The plaintiff, Gilbert M. Lisbon, had been a steward at the Bristol County house of correction for ten years when he suffered an incapacitating heart attack on March 30, 1989, while working at the facility’s cook house.2 Relying on the 1991 medical opinion of his personal physician and the certificate of a regional medical panel, Lisbon claimed accidental disability benefits, pursuant to G. L. c. 32, § 7(1), on the ground that his attack was precipitated by a prison riot and a confrontation with prisoners on March 30, 1989.
Citing Lisbon’s many health problems that predated his jail employment (which included coronary artery disease, three prior heart attacks, high blood pressure, diabetes, obesity, unhealthy eating habits, and a long-term smoking habit that lasted until 1988) as the more likely cause of his incapacity, the Bristol County Retirement Board (the board) rejected Lisbon’s claim. On Lisbon’s appeal to the Contributory Retirement Appeal Board (CRAB), an administrative magistrate recommended approval of his benefit application on the strength of his medical evidence. CRAB, however, concluded that Lisbon had failed to establish that his disabling heart attack was the result of an incident or a series of incidents occurring while in the performance of his duties rather than the product of the natural progression of his coronaiy arteiy disease and associated long-term illnesses.
Lisbon sought judicial review under G. L. c. 30A, § 14(7). *248A judge in the Superior Court found the magistrate’s analysis persuasive and ruled that CRAB had wrongly denied Lisbon accidental disability benefits. On the board’s appeal, we conclude that CRAB did not err in determining that Lisbon had failed to demonstrate his entitlement to those benefits and, accordingly, reverse the judgment of the Superior Court.
Background. The relevant factual background of this case is undisputed.3 Lisbon had, as noted above, a long history of medical problems, including “[arteriojsclerotic coronary heart disease.” During the time he was employed as a cook at the Bristol correctional facility, he had experienced an “inferior wall myocardial infarction” in 1983 and “coronary episodes” in 1986 and 1987. All of these “heart attacks” required hospitalization but were not (on this record) job-related. On March 30, 1989, while Lisbon was working at the cook house, a prison riot broke out elsewhere at the facility. Lisbon could see some of the riot from his cook house window, including inmates setting fires in the prison yard. At one point, “several” inmates entered the cook house, but left without incident when Lisbon ordered them out. There is no indication that the inmates said or did anything to menace or threaten Lisbon. He nonetheless locked the cook house. Correction officials shut down the jail and ordered all personnel, including Lisbon, to remain at their stations.
Soon thereafter, Lisbon began to have chest pain and was taken to the rear gate where emergency equipment and riot police were located. Lisbon was there examined by emergency medical technicians and then taken to St. Luke’s Hospital. He was diagnosed as having “coronary artery disease” and remained hospitalized for two and one-half weeks. His personal physician noted in a January, 1991, report to Lisbon’s attorney that Lisbon’s chest pain was “clearly related” to emotional and physical stress experienced at the jail during the riot and that Lisbon had suffered a “myocardial infarction” at work. Lisbon has not worked since March 30, 1989.
Lisbon applied for accidental disability retirement in July, 1990. On his application, Lisbon cited his March 30, 1989, “heart attack” as the “personal injury or hazard undergone *249by reason of which [he] claimed to be incapacitated.” His claim was confirmed, as noted above, by the written opinion of his personal physician, who also concluded that Lisbon was permanently disabled by reason of the job-related injury. Following a July 17, 1991, examination, a regional medical panel certified that Lisbon was permanently disabled and that his disability was “such as might be the natural and proximate result of personal injury sustained or hazard undergone on account of which retirement is claimed.”4
Because the medical panel suggested that it had, to some extent, relied on the “heart law,” G. L. c. 32 § 94,5 in reaching its decision, despite Lisbon’s ineligibility under that law, the board sought clarification from the medical panel. In October, 1991, the chairman of the panel responded, stating that “[djespite our statement to the contrary, Mr. Lisbon’s application for disability retirement was not considered under the ‘Heart Law’ but rather under Accidental Disability.” The panel also declined to amend its conclusions. The chairman further stated: “A detailed history regarding the confrontation [at the jail on March 30, 1989] was obtained and this convinced the Panel that it was highly stressful and productive of symptoms leading to his hospitalization. . . . The Panel finds there is a causal link to work in this case, despite the cited risk factors.”
On November 20, 1991, the board denied Lisbon’s application. Lisbon filed an administrative appeal to CRAB. An *250administrative magistrate of the Division of Administrative Law Appeals heard the case on March 4, 1993. The only testifying witness was Lisbon; all medical evidence was documentaiy. On May 15, 1993, the magistrate recommended that the board award Lisbon accidental disability retirement benefits, concluding as follows:
“The Appellant is an employee of the Bristol County House of Correction. He is not an employee of the Department of Correction and therefore is not entitled to the presumption under G. L. c. 32 sec. 94. While his duties may be the same as an employee within the Department of Correction, the statute is specific as to who is included and entitled to the ‘heart law’ presumption.
“The medical panel has found that the riot of March 30, 1989, with regard to the Appellant, was highly stressful and productive of symptoms leading to his hospitalization and a heart attack. It was their opinion, originally and upon request for clarification, that the riot was the causal link leading to Mr. Lisbon’s disability. The panel was aware of the Appellant’s past medical history and the risk factorfs] he had with regard to heart disease before they made their decision. No evidence was introduced to rebut the findings of the panel or to show that the panel employed an erroneous standard.
“Based on the above, I find that the Appellant has established that the riot of March 30, 1989, was the cause of his disability. This matter is remanded to the Respondent for the awarding of accidental disability benefits.”
The board objected to the magistrate’s decision and requested that CRAB deny Lisbon’s application for disability retirement. On October 23, 1993, CRAB adopted the magistrate’s factual findings, but denied Lisbon disability benefits on the following analysis:
“A condition precedent to the allowance of accidental retirement benefits is the affirmative certification by a *251majority of the Medical Panel that the applicant is substantially unable to perform his duties, that such disability is likely to be permanent, and that there is a medical possibility of a causal relationship between the disability and a personal injury or hazard undergone while in the performance of his duties.
“In this case, the members of the Medical Panel certified that in their opinion [Lisbon] is permanently disabled from performing the duties of a steward at the Bristol County House of Correction and further indicated that there is a medical possibility of a causal relationship between the Appellant’s disability and a personal injury or hazard undergone while he was in the performance of his duties. Nonetheless, G. L. c. 32, § 7, ‘gives to the local board [and subsequently on appeal to this Board], the determination whether, on the relevant evidence, the causal relationship exists.’ Noone v. Contributory Retirement Appeal Board, 34 Mass. App. Ct. 756 (1993).
“Here the facts indicate that the Appellant was working in the cook house when a riot broke out on March 30, 1989. He viewed the disturbance out a window and, when several inmates entered the cook house, he ordered them out. Later he experienced chest pain. The facts also indicate that the Appellant has a past history of hypertension, coronary problems, exogenous obesity and diabetes mellitus.
“The burden is on the Appellant to prove the causal nexus between his disability and an incident or series of incidents occurring while in the performance of his duties which resulted in that disability. The Appellee, consistent with its responsibilities, held that the Appellant failed to sustain that burden. Based upon the facts as found and the evidence presented herein, this Board concurs with that finding.
“The decision of the Appellee denying the Appellant’s application for accidental retirement benefits is, therefore, affirmed.”
*252On November 12, 1993, Lisbon filed this action for review of CRAB’s adverse decision. After a hearing, a Superior Court judge characterized CRAB’s determination that Lisbon had failed to sustain his burden of proving work-related causation as “conclusory . . . [and] against the clear weight of the evidence6 .... [It] reverses, without reason, the Administrative Law Judge’s [ALJ] well founded conclusion that [Lisbon] is entitled to benefits. The record overwhelmingly indicates that [Lisbon’s] exposure to the rioting prisoners caused him to suffer a heart attack.” The judge observed that CRAB provided no “adequate explanation [or] . . . basis for its disagreement with the ALJ, from which this court concludes that the agency’s findings are not supported by substantial evidence.” The instant appeal followed. We uphold CRAB’s disposition, however, for several reasons.
1. The judge, citing Vinal v. Contributory Retirement Appeal Bd., 13 Mass. App. Ct. 85, 100 (1982), reversed CRAB in substantial part because CRAB failed to “explain” its “disagreement” (as perceived by the judge) with the magistrate’s “rulings” regarding causation. That failing the judge regarded as fatally depriving CRAB’s “findings” of substantial eviden-tiary support. The judge’s conclusion in this regard was not, however, supported or mandated by Vinal. The only occasion requiring agency explanation for its disagreement with a hearing officer’s findings arises when those findings are based *253upon live witness testimony and credibility determinations made thereon. Vinal, 13 Mass. App. Ct. at 100-102. No such medical (or other) testimony was presented here with respect to the issue of causation.7
CRAB’s d. ision, agreeing with that of the board, explicitly concluded on the basis of medical and other evidence, including the magistrate’s subsidiary findings (which CRAB expressly adopted), that Lisbon had not sufficiently established causation. CRAB’s statement satisfied its responsibility to provide a reasoned articulation for its rejection of the magistrate’s recommendation so as to allow meaningful appellate review. See Maddocks v. Contributory Retirement Appeal Bd., 369 Mass. 488, 497 (1976); Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. 479, 483 n.2 (1985). The judge erred insofar as he held CRAB’s decision to be inadequately explanatory and for that supposed reason unsupported by substantial evidence.
2. In ruling that Lisbon was entitled to disability benefits, the judge rested on the magistrate’s decision, which he praised as being “well founded” on the basis of “overwhelming[ ] [evidence].” The magistrate’s decision was, unfortunately, doubly flawed. The magistrate explicitly based his conclusion upon the medical panel’s “findings” that the riot “was highly stress*254ful and productive of symptoms leading to [Lisbon’s] hospitalization and a heart attack” and the panel’s “opinion, originally and upon request for clarification, that the riot was the causal link leading to Mr. Lisbon’s disability.” The medical panel certification did not, however, include any such “findings.” As was its obligation, see G. L. c. 32, § 6(3); 840 Code Mass. Regs. § 10.10(8) (1989), the panel certified only whether or not the disability was “such as might be the natural and proximate result of the personal injury sustained or hazard undergone on account of which retirement is claimed” (emphasis added), in Lisbon’s case responding affirmatively.
Such a medical panel certification “is not conclusive of the ultimate fact of causal connection but stands only as some evidence on the issue.” Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 483. It is a mere statement of “medical possibility.” Noone v. Contributory Retirement Appeal Bd., 34 Mass. App. Ct. 756, 762 (1993). “The final determination . . . whether causation was proved [is] reserved to [CRAB], based on the facts found and all the underlying evidence, including both the medical and non-medical facts.” Blanchette, supra. See also Kelley v. Contributory Retirement Appeal Bd., 341 Mass. 611, 616 (1961). The existence of such an affirmative certification is not in itself a basis for concluding that an applicant has satisfied his burden of proving the causal connection between his disability and a work-related accident or incident, see id. at 614; Wakefield Contributory Retirement Bd. v. Contributory Retirement Appeal Bd., 352 Mass. 499, 502-503 (1967); Shrewsbury Retirement Bd. v. Contributory Retirement Appeal Bd., 5 Mass. App. Ct. 379, 381 (1977), particularly in light of “the strict causation standard imposed by [G. L. c. 32].” Blanchette, supra at 485. To the extent the magistrate’s causation decision in favor of Lisbon rested on the medical panel’s original certification, it could have been reasonably set aside by CRAB, and the judge’s conclusion relying on that decision is similarly vulnerable. See ibid.
The magistrate further improperly relied on the chairman of the medical panel’s subsequent letter of “clarification” stating that there “is a causal link” between the riot and Lisbon’s disabling condition. The chairman’s statement, even assuming it accurately reflected the other panelists’ opinions (as to which the record is silent), is entitled to no weight. “The *255panel has no statutory authority to express an unqualified . . . opinion as to causation, and such an opinion, if expressed, is a nullity . . . [T]he statute gives to the local board, not to the panel, the determination whether, on the relevant evidence, the causal relationship exists.” Noone v. Contributory Retirement Appeal Bd., 34 Mass. App. Ct. at 762. The chairman’s statement was the sort of “gratuitous” observation — going beyond the panel’s statutory duty by opining about actual causation and evaluating purported evidence of causation — that is inappropriate in disability retirement proceedings. See Kelley v. Contributory Retirement Appeal Bd., 341 Mass, at 613-614; Wakefield Contributory Retirement Bd. v. Contributory Retirement Appeal Bd., 352 Mass, at 502-503; Noone v. Contributory Retirement Appeal Bd., 34 Mass. App. Ct. at 761-762. The judge’s approbation of the magistrate’s dependence upon such ultra vires medical panel views is derivatively inappropriate.8
3. Lisbon had the burden of proof on the question whether there existed the requisite causal relationship between his incapacity and a work-related incident. See Wakefield Contributory Retirement Bd. v. Contributory Retirement Appeal Bd., 352 Mass, at 502. To discharge that burden, it was incumbent upon him to establish his case by a preponderance of the evidence. He had to show that it was more likely that his disabling heart attack was caused by his exposure to the circumstances of the riot on March 30, 1989, either directly or by aggravation of his preexisting heart disease, than by the natural, cumulative, deteriorative effects of his preexisting diseased condition and unhealthy habits. See Sargent v. Massachusetts Accident Co., 307 Mass. 246, 250 (1940); Josi’s Case, 324 Mass. 415, 417-418 (1949); Zavaglia v. Contributory Retirement Appeal Bd., 345 Mass. 483, 487 (1963); King’s Case, 352 Mass. 488, 489-492 (1967); Craven v. State Ethics Commn., 390 Mass. 191, 200 (1983); Adams v. Contributory Retirement Appeal Bd., 414 Mass. 360, 365-367 (1993); Robinson v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 638-639.
*256The judge’s description of Lisbon’s causation case as supported by “the clear weight of the evidence,” and “overwhelming evidence,” was hyperbolic. The only competent evidence in Lisbon’s favor on that issue was the medical panel’s certification and his personal physician’s subsequent opinion.9 As discussed in the preceding section, an affirmative certification cannot by itself discharge an applicant’s burden of proving the required causal connection; and neither the board nor CRAB was required to accept the documentary “conclusions [as to causation] expressed by [Lisbon’s] personal physician.” Campbell v. Contributory Retirement Appeal Bd., 17 Mass. App. Ct. 1018, 1019 (1984). The quantum of evidence presented by Lisbon on the causation issue — establishing the possibility of a causal connection between the riot and his disabling heart attack — was reasonably deemed insufficient by CRAB when measured against the extensive evidence of his long-standing health problems and proclivity to heart attacks. See Josi’s Case, 324 Mass, at 417-418 (1949); Tartas’s Case, 328 Mass. 585, 587 (1952); Cataldo v. Contributory Retirement Appeal Bd., 343 Mass. 312, 313-314 (1961); Campbell v. Contributory Retirement Appeal Bd., 17 Mass. App. Ct. at 1019 (“In the circumstances, and in light of the substantial evidence concerning the plaintiffs longtime smoking habit, there was no error in CRAB’s conclusions that the plaintiff had failed to sustain his burden of proof on the question of causation”); Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 483-484.
Lisbon’s presentation — supported by no medical testimony —was even weaker than that of disability applicants held to have failed to discharge their burden of proof despite expert testimony in their favor. See State Bd. of Retirement v. Contributory Retirement Appeal Bd., 342 Mass. 58, 65-66 (1961); Cataldo v. Contributory Retirement Appeal Bd., 343 Mass, at 313-314; Maddocks v. Contributory Retirement Appeal Bd., 369 Mass, at 495-497. Lisbon’s case stands in marked contrast to the one he most relies on, Robinson v. Contributory Retirement Appeal Bd., supra, because of the existence there of (1) “uncontradicted [medical] testimony” that the applicant’s myocardial infarction was caused by an activity on the job *257and (2) no contrary medical evidence. 20 Mass. App. Ct. at 636-641. See also Teacher’s Retirement Bd. v. Contributory Retirement Appeal Bd., 346 Mass. 663, 665-667 (1964).
4. As noted earlier (note 5, supra) a court may not set aside a CRAB decision, not otherwise legally erroneous, that is supported by substantial evidence. Contrary to the judge’s conclusion, CRAB’s denial of disability benefits to Lisbon was so supported. Judicial “review under the substantial evidence standard is circumscribed.” Cherubino v. Board of Registration of Chiropractors, 403 Mass. 350, 354 (1988). It is a standard of review “highly deferential to the agency”, which requires (as G. L. c. 30A, § 14[7] mandates) according “ ‘due weight to the experience, technical competence, and specialized knowledge of the agency, as well as to the discretionary authority conferred upon it.’ ” Flint v. Commissioner of Pub. Welfare, 412 Mass. 416, 420 (1992).10
In order to be supported by substantial evidence, an agency conclusion need not be based upon the “clear weight” of the evidence (see note 5, supra) or even a preponderance of the evidence, but rather only upon “reasonable evidence,” Medical Malpractice Joint Underwriting Assn. of Mass. v. Commissioner of Ins., 395 Mass. 43, 54 (1985), i.e., “such evidence as a reasonable mind might accept as adequate to support a conclusion,” after taking into consideration opposing evidence in the record. G. L. c. 30A, §§ 1(6), 14(8); New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 466 (1981). If the agency has, in the discretionary exercise of its expertise, made a “choice between two fairly conflicting views,” and its selection reflects reasonable evidence, “[a] court may not displace [the agency’s] choice . . . even though the court would justifiably have made a different choice had the matter been before it de novo.” Southern Worcester County Regional Vocational Sch. Dist. v. Labor Relations Commn., 386 Mass. 414, 420 (1982); Retirement Bd. of Somerville v. Contributory Retirement Appeal Bd., 38 Mass. App. *258Ct. 673, 677 (1995). See also Wheat v. United States, 486 U.S. 153, 164 (1988) (the essence of discretionary authority is the power to choose within a range of acceptable options, whether or not courts might reach differing or even opposite conclusions on the same record).
CRAB’s decision satisfies the “substantial evidence”standard. There is undisputed record evidence that Lisbon’s coronary artery disease, high blood pressure, diabetes, and excessive obesity long antedated his government employment. Notwithstanding repeated warnings from his physician, he refused to quit his life-long heavy smoking habit until shortly before his disabling incident. He would not abandon his unhealthy eating habits despite years of medical and counseling efforts. He suffered heart attacks in 1983, 1986, and 1987, none of which he suggests was related to his job or its unusual circumstances. His physician noted as early as 1986, in connection with one of his coronary episodes, that Lisbon exhibited “obvious severe risk factors,” and by 1987 presented with “increasing angina.”
This was not a case in which no or little contrary medical evidence existed detracting from the applicant’s medical evidence, as in Vinal v. Contributory Retirement Appeal Bd., 13 Mass. App. Ct. at 90-91, or Robinson v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 640. “Nor is this a case where the [CRAB] decision ... is not in accord with the facts found [and in the record].” McCarthy v. Contributory Retirement Appeal Bd., 342 Mass, at 49. Rather, despite “evidence tending to show that . . . [Lisbon’s disability] was a result of the [March 30, 1989, incident] ... it cannot be said that there was no evidence from which a reasonable mind could reach a contrary conclusion.” Id. at 48. CRAB might, in its discretion, have determined that Lisbon’s last heart attack was job-related. However, in applying its expertise to the substantial evidence in this record indicating the possibility of natural disease progression as. the cause of the disability, CRAB was not required to do so. It was reasonable for CRAB to base its adverse decision on the record of Lisbon’s predisposition to heart attacks, especially where the only evidence tending to show work-related causation was the inconclusive medical panel certification and the post hoc conclusion of Lisbon’s physician. See id. at 48-49; Cataldo v. Contributory Retirement Appeal Bd., 343 Mass, at 313-314; *259Campbell v. Contributory Retirement Appeal Bd., 17 Mass. App. Ct. at 1019; Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 4835 485-488.
Conclusion. Accordingly, we reverse the judgment below and order that a judgment be entered affirming CRAB’s denial of accidental disabibty retirement benefits to Lisbon.
So ordered.
We reject the board’s contention that Lisbon is not entitled to disability benefits because he was not “in the performance of his duties” but rather was “waiting to leave work for the day.” The board cites only distinguishable authority for this questionable proposition and ignores the fact that Lisbon was ordered to stay at his post until the riot was over.
The facts come from the findings of the administrative magistrate, which CRAB adopted as its own and which neither party challenges, and the underlying medical reports and certifications that were exhibits before the magistrate.
The medical panel explained: “The Panel considered his current state of health due to the diagnosis of coronary heart disease and diabetes mellitus and history of hypertension and came to the unanimous conclusion that he could not continue his present job as a cook in the correctional facility. The Panel accordingly has answered yes to Question One in Section F and yes to Question Two in Section G because of the progressive nature of [arteriosclerotic] coronary heart disease. The Panel carefully considered the history of confrontation with prisoners which necessitated hospital admission and the diagnosis of acute myocardial infarction was made on the April 4, 1989 hospital admission. The Panel has therefore answered yes to Question Three in Section H.”
General Laws c. 32, § 94, first enacted in 1950, presently creates a causation presumption in favor of fire fighters, police officers, and certain other public employees, including Department of Correction staff assigned to the supervision and custody of prisoners that any disabling or fatal heart condition experienced during employment shall be deemed to have been suffered in the line of duty. Lisbon did not qualify as a covered person under the statute.
In ruling that CRAB’s determination that Lisbon had “failed to sustain his burden goes against the clear weight of the evidence,” the Superior Court judge misstated the court’s obligation when reviewing a CRAB decision. Judicial review of a CRAB decision under G. L. c. 30A, § 14, is narrower than that. On appeal, courts are not called upon to determine whether the CRAB decision was based on “the weight of the evidence,” nor may they substitute their judgment for that of CRAB. Rather, they may set aside or modify a CRAB decision only if, after giving due regard to CRAB’s expertise in applying the public retirement statute to a variety of factual situations, see Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. 478, 487-488 (1985), they conclude, on the basis of the entire record, that “the [CRAB] decision is . . . [unsupported by substantial evidence,” G. L. c. 30A, § 14(7)(e), i.e., “such evidence as a reasonable mind might accept as adequate to support (CRAB’s) conclusion, G. L. c. 30A, § 1(6).” McCarthy v. Contributory Retirement Appeal Bd., 342 Mass. 45, 47 (1961). Since, however, the judge went on to restate his reasoning for reversing CRAB’s decision in terms of the “substantial evidence” test, we shall disregard his earlier anomalous phraseology and review his decision as one holding that CRAB’s denial of disability benefits for Lisbon was unsupported by substantial evidence.
The only testimony in any way relevant to the causation issue was that of Lisbon which formed the basis for the magistrate’s finding that Lisbon “began to experience chest pain” after the inmates had left the cook house and after he had been informed that all jail personnel were being held over because of the riot. CRAB did not disagree with but rather accepted this finding. The mere coincidence of Lisbon’s chest pain and the workplace incident, however, “do[es] not, standing alone, support [a] finding[ ] of causation . . . [which] ‘must rest upon expert medical testimony.’ ” King’s Case, 352 Mass. 488, 490 (1967). McCarthy v. Contributory Retirement Appeal Bd., 342 Mass, at 46-48; Teachers’ Retirement Bd. v. Contributory Retirement Appeal Bd., 346 Mass. 663, 666 (1964); Wakefield Contributory Retirement Bd., v. Contributory Retirement Appeal Bd., 352 Mass. 499, 502-503 (1967). No expert medical testimony was presented here. Contrast Robinson v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. 634, 636-637, 639-640 (1985). As to the findings of the magistrate not based on credibility, but rather on documentary evidence, which Vinal somewhat cryptically holds are only entitled to “some deference,” 13 Mass. App. Ct. at 99, 101, CRAB “was not bound by the [magistrate’s] recommendation . . . [not] based on credibility. . . . The record before [CRAB] was the [same] record before the [magistrate] and [CRAB] was ... in the same position as the latter to decide the application, including whether the burden of proof had been met.” Blanchette v. Contributory Retirement Appeal Bd., 20 Mass. App. Ct. at 483 n.2.
In light of the discussion in this section, as well as that in note 9, infra, the judge’s observation, that “Four doctors concluded that the stress of [the riot and prisoner confrontation] . . . caused plaintiff to suffer a heart attack,” did not accurately describe the competent evidence of causation in this record.
As explained above, the panel chairman’s “clarification” opinion and any panel conclusion stating the actual existence of a causal connection between disability and workplace incident were nullities.
The judge’s decision does not reflect the deference due CRAB’s expertise in retirement law matters. Such an oversight is especially questionable in this “difficult area of the law,” Blanchette v. Contributory Retirement Appeal Bd.., 20 Mass. App. Ct. at 487, which is “notoriously complex,” Namay v. Contributory Retirement Appeal Bd., 19 Mass. App. Ct. 456, 463 (1985), and in which courts accord “great weight” to CRAB’s interpretation and application of the statutory provisions it is charged with administering. Ibid.