BOUDIN, Circuit Judge.

Lawrence Schneider was convicted on multiple counts of mail and wire fraud and now appeals, presenting two difficult issues. One concerns the district court's refusal to allow Schneider's expert medical testimony proffered as pertinent to his state of mind; the other relates to the denial of Schneider's new trial motion based on jury-questionnaire information about one of the jurors. On both issues, we agree with the district court's outcome and affirm.

What Schneider did was largely undisputed at trial, although his state of mind was very much in dispute. Schneider ran two businesses from his home in Warwick, Rhode Island, engaged in buying and selling real estate and restoration of old houses. In mid-1990, Schneider was under financial pressure and unable to pay his bills. Starting in May 1990 and continuing through October 1990, he ordered on credit a variety of goods, including computer equipment and jewelry, the total value of which exceeded $200,000.

Beginning shortly thereafter, Schneider began to resell the same goods and pocket the money. Between June 1990 and October 1990, Schneider placed 25 classified advertisements in the Providence daily newspaper, offering to resell the goods in question at deep discounts. He gave purchasers false stories about the origin of the goods or how he acquired them. Schneider paid nothing to his suppliers. The government describes this as a classic "bust-out" scheme. See, e.g., United States v. DeVincent, 632 F.2d 147, 149 (1st Cir.), cert. denied, 449 U.S. 986 (1980).

In December 1994, Schneider was indicted and charged with six counts of wire fraud and three counts of mail fraud based on his use of such facilities in the conduct of his scheme. 18 U.S.C. Section(s) 1343, 1341. The government's theory of fraud was that Schneider never intended to pay for the goods but ordered them with the aim of reselling them immediately and bilking the sellers. In addition, the government also relied upon an affirmative misstatement, made on several of Schneider's credit applications, that he had 35 employees, which was untrue.

Schneider did not dispute the core events described by the government, but argued that he did not have the requisite specific intent to defraud. He pointed out that he had businesses, a home and a good credit rating, so his behavior made no sense as a rational criminal act; it followed, he argues on appeal, "that his behavior is inexplicable absent an understanding of the defendant's mental status." At trial, he sought to introduce medical evidence to explain his behavior.

The proffered evidence was testimony from two doctors. One was Dr. Wartenberg, who specialized in internal medicine and addiction medicine; he proposed to testify that Schneider's capacity and judgment were significantly impaired by misprescription and overprescription of medical drugs during the relevant period. A psychiatrist, Dr. Roth, was also prepared to testify to impaired judgment, based on chemical dependency and major depression with probable mania. This testimony, of which more will be said later, was taken as an offer of proof outside the presence of the jury.

In due course, the district judge ruled that the testimony would not be admitted. The court's explanation is summed up in the final sentences of a longer oral ruling:

It seems to me that the evidence does not suggest that the Defendant did not act purposefully, that to accept this evidence as a defense [of lack of] mens rea manipulates the concept of intent beyond the intent required and that under all the circumstances, having considered all of the evidence, the Court will sustain the government's objection to any offer of that proof before the jury and Defendant may have an exception.

As a preface to this conclusion, the district court described in some length the opinion in United States v. Pohlot, 827 F.2d 889 (3d Cir. 1987), cert. denied, 484 U.S. 1011 (1988), discussing the pertinence of psychiatric testimony in a murder-for-hire case.

After this evidentiary ruling, the case was tried in May 1995. The jury deadlocked on eight counts and acquitted on one count. In October 1995, Schneider was retried on the remaining eight counts. The district court stood by its earlier ruling that the medical testimony was not admissible. This time the jury convicted on all eight counts.

Following trial, two jurors contacted the district judge, expressing concerns about another member of the jury panel. On inquiry, the district judge discovered that after the verdict, the juror in question had said that she wanted to get out of jury service because she was "mental." A review of that juror's questionnaire revealed that the juror had affirmatively answered the question: "Do you have any physical or mental disability that would interfere or prevent you from serving as a juror?" On the reverse side of the questionnaire, the juror had written the following:

I have an upsetting emotional handicap since my children have been with the DCYF [a Rhode Island state agency], and the strain is sometimes too great for me and the walking to the bus stop especially in inclement weather would be kind of hard at this time. I am presently looking for housing for my family.

After an investigation of the juror and a personal interview by the district court, the court ruled that it was "completely satisfied that the juror was competent and mentally capable at the time of the trial . . . ." Schneider nevertheless sought a new trial on the ground that he would have used his peremptory challenges differently if he had known about the information in the questionnaire. The district court denied the motion.

In due course, Schneider was sentenced to 33 months in prison (he had a prior conviction for a "bust-out" offense some years before), and was ordered to make restitution of about $220,000. On this appeal, he claims as error the refusal to admit the medical testimony; and he argues that a new trial should have been granted because he could have used a peremptory challenge to dismiss the juror if he had known of the information in the questionnaire.

We begin with the harder of the two issues and ask whether the district court erred in refusing to admit the medical evidence. The standard of review depends upon what has been decided: rulings of law are reviewed de novo while review is typically deferential on so-called "mixed questions." Bergersen v. Commissioner of Internal Revenue, ___ F.3d ___, 1997 WL 120530, at *5 (1st Cir. Mar. 21, 1997). Here, a legal issue — the meaning of a federal statute — stands at the threshold.

In 1984, Congress enacted the Insanity Defense Reform Act, 98 Stat. 2057, redefining insanity and making it an affirmative defense to be proved by clear and convincing evidence. See 18 U.S.C. §(s) 17. The statute also states: "Mental disease or defect does not otherwise constitute a defense." Id. Section(s) 17(a). Elsewhere, the government has argued that this quoted language is meant to rule out any mental-condition defense or testimony that does not reach the heights of a properly pleaded insanity defense. E.g., Pohlot, 827 F.2d at 890.

In this case, Schneider did not offer an insanity defense, but he argues that his medical evidence is pertinent in helping a jury decide whether he had the requisite state of mind necessary for the offense charged. The government, in turn, says that a number of courts have admitted psychological evidence to negate specific intent but the First Circuit has suggested otherwise; that the First Circuit's view is consistent with the final sentence of the statute just quoted; and that it is nevertheless unnecessary to resolve the issue in this case because the evidence here was not relevant in any event.

Aside from the final sentence of section 17(a), in principle there should be no bar to medical evidence that a defendant, although not insane, lacked the requisite state of mind. As LaFave and Scott say:

The reception of evidence of the defendant's abnormal mental condition, totally apart from the defense of insanity, is certainly appropriate whenever that evidence is relevant to the issue of whether he had the mental state which is a necessary element of the crime charged.

1 LaFave Scott, Substantive Criminal Law Section(s) 4.7, at 530. The circuits that have considered the question have taken this view. See United States v. Marenghi, 893 F. Supp. 85, 89 (D. Me. 1995) (collecting cases). After all, if state of mind is a potential issue — as it is in most but not all criminal cases — why should expert medical evidence be excluded out of hand?

We doubt that the final sentence of section 17(a) was intended to exclude mental-condition evidence short of insanity. Pohlot canvassed the arguments and legislative history at length, and concluded (1) that the statute does not preclude a defendant from offering evidence to negate a requisite state of mind, 827 F.2d at 903, but (2) that — apart from such a negation — it does preclude any other new and different defense of diminished responsibility to excuse or mitigate the offense. Id. at 905-06. Pohlot's analysis seems to us persuasive on both issues.

Similarly, our own decisions in White and Kepreos were not intended to establish a general rule that mental-condition evidence is always inadmissible except in relation to insanity. In Kepreos, the court held that the specific psychiatric testimony involved was misleading and of doubtful utility, 759 F.2d at 964; in White, where no coercion defense was offered, the court rejected psychiatric evidence that a defendant engaged in a drug crime because "she was unable to resist her mother's request for assistance . . . ." 766 F.2d at 24. These cases largely turn upon their facts. See Marenghi, 893 F. Supp. at 88-91.

Once past the threshold of section 17(a), the situation becomes more difficult for the defendant. The specific medical evidence offered may still be irrelevant to the requisite intent, White, 766 F.2d at 24, or probative value may be substantially outweighed by confusion or delay. Fed.R.Evid. 403; Kepreos, 759 F.2d at 964. Finally, if the evidence is expert testimony, it must meet the further requisites of scientific reliability and helpfulness to the jury. Fed.R.Evid. 702; Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 589-91 (1993).

In deciding such issues — relevance, confusion, reliability, helpfulness — the district court has a comparative advantage over an appeals panel. The issues typically involve unique fact patterns and judgments of degree, and the district judge is closer to the case. Thus, so long as there is no misstatement of the legal standard and the result reached is not clearly unreasonable, the district judge's ruling is usually respected. United States v. Shay, 57 F.3d 126, 132 (1st Cir. 1995).

Against this background, we return to the testimony in this case. Dr. Wartenberg had examined Schneider for approximately an hour and a half and reviewed his medical records, including medicines prescribed for him over a three or four-year period. In his offer of proof, the doctor said that the prescriptions were medically inappropriate and that the drugs as prescribed "would impair intellectual function in a variety of ways," produce blackouts, roller coaster highs and lows, and permit misperception and delusion. Dr. Wartenberg summarized his view as follows:

My opinion is that to a degree of reasonable medical certainty that Mr. Schneider's intellectual capacity, cognitive function, ability to make executive judgments and decisions would have been impaired by that level of drug prescribing.

On cross-examination, government counsel secured an admission that Schneider during this period could engage in "activity that is planned to carry out a purpose." Quite properly, neither side asked Dr. Wartenberg the "ultimate" question whether Schneider had intended to defraud, such ultimate questions to experts now being forbidden as to mental state in a criminal case under Fed.R.Evid. 704(b), which is a companion amendment to section 17. See United States v. Meader, 914 F. Supp. 656, 658 (D. Me. 1996).

The expert psychiatrist, Dr. Roth, also offered testimony after interviewing Schneider and examining records. He said that Schneider suffered from chemical dependency and "major depression with probable mania." He said also that these conditions "impair[ed] . . . judgment." He declined to express a view on Schneider's ability to form a purpose, saying that he (Dr. Roth) was not there at the time, and that terms like "purposeful" posed a metaphysical question.

As we read the district judge's ruling, he deemed the medical testimony both irrelevant and misleading. Our own view is that the evidence may have been relevant but only to a limited degree, and that it had a substantial capacity to mislead the jury. And we think that the district court clearly would exclude the evidence under Fed.R.Evid. 403 even if told to view it as having limited relevance. To remand, in order to make the district judge say this explicitly, is a waste of time. Let us take these points in turn.

Relevance, the ordinary starting point for admissibility, Fed.R.Evid. 401, is a close issue. Here, the government urged that Schneider had ordered the goods, expecting never to pay for them but to resell them and pocket the money and thereby "intended to deceive." Schneider's answer was that he lacked this state of mind and the medical evidence helped him support this view, that is (in the language of the rule), that the medical testimony made his intent to defraud "less probable than it would be without the evidence." Fed.R.Evid. 401.

The sum of the doctors' evidence is that Schneider was depressed, that he had impaired judgment (due to his depressed state and overmedication), and that he was subject to blackouts. This might not appear at first to go very far in negating his capacity to deceive, especially as Schneider's scheme continued over several months. This explains the district court's view that the medical evidence did not negate an intent to deceive but instead amounted to a forbidden claim in mitigation.

Still, evidence may be "relevant" under Rule 401's definition, even if it fails to prove or disprove the fact at issue — whether taken alone or in combination with all other helpful evidence on that issue. In the latter instance, the judge could direct a verdict if the issue were essential, but not against the defendant in a criminal case. Schneider's best argument is, therefore, that his medical evidence did go some distance to negate intent to deceive and so was relevant.

Where evidence goes "some distance" but manifestly not far enough, it may be tempting to say that it is not relevant. Frankly, Pohlot appears to us to take this course by ruling that the psychological evidence of impairment was irrelevant because "by his own admission," Pohlot had "finalized an agreement to have his wife murdered and this 'purpose' to hire someone to kill his wife was enough regardless of whether he 'psychologically . . . understood the full consequences of this activity.'" 827 F.2d at 889.

But we have some doubt that this usage comports with Rule 401's definition quoted above. Nor do we think it helpful in a case like ours to ask, as Pohlot did, whether the conduct was "purposeful." Pohlot said that the ordinary mens rea requirement is satisfied "by any showing of purposeful activity, regardless of its psychological origins." 827 F.2d at 904. But it is quite possible purposefully to order goods on credit, and later not pay for them, without having an intent to deceive.

Pohlot's other theme is the capacity of evidence of this kind to mislead. Congress raised the hurdle for an insanity defense and barred a new diminished capacity defense that courts were beginning to invent. Yet the evidence offered, both here and in Pohlot, suggests that the defendant was temporarily out of his mind (even though not insane under section 17(a)) and that his crime was mitigated by his psychological condition. Such evidence tends to reintroduce the very concepts that Congress wanted to exclude and thereby to mislead the jury.

In weighing relevance against confusion, we will assume for the present purposes that this balance is to be struck under Rule 403. The government does not now challenge the evidence as unreliable under Daubert — only its pertinence to this case. Nor do we propose to decide here the issue explicitly reserved in Shay, namely, whether and when the judge may exclude otherwise relevant expert evidence on the ground that it will not "assist" the jury. See Shay, 57 F.3d at 132-33. It is enough that exclusion of the evidence here under Rule 403 is eminently justifiable.

The evidence, as we have said, is of limited relevance: showing "impaired" judgment might help piece out a lack of deceit claim but falls well short of sufficient proof. At the same time, the expert testimony offered here could easily mislead the jury into thinking that such a medical condition amounts to temporary insanity or ameliorates the offense. The instructions required here to guard the jury on this score would likely have gone very far to eliminate any use the evidence might otherwise have to the defendant.

Thus, we conclude that the district court was free to exclude this evidence on the ground that its capacity to mislead the jury substantially outweighed its limited relevance. Since exclusion was permitted but not required, we could remand for an explicit finding under Rule 403. Shay, 57 F.3d at 134. But given the district court's evident view of the matter, our limited disagreement with the district court's reasoning would hardly alter the district court's desire to exclude the evidence. A remand would serve no purpose. See United States v. Dolloph, 75 F.3d 35, 38 (1st Cir.), cert. denied, 116 S.Ct. 1866 (1996).

Although our concerns about such evidence are considerable, we shrink from any generic rule that would forbid the district courts from resolving admissibility case by case. Offenses differ from each other; the medical evidence, taken alone and in combination with other evidence, is going to vary widely; and this is an area in which everyone is still learning. In the spirit of Daubert, we rely heavily on the wise superintendence of the trial court. 509 U.S. at 592-93. In this case we have no quarrel with how it was exercised.

This brings us to Schneider's second claim of error, namely, that the district court erred in refusing to grant a new trial. The basis of the new trial motion was (so far as pertinent here) that the juror questionnaire would, if disclosed, have revealed mental or emotional problems pertaining to the juror in question, perhaps prompting the use of a peremptory challenge to remove the juror. Again, in considering the district court's action, our standard of review depends upon what the district judge decided.

The district court ruled that Schneider had waived his objection because he failed to move for disclosure of the questionnaire answers prior to empanelment. We took such a view in United States v. Uribe, 890 F.2d 554, 561 (1st Cir. 1989), but stopped just short of a definitive finding that the Rhode Island federal juror selection plan permitted such access to jury questionnaires. Id. In affirming, Uribe relied heavily on the alternative ground that the defendant had suffered no prejudice from the nondisclosure. Id. at 562.

On this appeal, the parties seek to litigate at length whether Schneider did have effective access to the jury questionnaires, which the governing statute treats somewhat differently than jury lists. Neither the statute nor the Rhode Island plan are crystal clear about access to questionnaires. See 28 U.S.C. Section(s) 1867(f); Davenport, 824 F.2d at 1515. We have previously interpreted the statute not to allow inspection of questionnaires "solely to aid in the voir dire process." Jewell v. Arctic Enterprises, 801 F.2d 11, 13 (1st Cir. 1986).

Under these circumstances, we are reluctant to rest our decision on waiver or forfeiture. Assuming arguendo that Schneider might have sought access to the questionnaire based on a showing of need, before trial he had no basis for such a motion. This is not a case where, prior to empanelment, the defendant knew or should have known of a potential problem with a juror and failed to ask for the questionnaire. E.g., United States v. Aponte-Suarez, 905 F.2d 483, 492 (1st Cir.), cert. denied, 498 U.S. 990 (1990).

On the other hand, Schneider has abandoned on appeal any effort to show actual prejudice (e.g., by pointing to characteristics of the juror that would likely have caused her to take an adverse view of Schneider or to fall under the sway of other jurors). Such contentions were made in the district court but were rejected and are not now pursued. Schneider's position instead is that prejudice is not required, and he relies on our own statement in United States v. Vargas, 606 F.2d 341, 346 (1st Cir. 1979):

[T]here is little doubt that if the court or prosecution deprives a defendant of his right to the effective exercise of peremptory challenges, it would, without more, be grounds for a new trial.

In a number of cases involving interference with peremptory challenges, the reviewing court has treated proof of prejudice as unnecessary. See United States v. Annigoni, 96 F.3d 1132, 1141 (9th Cir. 1996) (en banc) (collecting cases). But such cases normally involve a deliberate denial or interference. Without automatic reversal, such conduct could rarely be corrected; by definition, lawyers use peremptories where a challenge for cause will not work.

By contrast, a failure of the district court's screening apparatus to uncover some piece of potentially useful information in a raft of jury questionnaires is, if "error" at all, a mistake of quite a different kind. At worst, such a mistake might resemble the failure of the judge to ask adequate voir dire questions, where we have insisted on a showing of prejudice. United States v. Anagnos, 853 F.2d 1, 3-5 (1st Cir. 1988). See also 9A Wright Miller, Federal Practice Procedure 2482 at 115 n. 8 (2d ed. 1995).

We need not try to draw a definitive line between cases where a showing of prejudice is required and cases where it is not, as this one so clearly falls on the former side of the line. There was no direct inference with a peremptory challenge, cf. United States v. Cambara, 902 F.2d 144, 147-48 (1st Cir. 1990), and no indication that useful information was deliberately withheld. Here, as in Vargas (where a juror withheld information), "[a] new trial [without a showing of prejudice] would be a windfall for the defendant" without much countervailing benefit. 606 F.2d at 346.

Affirmed.