of Medical Malpractice Disputes, 4J. RISK & UNCERTAINTY 402, 415 (1991).
191. See Teresa M. Waters et al., Impact of the National Practitioner Data Bank on Resolution of Malpractice Claims, 40 INQUIRY 283, 283 (2003) (finding that physicians have been less likely to setde claims since introduction of the NPDB in 1990, especially for payments less than $50,000).
192. See, e.g., Brion v. Vigilant Ins. Co., 651 S.W.2d 183, 184 (Mo. Ct. App. 1983) (calling this a "pride" provision);
193. Kent D. Syverud, The Duty to Settle, 76 VA. L. REV. 1113, 1176 (1990) (stating that contracts with these clauses are 1% to 3% more expensive).
194. Gross & Syverud, supra note 151, at 361 n.103.
195. See id. (discussing "zero-offer cases in which a physician was the sole defendant at trial (9/23)").
196. Syverud, supra note 181, at 1178-79.
197. "Hard bargaining" is predicted by negotiation theory when a repeat player, like the liability insurer, faces non-repeat litigants. See Priest & Klein, supra note 175, at 24-29 (discussing the "likelihood of litigation" when parties have "[a]symmetric [s]takes");
198. see also Robert Cooter et al., Bargaining in the Shadow of the Law: A Testable Model of Strategic Behavior, 11J. LEGAL STUD. 225, 241-42 (1982) ("[O]ur model predicts that a repeat player whose opponents are not repeat players will adopt a hard bargaining strategy.").
199. For a discussion of the doctor's stakes, see Spurr & Simmons, supra note 161, at 338 (noting "the costs of trial to a defendant physician in terms of lost time, emotional strain, and the damaging publicity").
200. Metzloff, supra note 130, at 77 n.111. No settlement offer was made in twenty-eight of forty-eight cases, or 58.3%. Id. Metzloff also found that settlement offers had been made prior to trial in over 90% of the likely successful claims (ten of eleven), Id.
201. Id. at 77 n.111.
202. Id. at 78.
203. Id. at 77 n.111. No offer was made in sixteen of twenty anticipated unsuccessful claims, eleven of thirteen toss-ups, and one of eleven successful claims. Id.
204. Id.
205. Peeples et al., supra note 67, at 887.
206. Id. at 899.
207. See Gross & Syverud, supra note 151, at 378.
208. Id. at 346 tbl.3.
209. Id. at 343.
210. See Sloan & Hsieh, supra note 49, at 1018. This hard bargaining, including the frequent failure to make any offer at all, may partially explain why malpractice trials are twice as likely to occur than trials of other personal injury claims. Plaintiffs are left with a choice between dropping the case and proceeding to trial. This posture may also explain the high fraction of zero-dollar dispositions in medical malpractice cases.
211. Peeples et al., supra note 67, at 887 (noting that insurers resist the settlement of cases in which the standard of care was not breached).
212. Id.
213. Waters et al., supra note 180, at 290 (finding that settlements in low-odds cases had dropped about 30% and also finding that the settlement rate in high-probability cases had declined, though the change was not statistically significant).
214. See Priest Sc Klein, supra note 175, at 40.
215. See id.
216. Peeples et al., supra note 67, at 886. Of the twenty-two low-odds cases in which the defendant made no settlement offer, half were dropped without payment and ten of the remaining cases resulted in defense verdicts. Id. at 888.
217. Simmons, supra note 163, at 42.
218. The amount paid to settle all malpractice claims falls short of their total expected value. Thus, Figure 8 shows heavier discounting in the low-odds and toss-up cases than Figure 7. See supra Part II.B. If the discount implicit in this finding is not present in the cases with strong evidence of liability, then it presumably is concentrated in the toss-up and low-odds claims. The extra discounting of low-odds claims is also consistent with the evidence showing that insurers resist settlement in these cases.
219. Metzloff, supra note 130, at 78.
220. See, e.g., id. at 62-63; Priest & Klein, supra note 175, at 27. A risk-neutral plaintiff would be indifferent to the choice between accepting a $100,000 settlement offer and proceeding to trial in a case with a 50% chance of winning a verdict of $200,000 because each choice has the same expected value. By contrast, a risk-averse plaintiff would prefer the settlement as it substitutes a sure gain for a risky one. She will do so even though the "expected values" are the same and a trial would offer the possibility of an even larger gain (something that would appeal to a risk-seeking individual).
221. See Metzloff, supra note 130, at 62-63.
222. See Farber & White, supra note 25, at 208;
223. Chris Guthrie, Framing Frivolous Litigation: A Psychological Theory, 67 U. CHI. L. REV. 163, 167 (2000);
224. Sloan & Hsieh, supra note 49, at 998.
225. See Marc Galanter, Why the "Haves" Come Out Ahead: Speculations on the Limits of Legal Change, 9 L. & SOC'Y REV. 95, 99-100, 110 (1974) (noting that personal injury insurers typically are repeat players and that repeat players are more likely to be risk neutral, i.e., indifferent to uncertainty);
226. KOROBKIN, supra note 157, at 46 (illustrating why repeat players are risk neutral).
227. In theory, this negotiating advantage should not extend to the weakest malpractice claims because social scientists have found that the tendencies to accept a settlement are reversed when the probability of gain is very small. See Guthrie, supra note 207, at 168-69.
228. See generally Daniel Kahneman & Amos Tversky, Prospect Theory: An Analysis of Decision Under Risk, 47 ECONOMETRICA 263 (1979);
229. Amos Tversky & Daniel Kahneman, Advances in Prospect Theory: Cumulative Representation of Uncertainty, 51 J. RISK & UNCERTAINTY 297 (1992). That may explain why 10-20% of the low-odds cases result in some settlement.
230. See supra p. 124 fig.4, p. 125 fig.4A. In most low-odds cases, however, the plaintiffs advantage apparendy is outweighed by other factors that confer a negotiating advantage on defendants.
231. Farber & White, supra note 25, at 208. The authors ruled out lower trial costs and higher defendant optimism as explanations of this bargaining power. Id. at 208. For a description of how risk aversion affects litigants' decisions,
232. see W. Kip Viscusi, Product Liability Litigation with Risk Aversion, 17 J. LEGAL STUD. 101, 103 (1988) (concluding that claimants' risk aversion is greater than defendants' risk aversion). Patients who hold out until later in the negotiation process appear to get higher payments, thus rewarding their risk taking.
233. Sloan & Hsieh, supra note 49, at 1026.
234. See Spurr & Simmons, supra note 161, at 340 (attributing the difference between expected value and settlement amount to either plaintiffs' risk aversion or "their disadvantage in bargaining").
235. Peeples et al., supra note 67, at 887.
236. Simmons, supra note 163, at 48. Simmons also found that 15.6% of the settlements equaled the panel award. Id. at 51.
237. Farber & White, supra note 25, at 208.
238. See generally, Philip G. Peters, Jr., Doctors and Juries, 106 MICH. L. REV. 1453 (2007) (comprehensively reviewing the empirical literature that tests the correlation between jury verdicts and evidence of negligence).
239. Id.
240. See David M. Engel, The Oven Bird's Song: Insiders, Outsiders and Personal Injuries in an American Community, 18 LAW & SOC'Y REV. 551, 553, 559-61 (1984) (finding that citizens in a rural Illinois county disapproved of "cashing in" via personal injury lawsuits and characterized those who did sue as "people looking for the easy buck");
241. Edith Greene et al., Jurors' Attitudes About Civil Litigation and the Size of Damage Awards, 40 AM. U. L. REV. 805, 814 (1991) (finding that 91% of jurors studied thought "there were too many lawsuits");
242. Valerie Q. Hans & William S. Lofquist, Jurors' Judgments of Business Liability in Tort Cases: Implications for the Litigation Explosion Debate, 26 LAW & SOC'Y REV. 85, 94 (1992) (stating that "tort injuries approached their own cases with considerable suspicion about the plaintiff).
243. Hans & Lofquist, supra note 217, at 93. This study asked potential jurors about tort cases against corporations. Id.
244. Id. Some of the skepticism about plaintiffs may be the product of cognitive dissonance-we do not want to think poorly of the people to whom we entrust our lives and well-being. Another possible explanation has been suggested by medical sociologists who believe that the poor success of malpractice plaintiffs in court is a predictable consequence of their lower social status relative to physicians. See Jeffrey Mullis, Medical Malpractice, Social Structure, and Social Control, 10 SOC. F. 135, 137, 145, 149 (1995) (arguing that wealth and social status help defendants in medical malpractice cases).
245. NEIL VIDMAR, MEDICAL MALPRACTICE AND THE AMERICAN JURY: CONFRONTING THE MYTHS ABOUT JURY INCOMPETENCE, DEEP POCKETS, AND OUTRAGEOUS DAMAGE AWARDS 169 (1995).
246. Id.
247. Ellen L. Leggett, Identifying Juror Bias and Their Impact on Malpractice Cases (Sept. 7, 1999) (unpublished article, on file with the Iowa Law Review).
248. Id.;
249. see Greene et al., supra note 217, at 817 (finding that most jurors believe attorneys encourage people to file frivolous lawsuits).
250. See Leggett, supra note 222 (discussing her findings).
251. Engel, supra note 217, at 553, 559-61.
252. Id. at 553.
253. Id. at 564-65.
254. Stephen Daniels & Joanne Martin, It Was the Best of Times, It Was the Worst of Times: The Precarious Nature of Plaintiffs'Practice in Texas, 80 TEX. L. REV. 1781, 1783, 1796 (2002);
255. see also Stephen Daniels & Joanne Martin, The Strange Success of Tort Reform, 53 EMORY L.J. 1225 passim (2004) (arguing that the tort-reform debates have succeeded in changing the environment in which civil litigation occurs).
256. See HERBERT M. KRITZER, RISKS, REPUTATIONS, AND REWARDS: CONTINGENCY FEE LEGAL PRACTICE IN THE UNITED STATES 300 n.33 (2004) (reporting on anecdotal statements by counsel).
257. David A. Hyman et al., Do Defendants Pay What Juries Award? Post-Verdict Haircuts in Texas Medical Malpractice Cases, 1988-2003, J. EMPIRICAL LEGAL STUD. (forthcoming 2007) (manuscript at 2, on file with the Iowa Law Review), available at http://ssrn.com/abstract-914415 (finding that plaintiffs receive a mean "haircut" of 29% and an average "haircut" of 56%).
258. Id.
259. Id.
260. Id.
261. Id.
262. Hyman et al., supra note 230, at 2.
263. Id.
264. Lucian A. Bebchuk, Litigation and Settlement Under Imperfect Information, 15 RAND J. ECON. 404, 409 (1984).
265. Id.
266. Farber & White, supra note 38, at 788 (noting that thirty-seven of 355 claims were settled without a lawsuit when initiated by a complaint process).
267. Id. ("[T] hese empirical results are consistent with an information structure in which patients initially are poorly informed about the quality of medical care and the hospital initially is poorly informed about how litigious patients are.").
268. Id. at 795. At most, a hospital makes small settlement offers. Id. at 802. The goal is to avoid settling with the "peaceful" patients who will not file suit. Id. at 795.
269. Rosenblatt & Hurst, supra note 79, at 711 (noting that pre-suit settlement rarely occurred in cases of neonatal death).
270. See Peeples et al, supra note 67, at 886 (finding insurers almost always deposed experts for the plaintiff).
271. KOROBKIN, supra note 157, at 153.
272. See id. at 156 (noting that it is not always cost-effective to hire more expert witnesses).
273. Id. at 171-72.
274. I have observed this belief throughout my many years serving as defense counsel in malpractice suits.
275. Neil Vidmar, Are Juries Competent to Decide Liability in Tort Cases Involving Scientific/Medical Issues ? Some Data from Medical Malpractice, 43 EMORY LJ. 885, 902 (1994).
276. See SLOAN ET AL., supra note 54, at 92-93.
277. See, e.g., id. at 208, 216 (finding that specialists constitute a minority of plaintiffs' attorneys and recommending specialty certification);
278. Galanter, supra note 208, at 110 (noting that, in general, personal injury insurers are typically repeat players, while personal injury plaintiffs are not);
279. Catherine T. Harris, Ralph Peeples & Thomas B. Metzloff, Who Are Those Guys? An Empirical Examination of Medical Malpractice Plaintiffs' Attorneys, 58 SMUL. REV. 225, 241 (2005) (reporting that defense counsel in the study sample had handled an average of over twice as many malpractice cases as had their counterparts);
280. Sloan & Hsieh, supra note 49, at 997-98 (reviewing the literature relating to variability in medical malpractice payments).
281. See Galanter, supra note 208, at 97-98 (noting that repeat players will have greater expertise and better access to specialists).
282. SLOAN ET AL., supra note 54, at 76. The definition of "specialist" was based on a combination of factors and included the attorneys in firms that handled four or more malpractice cases during the period of the study and those listed on several specialty lists. Id. at 170 (noting that specialists represented 23% of the claimants).
283. Harris et al., supra note 250, at 243. Unfortunately, however, this study did not measure case quality and, thus, could not rule out the chance that the superior outcomes of experienced counsel were the product of a better case mix.
284. SLOAN ET AL., supra note 54, at 216 (holding liability and severity constant). "Holding other factors constant," specialists got payments that were 92% higher, Id. at 201. Interestingly, this was not because specialists were able to obtain settlements in excess of the case's expected value or even in excess of the patient's economic losses. The settlements were too small on average to cover even economic loss. Id.
285. Id. at 208.
286. Id. at 216. The cases in which no payment or very low payment resulted constituted 51% of the cases handled by nonspecialists and 29% of the cases handled by specialists. Id. at 197 tbl.9.4.
287. Id. at 208.
288. Metzloff, supra note 130, at 80. He observed that plaintiffs' attorneys more often made outlandish demands. Id. at 75-76. Similarly, Russell Korobkin notes that an inexperienced attorney may be more likely to mistake his aspiration price for his reservation price, thereby leaving a decent settlement on the table.
289. KOROBKIN, supra note 157, at 62-63. As to why this discrepancy in skill should be more powerful in medical malpractice cases than in other personal injury cases, Metzloff suggests that malpractice cases accentuate the need to locate experts and to understand a unique body of substantive issues.
290. Metzloff, supra note 130, at 80.
291. SLOAN ET AL., supra note 54, at 216. Specialists were no more or less likely to have a liable defendant but were slightly more likely to have an "uncertain" case. Id. at 185 app.8A. Their cases did, however, involve more severe injuries as measured by economic loss. Id. at 170.
292. See Peeples et al., supra note 67, at 895 (suggesting that strategic factors may explain why plaintiffs with "non-meritorious" claims pursue their claims).