Following Petros A. Palandjian’s (decedent’s) death from gastric cancer, his son, as executor of his estate, and his widow filed claims for wrongful death, medical malpractice, and loss of consortium against the decedent’s internist, Gerald S. Foster, and radiologists, Kathleen A. McCarthy and Carol A. Hulka, claiming that they were negligent in failing to diagnose the decedent’s cancer at an earlier stage. After trial in the Superior Court, the jury found that none of the physicians had been negligent. In an unpublished memorandum and order pursuant to its rule 1:28, the Appeals Court reversed the judgment in Foster’s favor, concluding that the trial judge had erred in excluding certain expert testimony related to the standard of care. Palandjian v. Foster, 63 Mass. App. Ct. 1104 (2005). We granted Foster’s application for further appellate review, limited to the issue whether Commonwealth v. Lanigan, 419 Mass. 15 (1994), applies to expert testimony on the standard of care, and we affirm the judgment of the trial court.

1. Background. Gerald Foster, an internist, served as the decedent’s primary care physician from 1979 to 1995. In February, 1991, the decedent visited Foster complaining of severe abdominal pain at night. Foster prescribed Zantac to treat possible peptic ulcers and ordered an upper gastrointestinal series of X-rays (upper GI series) to identify any abnormalities of the stomach or esophagus. Foster also referred the decedent to a cardiologist to determine whether his symptoms were heart related. These tests did not reveal any explanation for the decedent’s abdominal pain, which Foster concluded most likely was caused by stress.

The decedent next saw Foster in January, 1993, for a routine physical examination; records do not indicate any reports of abdominal pain at that visit. However, the decedent returned to Foster’s office six months later complaining of digestive problems, including foul-smelling belching and the sensation of food sticking in his throat. Foster ordered another upper GI series, which again did not detect any abnormality that could explain these symptoms.

In April, 1994, the decedent visited Foster with similar complaints and severe heartburn. This time, Foster referred the decedent for an endoscopy, which revealed multiple gastric[*102] ulcers.^[3] Biopsies of this tissue were negative, but Foster ordered another endoscopic examination because the ulcers looked suspicious. During the second endoscopy, doctors found a tumor in the decedent’s stomach; testing indicated that it was malignant. The decedent underwent surgery to remove the tumor and surrounding lymph nodes, as well as chemotherapy and radiation treatment. The cancer, however, recurred throughout the decedent’s abdomen, and he died in 1996.

Following his death, the decedent’s widow and son, as executor of his father’s estate, filed claims of wrongful death, medical malpractice, and loss of consortium, alleging that Foster was negligent in failing to detect the decedent’s gastric cancer sooner. Specifically, the plaintiffs maintained that the decedent’s gastric cancer could have been diagnosed and treated at a more curable stage if Foster had ordered an endoscopy in 1991 or 1993, after the inconclusive upper GI series. The plaintiffs also claimed that the radiologists who read the X-rays from the decedent’s upper GI series in 1991 and 1993 were negligent in failing to identify any abnormalities.

At the beginning of the trial, Foster, Hulka, and McCarthy (collectively, defendants) filed a motion in limine to exclude any reference to the fact that the decedent’s grandmother and aunt had died of gastric cancer, arguing that such information was inadmissible under Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993) (Daubert), and Commonwealth v. Lanigan, 419 Mass. 15 (1994) (Lanigan), because no scientifically reliable evidence established that a family history of gastric cancer in second degree relatives created a greater risk that a patient would develop gastric cancer himself.^[4] The trial judge initially ruled that the plaintiffs could refer to the decedent’s family history in their opening, but they could not argue that this family history actually increased the decedent’s risk of cancer.

At trial, the plaintiffs sought to establish the standard of care[*103] through an expert, who testified that Foster had deviated from the standard of care by not ordering an endoscopy in 1991 or 1993.^[5] The expert explained that he would have ordered an endoscopy earlier because the results of the upper GI series did not reveal the cause of the decedent’s persistent pain. The expert also testified that an endoscopic examination in 1991 or 1993 would have revealed the presence of stomach cancer.

On redirect examination, counsel for the plaintiffs asked the expert about the significance of the decedent’s family history of gastric cancer. The defendants’ objection to this line of inquiry as beyond the scope of cross-examination was sustained. The plaintiffs later made an offer of proof that the expert, if permitted to testify, “would have opined that the fact that [the decedent’s] grandmother [and paternal aunt] had died of gastric cancer was a factor that should have given the internist a heightened sense of suspicion . . . . [T]o an internist, that family history should have increased the index of suspicion that [the decedent] was at an increased risk for having gastric cancer.”^[6] The defendants objected to this testimony, claiming that there was no scientific or medical basis for this assertion and noting that it related back to the issue raised in their motion in limine. When the judge asked the plaintiffs’ attorneys for any Daubert-type evidence to support the expert’s opinion, counsel cited a study that found an increased risk of gastric cancer in patients with two or more first degree relatives who had the disease. However, the plaintiffs could not provide any data to support the contention that patients with a family history of gastric cancer among second degree relatives faced an increased risk of gastric cancer, and the judge excluded the evidence. The jury ultimately returned verdicts in favor of the defendants, and the plaintiffs appealed.

[*104] The Appeals Court affirmed the judgments in favor of the radiologists, but vacated the judgment in favor of Foster, reasoning that the judge erred in excluding the expert’s opinion about the decedent’s family history of gastric cancer and his increased risk of developing the disease. The Appeals Court concluded that “[t]he proffered expert testimony . . . merely reflects the expert’s opinion that the standard of care requires that internists exercise increased caution when dealing with a family history of gastric cancer.” Because the opinion related only to the standard of care, it was “not the sort of scientific evidence that must be preliminarily screened for reliability by a judge pursuant to Lanigan.” The Appeals Court determined that this testimony was material to the issue of Foster’s negligence, and its exclusion constituted reversible error.

2. Standard of review. The decision to exclude expert testimony rests in the broad discretion of the judge and will not be disturbed unless the exercise of that discretion constitutes an abuse of discretion or other error of law. See Commonwealth v. Pike, 430 Mass. 317, 324 (1999). Appellate courts similarly apply an abuse of discretion standard in reviewing the decision of a trial judge to exclude expert testimony as unreliable based on Commonwealth v. Lanigan, supra. Canavan’s Case, 432 Mass. 304, 312 (2000).

3. Discussion, a. Standard of care. To prevail on a claim of medical malpractice, a plaintiff must establish the applicable standard of care and demonstrate both that a defendant physician breached that standard, and that this breach caused the patient’s harm. See, e.g., Harlow v. Chin, 405 Mass. 697, 701 (1989). “The proper standard is whether the physician, if a general practitioner, has exercised the degree of care and skill of the average qualified practitioner, taking into account the advances in the profession. ... [A] specialist should be held to the standard of care and skill of the average member of the profession practising the specialty, taking into account the advances in the profession.” Brune v. Belinkoff, 354 Mass. 102, 109 (1968). Because the standard of care is based on the care that the average qualified physician would provide in similar circumstances, the actions that a particular physician, no matter[*105] how skilled, would have taken are not determinative.^[7] See id. See also Johnson v. Riverdale Anesthesia Assocs., 275 Ga. 240, 241-242 (2002) (because applicable standard of care derived from medical profession generally, “questions aimed at determining how the expert would have personally elected to treat the patient are irrelevant”); 5 D.W. Louisell & H. Williams, Medical Malpractice § 29.01, at 29-7 (2005) (“The standard is measured against what a reasonably prudent practitioner in the defendant’s position would do, not what any individual physician or physicians might do”). Commentators agree that this standard does not require physicians to provide the best care possible. See, e.g., W.L. Prosser & W.P Keeton, Torts § 32, at 187 (5th ed. 1984) (“it is not the middle but the minimum common skill” that determines applicable standard of care); Restatement (Second) of Torts § 299A comment e (1965) (“standard of skill and knowledge required” of physicians “is that which is commonly possessed by members of that profession [and] not that of the most highly skilled”). Furthermore, because the standard of care is determined by the care customarily provided by other physicians, it need not be scientifically tested or proven effective: what the average qualified physician would do in a particular situation is the standard of care.^[8] See Brune v. Belinkoff, supra.

Establishing the applicable standard of care typically requires[*106] expert testimony. See, e.g., Haggerty v. McCarthy, 344 Mass. 136, 139 (1962), quoting Bouffard v. Canby, 292 Mass. 305, 309 (1935) (“It is only in exceptional cases that a jury ... may without the aid of expert medical opinion determine whether the conduct of a physician toward a patient is violative of the special duty which the law imposes . . .”). In determining whether an expert is qualified to testify regarding the proper standard of care, “[t]he crucial issue is whether the witness has sufficient ‘education, training, experience and familiarity’ with the subject matter of the testimony.” Letch v. Daniels, 401 Mass. 65, 68 (1987), quoting Gill v. North Shore Radiological Assocs., 10 Mass. App. Ct. 885, 886 (1980). See Drevenak v. Abendschein, 773 A.2d 396, 418-419 (D.C. 2001) (factors relevant in assessing expert testimony on standard of care include expert’s training, board certification, specialized medical experience, attendance at seminars and meetings, familiarity with medical literature, and discussions with other physicians). Such qualifications ensure that the expert has sufficient knowledge of the practices of other physicians to assert that the average qualified practitioner would, or would not, take a particular course of action in the relevant circumstances. See Flagg v. Scott, 9 Mass. App. Ct. 811, 812 (1980) (“the expert’s qualifications [including his position as obstetrician-gynecologist-in-chief at one of Boston’s teaching hospitals] suggest that there is a firm basis for his opinion that ‘minimal standards of care were not met’ ”).

b. Application of Daubert-Lanigan. In Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993), the United States Supreme Court addressed the admissibility of scientific expert testimony under the Federal Rules of Evidence and held that a trial judge must conduct a preliminary assessment of such evidence, to ensure that it is both relevant and reliable. Id. at 592-593. In assessing the reliability of the expert’s opinion, the judge must determine “whether the reasoning or methodology underlying the testimony is scientifically valid and . . . whether that reasoning or methodology properly can be applied to the facts in issue.” Id. The Court, emphasizing the need for a flexible inquiry, provided a nonexhaustive list of factors to consider in evaluating the reliability of the expert’s testimony, including[*107] testing, peer review and publication, error rates, and general acceptance in the relevant scientific community.^[9] Id. at 593-595. This court adopted the basic reasoning of Daubert in Commonwealth v. Lanigan, 419 Mass. 15 (1994), although we suggested that general acceptance in the relevant scientific community likely would remain the most important factor in determining reliability. Id. at 26.

In Canavan’s Case, 432 Mass. 304 (2000), we extended the holding of Lanigan to apply to expert opinions based on personal observations and clinical experience, and concluded that medical expert testimony concerning diagnosis and causation should be subject to a Lanigan analysis. Id. at 313-316. “Observation informed by experience is but one scientific technique that is no less susceptible to Lanigan analysis than other types of scientific methodology.” Id. at 313. We based this conclusion in part on the reasoning of the United States Supreme Court in Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999), in which the Court held that Daubert's gatekeeping obligation applies to all expert testimony, whether based on scientific, technical, or other specialized knowledge. Id. at 149. In Canavan’s Case, supra at 314 n.5, we again emphasized that the test of reliability is flexible, and that a trial judge has broad discretion in determining which factors to apply to assess the reliability of proffered expert testimony.

Relying on the reasoning of Canavan’s Case, supra, Foster argues that Daubert-Lanigan should apply to medical expert testimony concerning the standard of care, just as it applies to medical opinions relating to diagnosis and treatment, to ensure that the judge admits only reliable expert opinions.^[10] Foster contends that the Appeals Court erred in holding that a Lanigan analysis was not necessary; even if the methodology employed[*108] by an expert is generally reliable, Daubert-Lanigan requires that the trial judge, acting as gatekeeper, determine whether the conclusions of an expert in a particular case have a reliable factual basis. Here, Foster claims, the expert’s opinion was based on the unsupported factual premise that a family history of gastric cancer in second degree relatives can increase one’s risk of developing the disease. Because the expert could not provide a reliable basis for this conclusion, that portion of his testimony was properly excluded.

The plaintiffs argue that the judge improperly used Lanigan to evaluate whether the practice called for under the standard of care had been proved effective, reiterating that the standard of care is custom based and therefore need not be validated scientifically. They maintain that the purpose of Daubert-Lanigan is to screen new or untested scientific evidence to prevent the admission of “junk science,” and note that courts in at least two other jurisdictions have held that Daubert does not apply to testimony about the standard of care because it does not involve novel scientific evidence.^[11]

We agree with the plaintiffs that expert testimony concerning the standard of care generally need not be subject to a Daubert-Lanigan analysis. Such testimony is based on the expert’s knowledge of the care provided by other qualified physicians, not on scientific theory or research: “How physicians practice medicine is a fact, not an opinion derived from data or other scientific inquiry by employing a recognized methodology.”^[12] Cramm, Ascertaining Customary Care in Malpractice Cases: Asking Those Who Know, 37 Wake Forest L. Rev. 699, 725 (2002). However, when the proponent of expert testimony[*109] incorporates scientific fact into a statement concerning the standard of care, that science may be the subject of a Daubert-Lanigan inquiry. Because expert opinion about increased risk, like diagnosis and causation, involves the application of science to patient care, Daubert-Lanigan would be applied to that portion of an expert’s testimony, requiring the proponent of such evidence, if challenged, to demonstrate its relevance and reliability. See Canavan’s Case, supra at 313-314. See also Sullivan v. United States Dep’t of the Navy, 386 F.3d 827, 833-834 (9th Cir. 2004) (applying Daubert to medical expert’s opinion about increased risk of infection associated with particular surgical procedure).

The plaintiffs contend that the expert’s testimony regarding family history was not offered to prove a scientific fact (i.e., that a history of gastric cancer in second degree relatives actually increases one’s risk of cancer), but only to show that the average qualified internist would consider a family history of gastric cancer in second degree relatives to be significant. We do not agree with this characterization of the proffered testimony. In their proffer the plaintiffs stated that Foster should have suspected that the decedent faced an increased risk of gastric cancer because of his family history.^[13] Even if we assume that part of this proffer could be characterized as a statement of the standard of care, it also includes the expert’s opinion that the decedent in fact faced an increased risk of gastric cancer. The plaintiffs thus attempt to admit scientific fact regarding increased risk without demonstrating the reliability of the science involved, by framing it as part of the standard of care.^[14] [*110] The proponent of expert testimony may not evade the reliability requirement of Daubert-Lanigan in this manner.^[15]

Therefore, when the defendants challenged the reliability of the expert’s testimony concerning the decedent’s increased risk of gastric cancer, the judge properly asked the plaintiff for any “Daubert-type evidence” that would support the expert’s contention. See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 149 (1999), quoting Daubert, supra at 592 (“where such testimony’s factual basis, data, principles, methods, or their application are called sufficiently into question . . . the trial judge must determine whether the testimony has ‘a reliable basis in the knowledge and experience of [the relevant] discipline’ ”); Lanigan, supra at 25-26. The plaintiffs cited a study demonstrating an increased risk of gastric cancer in patients whose first degree relatives had the disease, but they could not produce any data to show that patients like the decedent, with a family history of gastric cancer among second degree relatives, also would face an increased risk of developing cancer. The judge therefore acted well within his discretion in determining that there was no reliable basis for the expert’s assertion about the decedent’s increased risk of cancer, and excluding that portion of the expert’s testimony.

The plaintiffs further argue that the judge misapplied Daubert-Lanigan, both by requiring that the plaintiffs support their[*111] expert’s opinion with an epidemiological study and by failing to conduct a voir dire. Not all of the factors identified in Daubert will be applicable in every case; a trial judge accordingly has broad discretion to determine how to assess the reliability of expert testimony. See Canavan’s Case, supra at 314 n.5 (“Application of the Lanigan test requires flexibility. Differing types of methodology may require judges to apply differing evaluative criteria . . .”). We also note that the trial judge did not insist on an epidemiological study to demonstrate reliability; when he asked the plaintiffs’ counsel for any “Daubert-type evidence” to support their expert’s opinion about increased risk, counsel cited an article identifying an increased risk of gastric cancer among first degree relatives. Although the judge found that this article did not provide adequate support for the expert’s opinion, he did not refuse to hear other evidence, as the plaintiffs now claim.^[16] Furthermore, we are not aware of, and the plaintiffs do not identify, any case that requires a trial judge to conduct a separate hearing or a voir dire to determine the reliability of expert testimony. In fact, many Federal appellate courts have held that trial judges need not conduct separate evidentiary hearings to fulfil their gatekeeping function under Daubert, although a separate pretrial hearing or voir dire may be appropriate in some cases. See, e.g., United States v. Diaz, 300 F.3d 66, 74 (1st Cir. 2002); Oddi v. Ford Motor Co., 234 F.3d 136, 154 (3d Cir. 2000); United States v. Alatorre, 222 F.3d 1098, 1100 (9th Cir. 2000); Goebel v. Denver & Rio Grande W.R.R., 215 F.3d 1083, 1087 (10th Cir. 2000) (“district court may also satisfy its gatekeeper role when asked to rule on a motion in limine, on an objection during trial, or on a post-trial motion so long as the court has sufficient evidence” to assess relevance and reliability of expert’s opinion; separate Daubert hearing not required). Here, counsel for the plaintiffs had the[*112] opportunity to explain the basis of the proffered opinion, giving the judge adequate evidence to assess the relevance and reliability of the expert’s opinion concerning increased risk. In light of the broad discretion that Lanigan and its progeny provide trial judges in assessing the reliability of expert testimony, we cannot say that it was an abuse of discretion for the judge to exclude the proffered testimony as unreliable.^[17]

Finally, the plaintiffs claim that a 2003 study establishes that a history of gastric cancer in second degree relatives may increase a patient’s risk of developing the disease, thereby confirming that the proffered testimony should have been admitted because it is “the correct medical practice.” Even if this article does establish an increased risk of gastric cancer in second degree relatives, the plaintiffs’ argument overlooks that a “physician is held to the ‘standard of care and skill of the average member of the medical profession practicing his specialty’ at the time of the alleged negligence” (emphasis added). Brusard v. O’Toole, 429 Mass. 597, 607 (1999), quoting Stepakoff v. Kantar, 393 Mass. 836, 841 (1985). Here, Foster’s alleged negligence occurred in 1991 and 1993, some ten to twelve years before the publication of the article on which the plaintiffs rely. A physician cannot be held to a standard of care based on medical research and literature not in existence at the time of the alleged negligence. See Brusard v. O’Toole, supra (“experts cannot be cross-examined about discoveries or other knowledge developed after the acts complained of”).

4. Conclusion. The judgment of the Superior Court in favor of Foster is hereby affirmed.

So ordered.