In an action to recover damages for medical malpractice, etc., the plaintiffs appeal, as limited by their brief, from so much of a judgment of the Supreme Court, Bangs County (Jackson, J.), entered July 23, 2002, as, upon a jury verdict, and upon an order of the same court dated October 24, 2002, which denied their motion pursuant to CFLR 4404 (a) to set aside the verdict as against the weight of the evidence and for a new trial, is in favor of the defendant Arnold J. Stein and against them dismissing the complaint insofar as asserted against that defendant.
Ordered that the judgment is reversed insofar as appealed from, on the law, the complaint is reinstated insofar as asserted against the defendant Arnold J. Stein, the motion is granted, a new trial is granted with respect to that defendant, with costs to abide the event, and the order is modified accordingly.
The issue before the jury was whether the defendant Dr. Arnold J. Stein (hereinafter the defendant) committed malpractice which constituted a proximate cause of Menachem Lipschitz’s loss of his left eye due to endophthalmitis, an infection of the vitreous humor of the eye. The defendant performed a cataract operation upon Mr. Lipschitz’s left eye on November 10, 1997. Before the operation, the defendant applied Betadine, an antimicrobial solution, around the eye to avoid infection. After the surgery he placed a collagen shield soaked in cortisone and tobramycin over the eye.
The morning after the surgery, November 11, 1997, Mr. Lipschitz, suffering extreme pain," went to the defendant’s office for a postoperative office visit. The defendant treated Mr. Lipschitz for high eye pressure which he assumed was related to his glaucoma. When the defendant’s treatment failed, he sent *635Mr. Lipschitz to the hospital at approximately 1:45 em. An initial examination at the hospital showed, no light perception in the left eye and exploratory surgery confirmed “extremely aggressive endophthalmitis” requiring the evisceration of the eye on November 13, 1997.
Several theories of malpractice were submitted to the jury, which found, inter alia, that the defendant did not deviate from accepted standards of medical care in failing to administer broad spectrum antibiotics by injection at the end of the cataract surgery, that the defendant did not fail to see Mr. Lipschitz from approximately 9:00 a.m. until approximately 10:45 a.m. in his office on the day following the surgery, and that the defendant’s failure to diagnose endophthalmitis during that office visit was not a departure from accepted medical practice. We find that there were errors committed at the trial which could have affected the verdict and therefore cannot be considered harmless (see Garricks v City of New York, 1 NY3d 22, 27 [2003]).
According to the plaintiffs’ expert witness at the trial, the defendant deviated from the standard of care when he failed to inject a broad spectrum antibiotic into the conjunctiva underneath the eye immediately after the surgery, and the failure to do so was a substantial factor in causing permanent injury to the eye, requiring its evisceration.
To refute the testimony of the plaintiffs’ expert that failure to inject antibiotics immediately after the surgery was a deviation from the standard of care, the defendant’s infectious disease expert testified, over the plaintiffs’ objection, that an injection of antibiotics immediately after surgery “would not have made any demonstrable difference.” The defendant’s counsel asked “Why is that, Doctor?” and the expert replied “because there are . . . essentially no properly done randomized or controlled comparison studies of the ethicacy [sic] of any of these preventative approaches in the literature-.” Although opinion in a publication which an expert deems authoritative may be used to impeach an expert on cross-examination (see People v Feldman, 299 NY 153, 168 [1949]; Watkins v Labiak, 6 AD3d 426 [2004]; Labate v Plotkin, 195 AD2d 444, 445 [1993]; Mark v Colgate Univ., 53 AD2d 884, 886 [1976]), the introduction of such testimony on direct examination constitutes impermissible hearsay (see Kelly v St. Luke’s Hosp. of Kansas City, 826 SW2d 391, 396-397 [Mo App 1992]; Niagara Vest v Alloy Briquetting Corp., 244 AD2d 892, 893 [1997]). In any event, the expert testified on cross-examination that he did not consider any books or articles in the field of infectious diseases “authoritative.”
*636With respect to the delay in diagnosis, the plaintiffs’ expert testified that the delay in diagnosing endophthalmitis was a substantial factor in the loss of the eye because the bacteria involved was “quite sensitive . . . and . . . treatable in the first 12 to 24 hours.” The defendant’s infectious disease expert, in an attempt to minimize the significance of the delay, testified that by the morning of November 11,1997, the plaintiff s chance of preserving any useful vision had already been reduced to “ten percent or less.”
The defendant’s infectious disease expert was permitted to testify, over the plaintiffs’ objection, with respect to the results of an “Endophthalmitis Vitrectomy Study” (hereinafter the EVS). To refute the plaintiffs’ claim that the defendant’s failure to diagnose endophthalmitis constituted malpractice, the expert testified that the most common sign of endophthalmitis was a hypopyon (pus in the anterior chamber) which was experienced by 86% of patients in the EVS study. He further testified, over the plaintiffs’ objection, that of the 400 patients in the EVS, the median time between surgery and the presentation of endophthalmitis was six days and “only 24 percent presented in three days or fewer. A tiny number presented in one day.” The plaintiffs’ counsel’s request for a sidebar conference was denied and his objection on the ground that “[n]one of these things are in evidence” was overruled.
The EVS was not admitted into evidence and its reliability was not established. Accordingly, testimony with respect to the EVS should not have been admitted (see Niagara Vest v Alloy Briquetting Corp., supra at 893).
The time that Mr. Lipschitz arrived at the defendant’s office on November 11, 1997, was significant since the delay in diagnosis was attributed to the defendant’s alleged failure to attend to his patient in a timely manner. The plaintiffs’ expert testified that if Mr. Lipschitz arrived at the defendant’s office at 9:00 a.m. and was not seen until 10:45 a.m., that would be a deviation from accepted medical practice.
At the trial, the defendant acknowledged that he did not know when Mr. Lipschitz arrived in his office on November 11, 1997, for his postoperative visit. However, he recalled that Mr. Lipschitz was in unusual pain which warranted immediate attention.
Mr. Lipschitz testified that he arrived shortly after 9:00 a.m., and his daughter-in-law who drove him to the defendant’s office corroborated that testimony. As evidence that the defendant did not see Mr. Lipschitz from approximately 9:00 a.m. until approximately 10:45 a.m. on November 11, 1997, the plaintiffs *637referred to the defendant’s medical records, which indicated that drops were placed in Mr. Lipschitz’s eye at 10:45 a.m. and the defendant first examined Mr. Lipschitz at 11:00 a.m.
The defendant’s counsel stated in his opening statement that Mr. Lipschitz arrived at the defendant’s office “around 10 a.m.” The defendant’s receptionist testified that Mr. Lipschitz arrived at approximately 10:00 a.m. When she was asked whether that testimony was based upon “memory or from something else” she responded “[f]rom something else,” to wit, “[o]ur patient log.” The plaintiffs’ objection to that testimony was overruled and the trial court refused the plaintiffs’ counsel’s request for a sidebar conference. Her estimate of the time when Mr. Lipschitz arrived was based on the order of the names in the patient log. She acknowledged that she could not recall how many names appeared before Mr. Lipschitz’s name.
On cross-examination, the plaintiffs’ counsel asked the receptionist whether she knew when patients arrived at the office and the order in which they arrived. She replied, “I have my book” and that she knew what it said. The patient log was never produced or introduced in evidence notwithstanding that it was subpoenaed by the plaintiffs.
Since the defendant’s receptionist acknowledged that she had no independent recollection of when Mr. Lipschitz arrived at the office and her claim that he arrived at 10:00 a.m. was based upon a document never produced and which was not in evidence, the plaintiffs’ objection to her testimony should have been sustained. Her description of a document not in evidence and the conclusions she drew therefrom should have been stricken.
The references to the patient log were improper pursuant to the best evidence rule which “requires the production of an original writing where its contents are in dispute and sought to be proven” (Schozer v William, Penn Life Ins. Co. of N.Y., 84 NY2d 639, 643 [1994]). “[I]n modern day practice the rule serves mainly to protect against fraud, perjury and ‘inaccuracies . . . which derive from faulty memory’ ” (id. at 643-644, quoting Fisch, New York Evidence § 81, at 50 [2d ed]). Secondary evidence of the contents of an unproduced original document may be admitted if the proponent of the evidence has sufficiently explained why the original is not available (see Trombley v Seligman, 191 NY 400, 403 [1908]). In the instant case, the defendant’s counsel stated that “there was no explanation for why the record was not admitted.”
However, the record indicates that the defense provided an explanation. Prior to summations, the plaintiffs’ counsel stated *638that during the charge conference (which was apparently not recorded), “defense counsel conceded that he had the record, that the record was not produced intentionally because [the defendant’s receptionist] had altered that document. Specifically she recorded the times that she claims or that the defense claims Mr. Lipschitz arrived in his office. Now that is a stipulated fact he has admitted.” The plaintiffs’ counsel requested an instruction pursuant to PJI 1:77.1 (2004 Supp), that a fraudulent purpose can be inferred from the destruction of evidence.
The defendant’s counsel in response did not deny that the record had been altered and acknowledged that the defendant took “responsibility for the fact that the record was not admitted in evidence.” The defendant’s counsel argued that the question of when Mr. Lipschitz arrived in Dr. Stein’s office was a “collateral issue.” The trial court denied the plaintiffs’ request for an instruction pursuant to PJI 1:77.1 (2004 Supp) and denied the plaintiffs’ counsel’s request to “make a record” on the issue.
On appeal, the defendant argues that there is no evidence in the record that the patient log was destroyed or altered. However, the plaintiffs’ counsel’s claim that the alteration of the document was a “stipulated fact” that the defendant’s counsel admitted at the charge conference was never denied by the defendant’s counsel although he was afforded an opportunity to do so. It may be inferred from the defendant’s counsel’s silence that the document in fact was altered (see Woodward v City of New York, 119 AD2d 749 [1986]), giving rise to a permissive inference of a fraudulent intent warranting a jury instruction to that effect (see PJI 1:77.1 [2004 Supp]; Mylonas v Town of Brookhaven, 305 AD2d 561, 563 [2003]).
The totality of these errors warrants a new trial. Prudenti, P.J., Smith, Goldstein and Crane, JJ., concur.