UNITED STATES DISTRICT COURT EASTERN DISTRICT OF NEW YORK
July 15, 2011
JOSEPH S. AND STEVEN W., ET AL., PLAINTIFFS,
MICHAEL F. HOGAN, IN HIS OFFICIAL CAPACITY AS COMMISSIONER OF THE NEW YORK STATE OFFICE OF MENTAL HEALTH, ET AL., DEFENDANTS.
The opinion of the court was delivered by: Cogan, District Judge.
MEMORANDUM DECISION AND ORDER
Before me is a motion in limine to preclude the testimony of plaintiffs' expert, Dr. Brant Fries. I deny the motion because I find that defendants' arguments go to weight rather than admissibility and because defendants have failed to show that plaintiffs violated their expert witness disclosure obligations.
In this action plaintiffs allege that hundreds of mentally disabled individuals have been discharged from psychiatric hospitals into nursing homes instead of being integrated into their communities. To support their allegations, plaintiffs seek to introduce the testimony of Dr. Fries, a statistical expert. Dr. Fries devised a sample of nursing home residents who would then be interviewed by plaintiffs' psychiatric expert to determine their ability to live in a community setting.
Defendants contend that instead of selecting a random sample and abiding by it, Dr. Fries cherry-picked nursing homes on the advice of counsel in order to focus on the facilities that were more likely to support plaintiffs' allegations. He also changed the sample midstream, they explain, when he reshuffled the original sequence and picked new residents when the selected ones were deemed too impaired.
Although the motion seeks to exclude only Dr. Fries, defendants also contend that the interview process was corrupted because Dr. Factor and plaintiffs' counsel inappropriately excluded residents who were selected as part of Dr. Fries' sample. The result, defendants conclude, was "an interview population that was overly inclusive of the most functional residents, i.e., those most likely to be deemed capable living in a less restrictive placement, and under-inclusive of residents most likely to need the higher level of care that a nursing home provides."
After taking Dr. Fries' and Dr. Factor's depositions earlier this year, defendants previously moved to strike Dr. Fries' expert report, or in the alternative, for additional discovery. Defendants claimed that they learned for the first time at the depositions that plaintiffs skipped more nursing home residents in their interviews than Dr. Fries' report admitted. This missing information, defendants contended, would have further supported their argument that plaintiffs cherry-picked the targeted residents and that the sample chosen is not representative of the entire resident population. Plaintiffs opposed the motion, arguing that defendants had had ample time for additional discovery and that, in any event, the expert report and supporting documentation did not omit any information.
At the last hearing, I denied defendants' motion without prejudice to raise the substance of the argument in a Daubert motion that they had already proposed. See Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). At the hearing, defendants argued that the deposition testimony of the experts left open questions, such as why certain nursing homes were skipped, but defendants could not explain why additional discovery would have supported their Daubert motion or, stated differently, how lack of discovery prejudiced them; if Dr. Fries could not explain the purported gaps, it would only support exclusion of the expert's testimony. The instant motion raises this argument anew.
Rule 702 of the Federal Rules of Evidence permits expert testimony if it will "assist the trier of fact to understand the evidence or to determine a fact in issue." In Kumho Tire Co. v. Carmichael, 526 U.S. 137, 149 (1999), the Supreme Court extended the reliability inquiry it set forth in Daubert, 509 U.S. 579, to all expert testimony. Defendants do not contend that expert testimony of a statistical expert would not be helpful, but argue that Dr. Fries' methodology was unreliable, thereby rendering his testimony inadmissible.
In a jury trial, questions of reliability often go to weight rather than admissibility. See 29 C. Wright, et. al., Federal Practice and Procedure 1d § 6264 (2011); see also Daubert, 509 U.S. at 596 ("Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence."). Some expert testimony, however, is so unreliable that it is more likely to confuse the jury than to assist it. See 29 Wright, et. al., supra, § 6264 ("The analysis often is similar to the balancing of the benefits and costs of evidence that is undertaken under Rule 403."). Such evidence, notwithstanding any probative value it may have, must be "gate-kept" or excluded by the Court. Cf. Daubert, 509 U.S. at 597 ("We recognize that, in practice, a gatekeeping role for the judge, no matter how flexible, inevitably on occasion will prevent the jury from learning of authentic insights and innovations. That, nevertheless, is the balance that is struck by Rules of Evidence designed not for the exhaustive search for cosmic understanding but for the particularized resolution of legal disputes.").
The dynamic is slightly altered in a bench trial. See 4 Weinstein's Federal Evidence § 702.02[b] (Mathew Bender). In making a Daubert motion, defendants are essentially asking me to gate-keep expert testimony from myself. See New York v. Solvent Chem. Co., No. 83-CV-1401, 2006 U.S. Dist. LEXIS 65595, at *3 (W.D.N.Y. Sept. 12, 2006) ("The primary purpose of the holdings in Daubert and Kumho Tire is to protect juries from being bamboozled by technical evidence of dubious merit.") (citation and quotation marks omitted). Of course if the expert testimony amounts to pure speculation, it would have no probative value and would not assist the fact finder, be it the Court or the jury. See In re Air Disaster at Lockerbie Scotland on 12-21-88, 37 F.3d 804, 824 (2d Cir. 1994). But short of that, expert testimony should be admitted so that the Court could have the benefit of live testimony and cross-examination to determine how much weight, if any, to give to the expert's conclusions. See In re Salem, 465 F.3d 767, 777 (7th Cir. 2006) ("[W]here the factfinder and the gatekeeper are the same, the court does not err in admitting the evidence subject to the ability later to exclude it or disregard it if it turns out not to meet the standard of reliability established by Rule 702."); Victoria's Secret Stores Brand Mgmt. v. Sexy Hair Concepts, LLC, No. 07 Civ. 5804, 2009 U.S. Dist. LEXIS 30458, at *17 n.3 (S.D.N.Y. Apr. 8, 2009) ("where a bench trial is in prospect, resolving Daubert questions at a pretrial stage is generally less efficient than simply hearing the evidence").
Indeed, without the risk of poisoning the jury with misleading expert testimony of limited probative value, see Daubert, 509 U.S. at 595, the Court can take in the evidence freely and separate helpful conclusions from ones that are not grounded in reliable methodology. Cf. BIC Corp. v. Far Eastern Source Corp., 23 F. App'x. 36, 39 (2d Cir. 2001) ("admission of evidence in a bench trial is rarely ground for reversal, for the trial judge is presumed to be able to exclude improper inferences from his or her own decisional analysis"); Schultz v. Butcher, 24 F.3d 626, 632 (4th Cir. 1994) ("in the context of a bench trial, evidence should not be excluded under [Rule] 403 on the ground that it is unfairly prejudicial"); Gulf States Utils. Co. v. Ecodyne Corp., 635 F.2d 517, 519 (5th Cir. 1981). It follows then that in a bench trial, the risk is with exclusion of expert testimony rather than with its admission -- it is exclusion that has the potential for an indelible impact on the record; if the appellate court disagrees that the expert's testimony was unreliable, a review for harmless error will be thwarted. See 11 Wright, et. al., supra, § 2285; see also Van Alen v. Dominick & Dominick, Inc., 560 F.2d 547, 552 (2d Cir. 1977) ("it may be the more prudent course in a bench trial to admit into evidence doubtfully admissible records, and testimony based on them").
Another practical consideration is the need for a "Daubert hearing." Where the parties raise factual issues, like they do here with respect to the data on which Dr. Fries based his opinion, an evidentiary hearing is useful and may indeed be necessary to determine whether the expert should be presented to a jury. See Humphrey v. Diamant Boart, Inc., 556 F. Supp. 2d 167, 173 n.3 (E.D.N.Y. 2008) (finding a hearing unnecessary where objections to the testimony do not raise a factual issue); 4 Weinstein's Federal Evidence § 702.02[b]. With a bench trial, the hearing is unnecessary; if I am not satisfied with the parties' examination of the witness at trial, I will have the opportunity to ask my own questions, without the risk of influencing the jury, to determine whether the expert's opinion is sufficiently reliable and thus one on which I could base my findings of fact.
The foregoing is not meant suggest that Dr. Frier's testimony occupies the narrow universe of expert testimony that is admissible only in a bench trial. Nor should it imply that the Court will shirk its responsibility of performing a full Daubert analysis. See Metavente Corp. v. Emigrant Savings Bank, 619 F.3d 748, 760 (7th Cir. 2010) (although in a bench trial reliability determinations need not be made before evidence is presented, it must still be made at some point); see also 29 Wright, et. al., supra, § 6266. But at this stage, it sets up for a limited review.
Unaffected by this forgiving standard of review is plaintiffs' duty under Rule 26(a)(2)(B) of the Federal Rules of Civil Procedure to provide a detailed and complete expert witness report. See 6 James W. Moore et al.,Moore's Federal Practice § 26.23[b][ii]. Failure to make a full disclosure will preclude the expert's testimony unless the omission was harmless or justified. See Fed. R. Civ. P. 37(c)(1); Design Strategy, Inc. v. Davis, 469 F.3d 284, 297 (2d Cir. 2006). Defendants continue to maintain that Dr. Fries did not disclose the identities of all the residents that he initially selected but did not count towards the final sample. The defendants' statistical expert examined the residents who defendants knew were skipped and those who were interviewed and concluded that the chance of the two populations being the same in terms of their ability to live in a community setting to be only .6 percent. Defendants argue that they were prejudiced because had plaintiffs provided all the information, that number might have been even smaller.
As I noted above, at the last hearing, I denied defendants' motion for additional discovery but observed that if plaintiffs produce new evidence to explain any gaps, I would allow more discovery and may even consider sanctions. Plaintiffs have not revealed anything new. They are adamant that Dr. Fries' report and its annexed exhibit, along with several spreadsheets that were sent three days later, went beyond the required disclosure obligations as Dr. Fries did not have to disclose which residents were not available for interviews or show his pre-shuffled list.
Plaintiffs' contention is a stretch given that measuring the value of any statistical sample begins with comparing subjects that were not selected for the sample with those who were, and that the proponent of the expert testimony must disclose not only the data on which he relied but also that which he considered. See Schwab v. Philip Morris USA, Inc., No. 04-CV-1945, 2006 WL 721368, at *2 (E.D.N.Y. Mar. 20, 2006) (Weinstein, J.) (quoting Fed. R. Civ. P. 26(a)(2) advisory committee's note). But plaintiffs provided all the data, which defendants do not appear to dispute, arguing instead that they should not have to "decipher" the dozen spreadsheets that were turned over to them; the data, they appear to argue, must be contained within the report.
Reviewing the very manageable spreadsheets, it is apparent that even if not entirely intuitive, their production is not anything like a "document dump." Defendants did not serve plaintiffs with interrogatories to explain any portion of Dr. Fries' report or his supporting documents; they did not appear to ask clarifying questions at his deposition more than seven months after the expert's disclosures; nor do they claim to have mentioned their difficulties when they appeared for a conference before Magistrate Judge Gold a month after Dr. Fries' deposition to seek an extension of expert discovery. Instead, they rely on cases that confirm Rule 26(a)(1)(C)'s requirement to compute damages rather than merely provide "undifferentiated financial statements." Design Strategy, 469 F.3d at 295; see also Great White Bear, LLC v. Mervyns, LLC, No. 06 Civ. 13358, 2008 WL 2220662, at *2-3 (S.D.N.Y. May 27, 2008).
These cases are inapposite; they do not suggest that the expert must explain each data point in his report, and stand only for the limited proposition that the defendant must be aware of the damages that are asserted against him and the way in which they were computed. Nothing in Rule 26(a)(2) requires the expert report to contain all of the supporting data within its four corners so long as the report clearly explains its methodology, which Dr. Fries' report appears to do. Cf. Walsh v. Chez, 583 F.3d 990, 994 (7th Cir. 2009) (the purpose of expert reports "is not to replicate every word that the expert might say on the stand" but "to convey the substance of the expert's opinion . . . so that the opponent will be ready to rebut, to cross-examine, and to offer a competing expert"). Defendants have therefore failed to show that any sanction, let alone the "drastic" and "extreme" sanction of preclusion, is warranted. See Outley v. New York, 837 F.2d 587, 590-91 (2d Cir. 1988).
Defendants' remaining arguments, about the reliability of Dr. Fries' methodology and its execution, go to weight rather than admissibility. They catalog a number of purported deficiencies in Dr. Fries' report; I address only the main arguments. First, defendants argue that Dr. Fries intentionally picked certain nursing homes rather than random residents. These reasons, according to defendants, included a prior relationship that plaintiff's counsel had previously formed with the administrators of one of the facilities during which he learned about its population.
Plaintiffs dispute this contention, claiming that nursing homes were picked through a commonly used and accepted statistical technique known as geographical clustering, where the subjects are selected randomly from different regions. They explain that this method was employed in order to facilitate Dr. Factor's interview process. In his affidavit, Dr. Fries affirms that he had no information to intentionally select nursing homes, and that "[t]he only participation Plaintiffs' counsel had in this selection of nursing homes was providing advice as to which nursing homes were in the same geographic area."
Defendants nevertheless maintain that the process was rigged. In support, they cite Dr. Fries' deposition testimony where he concedes only that certain nursing homes were picked because of the cost of conducting interviews at geographically disparate locations. He explains that this decision was reached together with Robert Bearden -- plaintiffs' former counsel and, since his recent appointment by the Governor, Chairman of the New York State Commission on Quality of Care and Advocacy for Persons with Disabilities. Defendants have not produced an admission by Dr. Fries or plaintiffs' counsel that nursing homes were intentionally chosen because of the resident population. Dr. Fries' testimony and a Rule 30(b)(6) testimony of one of plaintiff's attorneys only confirms that it was practical considerations that played into any intentional selections of nursing homes.*fn1
Defendants concede that geographical clustering is not per se unreliable; they argue instead that the nursing homes, and more importantly, residents, should still have been selected at random after the clustering method was applied. Their point is well taken. Certainly, some of Dr. Fries' testimony is troubling, such as his explanation about choosing one of the facilities: "[I] spoke to Mr. Bearden about it and we said that [nursing home] would be good to pick in this area, and we looked at the facilities and said that appeared to be a good one." But without conclusive proof that the sampling was designed to select "favorable" nursing homes, I will not preclude Dr. Fries' testimony at this stage. The evidence suggests that there was at least some randomness to the selection of most of the facilities. It may be, however, that after further probing of the expert on the stand or after hearing the testimony of defendants' expert that the nursing homes picked, whether for convenience or less neutral reasons, is not reliable statistical evidence, I will conclude that Dr. Fries' testimony cannot be given any weight.
Next, defendants contend that Dr. Fries added nursing homes after plaintiffs had already begun the interviewing process and dropped at least one resident after having already interviewed him. In his deposition, Dr. Fries explained that certain nursing homes that were originally selected through random sampling had to be excluded because he was informed -- unclear by whom from the excerpts provided to the Court -- that certain facilities would not allow access for interviews. He also explained that reshuffling had to be done because unbeknownst to him, the software he used organized by age the list of residents from which the sample would be drawn. Because this compilation would not result in a random sample, Dr. Fries had to regenerate the sample after the interview process had already begun. Residents who were not on the reshuffled list whom Dr. Factor already interviewed were excluded.
As for other residents who were interviewed but excluded, plaintiffs explain -- although the testimony to which they cite is not clearly supportive -- that these "anecdotal" interviews were conducted to provide Dr. Factor with additional information about residents' lives in the nursing homes. Again, I have questions about an interview process that was purportedly designed to save cost through geographic clustering but sent Dr. Factor to a separate nursing home solely to conduct an anecdotal interview. But this is a question I expect to be addressed at trial and cannot be the basis for exclusion. The same is true of the reshuffling and responses to some nursing homes that turned plaintiffs away -- whether sufficient randomness was preserved should be inquired on cross-examination, which may well strip the expert's testimony of any probative value.
Defendants also complain that plaintiffs skipped residents who were deemed unable to be interviewed. Plaintiffs' Rule 30(b)(6) witness testified that nursing homes would not produce some residents because they were too "impaired." Defendants point to a question of fact when they dispute plaintiffs' explanation, citing the testimony of a social worker at one of the nursing homes who testified that plaintiffs did not face any resistance. They also raise the valid argument of strong non-response bias resulting from the exclusion of residents for a characteristic that bears directly on the inquiry for which the sample was designed. Finally, they highlight plaintiffs' failure to confirm that the non-response reason was accurate to properly assess the bias. Again, although they suggest it, defendants have not submitted conclusive proof that Dr. Fries' sample intentionally excluded residents who plaintiffs knew could not be integrated into the community because of their impairment.
Additional non-response bias, defendants argue, was introduced when plaintiffs notified residents about the lawsuit before they conducted the interviews. Dr. Fries affirms that the notification was consistent with the principle of informed consent and did not bias the study. Defendants' expert disagrees. The parties also disagree about the response rate of the study; defendants put it at below 50% while plaintiffs state that it was 69.2%. The difference is in what the parties consider to be the sample. Plaintiffs appear to have the better of this argument as it is the final, reshuffled list that they propound as their sample, and it is the response rate of those individuals that matters rather than any larger universe of residents from which Dr. Fries selected the sample. I will nevertheless consider both arguments at trial.
Finally, defendants misconstrue Dr. Fries' testimony when they argue that he deemed the order in which residents were interviewed (which was not followed) to be important. In the cited excerpt that defendants plainly submit out of context, Dr. Fries appears to explain why random, as opposed to ordered, selection of human subjects is necessary. In his affidavit, he confirms that "[t]he order in which Residents were interviewed did not matter at all." (emphasis added).
In sum, defendants' objections do not add up to preclude Dr. Fries' testimony in a bench trial. Some have more merit than others, and I expect it will be the former that will get more attention on cross-examination. It is worth noting, however, that the distinction I have drawn at various points, between the notion that subjects were intentionally picked for their known characteristic rather than for the more impartial reason of cost and convenience, may ultimately have no significance. I make that distinction now only because I assume that conclusive proof of gamesmanship would on its face show a purportedly representational sample to lack any representative value, and the expert's testimony to be clearly without value. But it is possible that even practical considerations can render statistical evidence completely unreliable, that given the circumstances -- non-cooperation, limited resources and New York's geography, impaired residents and inability to confirm their impairment, the need to provide informed consent and the revelation of a lawsuit -- a random survey of any probative value simply could not be conducted here. I will make that determination at trial.
For the foregoing reasons, defendants' Motion to Preclude Expert Testimony  is DENIED.
Brian M. Cogan U.S.D.J.