In January, 1994, Mary Kay Patterson, an anesthesiologist at the Brigham and Women’s Hospital (hospital), filed a claim with the Department of Industrial Accidents (DIA) for workers’ compensation benefits against each of her putative employers, the hospital and the Brigham and Women’s Hospital Anesthesia Foundation (foundation).2 She al*587leged that she had developed disabling respiratory problems (chiefly asthma) and “hyper-allergic sensitivity to . . . skin” because of exposure to airborne toxic substances in the course of her employment at the hospital. An independent medical examiner appointed by a DIA administrative judge subsequently found that she suffered, from “diffuse [or multiple] chemical sensitivity” syndrome (MCS). Following the proceedings described below, the DIA administrative judge awarded Patterson benefits for her medical expenses (plus costs and attorney’s fees) apportioned between the two organizations.
After unsuccessful appeals to the reviewing board of the DIA (which summarily affirmed the judge’s decision),3 the hospital (which is a self-insurer for workers’ compensation purposes) and Liberty Mutual Insurance Company (the workers’ compensation insurer for the foundation) here seek to reverse the judge’s determinations.4 They contend that the judge committed two major reversible errors:5 Liberty faults the judge for concluding that Patterson was an employee of the foundation as *588defined in the workers’ compensation statute.6 Both Liberty and the hospital (which conceded Patterson’s claimed status as its employee) challenge the judge’s adoption of an impartial medical examiner’s opinion as to a causal relationship between Patterson’s claimed medical problems and workplace conditions at the hospital.7 We conclude that the judge did err as asserted with respect to the causation issue and reverse the DIA decision on that ground.8
Factual summary. 9 Patterson applied for appointment to the professional staff of the hospital in May, 1981. In June, 1981, she was offered a staff appointment as anesthesiologist, at a fixed salary, in the hospital’s department of anesthesia. The offer came on hospital letterhead from the then-chairman of the hospital’s department of anesthesia, who was also the ex officio president of the foundation (by virtue of a pro forma appointment by the hospital’s board of trustees), but made no reference to the foundation. Each anesthesiologist appointed to the hospital staff by its board of trustees and to the Harvard Medical School faculty (as Patterson had already been while at the Massachusetts General Hospital) automatically became associated with the foundation. Patterson’s duties involved three days a week of clinical service in the operating room anesthetizing hospital patients undergoing operations.and instructing interns and residents in training there; two nonclinical days per week for research and “academic medicine”; and teaching duties at the Harvard Medical School. She neither admitted patients of her own to the hospital nor maintained a private office. Throughout the entire relevant period, Patterson received paychecks and benefits from both the hospital (approximately 16-20 per cent of her total remuneration) and the foundation (the remaining 80-84 per cent). The foundation was the contractually exclusive provider of anesthesia services for the hospital’s patients, whom it directly billed for such services.
*589Until late 1992 Patterson experienced no significant health problems. At that time, she began to develop sensitivity to the latex gloves she wore in the operating room. Switching to non-latex gloves and using skin ointment eliminated the problem. In March, 1993, however, she developed a cough in the operating room and later other respiratory symptoms (shortness of breath, wheezing, and asthma attacks) that abated when she ceased working at the hospital but intensified when she returned to work. Treatment from various specialists and efforts to create allergy-free environments lessened but did not extinguish her symptoms. By August, 1993, she could no longer work and over the remainder of 1993 began to have respiratory distress symptoms when exposed to an increasing number of extra-hospital agents and situations (e.g., books, fragrances, vehicle exhausts, dry cleaning, cosmetics, newspapers). Ultimately, she had to remake much of her home and restrict many activities in order to minimize such exposures.
After the hospital and the foundation rejected Patterson’s workers’ compensation claims, a conference before a DIA administrative judge ensued, pursuant to G. L. c. 152, § 10A. At that conference, Patterson submitted her treating physicians’ medical records, plus several engineering reports concerning environmental conditions at the hospital. The judge initially denied Patterson’s claims for temporary total disability (under G. L. c. 152, § 34), partial disability (§ 35), and medical expenses (§§ 13 & 30). Patterson appealed for a de novo hearing pursuant to §§ 10A(3) and 11 only as to her claims for medical expenses.10
A § 11A impartial medical examination was thereafter ordered by the judge and performed by Dr. Christiani (IME), an internist specializing in occupational environmental medicine (which is not recognized as a certifiable specialty by the American Board of Medical Specialists). The IME had no personal knowledge of or experience in the hospital’s operating rooms during the relevant period, nor any demonstrated expertise in the areas of toxicology, environmental engineering, or heating, ventilation, or air conditioning. The environmental *590reports submitted at the conference were transmitted by the judge to the IME, along with the medical records from Patterson’s treating physicians.
After examining Patterson, the IME filed his report in July, 1994. He opined that (1) Patterson suffered from latex hypersensitivity; asthma “triggered by exposure to agents, including latex and, in particular, by re-entry into the operating room areas at the Brigham and Women’s Hospital where latex antigen has been shown to be elevated (as expected) . . . [although] other agents may also be responsible for triggering her asthma”; perennial sinusitis and rhinitis; and “Diffuse Chemical Sensitivity, some of which is consistent with irritant and allergen induced asthma attacks”; and (2) “[Patterson’s] exposures at the Brigham and Women’s Hospital did causally contribute to the development and progression of her condition . . . [and s]he is unable to return to that hospital’s operating room and to any other operating room, utilizing similar materials.”
The judge then conducted three days of evidentiary hearings between August and early November, 1994, during which the IME’s report (but no other medical evidence) was admitted in evidence.11 An additional day of medical testimony was subsequently introduced in the form of the deposition of the IME, conducted on November 15, 1994. Motions by the hospital and Liberty to strike the IME’s report and deposition testimony and the environmental/engineering reports which the IME had consulted were later denied by the judge,12 who, in February, 1995, adopted the medical opinions of the IME regarding *591diagnosis and causation and determined that Patterson was concurrently employed by both the hospital and the foundation. As to the foundation’s denial of Patterson’s employee status, the judge found to the contrary, based upon the foundation’s payment of salary and fringe benefits to Patterson, the identification of Patterson as an employee on W-2 forms and other tax documents, and a 1993 workers’ compensation audit that identified the foundation as the policyholder and its affiliated anesthesiologists, including Patterson, as the insureds.13
*592 Discussion. It is axiomatic that an employee has the burden of estabhshing, by a preponderance of the evidence, all the elements of her claim for workers’ compensation benefits, including the fact of the requisite causal connection between her injury and workplace events or conditions; and that she cannot prevail if any critical element is left to surmise, conjecture or speculation or otherwise lacks evidential support. Sponatski’s Case, 220 Mass. 526, 527-528 (1915). Ruschetti’s Case, 299 Mass. 426, 431-432 (1938). See Phillips’s Case, 41 Mass. App. Ct. 612, 618 (1996).
In particular, the expert medical opinion as to that causal relation which the employee must obtain in order to prevail (when the matter is, as here, beyond the common knowledge of the ordinary layperson) has to be expressed in terms of probability, not mere possibility. Sevigny’s Case, 337 Mass. 747, 749-750 (1958). That opinion must further be based solely on the expert’s “direct personal knowledge” or admissible evidence in the record and not on assumptions that are not established by such evidence. Commonwealth v. Roman, 414 Mass. 235, 238 (1993). See Haley’s Case, 356 Mass. 678, 681-682 (1970); 452 *593Code Mass. Regs. § 1.11(5) & (6) (1993). Cf. Department of Youth Servs. v. A Juvenile, 398 Mass. 516, 531 (1986) (“an expert [may] base an opinion on facts or data not in evidence if the facts or data are independently admissible”). If the DIA factfinder rests his ultimate conclusion as to causality on an expert’s opinion which does not satisfy these standards, that conclusion is lacking in the required evidential support and cannot be affirmed. Sevigny’s Case, 331 Mass. at 749-754; Hachadourian’s Case, 340 Mass. 81, 85-86 (1959). Such is the instant situation, in which a deficient IME’s report was the sole evidence of causation relied on by the judge.
We note initially a structural defect in the IME’s report that (while not in and of itself fatal) was a portent of the underlying inadequacy of his causation opinion. General Laws c. 152, § 11A(2), as amended by St. 1991, c. 398, § 30, requires the IME’s report to “contain a determination [inter alia] . . . whether or not within a reasonable degree of medical certainty any . . . disability [of the claimant determined by the IME to exist] has as its major or predominant contributing cause a personal injury arising out of and in the course of the employee’s employment.” Here, however, the report failed to state any aspect of the IME’s opinion as being within a reasonable degree of medical certainty or that any particular workplace event or condition was the major or predominant cause of Patterson’s disabling respiratory problems or claimed MCS. His assertion that Patterson’s exposures to latex and unspecified operating room agents at the hospital “did causally contribute” to those problems fell far short of that requirement.14
Nor did the IME’s deposition testimony eliminate or *594compensate for his report’s inadequacies. He several times admitted that he could not identify and did not know what asthma-inducing allergen or toxin Patterson might have come in contact with at the hospital, other than latex. Even as to latex — which he did think “within a reasonable degree of medical certainty . . . was . . . more likely than not a problem” — he acknowledged that Patterson herself had testified that exposure to latex had produced a skin rash but not respiratory problems; that he had no actual evidence Patterson in fact had an allergy to latex which could produce a respiratory response (the establishment of which required special tests never administered to Patterson); and that he had no information regarding the levels of latex in the hospital operating rooms but rather relied on “reasonable assumption[s]” about hospital operations generally and on certain of the environmental reports he had received from the judge which “suggested] that latex allergy had been a problem” in the hospital operating rooms. His deposition opinions as to latex came down to his “thinking]” that a latex allergen “may develop, precipitate asthma attack in her” and “more likely than not latex was a problem with her.” Notwithstanding his deposition elaboration, the IME came nowhere close to opining that either latex or any other substances (none of which he could identify) was the, or even a, *595major or predominant contributing cause of Patterson’s disabling injury.
The facial shortcomings in the IME’s report and supporting testimony reflect the more fundamental defect that there was no competent evidentiary basis for his causation opinion. Most fatally, his conclusions regarding Patterson’s exposure to latex impermissibly rested on assumptions and information not established (as was required) by his own direct personal knowledge or by admissible evidence in the record. Patterson, on whom the burden lay, proffered no testimony, expert or lay, and no admissible exhibits regarding air quality or the presence of any potentially asthma-inducing agents in any part of the hospital. Instead, she relied entirely on a few nonmedical engineering reports — not made part of the administrative record — which apparently described certain environmental conditions at the hospital, including the operating room where she worked. Those reports were the only sources of information regarding latex problems aside from the IME’s claimed knowledge of hospital conditions generally.
Patterson submitted those reports at the informal § 10A conference (which is not subject to the rules of evidence) but never authenticated them or formally introduced them in evidence. They were entirely hearsay documents for which no recognized hearsay exception was ever claimed or proffered, or, so far as we can tell, exists.15 Nonetheless, at the § 11 stage of the process — which must be conducted as a de novo hearing solely on the basis of evidence therein introduced and admissible under the rules of evidence applied in the courts — the judge permitted those unadmitted and inadmissible documents to be sent to the IME. Despite the fact that DIA regulations specify that the IME should receive and rely on as documentary evidence only medical records and reports, see 452 Code Mass. Regs. §§ 1.10(2)(b), 1.11(6), 1.14(2) (1993) (which are documents within established hearsay exceptions, see G. L. c. 152, § 20; c. 233, § 79; O’Brien’s Case, 424 Mass. 16, 23 [1996]), *596the IME proceeded to review and, at least in part, rely upon those same hearsay reports.16
The result was a medical opinion, even were it not facially wanting, contrary to the governing regulations and to this Commonwealth’s rules of evidence as to the permissible bases for expert opinions. See discussion supra at 592-593.17 To the extent the IME claimed at deposition that his opinion would not have “substantially” changed had he lacked those reports — which claim the judge found credible — he effectively acknowledged that the opinion was based entirely on Patterson’s dermal latex sensitivity, on the post-March, 1993, reports of her treating physicians who diagnosed her asthmatic condition, and on his *597own claimed knowledge of typical conditions in hospital operating rooms and other medical areas (though not specifically at the hospital itself). None of those sources, however, provided an iota of expert or admissible evidence regarding the actual, as opposed to the hypothesized, presence of latex or any other asthma-causing substances in the hospital operating room or other hospital areas in which Patterson had worked. As such, it was merely an unsubstantiated opinion based on assumed facts not established by the admissible evidence and thus the product of surmise and conjecture as to the existence of any causal connection, much less the “major or predominant contributing cause.”18
The judge thus rested his causation decision on an IME’s report that was not only expressed in terms of mere possibility but was also unsupported by admissible evidence in the record or on any other proper basis. As such, the IME’s report was not entitled to any evidential weight, much less the status of prima facie evidence, see G. L. c. 152, § 11A(2), which the judge incorrectly accorded it. See 452 Code Mass. Regs. § 1.11(5) & (6) (1993) (medical report, or any part thereof, to be struck on motion when not based solely on the expert’s direct personal knowledge or admissible evidence in the record); Ruschetti’s Case, 299 Mass. at 431 (mere conjecture by expert as to causation based on equivocal facts “has ... no evidential value”); Sevigny’s Case, 337 Mass. at 749 (medical expert opinion as to causation expressed in terms of “possible causal relationship *598. . . would have to be disregarded”); Buck’s Case, 342 Mass. 766, 770-771 (1961) (expert causality opinion based upon misstatements or omissions of material facts is entitled to no weight); Bagge’s Case, 369 Mass. 129, 134 (1975) (same); Lettich’s Case, 403 Mass. 389, 393 (1988) (DIA reviewing board found that expert opinion based on facts not in evidence “must be rejected” and could not support a decision favorable to the employee); Scheffler’s Case, 419 Mass. 251, 257-259, 260 n.5 (1994) (prima facie effect is to be accorded only to medical opinion as to medical issues based on accurate and complete facts and a competent evidentiary foundation in the record); Collins’s Case, 21 Mass. App. Ct. 557, 563 (1986) (expert opinion based on either evidence not properly introduced or facts not proven should be struck on motion of the objecting party); Coggin v. Massachusetts Parole Bd., 42 Mass. App. Ct. 584, 588-589 (1997) (EME report is not to be admitted as prima facie evidence when based on inaccurate assumptions).19
The ultimate causality conclusion of the judge — who had no power to admit evidence at the hearing in a manner contrary to the DIA regulations or the rules of evidence (at least when such evidence is properly objected to) — consequently was itself not grounded, as it must be, on a competent evidentiary record. It thereby contravened the fundamental principles that the appellants were “entitled to a decision based [solely] on the evidence presented at the hearing . . . [, that n]othing can be considered or treated as evidence which is not introduced [and admitted] as such . . . [and that] the [judge’s] decision on a workers’] compensation claim may not be based on facts not introduced in evidence . . . .” Haley’s Case, 356 Mass. at 681-682. See 452 Code Mass. Regs. § 1.11(5) (1993); Hachadourian’s Case, 340 Mass. at 86.
Patterson in the end failed to present, as she was obligated to do, any expert medical evidence that would establish, by a *599preponderance of the evidence, that her diagnosed illnesses and disability were causally related to her employment in any way, and in particular that her disability had as its major or predominant contributing cause any specific workplace condition or particular irritant therein. It follows that the reviewing board’s decision in her favor must be reversed. See Sevigny’s Case, 337 Mass. at 753-754, and cases cited.
So ordered.