EAGLE-PICHER INDUSTRIES, INC.
v.
AMERICAN EMPLOYERS' INSURANCE CO.
Civ. A. No. 83-348-Z.
United States District Court, D. Massachusetts.
August 14, 1989.*1054 Charles R. Parrott, Brian T. Kenner, Nutter, McClennan & Fish, Boston, Mass., for plaintiff.
Richard H. Gimer, Hamel & Park, Stephen L. Humphrey, Washington, D.C., Ellen M. Martin, Boston, Mass., for American Employers.
Alice E. Richmond, Hemenway & Barnes, Boston, Mass., for Prudential Reinsurance.
MEMORANDUM OF DECISION
ZOBEL, District Judge.
This case represents, one hopes, the next-to-last installment of lengthy litigation regarding the liability of various insurance companies to indemnify Eagle-Picher Industries, Inc., a manufacturer of asbestos products, for sums paid to claimants suffering asbestos-related injuries. Eagle-Picher Indus., Inc. v. Liberty Mutual Ins. Co., 682 F.2d 12 (1st Cir.1982) ("Eagle-Picher I"), determined that the obligation of defendant insurers arose as of the date of manifestation (defined as the time at which the disease was "reasonably capable of medical diagnosis"). On appeal from the implementation judgment, the First Circuit adopted the presumption, rebuttable by clear and convincing evidence, that asbestosis is "reasonably capable of medical diagnosis" six years before the date of actual diagnosis and it remanded for findings as to the appropriate rollback period for other asbestos-related diseases. Eagle-Picher Indus., Inc. v. Liberty Mutual Ins. Co., 829 F.2d 227 (1st Cir.1987).
While the first appeal in Eagle-Picher I was pending, Eagle-Picher filed the instant case ("Eagle-Picher II") against a number of carriers not previously sued, including defendant, American Employers' Insurance Company ("American Employers'").[1] Defendant moved for summary judgment on issues of judicial, collateral and equitable estoppel and for partial summary judgment to the extent plaintiff's claim was based upon underlying asbestosis claims previously indemnified. Both motions were denied. Following the issuance of this Court's Memorandum of Decision in EaglePicher I on May 30, 1986, Eagle-Picher filed a Motion for Entry of a Final Judgment in this action seeking collateral estoppel effect to the second Eagle-Picher I decision, i.e., automatic application of the six-year rollback implementation formula. In a Memorandum of Decision, I denied the motion, noting that the Eagle-Picher II defendants could not be bound by the Court's implementation ruling in Eagle-Picher I, to which they were not parties. I then directed the parties to file written outlines of their implementation proposals. On the basis of these written submissions, plaintiff moved for partial summary judgment invoking the doctrine of stare decisis. It argued that the proposal submitted by defendant contained substantially, if not precisely, the same weaknesses that the Court had found in the Eagle-Picher I protocol and that the Court was therefore compelled to reject defendant's proposal and impose a six-year rollback formula. On May 31, 1988, that motion too was denied to afford defendant an opportunity to distinguish its implementation proposal from that presented in Eagle-Picher I. However, in view of the issues raised by the motion, I ordered that the trial proceed in two phases, the first to determine the feasibility of claim-by-claim review, and the second, if necessary, the manner of fixing the manifestation date.[2] What follows are my findings and conclusions in Phase I.[3]
*1055 The parties agree that defendant bears the burden of establishing that the "universe" of claims at issue in this case is so small that the claim-by-claim review it advocates is practicable. A necessary concomitant is a showing that existing claims files contain adequate information for such review and that a practical method exists for determining insurance attachment dates from the available information.
American Employers mounts a two-pronged attack. First, Dr. Paul Epstein, a physician who specializes in pulmonary diseases, offered a definition of disease and testified concerning his experiences with persons who present with an asbestos exposure history. Second, defendant offered its Draft Process for Determining when a Diagnosed Asbestos Disease was Clinically Evident, i.e., Reasonably Capable of Diagnosis, (hereinafter "the Process") and the results obtained by its use.
Dr. Epstein, a Professor at the University of Pennsylvania School of Medicine and a pulmonary specialist, defined disease as something that changes a person's life span or life style, that produces symptoms, or that causes a measurable change in function. Given that definition, he opined that the onset of asbestos-related disease is usually accompanied by symptoms such as shortness of breath or impairment of the person's lungs, and that persons with those symptoms seek medical attention within a relatively short period of time thereafter. Accordingly, the interval between the time when the disease is diagnosed and the earlier date when it was clinically evident or capable of diagnosis is relatively short: two years for asbestosis and lung cancer and one for mesothelioma. Dr. Epstein also stated that, with rare exceptions, claimants with only pleural changes are asymptomatic and they therefore do not have disease, as he defined it. Because they have no symptoms, the condition of these claimants is clinically evident only when observed by X-ray.
From this evidence, defendant argues first that at most the "universe of claims" includes only those Eagle-Picher claimants who were diagnosed on or before June 1, 1975, i.e., within two years after the expiration of defendant's policy on June 1, 1973. Second, it asserts pleural claims should be excluded altogether as not within defendant's policy since they do not constitute disease. It then approximates the number of claims subsumed by the period between June 1, 1973 and June 1, 1975, at 600, a number clearly allowing for individual review.
The second prong of defendant's case concerns the feasibility of determining insurance attachment dates based upon the information available about the underlying claims. The only such information is that contained in the files of Eagle-Picher, Liberty Mutual, its claims administrator and defense counsel. Defendant acknowledges that the first two may be inadequate, but offered the testimony of Edward J. David that defense counsel files are fully sufficient for this purpose. Mr. David served as defense counsel in the underlying litigation for a number of manufacturers, not including Eagle-Picher, prior to 1985. Although he apparently collected more medical information on behalf of his clients than did plaintiff and its counsel, he agreed that only a tiny fraction of cases is intensively prepared for trial. He also agreed that the quantum of evidence in counsel files varied *1056 from time-to-time and from jurisdiction-to-jurisdiction.
The procedure by which American Employers' proposes to establish an insurance attachment date is divided into two phases. Phase I consists of file review and data extraction; the actual claims analysis occurs in Phase II. During the first phase, data recorders review the claim files presented by plaintiff and extract medical information therein onto data-collection forms. During Phase II a team of claims analysts, using the data so collected, applies the Process to determine the insurance attachment date.
The Process provides for three approaches. As described by defendant, Approach I focuses on prediagnostic signs and symptoms of asbestos-related disease which, taken alone or in combination, indicate the time when, prior to actual diagnosis, the particular disease involved was diagnosable. According to defendant, it is to be applied with some flexibility in order to address the greatest possible number of scenarios of available prediagnostic information. Approach II assumes the absence of sufficient information to apply Approach I and looks, instead, for evidence of good chest health as a means of circumscribing the time in advance of diagnosis when an asbestos-related disease was diagnosable. That is, when a claimant's file evidences a healthy chest after the expiration of defendant's policy, defendant argues, he could not have had diagnosable disease during the covered period. Approach III is to be used when Approaches I and II cannot be employed, i.e., when the file contains insufficient evidence of prediagnostic symptoms or prediagnostic indications of "normal" health but it does contain "information indicating the mildness of the disease at the time of diagnosis." Approach III apparently applies only to asbestosis claims, and the "degree" of mildness dictates whether the disease was diagnosable one or two years before the appearance of symptoms.
Plaintiff relied on the testimony of Drs. Hans Weill, John Diem and Gary Epler to establish that a large number of claims diagnosed well after the policy years, 1967 to 1973, were in fact reasonably capable of medical diagnosis during that period. Drs. Weill and Diem are the authors of the epidemiologic study described in Eagle-Picher I. Dr. Epler is a physician specializing in pulmonary disease as clinician, researcher, author and academic. He serves as Chairman of the Department of Medicine and as Director of Respiratory Therapy at the New England Baptist Hospital in Boston, Massachusetts, and is an Associate Clinical Professor of Medicine at the Boston University School of Medicine. He testified that asbestosis and/or pleural conditions have long latency periods after exposure, at least fifteen years, that both progress slowly, and that a percentage of diagnoses made in the 1980s could have been made at least fifteen years earlier. He concluded that the meaningful possibility of diagnosability begins once the latency period has been met.
For purposes of Eagle-Picher I, the Weill/Diem study showed that it is more probable than not that a currently diagnosed case of asbestosis could have been diagnosed approximately six years ago. For this litigation, Drs. Weill and Diem were asked to extrapolate the findings of the initial study, which involved the observation of the progression of asbestosis among workers over a ten-year period to a twenty-one-year period. The result of that exercise is the statistical likelihood that disease found well after the policy period in issue could have been diagnosed within the period. Dr. Weill testified that projections of this sort beyond the actual period studied are frequently done, although he was unaware of any instance where projections had been made over double the observation time.
Finally, plaintiff offered the testimony of Lawrence G. Cetrulo, counsel for a number of manufacturers of asbestos products and for the Asbestos Claims Facility, to rebut Mr. David's description of defense counsel files. Mr. Cetrulo found that the available files often contain conflicting medical information, that typically they include only a small amount of prediagnostic material, and that it is difficult to recreate accurately the state of the medical record. Roger *1057 L. Prickett, plaintiff's general counsel, through his earlier testimony, confirmed that a large number of cases have settled at relatively early stages in the litigation process and that those files may contain no medical information beyond the allegations of the complaint. Indeed, because of potential statute of limitations problems, it is against a claimant's interest to obtain such historical information. Mr. David did agree with this observation and further acknowledged that defense counsel does not make a conscious effort to obtain all prediagnostic medical information.
The evidence requires resolution of two underlying factual disputes: the amount and quality of available medical information as to particular claimants, and the nature and rate of progression of asbestos-related disease, and, thus, the size of the universe.
As to the first, I find that, contrary to defendant's assertions, the earlier case files do not contain more information than do the later ones. The evidence in this case simply does not contradict the finding in Eagle-Picher I that the aggregation of files as to any one claimant generally contains little medical information. Although Mr. David suggested that any defendant in the underlying litigation "could, on a routine basis, with very little if any additional effort or expense," obtain a variety of additional medical information, the issue is not what might be or might have been obtained when these cases were pending, but rather what is now contained in the existing files of those cases. On that issue I credit the testimony of Mr. Cetrulo that the vast majority of cases are not tried or even prepared for trial, that the files of the settled cases generally contain current diagnostic information but no historical medical information prior to the claimant's diagnosis. This is so because many of these cases result from mass X-ray screenings at occupational locations conducted by unions and/or plaintiffs' attorneys, and many claimants are functionally asymptomatic when suit is filed. Moreover, many diagnoses are made by physicians not well schooled in the American Thoracic Society's criteria for the diagnosis of asbestosis, and the medical literature does not provide standards for judging the diagnosability of asbestos-related disease. In addition to being inadequate, the files thus also frequently contain conflicting information.
As to the second underlying issue, I credit Dr. Epler's testimony that asbestos-related disease has a long latency period, that asbestosis and pleural conditions progress slowly and at variable rates, but that a significant possibility of diagnosability exists from the expiration of the latency period and that, indeed, a percentage of claimants diagnosed in the 1980s were diagnosable fifteen years earlier. Defendant's expert, Dr. Epstein, agreed with most of Dr. Epler's premise, but he denies the possible existence of pleural conditions in such claimants and he fixes the diagnosability of asbestosis at two years before diagnosis in each instance, based on his definition of disease. However, Dr. Epstein's definition is at variance with common sense[4] as well as the experience of the parties and the court in the underlying cases.[5]
As I did in Eagle-Picher I, I note the imperfections of the Weill/Diem study. Notwithstanding those imperfections, I accept the study as adequate for the limited purpose of establishing the probability of prior diagnosability of asbestosis. Although the projections covered an extended period, Dr. Epler's experience supports the Weill/Diem analysis. I note also in this connection that Exhibit E, Frequency Distribution of Intervals between Actual Diagnosis and Attachment Date, a document submitted with the direct testimony of defendant's computer expert, shows that by defendant's own calculation half of the cases processed under defendant's Process have intervals of greater than two years. In fact, according the Exhibit E approximately *1058 90% of the claims reviewed by defendants were assigned attachment dates within six years of actual diagnosis. Accordingly, the number of claims subject to indemnification by defendant's policy far exceeds defendant's estimate of 600.
Although not necessary to the decision of this phase of the case, it is appropriate to address the evidence of defendant's Process and to record the problems inherent in its use. That is, even were the universe sufficiently small to permit claim-by-claim review, such review with the use of the Process would yield no more accurate and fair result than does the use of the rollback sanctioned by Eagle-Picher I.
First, the Process is unworkable. It does not define how conflicts in medical opinions are to be resolved. It articulates no standard for making a whole series of judgments that result ultimately in the decision as to attachment date. It arbitrarily ignores certain events in a claimant's medical history and elevates others to assume greater importance than that assigned by the claimant's physician. By establishing its own definitions of diagnostic events, it sometimes fails to accept the conclusion of the treating physician. It fails to explain the means by which an attachment date is assigned to a claim without prediagnostic information or evidence of normal health but with indications of severe asbestosis at the time of diagnosis.
Second, and more important, the Process is unreliable. It is unreliable because at each level of execution all of the results obtained by its use are based on subjective judgments which are necessarily carried forward to subsequent phases of analysis and review. This is particularly troubling as these judgments are made largely by persons lacking even rudimentary medical training. During data extraction, paralegal assistants initially judge the completeness of the file and determine whether additional information is needed. Upon obtaining such additional information, they decide whether or not it is sufficient and, finally, what is to be extracted and recorded on the prescribed data-collection forms.[6] To the extent that the files contain descriptive terms, e.g., severe, mild, normal, abnormal, possibly caused by, impression that, early, late, etc., the paralegals decide how these terms are recorded on the data-collection forms. The paralegals thus decide which diagnostic criteria will be available to the Phase II Team. The Phase II Team, comprised of experienced claims analysts, decides whether enough information exists, pre-diagnostic or otherwise, to proceed under Approach I or whether Approach II or III should be applied. They do not review the data extraction forms for reliability. However, they interpret the descriptive terms memorialized on the data extraction forms without reference to any established medical conventions and without accounting for any possible bias in the files. At the same time, the Process permits the claims analysts to override a treating physician's diagnosis and to establish a date of diagnosis different from that of the doctor. In the case of conflicting medical information or diagnoses, the claims adjusters resolve the difference. Defendant has conducted no studies to determine whether all members of the Team apply the stated criteria consistently. These subjective decisions are layered on top of the grander underlying assumptions of the Process, including the definition of disease and the diagnostic criteria employed under the Process.
Given the numerous subjective decisions made in the course of executing the Process and the absence of any evidence indicating *1059 the level of intervariability of either data collectors or claims analysts, the consistency of its application and interpretation are called into serious question. In light of these problems, it is also noteworthy that the Process provides no mechanism for resolving any dispute with plaintiff concerning the use of particular data or the application of the Process.
Finally, as was true in Eagle-Picher I, the speed with which this defendant would review claims is relevant to the issue of feasibility. From mid-May until November 22, 1988, the Phase II Team reviewed only 594 claims. Not all were assigned an attachment date. At this rate of approximately twenty claims per week the speed of review becomes a procedural barrier effectively denying insurance coverage to plaintiff.
Because defendant has failed to sustain its burden of demonstrating the feasibility of claim-by-claim review, the doctrine of stare decisis applies. Accordingly, pursuant to the decision of the Court of Appeals in Eagle-Picher I, a six-year presumptive rollback from the date of diagnosis, as established therein, will determine the attachment date for all asbestosis and pleural cases. The presumptive rollback for other asbestos-related diseases will be determined after a further trial. A pretrial conference will be held on September 21, 1989 at 3:00 p.m.
Given the length of these proceedings and the extensive discussion of the issues in this Court and in the Court of Appeals, there is no justification for further delay. Accordingly, judgment shall be entered pursuant to Rule 54 declaring that implementation for asbestosis and pleural cases shall be in accordance with the rule announced in Eagle-Picher I. Plaintiff shall submit a form of judgment by August 23, 1989. Any indemnity payable for asbestosis and pleural claims shall carry interest from September 16, 1987.
NOTES
[1] All defendants but American Employers' have settled.
[2] The Order states in relevant part:
Phase I will address the feasibility of individual review of claims. Resolution of that issue hinges on the definition of the relevant universe, the number of claims encompassed thereby and the quality of the underlying claim files.
Phase II will address the method for determining the manifestation date in light of the decision in Phase I and to the extent that decision does not call for the application of stare decisis.
[3] After the close of the evidence of Phase I, defendant submitted a motion for reconsideration of this Court's denial of its motion for partial summary judgment concerning pleural claims without functional significance. The original motion for partial summary judgment sought to exclude claims for pleural conditions without functional significance, because those conditions do not amount to disease and are not therefore within the coverage of its policies. Defendant also now moves for summary judgment based on a letter which, defendant argues, constitutes an agreement between plaintiff and Liberty Mutual Insurance Company ("Liberty Mutual") attesting to a "first medical diagnosis" trigger. Defendant claims that the comprehensive general liability insurance policies between plaintiff and Liberty Mutual governs all excess coverage policies in effect at that time, including policies provided by defendant, and that a "first medical diagnosis" trigger should therefore be imposed in this case. The merits of both motions are dealt with in the body of this opinion. With respect to defendant's Motion for Summary Judgment, the Court adds that the law of the case is diagnosability and that the letter is therefore irrelevant. Further, a letter between plaintiff and Liberty Mutual cannot alter a contract between plaintiff and American Employers'.
[4] Dr. Epstein testified that a person who has hypertension without functional impairment and a sprained wrist suffers from one disease the sprained wrist.
[5] Claimants without functional impairment are being diagnosed with asbestos-related disease on the basis of X-ray and an exposure history and they bring lawsuits and recover damages.
[6] Several data-extraction forms in evidence were shown to have been prepared with significant deletions. Defendant's computer expert testified that during the data-extraction process 35,000 separate indications were recorded for each of 2,700 claimants, totalling over 28-million characters of information and rendering the deletion of any one item of information statistically insignificant. However, each piece of information is itself composed of a number of characters, and the information extracted is of varying importance to the validity of the process. For example, failure to record the presence of one of the diagnostic criteria or a description of the medical condition as stated by a treating physician, both of which had occurred, may well skew the conclusion as to that claim, whereas an error in the claimant's address may indeed be unimportant.