This is an appeal from a judgment entered by a single justice of this court affirming a decision of the Department of Industrial Accidents reviewing board (reviewing board), *756which in turn affirmed the decision of an administrative judge to admit in evidence certain medical opinion testimony, and reversed the decision of a second administrative judge who had concluded that the employee had failed to prove that her injury arose out of and in the course of her employment.
Background. On October 15, 1996, the Boston Medical Center (BMC) offered free influenza vaccinations (flu shots) to its employees and the general public. On that date, fifty-four year old Carolyn R. Hicks (the employee), an electrocardiogram (EKG) technician employed by the BMC, went to the BMC lobby during her lunch hour and received a flu shot. The following day, the employee felt like she had sand in her eyes. She was seen at the Brigham and Women’s Hospital on October 18, 1996, where she had a computed tomography (CT) scan and was told to return if she experienced more problems.
While dressing for work on October 22, 1996, the employee’s vision began to fail and she returned to the Brigham and Women’s Hospital where she was seen by an ophthalmologist. She was admitted to the hospital and remained there for the next three days while she received treatment consisting of large-dose Prednisone injections. While she was in the hospital, the employee was seen by a neurologist, Dr. Bluth, and a neuroophthalmologist, Dr. Iwamoto. During that admission, she was diagnosed as having optic neuritis.
Dr. Bluth and Dr. Iwamoto continued to treat the employee after she was discharged. Although the employee’s vision initially improved, when her physicians later attempted to reduce the amount of Prednisone, her vision worsened and she was diagnosed as being legally blind. She never recovered her eyesight and has remained legally blind at all times relevant to this action.
Don Bienfang, a board certified ophthalmologist who specializes in neuroophthalmology (the body of medicine that deals with disorders of the brain that are reflected in either eye movement or abnormalities of vision) eventually began treating the employee for her vision problems. As of the date of the hearing, the employee continued to be treated by Dr. Bienfang. Dr. Bienfang eventually ruled out several possible illnesses, including *757multiple sclerosis, that could cause blindness. He concluded that the employee’s blindness was precipitated by an autoimmune reaction to the flu shot she had received on October 15, 1996.1
The employee filed a workers’ compensation claim against the self-insurer, BMC, seeking temporary total incapacity and medical benefits, commencing October 16, 1996, and continuing, on the ground that the flu shot caused her optic neuritis, vision loss, and blindness. BMC denied the employee’s claim, contesting liability, causal relationship, disability and the extent thereof, as well as denying entitlement to medical benefits. At conference, BMC was ordered to pay the employee’s claim, and both parties appealed. The matter ultimately proceeded to a hearing before an administrative judge. See G. L. c. 152, § 11. In conjunction with that hearing, the employee was examined by an impartial medical examiner, James Lehrich, who is a neurologist. G. L. c. 152, § 11A(2).
Dr. Lehrich diagnosed the employee as having bilateral optic neuritis and myelopathy and opined that she was permanently disabled because of her blindness. However, Dr. Lehrich opined that the cause of the employee’s neurological disease was uncertain but may have been caused by multiple sclerosis. He further opined that a causal relationship between optic neuritis and the flu shot had not been shown by scientific or epidemiological studies and that the onset of the employee’s symptoms within two to three days after she received the flu shot was inconsistent with a “post-vaccination autoallergic central nervous system disease.”
To refute the employee’s claim that her disability was caused by the flu shot, BMC submitted the report of Carolyn Langer, a doctor board certified in occupational and environmental medicine. Dr. Langer agreed with the impartial physician’s diagnosis and also concurred that the employee’s loss of vision *758was not the direct result of, or causally connected to, the flu shot.2
BMC also submitted the report of Vincent Patalano, a board certified ophthalmologist, who examined the employee on November 3, 1997. Dr. Patalano’s report stated that he did not have any of the employee’s medical records available for review, and he suggested that a neurologie examination, specific laboratory work, and neuro imaging be done. The report also stated that if the employee reported that her vision loss became severe the day after the vaccination, then, based on the case report he reviewed, her optic nerve disease was not likely related to the vaccination.
In contrast, the employee’s treating physician, Dr. Bienfang, opined that the employee has “an autoimmune induced central nervous system disease which is related to a vaccination exposure.” Dr. Bienfang testified that he had considered all of the other known causes of optic neuritis and had excluded them, and that his opinion regarding causation had remained unchanged.
BMC moved to strike Dr. Bienfang’s testimony, claiming that his medical opinion on the issue of causal relationship, was not generally accepted in the relevant scientific community, and that the methodology used by Dr. Bienfang in arriving at his opinion in this case was unreliable. See Commonwealth v. Lanigan, 419 Mass. 15, 24-26 (1994). The administrative judge denied that motion and adopted Dr. Bienfang’s testimony that the employee’s optic neuritis was causally related to the flu shot.
The administrative judge also made a general finding that the employee’s injury arose out of and in the course of her employ*759ment with BMC, and ordered BMC to pay the employee temporary total incapacity benefits through their exhaustion on October 15, 1999, followed immediately thereafter by payment of partial incapacity benefits at the maximum rate “to date and continuing.” G. L. c. 152, §§ 34, 35. BMC appealed, arguing that the decision was contrary to law because (1) the administrative judge had abused her discretion in admitting the testimony of Dr. Bienfang, as his opinions lacked a reliable scientific foundation, and (2) the employee failed to prove that her injury arose out of her employment with BMC.
On the first issue, the reviewing board concluded that the administrative judge had not abused her discretion in admitting Dr. Bienfang’s testimony and affirmed that portion of her decision regarding the issue of causation. Hicks v. Boston Med. Center, 15 Mass. Workers’ Comp. Rep. 1, 9 (2001). However, on the second issue, because the administrative judge had not made any subsidiary findings of fact regarding the relationship between the employee’s work and the flu shot, the reviewing board recommitted the matter for further findings “on the question of whether, and if so, how the flu shot was for the mutual benefit of [the employee] and [BMC].” Id. at 5.
A second administrative judge3 denied the employee’s claim, ruling that she had failed to prove that her injury arose out of and in the course of her employment because BMC did not derive any benefit from offering her the flu shot. The employee appealed, and the reviewing board reversed the decision and reinstated the original award, concluding that the employee’s receipt of the flu shot was an incident of her employment and that it was also a benefit to BMC. Hicks v. Boston Med. Center, 16 Mass. Workers’ Comp. Rep. 426, 428 (2002). BMC appealed to a single justice of this court, who affirmed the reviewing board’s decision to reinstate the benefits awarded by the first administrative judge. See id. at 429.
BMC now seeks further review by a panel of this court, arguing that the first administrative judge abused her discretion in *760admitting the opinions of Dr. Bienfang and that in regard to the second administrative judge’s decision, the reviewing board exceeded the scope of its authority and erred in concluding that BMC benefitted by offering the flu shot to its employees. We discuss each issue in turn.
1. Admissibility of Dr. Bienfang’s testimony. BMC argues that Dr. Bienfang’s expert opinion on causation was inadmissible because (1) the theory that optic neuritis can be caused by a flu vaccination is not generally accepted in the scientific community and its reliability has not been shown by other means, and (2) the methodology used by Dr. Bienfang in determining that the employee’s optic neuritis was caused by the flu shot was unreliable.
The judge serves as a gatekeeper on the admission of expert opinion testimony; a ruling on the admissibility of challenged testimony “entails a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue.” Commonwealth v. Lanigan, 419 Mass. at 26, quoting from Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 592-593 (1993).4 Recognizing the “inherently fact-intensive” nature of these determinations and the flexibility of the Lanigan analysis, we apply the abuse of discretion standard in reviewing the administrative judge’s decision to admit Dr. Bienfang’s opinion testimony. Canavan’s Case, 432 Mass. 304, 312 (2000).
“The ultimate test [for the admissibility of expert testimony] is the reliability of the theory or process underlying the expert’s testimony.” Commonwealth v. Lanigan, supra at 24. Factors such as general acceptance in the scientific community, peer review, and publication of the theory are relevant, but not “indispensable predecessor^] of admissibility.” Id. at 25.
In this matter, it is undisputed that there have been no scientific or epidemiological studies showing a statistically significant relationship between receipt of the flu vaccine and optic neuritis. However, the absence of such studies does not compel a finding that Dr. Bienfang’s theory is unreliable. Vas*761sallo v. Baxter Healthcare Corp., 428 Mass. 1, 13 (1998). Rather, “a party seeking to introduce scientific evidence may lay an adequate foundation... by showing that the evidence is reliable or valid through an alternate means.” Canavan’s Case, 432 Mass. at 310, citing Commonwealth v. Sands, 424 Mass. 184, 185-186 (1997).
Here, the first administrative judge found, based on Dr. Bienfang’s deposition, that case studies published in ophthalmology and medical journals — some authored by Dr. Bienfang — suggested a causal relationship between the receipt of a flu shot and optic neuritis, vision loss, and blindness. The administrative judge found the case studies were sufficiently reliable to support Dr. Bienfang’s theory.
In addition to the reliability of Dr. Bienfang’s theory, the administrative judge found that Dr. Bienfang’s methodology was also reliable.5 Dr. Bienfang’s methodology, known as “ [differential diagnosis, or differential etiology, is a standard scientific technique of identifying the cause of a medical problem by eliminating the likely causes until the most probable one is isolated.” Westberry v. Gislaved Gummi AB, 178 F.3d 257, 262 (4th Cir. 1999). See Baker v. Dalkon Shield Claimants Trust, 156 F.3d 248, 252-253 (1st Cir. 1998); In re Paoli R.R. Yard PCB Litigation, 35 F.3d 717, 758 (3d Cir. 1994), cert. denied sub nom. General Elec. Co. v. Ingram, 513 U.S. 1190 (1995) (“differential diagnosis generally is a technique that has widespread acceptance in the medical community, has been subject to peer review, and does not frequently lead to incorrect results”). See also Vassallo v. Baxter Healthcare Corp., 428 Mass. at 12 (finding that expert opinion that relied, inter *762alia, on the “differential diagnosis method of identifying a patient’s symptoms as being among those associated with silicone gel breast implants and eliminating any other causes for these symptoms” was properly admitted). “A reliable differential diagnosis typically ... is performed after ‘physical examinations, the taking of medical histories, and the review of clinical tests, including laboratory tests . . . .’ ” Westberry v. Gislaved Gummi AB, 178 F.3d at 262, quoting from Kannankeril v. Terminix Intl., Inc., 128 F.3d 802, 807 (3d Cir. 1997). Because Dr. Bienfang followed those procedures, the administrative judge could reasonably conclude that Dr. Bienfang’s methodology was rehable.6 See Westberry v. Gislaved Gummi AB, supra at 263 (“a reliable differential diagnosis provides a valid foundation for an expert opinion”).
We conclude that on this record, it was not an abuse of discretion to admit Dr. Bienfang’s expert opinion on causation.
2. Arising out of and in the course of the employment. In order for an injury to be compensable, the injury must not only arise out of, but also occur in the course of employment. G. L. c. 152, § 26. BMC concedes that the injury occurred in the course of employment, but disputes the determination that the injury arose out of the employment. “Generally, the determination whether an employee’s injury arises out of her employment is a question of fact. . . .” Corraro’s Case, 380 Mass. 357, 359 (1980).
On appeal, BMC argues that (1) cases concerning employer-sponsored inoculations require the employee to prove elements of compulsion and an employer benefit from the inoculation; (2) the second administrative judge properly concluded that the employee failed to meet her burden because BMC did not compel or strongly urge her to receive the flu shot and the receipt of the vaccine was a purely personal activity; and (3) the reviewing board exceeded its authority in reversing the administrative judge because its conclusion was based on facts not found by the administrative judge.
*763According to G. L. c. 152, § 11C, as amended by St. 1991, c. 398, § 31, “[t]he reviewing board shall reverse the decision of an administrative judge only if it determines that such administrative judge’s decision is beyond the scope of his authority, arbitrary or capricious, or contrary to law.” “Our review of the board’s action is also limited. See Scheffler’s Case, 419 Mass. 251, 258 (1994). With these principles in mind, we review the findings of the administrative judge and the conclusions of the reviewing board.” Aetna Life & Cas. Ins. Co. v. Commonwealth, 50 Mass. App. Ct. 373, 377 (2000).
We begin our analysis by acknowledging that “[a]n injury arises out of the employment if it arises out of the nature, conditions, obligations or incidents of the employment; in other words, out of the employment looked at in any of its aspects.” Caswell’s Case, 305 Mass. 500, 502 (1940). “[T]he employee. . . need not necessarily be engaged in the actual performance of work at the moment of injury. It is enough if he is upon his employer’s premises occupying himself consistently with his contract of hire in some manner pertaining to or incidental to his employment.” Souza’s Case, 316 Mass. 332, 335 (1944).
If an employee is injured on the employer’s premises while engaged in a purely personal activity, the injury will not be considered to arise out of the employment. See, e.g., Ritchie’s Case, 351 Mass. 495 (1966) (employee fell while reaching over to pet a cat). Conversely, an injury sustained while the employee is off the employer’s premises will be compensable if that activity benefits the employer. See Corraro’s Case, 380 Mass. at 360-361. Here, the employee’s injury was sustained on BMC’s premises. We, therefore, focus on whether the activity — receipt of the flu shot — was a purely personal activity, or whether it was an incident of employment. See Souza’s Case, supra.
We recognize that whether an employee’s injury received as a result of the administration of an employer-sponsored inoculation arises out of the employment is a matter of first impression in this Commonwealth. Nonetheless, the issue has been considered by legal scholars and by appellate courts of other jurisdictions.
In most instances, the administration of an inoculation that an *764employer has neither required nor urged its employees to receive has little, if anything, to do with the employment. While the employer generally may be seen as deriving some benefit in the form of decreased employee absenteeism, that benefit, standing alone, is insufficient to justify imposing an economic burden upon the employer for an injury sustained as a result of the employee’s personal decision to receive the inoculation. In other words, under those circumstances, the activity is of such a personal nature that no real work connection may be shown.
Because the issue whether an injury arises out of a claimant’s employment is generally a question of fact, as the facts change, so do the results of the work connection calculus. “When the inoculation is not . . . strongly tied to the employment either by employer compulsion or by the special risks of the assignment, it may still be covered if there is a combination of strong urging by the employer and some element of mutual benefit in the form of lessened absenteeism and improved employee relations.” 2 Larson, Workers’ Compensation § 27.03[2], at 27-30 (1999).
In the present case, the link between the activity and the employment is particularly strong, as recognized by the reviewing board. See Monette v. Manatee Memorial Hosp., 579 So. 2d 195, 197 (Fla. Dist. Ct. App. 1991), quoting from Suniland Toys & Juvenile Furniture, Inc. v. Karns, 148 So. 2d 523, 524 (Fla. 1963) (injury from flu shot “flow[ed] as a natural consequence of the [claimant’s] employment” as a housekeeper in a hospital setting). Here, the employee was a health care worker employed by a hospital. In contrast, none of the cases requiring the employee to show either employer compulsion or that the employer strongly urged its employees to receive the inoculation involved situations where the employer and the employee provide health care services to the public. See, e.g., Smith v. Seamless Rubber Co., 111 Conn. 365 (1930) (factory worker); Smith v. Brown Paper Mill Co., 152 So. 700, 703-704 (La. Ct. App. 1934) (mill worker); Lampkin v. Harzfeld’s, 407 S.W.2d 894, 897 (Mo. 1966) (clothing store sales person); Saintsing v. Steinbach Co., 1 N.J. Super. 259, 264, aff'd, 2 N.J. 304 (1949) (department store employee); King v. Arthur, 245 N.C. 599 *765(1957) (dairy worker); Austin v. Smith, 579 S.W.2d 84 (Tex. Civ. App. 1979) (firefighter).7
In the present case, the second administrative judge found the following relevant facts. The employee was an EKG technician in BMC’s “Heart Station.” BMC offered to its employees and to the general public yearly flu shots free of charge. It advertised, via radio and printed leaflets, that it was providing these inoculations “to encourage” people to receive the flu shot. The flu shots were administered by hospital employees on the hospital’s premises and the employee received her flu shot during her lunch hour, on a day that she was working for the employer. The administrative judge rested his determination that BMC did not “derive any benefit” from offering a free flu shot to the employee on his finding that BMC did not “actively” seek to inoculate all of its employees in order to reduce absenteeism due to the flu, and the employee was not an “indispensable or difficult-to-replace employee.”
The reviewing board noted that, in reaching his conclusion, the administrative judge had failed to consider both the employee’s job duties and the nature of BMC’s business. Thus, *766the administrative judge’s decision was arbitrary and capricious. Moreover, it was contrary to law because, as a result of his truncated review of the facts, the administrative judge failed to consider “the employment looked at in any of its aspects.” Caswell’s Case, 305 Mass. at 502. The reviewing board was therefore well within its authority in reversing the administrative judge’s decision. See G. L. c. 152, § 11C.
The reviewing board concluded that the employee’s claim was compensable because, given her direct contact with high risk patients, the employee’s receipt of a flu shot was an incident of her employment, and the provision of flu shots was a “clear benefit” to BMC “by virtue of the avoidance of contagion caused by hospital employees.” We accord deference to the reviewing board’s interpretation of the requirements of G. L. c. 152. See Aetna Life & Cas. Ins. Co. v. Commonwealth, 50 Mass. App. Ct. at 379, and cases cited.
As an EKG technician in the heart station, the employee’s job required direct contact with patients.8 Moreover, in this case, the employer is a hospital that, as a matter of law, has an interest in, and commitment to, promoting the public health. Boston Public Health Act of 1995, St. 1995, c. 147, § 1(b). In light of the nature of BMC’s business, the prevention or limitation of the potential that its own employees might spread a contagious illness necessarily benefits BMC as a matter of law. See Washington Hosp. Center v. District of Columbia Dept. of Employment Servs., 821 A.2d 898, 902 (D.C. 2003) (hospital “obviously benefitted from having employees . . . immunized against diseases communicable in the hospital setting”). The employee’s receipt of the flu shot, which was both encouraged by BMC and offered on its premises, was consistent with her status as a health care worker providing direct patient care so as to warrant the conclusion that her injury arose out of her employment. Moreover, because the employee’s receipt of the flu shot plainly furthered her employer’s business interests, the benefit BMC received in this case was separate from and beyond “some element of mutual benefit in the form of lessened *767absenteeism and improved employee relations,” 2 Larson, Workers’ Compensation § 27.03[2], at 27-30 (1999). Compare Suniland Toys & Juvenile Furniture, Inc. v. Karns, 148 So. 2d at 525 (“preventive treatment [typhoid inoculation offered after employee drank contaminated water at work] sufficiently served the employer’s special interests [avoidance of employee’s potential disability resulting from exposure that occurred within course of employment] to bring the accidental reaction injury within the employment scope”). Under these unique circumstances, no greater encouragement or compulsion by the employer was required in order to conclude, as a matter of law, that the ñu shot arose out of and in the course of the employment. See Monette v. Manatee Memorial Hosp., 579 So. 2d at 197.9
Accordingly, we affirm the judgment entered by the single *768justice affirming the November 19, 2002, decision of the Department of Industrial Accidents reviewing board.
So ordered.
While being treated by Dr. Bluth, the employee underwent two CT scans that were negative for multiple sclerosis; sarcoidosis, a lung inflammation that can spread to other tissues, was also ruled out as a cause of the employee’s blindness.
Dr. Langer based her conclusion on the following: She opined that there was a poor temporal relationship between the administration of the vaccine and the reported eye problem; that none of the employee’s immunized coworkers reported the development of any symptoms; that the employee had received flu shots without incident on a yearly basis for many years prior to 1996; that it was far more likely that the employee’s optic neuritis was caused by multiple sclerosis or other risk factors; and that, given the low number of adverse events reported to the United States Food and Drug Administration associated with the flu shot, the data did not suggest any additional risk of optic neuritis from the flu shot.
The first administrative judge had ended her service with the Department of Industrial Accidents. Hicks v. Boston Med. Center, 16 Mass. Workers’ Comp. Rep. 426, 426 (2002).
Expert opinions regarding medical causation are subject to the Lanigan test. Canavan’s Case, 432 Mass. 304, 316 (2000).
BMC argues that Dr. Bienfang’s methodology was flawed because only a magnetic resonance image (MRI) could rule out the possibility that the cause of the employee’s condition was multiple sclerosis, but she was not able to undergo an MRI examination because she had a pacemaker.
Dr. Bienfang did indeed rule out multiple sclerosis, based on the following: the employee’s medical course was not typical of multiple sclerosis, two CT scans failed to show any of the typical brain lesions that would support the diagnosis, and her age at the time of onset of symptoms was at the end of the bell curve for a new onset of the disease. Contrary to BMC’s representations during oral argument, those opinions of Dr. Bienfang were part of the evidence in this matter.
We are aware that Dr. Bienfang’s conclusions as to causation differed from those of the impartial physician and BMC experts, but we emphasize that “[cjhallenges to the conclusions themselves go to the weight, not the admissibility, of the evidence.” Vassallo v. Baxter Healthcare Corp., 428 Mass. at 13.
We note that Larson has criticized cases denying workers’ compensation benefits where the courts “appear[] to assume that the employment must be the sole cause of the test or inoculation.” 2 Larson, Workers’ Compensation § 27.03[2], at 27-32 (2001). See, e.g., Smith v. Seamless Rubber Co., 111 Conn. at 369-370 (injury not compensable where employer offered small pox vaccine as a public service to assist board of health in preventing epidemic); King v. Arthur, 245 N.C. at 603 (dairy worker required by law to have a Wasserman test was denied compensation for injury sustained as a result thereof because the medical testing was required by State law, not the employer). Larson suggests instead that “[i]t would be more correct to say that the employment need only be a concurring cause.” Larson, supra. See, e.g., Washington Hosp. Center v. District of Columbia Dept. of Employment Servs., 821 A.2d 898, 902 (D.C. 2003) (citing Larson with approval and holding that injury resulting from measles, mumps, and rubella vaccination, required by law before commencing employment at a hospital, did arise out of and in the course of employment).
We agree with Larson’s view, as it appears to be more consistent with our long-standing case law. Cf. Kilcoyne’s Case, 352 Mass. 572, 575 (1967), quoting from Sylvia’s Case, 298 Mass. 27, 28 (1937) (employer benefitted from employee living on premises; “[t]he occupancy of the room. . . had ‘elements of convenience and advantage on both sides”; thus, injury sustained on day off while carrying groceries arose out of and in the course of employment).
Although the administrative judge made no findings regarding the employee’s job duties, there is ample undisputed evidence in the record that the employee’s duties involved direct contact with patients.
The BMC’s attempt to distinguish Monette from the present set of facts is unpersuasive. That the vaccine was provided to the general public in the instant matter, whereas in Monette it was offered only to employees, is of no moment, where the nature of BMC’s business is to provide such care to the public. The fact that in the present case, the employee received the vaccine during her lunch hour, whereas in Monette, the shot was administered during work hours, also matters little. There is undisputed evidence in the record that the employee’s supervisor had scheduled a time during the morning hours for her to receive the vaccine but that she did not go at the scheduled time, opting instead to receive the flu shot during her lunch hour. Moreover, even if the shot had been available only during the lunch hour, that would not necessarily preclude the employee’s claim. See Holmes’s Case, 267 Mass. 307, 308-309 (1929).
In addition, the employee’s testimony that she took the flu shot for her own protection does not negate any benefit BMC derived from offering the shot, nor does BMC point to any case that stands for the proposition that an employee’s subjective intent is the sole determinative of whether her acts benefit the employer. Furthermore, BMC’s assertion that the testimony in Monette was that “vaccinations were made available to employees at the recommendation of [the] National Advisory Council to avoid transmitting [influenza] to patients” misstates the testimony in that case. The testimony was that the flu vaccine was recommended “for persons in high risk categories, [that] the hospital has many employees who fit the high risk classification in terms of age and susceptibility to influenza. . . [and that the employee’s] health record shows that claimant has chronic pulmonary disease, placing her within the high risk category of susceptibility to influenza.” Monette v. Manatee Memorial Hosp., 579 So. 2d at 196. Similarly, in the present case, the employee’s physician recommended that she receive the flu vaccine because of her “several medical conditions including asthma and high blood pressure” as well as a “pacemaker.”