Defendant appeals from the judgment for plaintiff in this medical malpractice action. We reverse in part.

Plaintiff1 suffered from AIDS, and defendant was his treating physician. In the course of treatment, defendant prescribed a sulfa-based medicine for plaintiff, despite the fact of which defendant reasonably should have been aware that plaintiff was allergic to sulfa preparations. Plaintiff suffered severe and protracted diarrhea and weight loss as a result. An extensive period of hospitalization was required before the condition could be remedied, and defendant at one point advised plaintiff that he would die in a short time if the disorder could not be controlled.

Plaintiff stated claims for negligence and failure to obtain informed consent. Among the damages that he sought were for emotional distress, resulting from his fear of imminent death, his weight loss and the deterioration of his underlying condition, and his concern over his young daughter’s future if he died. Plaintiffs wife had died of AIDS a short time before. Judgment was entered for plaintiff on the jury’s verdict, awarding him $75,000 in noneconomic damages and economic damages in excess of $100,000. The damages were not further segregated.

Defendant appeals. His only assignments of error are that a circuit court judge erred in a pretrial ruling, and that the trial court erred in ratifying the earlier ruling, refusing to compel production of the records of plaintiffs psychologist, Dr. Drucker. Defendant sought the records principally to assist in his “investigation of the causes and measure of plaintiffs alleged emotional damages.” Defendant’s theory is that there were factors that predated or coexisted with the consequences of the medical negligence that could have independently caused or contributed to plaintiff’s emotional distress.

Although defendant argues that the pretrial ruling and the ruling by the trial judge were erroneous for independent reasons, we understand the later ruling to have been a *46more or less pro forma ratification of the earlier one. Consequently, if the pretrial ruling was erroneous, no independent analysis of the trial judge’s decision is necessary.

ORCP 36 B(l) provides:

“For all forms of discovery, parties may inquire regarding any matter, not privileged, which is relevant to the claim or defense of the party seeking discoveiy or to the claim or defense of any other party, including the existence, description, nature, custody, condition, and location of any books, documents, or other tangible things, and the identity and location of persons having knowledge of any discoverable matter. It is not ground for objection that the information sought will be inadmissible at the trial if the information sought appears reasonably calculated to lead to the discovery of admissible evidence.”

The court explained its denial of the motion to compel in a letter to counsel, stating as material:

“I have reviewed the records of Dr. Drucker submitted to me by [plaintiffs attorney]. The records appear to cover a period of time from May 22, 1990 to November 14, 1990. Much of the material appears to address the marital difficulties Mr. and Mrs. Baker experienced before her death on August 19,1990. Some of the records deal with the aftermath of the death.

“The complaint in the present case alleges damages suffered by Mr. Baker following the prescribed drug being [taken] by Mr. Baker commencing on November 28, 1990. Although one could conclude that Mr. Baker was experiencing some depression it appears that Dr. Drucker was dealing with issues of anger, conflict resolution, adjustments with HIV disease progression, and issues of their daughter for future custody etc. I have also concluded that some of the materials that may be relevant would be highly prejudicial to the plaintiff and such prejudice would outweigh the probative value of the evidence.”

Although our review of the court’s decision is for abuse of discretion, the court’s discretion must be exercised in accordance with, and our review must necessarily take into account, the applicable legal principles governing the right to discovery. There is no dispute that Drucker’s records, as they pertained to plaintiff himself, were not privileged, because plaintiff put his own psychological condition into question by *47claiming emotional distress damages. It is also clear, as defendant argues, that the information sought was reasonably calculated to lead to the discovery of admissible evidence relevant to the nature and cause of plaintiffs psychological and emotional problems. Doran v. Culver, 88 Or App 452, 745 P2d 817 (1987), rev den 305 Or 102 (1988).

Defendant argues that the trial court failed to follow the basic principles relating to whether discovery is to be allowed. Among the problems that defendant identifies is that the court erred by denying discovery on the basis of its assessment under OEC 403 of whether material in the records would be admissible as evidence at trial. That approach, according to defendant, was not only premature, but also ran afoul of the express provision of the rule that production is not to be denied on the ground that the information sought may not be admissible in itself, if it is reasonably calculated to lead to the discovery of admissible evidence. We agree.

Plaintiff argues that the records contain non-discoverable material. In particular, he contends that there is privileged material in them relating to plaintiffs wife and daughter, and material that the parties agree is not to be part of the case pertaining to plaintiffs sexual orientation and the manner in which he became infected with the HIV virus. However, defendant argues that he does not seek materials of those kinds, and that they can be “redacted” from the records that are produced. We agree that the described material can and should be redacted. See ORCP 36 C. However, we conclude that the denial of any discovery of the records was error.

Plaintiff argues that any error was harmless. Plaintiff explains that, although defendant was aware of his claim and his physical and mental problems and had information and access to many alternative sources of information relating to those problems, defendant did not meaningfully pursue the theory before or at trial that he contends he sought the records to assist him in developing. Plaintiff points out that defendant did not call Drucker as a witness or depose him, did not take any responsive action to available and known reports and evidence relating to plaintiffs mental condition, and did not inquire about plaintiffs antecedent emotional problems in his cross-examination of plaintiffs friends and family *48members who offered testimony relevant to plaintiffs emotional state on direct examination. Further, at plaintiffs deposition, taken to perpetuate testimony, plaintiff gave some information about psychological problems that predated the use of the medication, which would have supported defendant’s theory. However, plaintiff contends, defendant made no significant use of that information at trial.

Defendant responds that plaintiffs argument “is based upon second guessing of defendant’s trial strategy,” although the records were sought for the very purpose of enabling defendant “to marshall and assess all evidence potentially relevant to this case prior to developing a trial strategy. ’ ’ Defendant concludes that “ any finding of harmless error would have to be based upon pure speculation as to the potential effect of the documents.”

Although the question is close, we agree on balance with defendant. This is a classic chicken-and-egg problem. It is equally possible that defendant did not pursue the theory because he did not obtain vital information that the records might have contained, or that nothing in the records could have affected a tactical decision not to pursue the theory. The answer can only be based on speculation. However, the very fact of that speculativeness militates in defendant’s favor: he was denied discovery to which he was entitled, and there is no basis for concluding one way or the other whether that denial might have affected his trial strategy or the result of the trial. Consequently, we cannot affirmatively conclude that the error was unlikely to have affected the result and cannot hold that it was harmless.

The question that remains is the appropriate scope of our remand. Plaintiff argues that, in the event of a remand, it should be limited to the issue of noneconomic damages. Defendant argues that it should be plenary, applying to all issues. The parties cite and seek to analogize this case to myriad decisions of the Supreme Court and this court dealing with the appropriate scope of remands. Those cases culminate in and are discussed by the majority and dissenting opinions in Wells v. Marleau, 79 Or App 784, 720 P2d 409, rev den 302 Or 159 (1986). The majority noted there that the “[ajppellate courts have given diverse treatment to the questions of whether and what issues must be retried on remand *49other than the specific issue on which the trial court erred.” Id. at 787. We concluded in Wells that there was not a sufficient connection between the punitive damages issue, which gave rise to the error, and the issues of liability and compensatory damages to necessitate that the latter be retried. We accordingly limited our remand to the punitive damages issue.

In Wells, we implicitly questioned the continuing vitality of Maxwell v. Port. Terminal RR. Co., 253 Or 573, 456 P2d 484 (1969), where the court stated that the general rule was to remand on all issues, with limited remands being reserved for exceptional cases. Even in Maxwell, however, the court defined the exceptional situations as including those where “the issue or issues to be eliminated are no longer viable issues in the case.” Id. at 577.

We conclude that, under either the Wells or the Maxwell formulation, this case is an appropriate one for a limited remand. The denial of discovery bears directly only on emotional distress damages, which could only have been a part of the jury’s award of noneconomic damages. The liability and economic damages findings are not connected with the emotional distress or discovery issues, and the issues of liability and economic damages, to which no independent error was assigned by defendant on appeal, “are no longer viable issues in the case.”

Defendant attempts to link the discovery denial to the issues other than noneconomic damages, by asserting that Drucker’s records might have revealed bases for challenging plaintiffs credibility and, hence, have had a possible effect on the liability finding. However, defendant does not specifically explain the possible connection in his argument to us. Particularly in view of the fact that these records do not concern plaintiffs treatment by defendant, we cannot identify any reason beyond sheer speculation to believe that the records might have led to the discovery of bases for a credibility challenge affecting the liability issues. It is true, as we indicated in discussing the affect that the records could have on the issue of emotional distress damages, that defendant could not know what the records might contain or lead to without having had the opportunity to review them. However, the possible connection between the psychologist’s *50records and defendant’s emotional state is palpable and self-evident. That is not true of the relationship that defendant hypothesizes between the records and plaintiffs credibility. We remand only on the issue affected by the error.

Reversed and remanded on issue of noneconomic damages; otherwise affirmed.

dissenting.

I agree with the majority that the trial court erred in denying the motion to compel production of the records. I do not agree that the error warrants reversal, and I respectfully dissent.

To begin with, the single error that defendant contends occurred1 should be put into perspective. The jury found that defendant was liable. That finding was abundantly supported by the evidence, and defendant makes no assignment of error that bears in any direct way on liability. The jury also found that defendant’s negligence caused plaintiff economic damages of more than $100,000. That finding is also not the subject of or affected by the error that defendant assigns. Finally, the jury awarded noneconomic damages of $75,000, which could have included a component for the mental distress that plaintiff alleged that the malpractice caused. The error that defendant asserts relates only to that aspect of the judgment. It is difficult to imagine a more cleanly tried case; by the losing party’s own count, only one error occurred in a case that took six days to try.

Obviously, the absence of other errors does not mean that the error that was committed should not be rectified on appeal if there is a likelihood that it might have affected the part of the result to which it relates. However, I cannot agree with the majority that there was any such likelihood, or that we are unable to affirmatively determine that there was not.

The majority ably summarizes plaintiffs harmless error argument, 134 Or App at 47-48, and it is unnecessary for me to repeat the particulars of the argument here. It suffices to say that there were many ways other than obtaining access to Dr. Drueker’s records by which defendant could *51have attempted to develop and present his theory at trial that plaintiffs emotional problems were caused, in whole or in part, by something other than the drug reaction that resulted from defendant’s malpractice. Indeed, Drucker himself was available and could have been called as a witness both at trial and for a discovery deposition. Nevertheless, by choice or inadvertence, defendant allowed the theory to lie fallow.

After describing the many avenues that defendant did not explore, the majority proceeds to accept defendant’s proposition that we cannot determine whether, if one more avenue to the same potential destination had been available to defendant, he might have found reason to develop and present the defense that an abundance of other available routes had not induced him to pursue.

Although I agree with the majority that it is impossible to say that defendant might not have found something in Drucker’s records that could have bolstered the theory, I do not think that that is the relevant inquiry. The question in connection with the harmless error issue is not whether the records could have been useful if defendant had chosen to develop and present the theory for which he sought them; rather, the question is whether the record before us demonstrates a convincing probability that defendant would not have relied on the theory at trial to any cognizable extent, no matter what evidence he might have found to support it. I think that the record makes that demonstration. For that reason, I do not agree with the majority that this appeal presents a “classic chicken-and-egg problem.” 134 Or App at 48. Rather, in my view, defendant neglected or chose not to go anywhere near the henhouse.

Although we cannot know whether that course was a . product of choice or neglect, nor does it matter to the proper disposition of the appeal which it was, a tactical decision not to make an evidentiary presentation on the theory probably would have been a good one. As I have noted, emotional distress was not the only issue in this case. Defendant fervently contested liability at trial, although he assigns no error pertaining to it here. An extensive presentation refuting the cause of one of the several types of damages that plaintiff sought could well have been regarded by defendant’s counsel as a potential distraction to the jury, and one that could have *52implied that defendant was not optimistic about prevailing on the liability issues. Moreover, direct evidence of the causes of plaintiffs emotional problems could have been perceived as difficult to produce without touching on other matters, such as plaintiffs sexual orientation, that defendant did not wish to present to the jury. Finally, counsel could more than plausibly have considered that the evidence was unnecessary: The jury was aware without any direct evidence from defendant that, independently of the consequences of the drug reaction, plaintiff was subject to other potential causes of emotional problems, including the fact that he was afflicted with a deteriorative and terminal disease.

I do not imply that the wisdom of the tactical choice, if such it was, has a direct bearing on the harmless error calculus. The point is that the inherent reasonableness of a tactical decision not to present evidence on the theory contributes to the likelihood that such a decision was made and, along with the fact that defendant did not pursue the theory through other known and available means, adds to the likelihood that nothing in the psychotherapist’s records would have altered his defensive approach.

Determining whether an error is harmless is not an exact science. It entails an assessment of probabilities, which must be guided in substantial part by intuitive responses to the circumstances of cases. Indeed, the majority opinion illustrates something akin to that proposition by holding, correctly, that there is so little likelihood that anything in the records would have provided a basis for attacking plaintiff s credibility that our remand should not extend to the liability issues on which his credibility might have bearing. I would apply a similarly realistic view to the principal issue in the appeal, and would hold that the error was extremely unlikely to have affected defendant’s failure or choice not to rely at trial on the theory he now stresses, regardless of where the evidence to support that theory might have been looked for or found.

This is a case that should be over.