LEVIN H. CAMPBELL, Circuit Judge.

Plaintiffs brought this diversity action for medical malpractice against Dr. Davis T. Gallison. The jury found for defendant, and plaintiffs appeal, alleging as error the exclusion of certain evidence contained in James Petrocelli’s medical record at Massachusetts General Hospital.

Dr. Gallison performed a hernia operation on James Petrocelli at Tobey Hospital on March 18, 1975. Mr. Petrocelli subsequently suffered a great deal of pain in his groin area and some months later visited Massachusetts General Hospital where he saw a Dr. Swartz. Dr. Swartz diagnosed a recurrence of the hernia and operated on Petrocelli on September 25, 1975. This effort failed, and Petrocelli ultimately underwent a third operation by another physician to solve his hernia problem. Throughout this period Petrocelli suffered from continuous, intense groin pains. Subsequent to 1975, Petrocelli has undergone several surgical procedures directed at reducing his pain by deadening the nerves in the affected area.

Plaintiffs’ malpractice claim against Dr. Gallison rested on the premise that in the first hernia operation Dr. Gallison had severed Petrocelli’s ilioinguinal nerve, a major nerve in the groin area. At the trial, plaintiffs’ witnesses consisted of themselves and a medical expert. James Petrocelli testified to the intense pain he had suffered and chronicled his various operations. His wife, Beverly, alleged that she had called Dr. Gallison soon after the first operation to ask him if there was anything he could do about her husband’s pain. She testified that Gallison had replied, “I could give him Darvon, but it is not going to do anything because I cut a nerve. What do you expect?”. Plaintiffs’ medical expert, Dr. Robert MacIntyre, who had neither treated nor examined Petrocelli but had examined the hospital and medical records, testified that, in his opinion, Petrocelli’s ilioinguinal nerve was “injured” or “traumatized” during Petrocelli’s first operation, though he would not say definitely that the nerve was severed. He based his opinion on Petrocelli’s complaints following the operation and the distribution of Petrocelli’s pain in the groin area. He also stated that, given scar tissue present from prior operations, it would be “tantamount to impossible to find [a severed nerve]” during a subsequent operation, though such an injury would be “easy to identify” if seen.

By way of rebuttal, defendant brought out that Dr. MacIntyre had never personally examined Petrocelli, that he had socialized on occasion with plaintiffs’ attorney and that he had not performed a hernia operation in some 16 years. It was brought out that Dr. MacIntyre was a thoracic (chest) surgeon rather than a specialist in neurology. Without objection, defendant had both Dr. MacIntyre and Dr. Gallison read aloud from a letter sent by a neurological specialist who had examined Petrocelli after the first operation. This specialist, a Dr. Gifford, stated in his letter that “the sensation appeared intact in [the ilioinguinal] area .... ” Dr. Gallison testified and denied having severed Petroeelli's ilio nerve. He also denied having told Mrs. Petrocelli after the operation that he had “cut a nerve.”

The only other evidence offered by either party on the critical issue of whether or not Dr. Gallison had actually severed the ilio nerve were two references in Petrocelli’s medical record from Massachusetts General Hospital. These references were offered by plaintiffs to prove that the nerve had been severed, but they were excluded as hearsay by the district court on timely objections by defendant. The exclusion of these two pieces of evidence — a sentence in Dr. Swartz’s September 26 post-operative report and a surgical note by another physician dated the following October 28 — now form the basis of plaintiffs’ appeal.

The first item of excluded evidence, noted below in italics, was contained in Dr. *289 Swartz’s report filed the day after he performed Petrocelli’s second hernia operation. That report is divided into two sections, one labeled “Indications” and one labeled “Procedure.” The first of these reads, in its entirety, as follows:

INDICATIONS: This 37 year old man had a left inguinal hernia repair at an outside hospital 5 mos. prior to admission. During the course of that surgical procedure, the left ilioinguinal nerve was severed. A recurrence of the hernia was noted in the immediate postoperative period. He presented to this hospital for repair of the recurrence. [Emphasis added.]

There follows the section entitled “Procedure” which at great length details what Dr. Swartz himself did and observed during his repair operation. There is no mention whatever of the ilioinguinal nerve in this section.

The other excluded portion of the medical record was an entry made by a different physician at a Massachusetts General Hospital surgical clinic on October 28,1975. In his report of Petrocelli’s visit to the surgical clinic, this doctor noted, “Hernia well healed but very worried about pain from transected ilio femoral nerve .... ” Plaintiffs assert on appeal that both of the above statements, while hearsay, should nevertheless have been admitted under the exception to the hearsay rule codified in Federal Rule of Evidence 803(6).

Rule 803(6), commonly called the business records exception, governs admissibility of “Records of regularly conducted activity.” 1 It provides that any report of “acts, events, conditions, opinions, or diagnoses, made at or near the time by, or from information transmitted by, a person with knowledge, if kept in the course of a regularly conducted business activity” should be admitted “even though the declarant is available as a witness.” Fed.R.Evid. 803(6). Plaintiffs argue that the notations describing the ilioinguinal nerve as having been “severed” at the earlier operation in a different hospital, and mentioning Petrocelli’s worry about pain from “transected ilio femoral nerve,” were both contained in reports kept by the hospital in the regular course, that they were made by doctors with knowledge of Petrocelli’s condition, and that they should therefore have been admitted as the “opinions or diagnoses” of Petrocelli’s attending physicians.

We think the district court did not abuse its discretion in excluding the parts of this hospital record which indicated that the nerve had earlier been severed. We reach this result primarily because of the complete absence of any indication as to where this information — relating to something that had happened six months ago in another hospital — came from. To be admissible as “business records” under Rule 803(6), the referenced notations would have to represent either the opinions or diagnoses of the Massachusetts General Hospital doctors who made the notations or the diagnoses of some other “person with knowledge” (such as a medical colleague) who reported to the maker of the record as part of the usual business or professional routine of Massachusetts General Hospital. If the entries were merely relaying what Mr. Petrocelli or his wife told the reporting physicians, when providing a medical history, the *290 matter would not be admissible solely under Rule 803(6) (although, as discussed below, it might have been admissible through Rule 803(6) when combined with Rule 803(4) as patient history told to a doctor with a duty to record it). Rule 803(6) requires that information in a business record be “transmitted by a person with knowledge” acting “in the course of a regularly conducted business activity.” The Advisory Committee Notes make clear that this encompasses only declarants — like nurses or doctors in the case of hospitals — who report to the recordkeeper as part of a regular business routine in which they are participants. Where the declarant is a hospital patient, his relating of his own history is not part of a “business” routine in which he is individually a regular participant. McCormick on Evidence § 313 at 731 (2d ed. 1972).

Plaintiffs argued vigorously both here and below, that the notations concerning the severed nerve must be taken to reflect the medical opinions of the reporting doctors. But we think it entirely uncertain whether the reporting doctors themselves determined from ascertainable symptoms or observations, that the nerve had been “severed” or “transected” in the previous operation, or whether instead these doctors were simply recording what the patient or his wife had reported. The statements are not obviously diagnostic in quality. There is no mention of symptoms leading to a conclusion that the nerve had previously been cut. See Buckminster’s Estate v. Commissioner, 147 F.2d 331 (2d Cir. 1944) (admitting medical report diagnosing party as exhibiting symptoms of cerebral hemorrhage). Nor are the statements in the nature of opinions based upon observations by the physicians. See Reed v. Order of United Commercial Travelers of America, 123 F.2d 252 (2d Cir. 1941) (statement by admitting physician in medical record that party was “still apparently well under the influence of alcohol” admitted to prove intoxication).

Particularly since Mrs. Petrocelli believed that Dr. Gallison had confessed to severing the nerve, it is altogether possible that the Petroeellis presented the severance of the nerve to Dr. Swartz in their description of patient history as an acknowledged fact. The district court was in no position to know whether a surgeon like Dr. Swartz, operating six months later, could or did independently determine whether the ilioinguinal nerve had earlier been severed. We note, moreover, that plaintiffs’ expert testified that such a nerve would be very difficult to find during a subsequent surgical procedure given the scarring in the area from a prior operation. And in the detailed narrative of his own operation, Dr. Swartz did not once mention the ilioinguinal nerve. This strongly suggests that Dr. Swartz did not independently confirm the condition. Nor could the trial court know whether, apart from direct observation, such a conclusion was possible to derive from the symptomatology. Dr. Swartz’s statement is on its face merely a bare, conclusory sentence in a preliminary section of the report devoted to patient history which could likely be entitled to no weight as reflecting the doctor’s own medical opinion.

The mention of the ilio nerve in the surgical clinic report made on October 28, a month later, is even more cryptic. It merely states that Petrocelli is “still worried about pain from transected ilio femoral nerve” without indicating in any way that the writer had investigated and agreed with the apparent assumption that the nerve had been transected. 2 Without more conclusive contextual corroboration or qualifying testimony by the doctors themselves, therefore, it is simply impossible to tell if the statements are actually the “opinions or diagnoses” of the reporting physicians as urged by the plaintiffs. Fed.R.Evid. 803(6).

In judging the reasonableness of exclusion under Rule 803(6), it is relevant that the excluded, elliptical reports could easily have been misconstrued by the jury as definitive opinion testimony on the most critical issue in the entire case — whether the nerve had, in fact, been severed. See Fed. *291 R.Evid. 403. Given the impossibility of determining from the records themselves whether these reports reflected medical judgments, and the lack of any corroborative evidence or testimony offered by the plaintiffs to assure the court that these were professional opinions, the district court could reasonably determine that the notations were simply too inscrutable to be admitted, bearing in mind that, if admitted under Rule 803(6), they would be admitted for their truth without any opportunity to cross-examine the physicians who made them. Pretrial discovery rules made it possible for plaintiffs to have deposed or otherwise sought clarification from the various physicians, including Dr. Swartz, which they apparently did not do. The trial court could thus entertain legitimate doubts as to whether the doctors who recorded these statements were actually rendering professional judgments. We emphasize that this case involves a rather narrow issue regarding the admissibility under Rule 803(6) of a business record which is so cryptic that pure guesswork and speculation is required to divine the source of the cited information. Under such circumstances the district judge acted within his discretion in determining that the hearsay statements could not be admitted as business records reflecting medical “opinions or diagnoses” under Rule 803(6). 3

There is, to be sure, another basis for admission if we assume that Petrocelli himself or his spouse was the source of the “cut nerve” information. While the business records exception embodied in Rule 803(6) does not provide a basis for admitting patient statements regarding medical history or prior treatment, see ante, Rule 803(4) 4 provides for admission of patient or family statements “describing medical history, or past or present symptoms, pain, or sensations, or the inception or general character of the cause ...” of a condition, provided the statements are “reasonably pertinent to diagnosis or treatment.” Fed. R.Evid. 803(4). Thus, if not admissible under Rule 803(6), the Petrocellis’ recorded statements about the cut nerve were arguably admissible as part of patient history. 5 For the reasons which follow, however, we do not think such a theory warrants either a reversal or a new trial in this case.

First, the Petrocellis did not argue either to the court below or in their briefs on this appeal that the references in these medical records derived from statements made by Petrocelli or his wife to doctors at Massachusetts General Hospital and were, therefore, admissible as patient history. We do not ordinarily decide cases on the basis of theories never presented below nor argued to us. See Johnston v. Holiday Inns, Inc., 595 F.2d 890, 894 (1st Cir. 1979).

One might surmise, moreover, that plaintiffs’ reluctance to claim admissibility under Rule 803(4) was deliberate and strategic rather than inadvertent, making it even less appropriate for us to upset the lower court on the basis of a theory never advanced. The jury knew from the testimony of Petrocelli’s wife that the Petrocellis believed the nerve had been severed by *292 Dr. Gallison. A mere reiteration via the hospital record that Petroeelli or his wife had told Dr. Swartz that his nerve was cut during Dr. Gallison’s operation would have added little beyond corroborating what Mrs. Petroeelli had already told the jury. Yet were these statements to have been admitted under Rule 803(4), fairness would have dictated that the judge instruct the jury, if requested, that the statements were admitted for their truth solely as matters related by the patient or a member of his family, not as professional opinion. Otherwise the jury was open to the misimpression that the notations were admitted as the medical judgments of the two Massachusetts General Hospital doctors who wrote the entries or of the hospital itself. As it would have been manifestly improper for statements admitted as patient history under Rule 803(4) to be presented to the jury as something else, 6 plaintiffs’ attorney understandably sought admission exclusively under Rule 803(6).

In short, we believe that the Rule 803(4) issue was not, for perhaps strategic reasons, presented to either the district court or to us and that, even had Petroeelli sought admission of the statements under Rule 803(4), an exclusionary ruling by the district court on these facts would have constituted but harmless error. Fed.R.Civ.P. 61.

Affirmed.

. That rule reads in full as follows:

Rule 803. Hearsay Exceptions; Availability of Declarant Immaterial The following are not excluded by the hearsay rule, even though the declarant is available as a witness:

(6) Records of regularly conducted activity. A memorandum, report, record, or data compilation, in any form, of acts, events, conditions, opinions, or diagnoses, made at or near the time by, or from information transmitted by, a person with knowledge, if kept in the course of a regularly conducted business activity, and if it was the regular practice of that business activity to make the memorandum, report, record, or data compilation, all as shown by the testimony of the custodian or other qualified witness, unless the source of information or the method or circumstances of preparation indicate lack of trustworthiness. The term “business” as used in this paragraph includes business, institution, association, profession, occupation, and calling of every kind, whether or not conducted for profit.

. At oral argument it was plausibly suggested that the physician who wrote the second note was making reference to Dr. Swartz’s report— which could have been in the same file.

. It may also be argued, although we need not decide the issue, that the patent ambiguity as to the true source of information in this hospital record created such a “lack of trustworthiness,” see Fed.R.Evid. 803(6), as to authorize the trial judge to exclude, on that separate ground, the references to the severed nerve. See Meder v. Everest & Jennings, Inc., 637 F.2d 1182, 1187 n.6 (8th Cir. 1981) (lack of “trustworthiness” where there is no way of divining basis of description of cause of accident contained in a police report). See also 4 Weinstein’s Evidence j[ 803(6)[06] at 803-170 (1979).

. Rule 803(4) provides that such statements are not excluded by the hearsay rule as follows:

(4) Statements for purposes of medical diagnosis or treatment. Statements made for purposes of medical diagnosis or treatment and describing medical history, or past or present symptoms, pain, or sensations, or the inception or general character of the cause or external source thereof insofar as reasonably pertinent to diagnosis or treatment.

. Since the medical history arguably asserted here was recorded in a hospital record, the plaintiffs would actually have had to qualify the notations under a combination of Rule 803(6) (to admit the hospital record) and Rule 803(4) (to admit the patient history). See Fed. R.Evid. 805.

. It may, of course, be urged that since it is fairly certain the information in the report came from one or the other of two sources— the doctors or the Petrocellis — either of which might have qualified the statements for admission under an applicable hearsay exception, the district court should have let the information in and allowed the jury to decide whether the statements were, in fact, the diagnoses of the attending physicians or the statements of the Petrocellis. No such theory was posed by plaintiffs, however, and such a procedure would run counter to Fed.R.Evid. 104(a), which gives to the judge the preliminary duty of determining whether — and under what theory— an item of evidence is to be admitted. See McCormick on Evidence § 53 (2d ed. 1972). More importantly, where, as here, an item in a medical record is so ambiguous as to its source as to leave a jury with no clue as to how to evaluate it (other than the fact that it appears in an official record), a trial judge would be entitled in a case like this to exclude the evidence under Fed.R.Evid. 403 on the ground that the danger of unfair prejudice from jury confusion substantially outweighed the record’s probative value. To allow medical evidence such as this to be admitted “for what it was worth” would be to invite the jury to speculate on the weight to be given crucial statements, without any evidentiary basis whatever to guide them. Such a practice would seem unlikely to “secure fairness” or to further “the end that the truth may be ascertained.” Fed.R.Evid. 102.