OPINION

I. INTRODUCTION

The Anchorage Police and Fire Retirement Board denied Geoffrey Palmer occupational disability benefits for heart disease that worsened during his twelve-year tenure as a police officer for the Municipality of Anchorage. Palmer appealed, and the superior court affirmed. Because a board member’s motion to award Palmer benefits did not receive the requisite five votes, we hold that the board correctly rejected Palmer’s claim, even though four board members — a majority of the seven deciding his case— voted for the motion.

We also affirm the board’s decision on the merits. There was substantial evidence that Palmer’s disability was solely the result of his preexisting coronary artery disease, and was not aggravated by his 1990 work-related heart attack. There was also substantial evidence that work-related stress was not a substantial factor in bringing about or aggravating Palmer’s heart disease. We therefore affirm the superior court judgment upholding the board’s decision.

*836II. FACTS AND PROCEEDINGS

A. Palmer’s Heart Disease

Geoffrey Palmer suffered a minor heart attack during the course of his employment as an Anchorage police officer in April 1990. Palmer was diagnosed with coronary artery disease, or atherosclerosis. Dr. Mohammed Sarwar performed a double coronary artery bypass about a week after Palmer’s attack.

Palmer returned to unrestricted duty with the Anchorage Police Department. He experienced no symptoms of his disease for the next six years.

In October 1996 a fellow police officer was killed in the line of duty while executing an arrest warrant. The next day, Palmer experienced chest and arm pains and had trouble breathing shortly after discussing the officer’s death with his wife.1 Palmer went to the emergency room and was diagnosed with a possible heart attack. Three days later a cardiologist diagnosed unstable angina and a possible mild heart attack. Tests revealed seventy to one hundred percent occlusions of several arteries, including a seventy percent occlusion of one of the two vein grafts used in the 1990 bypass surgery.

Palmer developed recurring angina pector-is after release from the hospital. His treating physician, Dr. Thomas Kramer, performed two coronary balloon angioplasties and placed a stent in Palmer’s mid-left anterior descending coronary artery in November 1996. Palmer has not experienced significant chest pains since these procedures. Palmer testified at his 1998 occupational disability hearing that he had been on leave from the police force since his 1996 hospital admission.

B. The First Board Decision Denying Occupational Disability Benefits

Palmer filed an application for occupational disability benefits under Plan III of the Anchorage Police and Retirement System in April 1997. The Police and Fire Retirement Board (“PFRB” or the “board”) denied his claim, and scheduled a hearing to review this denial. The only issue at the hearing was whether Palmer’s disability was work-related.

Palmer presented two theories under which he argued that he was entitled to benefits. First, he introduced expert medical evidence that his 1990 heart attack and subsequent bypass operation were precipitated by work-related activities. He also introduced evidence that one of the vein grafts had become severely occluded by 1996, and that this occlusion occurred for reasons unrelated to the progression of his underlying coronary artery disease. He concluded that he was entitled to occupational disability benefits because this occlusion aggravated or contributed to his disability. Alternatively, Palmer argued that he suffered a great deal of stress on the job that aggravated or contributed to his underlying disease.

Only three of the seven participating members of the board voted for a motion to award Palmer occupational disability benefits. Accordingly, the board treated the motion as denied under former Anchorage Municipal Code Regulation (AMCR) 3.85.010(B).2 The board relied extensively on the testimony of Dr. William Breall, the independent medical examiner hired by the board for this case, in concluding that Palmer’s disability was not work-related under either of his two theories.

On appeal by Palmer, the superior court held that the board “erred in adopting Dr. Breall’s opinions as ‘substantial’ evidence and erred in failing to resolve inconclusive and doubtful medical testimony in favor of Palmer.” The superior court remanded the case to the board for “reconsideration of the record, applying the proper legal standard,” and ordered the board to “reconsider its reliance on Dr. Breall’s opinions.”

*837C. The Second Board Decision Denying Occupational Disability Benefits

Seven board members, including two who did not participate in the original decision, reconsidered the record and voted four to three to award Palmer occupational disability benefits. But because the motion to award benefits did not receive five votes, the motion failed under AMCR 3.85.075(B).3

Accordingly, the board issued a second ruling explaining its denial of Palmer’s claim. The board relied on the testimony and reports of Dr. Breall and Dr. Werner Samson as well as “other objective medical evidence” in the record.4

Palmer appealed, and the superior court affirmed the board’s second decision. The court found that the board properly reconsidered its reliance on Dr. Breall’s testimony in accordance with the first superior court order, and that the board relied on the testimony of Dr. Samson and objective medical evidence as well as Dr. Breall’s testimony in reaching its decision on remand. The court concluded that “substantial evidence in the record supports the Board’s denial of occupational disability benefits.”

Palmer appeals the superior court’s ruling, contesting the board’s procedures as well as its factual and legal conclusions.

III. STANDARD OF REVIEW

This case first requires us to consider the constitutionality of an administrative regulation. We conduct this review using our independent judgment,5 adopting “the rule of law that is most persuasive in light of precedent, reason, and policy.”6 We review an agency’s interpretation of its own regulations using our independent judgment, so long as that interpretation does not implicate the agency’s area of expertise or questions of fundamental policy committed to the agency’s discretion.7

Next, we must consider a merit appeal of an agency adjudication. “In considering an administrative appeal from a decision issued by the superior court [sitting] as an intermediate court of appeal, we review the agency’s action directly.”8 We review questions of law not involving agency expertise using our independent judgment.9 We review the agency’s factual determinations under the substantial evidence standard.10 Whether the quantum of evidence is sufficient to constitute “substantial evidence” supporting the agency’s conclusion to deny benefits is a legal question to which we apply our independent judgment.11

*838IV. DISCUSSION

A. The Board’s Voting Procedures Were Neither Invalid Nor Unconstitutional.

Palmer argues that several flaws in the PFRB proceedings conflicted with either his constitutional rights or municipal laws governing the PFRB’s adjudication of his claim. Only two of these challenges deserve extended comment.

1. AMCR 3.85.075(B)’s five-vote requirement did not violate Palmer’s right to a fair adjudication of his claim.

First, Palmer argues that AMCR 3.85.075(B)’s five-vote requirement infringed on his due process right to a fair adjudication of his claim because it enabled his claim to be denied by a minority of the participating board members.12 We reject this contention because the regulation serves the legitimate purpose of ensuring that only a majority of the full board and a coalition of the former board’s management-side and labor-side appointees can decide to change the status quo.

Former chapter 3.85 of the Anchorage Municipal Code13 established an eight-member Police and Fire Retirement Board to preside over disability claims brought by members of the Police and Fire Retirement Plan,14 and authorized the board to promulgate regulations consistent with its mandate.15 Anchorage Municipal Code Regulation 3.85.075(B) provides that “[i]n order for any motion to be passed by the board, it must be supported by not less than five votes of board members.” As a matter of practice, the board makes all of its benefits decisions by motion. Following a hearing, a member makes a motion to award benefits. Benefits are awarded if five or more members vote in support; otherwise, benefits are denied. Anchorage Municipal Code Regulation 3.85.075(A) provides that five members of the board constitute a quorum, and that absent a quorum the board “shall not transact any business other than adjournment of [a] meeting.”

Palmer argues that because of the five-vote requirement, the burden of persuasion fluctuates arbitrarily depending on the number of board members who actually participate in a given case. If eight participate, then the claimant simply has to persuade a majority that he is entitled to occupational disability benefits. But if only five participate, the claimant must persuade every participating board member. In Palmer’s case, four of the seven participating members voted to grant him occupational disability benefits, but his claim was nonetheless denied because of the five-vote requirement. Palmer insists this procedural rule violated his due process right to a fair adjudication of his claim, and asks us to hold AMCR 3.85.075(B) unconstitutional. Alternatively, he impliedly asks that we allow him to re-argue his ease before a full board.

The board responds that because Palmer received notice and an opportunity to be heard, he cannot claim the benefit of any further procedural protections.16 But this *839court has never held that such minimal protections are categorically sufficient to insulate an administrative adjudication against due process challenges. Rather, we conduct a full review to ensure that the agency’s adjudicative procedures are “consistent with a fair trial.”17

To determine whether specific procedures should be added or substituted, we apply the three-part analytical framework announced by the United States Supreme Court in Mathews v. Elckidge,18 First, we look to the private interest in question; second, we assess the risk of an erroneous deprivation of that interest posed by current procedures and the probable value of additional or substitute procedural safeguards; and finally, we examine the government’s interest, including the function involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail.19

There is no dispute that occupational disability payments constitute a substantial benefit. An officer with Palmer’s qualifications would have received fifty percent of his final average compensation if entitled to occupational disability benefits, as opposed to thirty percent if entitled only to non-oecupational disability benefits.20 Accordingly, the private interest at stake is significant.

Palmer is arguably correct that his chances of prevailing on his claim are lowered by each missing board member, since absent board members are effectively counted as nay votes when the board votes on a motion to award benefits. One of the chief purposes of the due process clause is to “ensure that individuals who have property rights are not subjected to arbitrary governmental deprivation of those rights.”21 The board’s procedures expose occupational disability claimants to the risk of arbitrary deprivation.

Further, the solutions to this problem would be simple and effective. Claimants can be given the right to require the full board to hear their eases. Alternatively, when members cannot participate in a particular case due to conflicts of interest or other compelling reasons, the five-vote requirement could be suspended, and motions to award benefits could be passed by a simple majority of participating board members.22

But we must also consider the government interest affected by the proposed corrective procedures. At all times relevant to this case, the board was comprised of four members appointed by the mayor and four members chosen by the retirement plan participants.23 This corresponds to an even split between management-side and labor-side board members. The five-vote minimum ensured that any board member could not change the status quo without persuading at least one member whose professional background differed.24 By requiring a cross-section of the board to approve any change, the five-vote minimum may have had the effect of encouraging a spirit of cooperation and compromise. Allowing fewer than five mem*840bers to award benefits might frustrate this legitimate goal.25

Likewise, a rule allowing a claimant to postpone a hearing if less than a full board was present would constitute substantial judicial interference with the board’s ability to conduct its business. While this approach would leave AMCR 3.85.075(B) intact, it would effectively substitute an eight-member quorum rule for AMCR 3.85.075(B)’s five-member quorum rule. The board has a significant interest in being able to set its own calendar and hear claims in an orderly manner regardless of whether all of its members participate in a given claim.26

At least one court has rejected a due process challenge in a very similar procedural context. In Stcmson v. San Diego Coast Beg’l Comm’n, a property owner was denied a coastal development permit despite the fact that a majority of the participating regional commissioners voted to grant the permit.27 Only eight of eleven regional commissioners attended his hearing; because only five of these eight voted to grant a permit, the developer did not obtain the support of a majority of the total appointed membership of the regional commission as required by the relevant statute.28 The court noted that the developer arguably bore a “more onerous burden” because he did not have a chance to persuade the missing regional commissioners.29 But the court concluded that this did not violate due process, because the process ensured that “the issuance of a permit will be based on the judgment of a broader cross-section of the commission” than would be the case if permits could be granted by less than a majority of the commission, and the government had a “substantial interest in protecting the integrity of the decision-making process.”30 The same reasoning applies here. Accordingly, Palmer’s challenge to AMCR 3.85.075(B) fails.

Furthermore, full-body voting provisions and supermajority requirements are both common and useful administrative tools. The common law rule that a majority vote requires a majority of those voting, in the presence of a quorum, excluding blanks and abstentions, is merely a default presumption.31 It can be modified and often is.32 Full-body provisions and supermajority requirements help ensure a broad consensus is reached for important governmental action. For this reason, many states recognize full-body voting provisions for municipal entities in abrogation of the common law presumption.33 Indeed, the Alaska Constitution includes provisions requiring enhanced majori*841ties among defined voting bodies.34 In this case, the legislature clearly delegated responsibility for designing rules of order and procedure to the board, and the board permissibly adopted voting procedures designed to guard against the expenditure of system funds on the vote of a minority of the full membership. Similar rules are common among governmental entities, and provide a useful procedural safeguard.

The dissent, having analogized the PFRB to a jury, would hold that the PFRB’s former procedure violated due process. But the PFRB is not analogous to a jury. Unlike a jury, its constituent representative makeup was specified. And as an administrative agency, it differs fundamentally from a jury: it has repeat business and collective expertise; its members bring individual expertise and different professional perspectives that would probably preclude them from sitting as jurors if a jury were somehow trying Palmer’s claim; and it even has some policy-setting capability entitling it to deference when it uses its expertise to interpret its enabling provisions.

The dissent asserts that the “full-body” cases the court cites above are distinguishable because it claims they do not involve tribunals adjudicating rights. But the dissent would hold that the municipality’s procedure violates due process even though the dissent cites no case holding that applying a full-body voting requirement to a tribunal like the PFRB violates procedural due process.

In short, we are unconvinced that requiring Palmer to persuade five of the seven sitting board members denied him due process.

2. The board’s practice of deciding cases by motions framed to require the claimant to obtain votes of a majority of the board is not inconsistent with the presumption of compensability.

Palmer also argues that the board’s practice of requiring a majority of the board to vote in favor of awarding benefits is inconsistent with the rebuttable presumption of com-pensability applicable to his occupational disability claim. Palmer argues that benefits should be awarded unless five members support a motion to deny them. Because only three members of the board voted to deny benefits in the board’s second vote on his claim, Palmer argues that he should prevail.

The board argues that framing the motion for benefits in a way that requires the claimant to establish his claim is not inconsistent with the presumption of compensability. The board reasons that regardless of the legislature’s decision “to lighten [Palmer’s] burden [of production] via a rebuttable presumption,” Palmer is not entitled to occupational disability benefits until a five-member majority of the board decides that he is.

The board is correct. Palmer overstates the effect of the presumption of com-pensability. The presumption simply shifts the burden of production from the claimant to the employer with respect to certain types of favored claims. The presumption bears no relationship to any aspect of the burden of persuasion — in this case, how many board members Palmer must persuade in order to prevail. The board’s practice of framing motions to award benefits rather than to deny them does not contradict former AMC 3.85.230(C)(2)’s rebuttable presumption.

3. Palmer’s remaining procedural challenges are meritless.

We also reject Palmer’s remaining contentions of procedural error. First, the participation by two new board members on remand did not violate Palmer’s due process rights or the board’s regulations, despite Palmer’s strained argument to the contrary. Palmer overstates the effect of AMCR 3.85.075(C), which provides that board members “who miss part of a hearing shall not participate in the adjudication” unless the parties agree to the members’ participation *842and the members review “the recorded testimony taken in their absence.” Neither this provision nor any other suggests that only those board members who originally decided Palmer’s claim may participate on remand. The board correctly reasons that such a requirement would be impractical given that the board’s composition changes over time. The two board members who participated for the first time in the reconsideration of Palmer’s claim on remand attested that they “had reviewed the record and would be able to make an informed decision on [Palmer’s] claim.” Neither the regulations nor the due process provisions of the state and federal constitutions require anything further.

Finally, Palmer alleges that the board’s option to frame motions as either motions to award benefits or motions to deny them constitutes an equal protection violation, because the board can vary the burden of similarly situated claimants depending on how the motion is framed.35 But Palmer does not allege that the board has ever made a motion to deny as opposed to award benefits. There is no ground on which to begin an equal protection analysis in the absence of some allegation of unequal treatment.

B. The Board Correctly Refused To Give Preclusive Effect to the Alaska Workers’ Compensation Board’s Ruling that Palmer’s Disability Was Work-Related.

The PFRB originally determined that Palmer’s injury was not work-related. While the first PFRB decision was pending on appeal to the superior court, the Alaska Workers’ Compensation Board (AWCB) concluded that Palmer’s disability was work-related under the Alaska Workers’ Compensation Act. The superior court reversed and remanded the PFRB’s first decision, holding that the board’s reliance on Dr. Breall’s opinion was misplaced. The board held on remand that the AWCB’s opinion was not entitled to preclusive effect,36 and again determined that Palmer’s injury was not work-related. In affirming the board’s second decision, the superior court held that the AWCB’s contrary determination had no pre-clusive effect because the PFRB’s initial decision remained the first final judgment despite the intervening reversal and remand by the first superior court order.

The board reiterates this reasoning on appeal, and further contends that collateral es-toppel is inapplicable because the real parties in interest in the two proceedings were neither identical nor in privity with each other. Because we agree with the latter argument, we need not address the superior court’s reasoning regarding the final judgment issue.37 Following the reasoning in Holmberg v. State, Division of Risk Management,38 we conclude that the defendants before the AWCB and the PFRB were not in privity.

In Holmberg, we held that the Public Employees Retirement Board’s (PERB’s) factual determination that a former employee was physically unable to perform her duties did not preclude relitigation of this issue before the AWCB because the real parties in interest were not in privity.39 The real party before the AWCB was the state as an employer, while the real party before the PERB was the state as administrator of the Public Employees Retirement System (PERS).40 The state represents the interests of the PERS in proceedings before the PERB, and as the court explained, PERS’s interests are *843not identical to the state’s interests.41 Specifically, PERS awards are paid from the trust fund to which all PERS members and employer-participants contribute, whereas any amount awarded by the AWCB would come directly from the state treasury42 The state is only nominally represented in the PERB proceeding to the extent it has an interest in the trust fund.43 Accordingly, we held that the real parties in the separate proceedings were not sufficiently related to assure that the state’s interests were adequately protected in the PERB proceeding.44

Holmberg’s reasoning applies here. First, in proceedings before the board, the staff of the PFRB or its appointed attorney represents the interests of the retirement system, not the interests of the municipality as employer.45 Likewise, occupational disability benefits are paid from the retirement system’s trust fund, not from the municipality’s general coffers.46 Finally, the interests of all members of the Police and Fire Retirement System other than the municipality were not represented at all in the AWCB proceedings.47

Therefore, the real parties in interest before the AWCB and the PFRB were not in privity, and the AWCB determination that Palmer’s disability was work-related did not preclude PFRB’s reconsideration of that issue.

C. Substantial Evidence Supports the Board’s Conclusions that the Presumption of Compensability Was Successfully Rebutted and that Palmer Failed To Prove by a Preponderance of the Evidence that His Injury Was Work-Related.

As a member of Plan III of the Police and Fire Retirement System, Palmer enjoys a rebuttable presumption that his heart condition is work-related.48 The PFRB recognized that the Alaska Workers’ Compensation Act (AWCA) contained a similar re-buttable presumption,49 and formally adopted this court’s interpretation of the AWCA’s rebuttable presumption in Resolution 85-1.

Accordingly, a three-step analysis governs our review of the board’s denial of Palmer’s occupational disability benefits claim. The first step requires Palmer to offer “some evidence” that his disability claim arose out of his employment.50 The board found that Palmer met this test with respect to both of his theories that his disability was work-related, and the board does not challenge these findings on appeal.51

The second step requires us to examine whether the board correctly determined that substantial evidence was presented rebutting the presumption of compensability with respect to each of Palmer’s theories.52 Because the presump*844tion shifts only the burden of production and not the burden of persuasion, we review any evidence tending to rebut the presumption in isolation, without reweighing the rebuttal evidence against the evidence tending to establish causation.53

Finally, the third step requires us to review whether substantial evidence supports the board’s decision that Palmer failed to prove his claim by a preponderance of the evidence under either of his proposed theories.54 As with the second step, we do not reweigh the evidence or choose between competing inferences.55

Palmer contends that the board’s findings regarding the second and third steps were erroneous, and that the board failed to apply the correct “substantial factor” test in both steps. But all of the board’s findings were supported by substantial testimonial and objective medical evidence, and the board explicitly applied the proper legal test in reaching its conclusions. Therefore, Palmer’s challenge on the merits fails.

1. The board properly relied on Dr. Breall’s testimony in conjunction with other objective evidence and expert testimony.

The board relied extensively on Dr. Breall’s testimony to determine that the presumption of compensability, was successfully rebutted and that Palmer failed to prove his disability was substantially work-related. Palmer argues that the superior court erred by accepting Dr. Breall’s opinions as substantial evidence supporting the board’s conclusions.

a. Dr. Breall’s testimony supported a conclusion the disability was not work-related.

Dr. Breall was hired by the board to conduct an independent review of Palmer’s medical history, offer his diagnosis, and address whether he thought “any industrial factors” might have been a “substantial factor” in bringing about Palmer’s disability. Dr. Breall interviewed Palmer over the telephone, reviewed his medical records, and issued a report.

Dr. Breall concluded that Palmer’s disability is solely the product of his underlying “severe triple vessel atherosclerotic coronary artery occlusive disease.” Dr. Breall found that this disease was the result of “a number of well-recognized, but non-industrial, risk factors.”56 Dr. Breall concluded that “[a]ny disability which [Palmer] currently has is due to the progressive non-industrial atheroscler-otic development within the coronary arteries and within his bypass grafts.”

Dr. Breall found that Palmer’s 1990 work-related heart attack “resulted in negligible damage to the left ventricle,” but that the occlusions following the 1990 bypass surgery had nothing to do with this attack. He explained that an angiogram performed just after the 1990 attack revealed that Palmer’s left ventricle ejection fraction was within the normal range. That is, despite the minor damage caused to the left ventricle by the heart attack, Palmer “had no disability whatsoever ] with respect to the heart muscle itself. Therefore, any and all disability that he had, had to be because of impairing the blood flow through the coronary arteries.” Accordingly, Dr. Breall concluded that the 1990 heart attack “played no role in Mr. Palmer’s current disability.”

Dr. Breall also concluded that “[i]n all likelihood,” Palmer did not suffer a second heart attack in 1996. He explained that Palmer’s 1996 episode was not a second heart attack because neither of the two indicia of a myocardial infarction — changes in the electrocardiogram or elevated levels of certain enzymes — was present. Dr. Breall concluded that the 1996 event was unstable angina or preinfarction angina. He further noted that *845treadmill tests taken after Palmer’s hospitalizations in 1990 and 1996 revealed significant improvement in aerobic function since the 1990 attack. Dr. Breall inferred from this data that neither the 1990 heart attack nor the 1996 episode caused any lasting disability.

Dr. Breall also found that job-related stress did not contribute to Palmer’s disability. Dr. Breall testified that “Type A” behavior is the only emotional stress factor that has been proven to cause, aggravate, or accelerate coronary artery disease.57 Dr. Breall testified that while Palmer has “a severe degree of [T]ype A behavior[,] .... this has absolutely nothing at all to do with his job.”58 Accordingly, Dr. Breall concluded that Palmer’s Type A, or coronary-prone, behavior was just another nonindustrial risk factor that may have contributed to the development of Palmer’s disease and resulting disability.59

Dr. Breall was asked on cross-examination about his failure to perform a physical exam. Dr. Breall testified that his telephone interview and review of Palmer’s extensive medical records were sufficient for purposes of evaluating whether Palmer’s injury was work-related. Dr. Breall also acknowledged that he had previously published an opinion that doctors conducting evaluations for litigation purposes should not rely on examinations, medical histories, or diagnostic testing obtained or performed by non-medical or technical medical personnel.

b. The first superior court order did not require the board to ignore Dr. Breall’s testimony.

Palmer argues that the superior court unequivocally ordered the board to ignore Dr. Bread’s report and testimony on remand, but the court’s opinion does not support Palmer’s interpretation.

The court first noted Palmer’s concerns that unlike Dr. Kramer, Dr. Breall never physically examined him and did not have an ongoing professional relationship with him. The court further noted that Dr. Bread’s testimony was potentially inconsistent with his prior published opinions. The court found “that the Board’s reliance on Dr. Bread’s opinions with respect to the relationship between the [effects] of Palmer’s coronary artery bypass of 1990 and his present disability does not constitute ‘substantial evidence,’ as required by law upon which the Board can properly rely in reaching its decision.” The import of this finding is best deciphered in light of the court’s subsequent statement that “the Board must reconsider its reliance on Dr. Bread’s opinions applying the proper legal ‘substantial factor’ test.” The latter statement resolves any doubts created by the former: the court would not have ordered reconsideration of testimony it intended the board to ignore.

The board adopted the correct interpretation of the first superior court order; namely, that it should reconsider Dr. Bread’s testimony in Hght of Palmer’s concerns, and that the doctor’s testimony and medical report was insufficient by itself to clear the “substantial evidence” hurdle.

*846c. Substantial evidence supports the board’s conclusion that Dr. Breall’s testimony was conclusive and consistent with objective medical data and other expert testimony.

Palmer next argues that Dr. Breall’s testimony cannot support the board’s conclusion because it conflicted with his prior published opinions and the testimony of Dr. Kramer, Palmer’s treating physician. The board acknowledged our rule that doubtful or inconclusive evidence must be resolved in favor of the claimant.60 However, this rule is not triggered simply by a lack of unanimity among experts, which clearly exists in this case, but only when “the substance of a particular witness’ testimony is in doubt.”61 The board correctly explained that Dr. Breall’s testimony was conclusive and consistent with both his earlier published opinions and other evidence presented to the board in this case.

The board explained that Dr. Breall’s opinions expressed in the professional publications brought to the board’s attention by Palmer were not inconsistent with his opinions in or preparation for this case. Specifically, the board found that Dr. Breall’s opinion that non-medical personnel should not perform medical examinations for litigation purposes was completely consistent with the fact that in this case all of the records Dr. Breall relied on were produced by other doctors, and mostly by other cardiologists. Likewise, the board found that Dr. Breall’s conclusion that Palmer’s work-related stress did not aggravate his coronary artery disease was consistent with his previously published opinions.

Palmer argues that even if Dr. Breall’s testimony is conclusive and internally consistent, the board was obliged to ignore it by our decision in Black v. Universal Services, Inc.62 We held in Black that the medical testimony of a doctor whose examination consisted solely of a twenty-minute interview with the claimant and a brief physical examination did not constitute substantial evidence, especially in light of the fact that the doctor’s conclusions were “contrary to those of the numerous physicians who treated her.”63 Palmer argues that we should reach the same conclusion regarding Dr. Breall’s testimony because he did not physically examine Palmer and his testimony was contrary to that of Dr. Kramer, Palmer’s treating physician since 1996.

Palmer’s reliance on Black is misplaced. We have limited our holding in that case by consistently refusing to reverse a board’s decision “where the reviewing physician’s statement did not stand alone and was consistent with other evidence presented.”64 We have held that a physician’s reliance on medical records rather than a physical exam is not fatal where the testimony is consistent with other medical expert testimony and objective test results.65 In this case, Dr. Breall’s testimony is a reasonable interpretation of Palmer’s medical records and objective medical data.66 Additionally, Dr. Breall’s opinion is shared in substantial part by Dr. Samson, who physically examined Palmer in 1997 for the parallel AWCB proceeding and whose report was admitted as evidence in the PFRB proceeding.67

*847Dr. Samson interviewed Palmer and conducted a physical examination, treadmill test, electrocardiogram, and chest X-ray. Dr. Samson concluded that Palmer’s 1990 work-related heart attack caused no lasting damage and that Palmer’s disability was not caused by this attack.68 Dr. Samson took “strong issue” with Dr. Kramer’s opinion that the 1990 coronary bypass was necessitated by the heart attack. Dr. Samson also found that the 1990 bypass had “absolutely nothing” to do with Palmer’s angina and 1996 angioplasty. Rather, Dr. Samson determined that Palmer’s angina pectoris was the result of his underlying coronary artery disease, which he thought was “undoubtedly” caused by the same nonindustrial risk factors identified by Dr. Breall. Like Dr. Breall, Dr. Samson found that Palmer did not suffer a second heart attack in 1996. Dr. Samson concluded that Palmer’s angina pectoris and angioplasty “in no way were related to his employment.”

Accordingly, the board did not violate the superior court’s order or otherwise err by relying in part on Dr. Breall’s testimony.

2. The board applied the correct “substantial factor” test in both the second and third steps of its analysis.

Palmer contends that the board failed to properly apply Tolbert v. Alascom, Inc.’s “substantial factor” test in the second and third steps of its analysis — i.e., its determinations that substantial evidence rebutted the applicable presumption of compensability and that Palmer failed to prove that his disease was work-related by a preponderance of the evidence.69 But the board explicitly stated that it would award benefits “where the work-related injury is a substantial factor in the employee’s disability regardless of whether a non-work-related injury could independently have caused [the] disability.” This is precisely the standard we adopted in Tolbert to govern cases in which two or more causal factors may operate in concert to produce the disability.70 As discussed below, substantial evidence supports the board’s decision in both the second and third steps of its analysis.

3. Substantial evidence supports the board’s findings regarding the second and third steps of the analysis.

The “substantial factor” test announced in Tolbert is consistent with our recognition in Grainger v. Alaska Workers’ Compensation Board that often “no single factor can be isolated as the ‘cause’ of ... arteriosclerosis.”71 Grainger also recognized the corollary that “several risk factors often operate together to precipitate or accelerate the development of [arteriosclerosis].”72 Thus, whether nonindustrial risk factors associated with Palmer’s lifestyle and physiological characteristics were the original cause of his coronary artery disease is not determinative. He may nonetheless qualify for occupational disability benefits if either his 1990 work-related heart attack or work-related stress substantially combined with, aggravated, or accelerated his disease.73

Given the evidence discussed in Part IV. C.I., the board’s findings that the presumption of compensability was rebutted with respect to each of Palmer’s theories were not *848erroneous. As to each of Palmer’s theories, the board was presented with an alternative explanation of Palmer’s disability that excluded work-related factors as a substantial cause.74

Palmer’s first theory is that his occluded vein graft, which was necessitated by his 1990 work-related heart attack, aggravated or contributed to his disability. But Dr. Breall and Dr. Samson both concluded that Palmer’s disability was caused solely by his underlying coronary artery disease, and that Palmer’s work-related heart attack in 1990 did not aggravate or contribute to that disease. Further, Dr. Samson took “strong issue” with Dr. Kramer’s opinion that the 1990 bypass operation was necessitated by Palmer’s heart attack. The board could reasonably infer that Palmer’s bypass operation was necessitated by his preexisting, nonindustrial coronary artery disease, not his 1990 work-related heart attack. Following this view, whether occlusion of one of the 1990 vein grafts contributed to Palmer’s disability is irrelevant — the vein grafts were required to address the underlying disease, not the negligible damage caused by the 1990 work-related heart attack. Finally, both doctors agreed that the 1996 episode did not constitute a second heart attack or contribute to Palmer’s disability. Thus, whether the 1996 episode should be considered work-related is likewise irrelevant.75

Palmer’s second theory is that work-related stress contributed to his disability. But Dr. Breall testified that Palmer’s work did not aggravate his predisposition toward coronary-prone or Type A behavior.76 Again, the fact that Dr. Pécora was of a different view is unimportant. The board could permissibly choose to credit the testimony of a well-credentialed expert in the relationship between emotional stress and coronary artery disease over that of Dr. Pécora, a friend of Palmer’s parents. Conversely, the board was not required to credit Palmer’s testimony that his stress was largely work-induced.

Despite Palmer’s arguments to the contrary, our opinion in Grainger does not render Dr. Bread’s testimony insufficient to rebut Palmer’s work-related stress theory. We held in Grainger that the AWCB improperly relied on the inconclusive testimony of the claimant’s treating physicians in determining that the employer had rebutted the presumption that job-related stress was a factor in causing Grainger’s disability.77 As explained above in Part IV.C.l, Dr. Breall’s testimony is not inconclusive. Unlike the physicians in Grainger,78 Dr. Breall explicitly testified that Palmer’s work-related stress did not aggravate his underlying coronary artery disease. Accordingly, we hold that substantial evidence supports the board’s decision that the presumption of compensability was rebutted.

Based on all of this evidence, we further hold that substantial evidence supports the board’s decision that Palmer failed to prove by a preponderance of the evidence that his disability was work related.79

*849V. CONCLUSION

For these reasons, we AFFIRM the decision of the superior court upholding the board’s second decision denying Palmer’s claim for occupational disability benefits.

BRYNER, Jftstice, with whom FABE, Chief Justice, joins, dissenting.

with whom FABE, Chief Justice, joins, dissenting.

I disagree with the opinion’s due process analysis. In my view the opinion misapplies Mathews v. Eldridge’s balancing test1 by treating it as if it were a rational basis analysis. Without any meaningful attempt at balancing the competing private and governmental interests, and despite conceding the importance of Palmer’s retirement benefits, the opinion summarily concludes that his right to a rational and non-arbitrary ruling on his entitlement to these benefits must yield to the city’s conjectural interests in the PFRB’s five-minimum-vote requirement — illusory interests that the city itself has not even bothered to argue and that are incapable of withstanding even rational basis scrutiny.

The opinion identifies two supposedly legitimate governmental interests that a “full-body” voting requirement might further: (1) to “help ensure a broad consensus is reached for important governmental action;” 2 and (2) to encourage “a spirit of cooperation and compromise” among participating PFRB members.3 Upon examination, however, neither interest proves legitimate when a full-body voting requirement applies to a panel like the PFRB.

A full-body voting provision can be eminently sensible when applied to political or administrative bodies charged with making, planning, or implementing public policy on a community-wide basis for the good of the public as a whole. As the opinion correctly observes, when used by public bodies that make policy-level decisions of this kind — bodies like legislatures or assemblies, zoning boards, and various planning or regulatory commissions — a voting provision requiring broad-based consensus can serve a legitimate governmental interest by ensuring that the body’s actions do what they are supposed to do: promote the general public interest by reflecting community-wide perceptions of sound policy.

Yet no comparable governmental interest is readily apparent when a public board performs judicial, rather than political, junctions: when its actions are guided not by broad notions of public policy meant to advance community interests as a whole, but the traditional principles of case-specific adjudication. When a panel’s core duties require it to resolve individual disputes concerning actionable rights and duties through a formal process that entails an evidentiary hearing, factfinding, and a binding decision based on applicable law, the government has no legitimate interest in seeking broad-based political consensus.4

Reflecting this disparity of interests, every example of full-body voting discussed in the court’s opinion involves a public body engaged in political decision-making rather than case-specific adjudication. The court cites no authority supporting the proposition that full-body voting has any legitimate place in administrative adjudication of legal claims.5

*850Here, the PFRB’s decision indisputably involved an adjudicative process, not a political process. Palmer is not asking for discretionary public benefits. He is asserting a straightforward contract claim. As the court’s opinion acknowledges, Palmer had a significant financial stake in receiving the benefits guaranteed by his retirement plan. His interest in those benefits arose under the terms of his municipal employment contract; it accrued over his many years of service as a city police officer; and, by the time he applied for retirement, it had ripened into a fully vested and enforceable property right. In asserting administrative jurisdiction over Palmer’s legal claim of right, the PFRB was obliged to resolve the claim by finding the facts of Palmer’s case through a fair and impartial evidentiary hearing and by determining the legal significance of those facts under his retirement plan and applicable law. The board’s duties were thus case-specific and guided by legal principles that leave no legitimate room for individual board members to assert partisan interests or make general policy decisions for the broader public good.

Given the narrow adjudicative role played by the PFRB, the city’s ostensible interest in using full-body voting to ensure broad-based consensus rings false: in this distinctly nonpolitical context, counting two absent board members as voting against a claim is functionally indistinguishable from — and no more defensible than — excusing two jurors from attendance at trial and counting their absences as votes for the defendant. The analogy between the PFRB and a jury is of course somewhat inexact'but is nonetheless apt. Its accuracy can be confirmed by comparing the PFRB’s voting rule to those used by other quasi-judicial tribunals in Alaska whose composition and duties are similar to the PFRB’s. For example, Alaska’s Teachers’ Retirement Board, Public Employees’ Retirement Board, and Workers’ Compensation Board all have memberships with a specified makeup; all sit as multi-member tribunals, engage in repeated adjudication, and develop collective expertise; and all have individual members who bring different expertise and professional perspectives to the adjudicative process.6 Yet ttíeir voting rules reflect no legitimate need for full-body voting: each of these boards decides each case it considers by majority vote based upon those members present and voting.7

The second purportedly legitimate government interest in full-body voting — encouraging a “spirit of cooperation and compromise” on the PFRB — is simply a variation on the first and fares no better. A “spirit of cooperation and compromise” may be a laudable and important goal when dealing with a governmental body like a coastal regulatory commission, which addresses broad issues of policy and renders decisions based on its individual members’ subjective perceptions of public interests. But pressing for compromise becomes far more questionable when the pressure is applied to an adjudicative body whose primary obligation is similar to a jury’s — to decide individual cases fairly and impartially by hearing evidence, finding facts, and applying settled legal rules to their findings. A policy encouraging PFRB members to “cooperate and compromise” their individually held views seems no more acceptable, and no worthier of judicial deference, than would be a comparable policy encouraging juries to compromise in judicial proceedings.

As applied in this case, then, the five-minimum-vote requirement is fundamentally arbitrary and serves no legitimate purpose. Moreover, as the court admits, the rule could be easily be cured; in fact it has already been discarded by the city. Given the countervailing importance of Palmer’s right to *851retirement benefits,8 I would hold that Mathews v. Eldridge’s balancing test compels the conclusion that the PFRB’s voting rule violated Palmer’s right to due process.

I therefore dissent from the court’s opinion affirming the superior court’s judgment.