dissenting:
I am unable to agree with the court’s opinion affirming the hearing examiner’s decision that McKinley’s disability did not “aris[e] out of and in the course of employment.” D.C.Code § 36-301(12) (1997 Repl.). The hearing examiner discerned, from McKinley’s testimony two, alternative theories of recovery: (1) work stress attributable to problems with his supervisor caused an attack of ventricular tachycardia, which led to a totally disabling psychological disorder; and (2) the work stress directly caused his psychological disorder. McKinley lost on both theories but has appealed only as to the first.
On appeal, McKinley challenges the examiner’s finding that his attack of ventricular tachycardia was not sufficiently work-related to qualify as an injury under the Workers’ Compensation Act. I conclude that, although properly granting McKinley the benefit of the statutory presumption of compensability under D.C.Code § 36-321(1) (1997 Repl.), the examiner erred in deciding that the employer had satisfied its burden of producing substantial evidence to show that McKinley’s disability was not work-related. Because the examiner’s decision does not comport with the law, I respectfully dissent.
I.
McKinley’s principal theory, proffered to demonstrate that his psychological disorder was causally related to his employment, is expressed in three steps. McKinley argued, first, that he experienced considerable stress resulting from confrontations with his supervisor, Robert Currenee, and from the fact that McKinley was required to return to work with Currenee pending resolution of McKinley’s formal complaint against him. Next, according to McKinley, that stress (coupled with anxiety over the impending birth of a child) aggravated his preexisting, asymptomatic heart condition — hypertrophic cardiomyopathy — and thereby caused the attack of ventricular tachycardia that he experienced while driving to work on January 10, 1991. Finally, said McKinley, this traumatic, near-death experience on January 10 — a heart-related attack caused primarily by work stress — in turn caused a psychological condition, post-traumatic stress disorder, resulting in temporary total disability. In short, he says, work stress caused a heart-*1387related attack which in turn caused a temporarily totally disabling psychological condition that entitled him to worker’s compensation benefits.
The only issue the examiner decided in rejecting this theory of recovery was whether McKinley’s attack of ventricular tachycardia “ar[ose] out of and in the course of employment.” D.C.Code § 36-301(12). The examiner agreed that McKinley presented evidence of work-relatedness sufficient to trigger the presumption of compensability found in D.C.Code § 36-321(1). But, according to the examiner, the employer rebutted this presumption by demonstrating that the heart-related attack occurred off the employer’s premises. The examiner, therefore, ultimately rejected McKinley’s argument, concluding that: “[McKinley’s] ventricular tachycardia did not arise out of and in the course of his employment. As the tachycardia was not work-related, any psychological condition attributable thereto was not work-related.” McKinley’s theory therefore failed, as the examiner saw it, before completing its second step.1
The hearing examiner concluded that McKinley’s testimony had suggested a second, alternative theory: that his disabling psychological condition was directly attributable to his employment, without any intervening causal step such as his attack of ventricular tachycardia. The examiner also rejected this argument, noting that all three psychiatrists called by the parties had “opined that the cause of claimant’s [psychological] disorder was the attack of ventricular tachycardia on January 10, 1991,” not McKinley’s “work or work environment.” The examiner noted, moreover, that while “claimant fixed the cause of his [psychological] problem to Mr. Currence and the employer, claimant lacked the requisite medical expertise to make this determination.” This second theory is not at issue on appeal.
II.
I am unable to agree with the hearing examiner’s application of the law governing compensation claims. The workers’ compensation statute establishes a presumption that a claim falls within its coverage “in the absence of evidence to the contrary.” D.C.Code § 36-321(1). We repeatedly have held that this presumption applies to the issue of whether an employee’s disability is sufficiently work-related to fall within the definition of “injury” provided by D.C.Code § 36-301(12). See, e.g., Ferreira v. District of Columbia Dep’t of Employment Servs., 531 A.2d 651, 655 (D.C.1987). This presumption effectively implements the statute’s “humanitarian purposes” and “reflects a ‘strong legislative policy favoring awards in arguable cases.’ ” Id. (quoting Wheatley v. Adler, 132 U.S.App. D.C. 177, 183, 407 F.2d 307, 313 (1968) (en banc)).
To gain the benefit of the presumption, the employee must make an “ ‘initial demonstration’ of the employment-connection of the disability” by showing “two ‘basic facts’: [1] a death or disability and [2] a work-related event, activity, or requirement which has the potential of resulting in or contributing to the death or disability.” Id. The employer may rebut the presumption by “bringing] forth substantial evidence showing that the death or disability did not arise out of and in the course of employment.” Id. (internal quotation marks omitted). Thus, we have said that for the employer to satisfy this burden, it must introduce “ ‘circumstantial evidence specific and comprehensive enough to sever the potential connection between a particular injury and a job-related event.’ ” Id. (quoting Swinton v. J. Frank Kelly, Inc., 180 U.S.App. D.C. 216, 224, 554 F.2d 1075, 1083, cert. denied 429 U.S. 820, 97 S.Ct. 67, 50 L.Ed.2d 81 (1976)).
The examiner correctly found that McKinley offered sufficient evidence to trigger this *1388statutory presumption. The examiner credited McKinley’s testimony that he had “experienced lightheadedness, heart palpitations and dizziness” while driving to work on January 10, 1991. Dr. Cowan, McKinley’s treating cardiologist, opined that these symptoms were consistent with an attack of ventricular tachycardia, and that such an attack could be brought on by stress. The examiner concluded that, taken together with McKinley’s testimony that he was suffering from a great deal of stress related to his problems with his supervisor, this evidence from Dr. Cowan sufficiently demonstrated that McKinley’s cardiac episode was work-connected for purposes of triggering the presumption of com-pensability.
The examiner further concluded, however, that the employer rebutted the presumption by pointing to the undisputed fact that McKinley’s heart-related attack occurred on Interstate 270 while McKinley was driving to work, and then noting the general rule that an employee cannot recover for injuries suffered while coming and going to work. See Grayson v. District of Columbia Dep’t of Employment Servs., 516 A.2d 909, 911 (D.C.1986) (noting existence of general rule).
This analysis misses the point and misapplies DOES’s own precedents for assessing the compensability of cardiac injuries sustained off the job site yet attributable to employment-related stressors (either physical or emotional). The mere fact that McKinley suffered the attack of ventricular tachycardia while driving to work does not constitute “evidence such as a reasonable mind might accept as adequate to support a conclusion” that the cardiac episode was not causally related to McKinley’s employment. 1 Arthur Larson & Lex K. Larson, The Law of Workmen’s Compensation § 10.83(b), at 3-188 (1997) (internal quotation marks omitted).
In Austin v. Eichberg Constr., H & AS No. 88-311 (Feb. 16, 1990) (remand order), the claimant was assaulted off the work site by a co-worker. See id. at 3. Austin argued the assault was work-related, but the hearing examiner denied compensation, applying the general rule that injuries taking place off the employer’s premises are not compensable. The Director reversed, concluding:
The matter of compensability of claimant’s injury is not decided by the mere application of the so-called “coming and going” rule emphasizing the situs of the injury.... There was substantial evidence ... which indicated that the claimant’s assailant was a co-worker who had quarreled with the claimant concerning work-related matters only a short time before the injury, a quarrel which clearly resulted in the assailant’s dismissal. Thus, the origin of claimant’s injury arose out of a work-related incident.
Id.
According to the statute, the work-connection or work-relatedness required for com-pensability is not determined solely by reference to whether the injury occurred on site or off site; rather, work-relatedness is established through a more complex inquiry by demonstrating that the disability “ar[ose] out of and in the course of employment.” D.C.Code § 36-301(12). While this language includes two concepts, they are closely related. As to the first, the Director has held that “[f|or an employee’s injury to have arisen out of the employment the obligations or conditions of employment must have exposed the employee to the risks or dangers connected with the injury.” Grayson, 516 A.2d at 911 (internal quotation marks omitted) (emphasis added). As to the second, “[a]n injury is said to arise in the course of employment when it takes place within the period of employment, at a place where the employee reasonably may be, and while he is fulfilling his duties or engaged in doing something incidental thereto.” 1 Larson & Larson, supra, § 14.00, at 4-1 (emphasis added). As Austin makes clear, this latter test does not automatically preclude compensation for an injury that occurs off the job site if the origin of that injury is demonstrably work-related.
Although the examiner concluded that McKinley’s employer successfully had rebutted the presumption of compensability, the examiner at the same time acknowledged that, in Austin, the DOES Director accepted *1389Professor Larson’s view of the interdependence of these two concepts:
In practice, the “course of employment” and “arising out of employment” tests are not, and should not be, applied entirely independently; they are both parts of a single test of work-connection, and therefore deficiencies in the strength of one factor are sometimes allowed to be made up by strength in the other.
2 Larson & Larson, supra, § 29.00, at 5-476, quoted in Austin, supra, at 3. In considering the two tests in concert, a useful way of conducting the inquiry recognizes the importance of both medical causation and legal causation. Larson’s treatise argues persuasively that, particularly in “heart cases” — of which this case is one — there are two distinct causation issues: (1) whether the exertion actually caused, from a medical point of view, the claimant’s injury; and (2) whether the particular exertion is deemed sufficient by the courts to establish legal causation. See 2 Larson & Larson, supra, § 38.83(a), at 7-312 to -319. While neither form of causation is precisely coextensive with the “arising out of’ and “arising in the course of’ tests, respectively, these concepts of medical and legal cause taken together offer a comprehensible mechanism for looking at the factors underlying each statutory test. In a case like this one, where the heart-related attack occurs off the employment premises but allegedly is attributed to work-related conduct or stimuli, the medical cause inquiry essentially corresponds to evaluating the risks and dangers to which the employment exposes the employee (“arises out of’), whereas the legal cause inquiry covers the scope of activities which the employment can be said to embrace (“in the course of’).
With this medical/legal distinction in mind, I believe it becomes apparent that the hearing examiner’s decision should not be sustained. First, as to medical causation, the employer has provided no evidence whatsoever to overcome the presumption of medical causality. It did not rebut the testimony of McKinley and his cardiologist that linked work-related stress to McKinley’s attack of ventricular tachycardia. The employer offered no expert medical testimony to the effect that stress does not contribute to attacks of ventricular tachycardia; nor did any expert for the employer testify that McKinley had not been under a sufficient amount of work-related stress to trigger such a cardiac episode. The employer, in fact, offered no testimony from any cardiologist.
Nor does the examiner’s cryptic observation unrelated to the expert testimony — that “[tjhere was evidence that the hostility had subsided” — prove sufficient to defeat the presumption of causation. The only evidence the examiner cited to support this finding was: (1) the general timeline— McKinley’s quarrel with Currence occurred on January 6, while the cardiac episode took place on the highway on January 10; (2) McKinley “had filed a grievance, a recognized outlet for disagreements with management”; and (3) McKinley “had dinner with his family and went to church.” In evaluating medical causation, the relevance of these factors is speculative. Absent medical testimony in support of the examiner’s analysis, the separation in time does nothing but establish a mere possibility that McKinley’s anxiety had waned during the three days between his quarrel with Currence and his attack. McKinley’s cardiac episode, moreover, was not isolated from the workplace; it occurred on his way to work after his employer had told him he must continue to work with Currence. Furthermore, McKinley testified that he was thinking about the impending confrontation with Currence when he suffered the attack. This brief (three-day) lapse of time, therefore, even when coupled with the fact that McKinley had filed a grievance rather than taking less civilized action, and had eaten dinner with his family and gone to church before returning to work, does not — without supporting expert testimony — negate medical causation. It is not the kind of “ ‘circumstantial evidence specific and comprehensive enough to sever the potential connection between a particular injury and a job-related event.’ ” Ferreira, 531 A.2d at 655 (quoting Swinton, 180 U.S.App. D.C. at 224, 554 F.2d at 1083).
The next issue is legal causation. While the examiner correctly observed, by citing a general legal rule, that an employee typically *1390cannot recover for injuries that occur on the way to work, this general rule has no application to the present case. Under the Austin test explained above, the “main inquiry here is whether the work-related dispute caused, or contributed to, claimant’s injury.” Austin, supra, at 4 (emphasis added). Rather than reading the examiner’s conclusion about the subsidence of hostility as an observation about the lack of medical causation, I believe her conclusion is better understood as the reflection of a legal judgment — a limitation of legal cause — to the effect that McKinley should not be permitted to recover compensation, given (1) the separation in time between the last confrontation with Currence and the onset of the heart-related attack, (2) his presumably peaceful time at dinner and at church with his family, and (3) the fact that he filed an employee grievance — “a recognized outlet for disagreements with management” — that supposedly reflected a calm, not anxiety-provoked, response to workplace issues.
I do not believe this reasoning can withstand scrutiny. It would be a peculiar rule of legal causation indeed that prevented McKinley from recovering for a disability unquestionably attributable, medically, to workplace stress simply because he was away from work for a few days, peacefully ate and worshipped with his family, and used available mechanisms of resolving employment disputes — especially when he was forced to continue to work with the supervisor whose conduct he complained of, and was on his way to work where he would have to face that supervisor when the incident occurred.
In Austin, the Director quoted with approval Professor Larson’s analysis of a New York case on the work-relatedness of an assault: “[SJince the ultimate test applied by Judge Cardozo was whether ‘the quarrel from origin to ending must be taken to be one,’ it should make no difference how widely separated the assault was from the employment in time and space if it remained an inherent part of an employment incident.” 2 Larson & Larson, supra, § 29.21, at 5-488 (quoting Field v. Charmette Knitted Fabric Co., 245 N.Y. 139, 156 N.E. 642, 643 (1927)), quoted in Austin, supra, at 3. The same analysis should apply here to the issue of legal causation. If the heart-related attack on the way to work was, in medical fact, related to work stress, the mere fact that the attack occurred off the job site does not preclude recovery as a matter of law.
Nor does McKinley’s preexisting heart condition, hypertrophic cardiomyopathy, necessarily defeat his claim, since work-caused aggravation of a preexisting condition can serve as a basis for worker’s compensation just as a first-time, work-caused injury can do so. See Ferreira, 531 A.2d at 660. I think Larson gets it right when he proposes the following test for legal causation:
If there is some personal causal contribution in the form of a previously weakened or diseased heart, the employment contribution must take the form of an exertion greater than that of nonemployment life.
2 Larson & Larson, § 38.83(b), at 7-320 (emphasis in original). McKinley’s testimony was certainly sufficient to trigger the presumption of compensability, because his testimony clearly established that the level of stress he had experienced as a result of Currence’s escalating attacks on him was unusual as compared to the ordinary stresses of nonemployment life.2 I would allow *1391DOES on remand to determine whether any evidence already introduced in the record overcomes this presumption by demonstrating “usualness.”3 I would then provide that if the evidence of “unusualness” fails to overcome this presumption, DOES must consider the issues not previously addressed by the hearing examiner’s order, namely, whether: (1) McKinley’s psychological condition, which all of the psychiatrists attributed to his attack of ventricular tachycardia, see supra note 1, arose out of and in the course of employment; (2) his psychological condition was totally disabling; and (3) he voluntarily limited his income.
III.
The majority concludes that “there was substantial evidence in the record to support the hearing examiner’s conclusion that workplace stress did not trigger Mr. McKinley’s V-tach on January 10, 1991.” Ante at [21]. That conclusion, however, is expressed without coming to grips first with the presumption of compensability.
It is important to recall that the hearing examiner herself concluded that McKinley’s testimony about his work stress, when coupled with Dr. Cowan’s testimony that such stress could trigger an attack of ventricular tachycardia, reflected enough work-connectedness to trigger the presumption of com-pensability, and thus shifted to the employer the burden of introducing “evidence specific and comprehensive enough to sever the potential connection between”- the stress and the heart-related attack. Ferreira, 531 A.2d at 655 (internal quotation marks omitted). The majority not only fails to acknowledge that the burden here has shifted but also neglects to recognize that the only evidence the examiner relied on to discredit McKinley’s claim of a work-related episode is its off-site location (Interstate 270) — evidence that pertains only to legal, not medical, causation. The majority, however, does not even discuss legal causation, preferring simply to repeat the hearing examiner’s rejection of Austin which, as I have tried to show, is legally incorrect.
I recognize that Dr. Cowan, the cardiologist, testified it was possible for ventricular tachycardia to occur without unusual stress, but his bottom-line opinion included McKinley’s stress level as a precipitating factor. As a result, the examiner was satisfied that McKinley and Dr. Cowan had combined to: present a prima facia case of medical causation, trigger the presumption of compensability, and shift the burden of production to the employer.
Where I fundamentally differ with my colleagues, therefore, is in their acceptance of the examiner’s purely legal — not medical— basis for saying the presumption of compens-ability did not survive, namely, that McKinley’s heart-related attack occurred off the job site, after a meal and church with his family, during a period after he had filed a grievance against his supervisor. That, I suggest, is not enough to rebut the presumption of com-pensability, based on medical causation which the hearing examiner herself did not find refuted.
* * # * # *
*1392Because the majority fails to apply the statutory presumption of compensability and muddles the analysis by failing to recognize the different roles of medical and legal causation — and because the examiner’s order does not comport with the law — I respectfully dissent from the court’s opinion and judgment.
. There was no dispute that McKinley's proof satisfied the third step of his analysis: the ventricular tachycardia caused McKinley's psychological disorder. McKinley’s psychiatric expert. Dr. Jeffrey Mendell, opined that McKinley suffered from "post-traumatic stress disorder.” The employer's experts, Dr. Lawrence Brain and Dr. Brian Shulman, opined that McKinley had an "anxiety disorder.” However the psychological disorder should be characterized, the hearing examiner noted that each psychiatrist "opined that the cause of claimant's disorder was the attack of ventricular tachycardia on January 10, 1991.”
. The majority's reliance on Spartin v. District of Columbia Dep’t of Employment Servs., 584 A.2d 564 (D.C.1990), and thus on the objective test discussed in that case for determining the com-pensability of emotional injury attributable to employment stress, is perplexing. See ante at 1384-1385 nn. 7-8. McKinley alleges that work-related stress caused his attack of ventricular tachycardia, which in turn caused his anxiety disorder. These distinct causal steps should not be lumped together to conclude that McKinley seeks compensation for emotional injury caused, simply, by job-related stress. The first question, therefore, is whether McKinley’s cardiac episode was work-related. Spartin has nothing to say about the work-connectedness of a cardiac injury, and thus we should apply Larson's legal causation formula for cardiac episodes partially caused by a diseased or weakened heart: was McKinley's job stress an "exertion greater than that of unemployment life”? If the answer is yes, the second question is whether McKinley's psychological condition satisfies the work-relatedness test because it resulted directly from a compensable injury, namely McKinley's near-death experience on Interstate 270. Spartin offers nothing here. As the majority itself notes, *1391see ante at 1385 n. 8, Spartin's test applies only to "non-traumatically caused mental injury.” 584 A.2d at 569 (internal quotation marks omitted). In sum, because McKinley's complaint alleges that his psychological condition was caused by a traumatic, stress-related, heart-related attack — not simply by work stress itself — -Spartin is altogether inapplicable.
. In Capital Hilton Hotel v. District of Columbia Dep't of Employment Servs., 565 A.2d 981 (D.C.1989), we noted the Director has concluded that "an employee with a preexisting arteriosclerotic condition can meet the 'unusual exertion’ test merely by proving job-related stress that is unusual compared to other occupations or to his daily life; he [or she] need not show a job stress unusual for him [or her].” Id. at 986 (citing Trilling v. Washington Area Metro. Area Transit Auth., H & AS No. 86-449 (Sept. 16, 1988)). This test substantially modified the Director’s more stringent formulation of unusualness articulated in Rose v. George Hyman Constr. Co., H & AS No. 83-226 (Aug. 27, 1984), aff'd, sub nom. George Hyman Constr. Co. v. Rose, 497 A.2d 103 (D.C.1985). See Capital Hilton Hotel, 565 A.2d at 985-86 & n. 7 (noting change and collecting cases where this court declined to reach merits of validity of Rose rule).