The North Dakota Workmen’s Compensation Bureau appeals from a district court judgment which reversed the Bureau’s order denying Robert D. Syverson’s claim for benefits. We affirm the district court judgment.
Syverson filed a claim with the Bureau for compensation and medical expenses in connection with an alleged injury to his shoulders during his employment as a delivery person with Dodgson, Inc., in Jamestown. Syverson first sought medical treatment for pain in his shoulders from Dr. Ard Mardirosian, an orthopedic surgeon, on July 30, 1984. At that time, Syverson, age 37, had been handling and moving heavy appliances and furniture for five years while employed by Dodgson. He had not previously suffered from or sought medical attention for a shoulder disability.
After a number of examinations and treatment attempts, including surgery, Dr. Mardirosian determined that Syverson had bilateral degenerative arthritis of the acro-mioclavicular (AC) joints. When Syverson was allowed to return to work, limitations were placed on the amount of lifting he could do.
Syverson asserted to the Bureau that his shoulder condition, in part caused by osteo-arthritic changes, was brought on or worsened through his employment as a furniture and appliance handler. The Bureau dismissed Syverson’s claim. Noting that “arthritis, in various forms, is a disease common to the general public,” the Bureau determined that Syverson did not receive an injury by accident arising out of and in the course of employment; that his arthritic condition was not causally related to an employment injury; and that the arthritic condition was not fairly traceable to his employment. The district court reversed the Bureau’s decision, concluding that “a causal link between the employment and injury is the only reasonable finding that could be made.”
The Bureau asserts in this appeal that Syverson failed to prove by a preponderance of the evidence that the acromioclavi-cular arthritic condition in his shoulders was a compensable injury.
A “compensable injury” is defined in § 65-01-02(7), N.D.C.C., as “an injury by accident arising out of and in the course of employment.... Such term, in addition to an injury by accident, includes: a. Any disease which can be fairly traceable to the *690employment.” 1 In Satrom v. North Dakota Workmen’s Compensation Bureau, 328 N.W.2d 824 (N.D.1982), we discussed the “injury by accident” requirement of this statute in relation to a claimant’s gradual development of a back injury. In that case, a hairdresser sought benefits from the Bureau for an acute disc syndrome she claimed was caused by six years of repeated bending and standing on a hard floor while working at the salon. The Bureau ruled that she failed to prove an injury by accident and that her condition was not causally related to her employment.
This court reversed, concluding that a back injury, such as the claimant’s acute disc syndrome, “which can be medically related to repeated stress or strain in a claimant’s usual work exertions, is a com-pensable injury” even if the cause was not of an “accidental character.” Satrom, supra, 328 N.W.2d at 828. We rejected the Bureau’s argument that the employment must be the sole cause of the claimant’s injury:
“It is sufficient if the work-related stress is a ‘substantial contributing factor’ to the injury. Nelson v. North Dakota Workmen’s Comp. Bureau, 316 N.W.2d 790, 795 (N.D.1982). Thus, ‘the fact that an employee may have physical conditions or personal habits’ which make him or her more prone to such an injury does not constitute a sufficient reason for denying a claim if the preponderance of the evidence indicates that the injury ‘was causally related to the worker’s employment, with reasonable medical certainty’, and was precipitated by usual stress. Nelson, supra.” Satrom, supra, 328 N.W.2d at 831.
The same rationale applies to a claimant’s development of an arthritic condition. See IB Larson, Workmen’s Compensation Law § 39.10 (1987). Thus, the focal issue in this case is whether Syverson proved by a preponderance of the evidence that the heavy lifting he performed at work was causally related to the development of his arthritic condition.
In an appeal of the decision of an administrative agency, we review the decision by the agency rather than the determination of the district court. Grace v. North Dakota Workmen’s Compensation Bureau, 395 N.W.2d 576, 579 (N.D.1986). We consider whether a reasoning mind reasonably could have determined that the factual conclusions reached were proved by the greater weight of the evidence. Ganske v. North Dakota Workmen’s Compensation Bureau, 355 N.W.2d 800, 802 (N.D.1984). The Workmen’s Compensation Act is to be construed liberally with the view of extending its benefit provisions to all who can fairly be brought within them. Claim of Bromley, 304 N.W.2d 412, 415 (N.D.1981). Moreover, because the adversary concept has only limited application to claims for workmen’s compensation benefits, the Bureau may not rely only upon that part of inconsistent medical evidence which is favorable to the Bureau’s position without attempting to clarify the inconsistency. Roberts v. North Dakota Workmen’s Compensation Bureau, 326 N.W.2d *691702, 706 (N.D.1982). We look to the entire record compiled by an administrative agency. Geiger v. Hjelle, 396 N.W.2d 302 (N.D.1986).
Syverson’s employment with Dodgson entailed years of lifting and moving heavy appliances and furniture and he had not previously suffered from any shoulder disability. In the “Attending Physician’s Report” submitted after one month of treatment, Dr. Mardirosian indicated that “[h]eavy lifting at work” was the claimant’s history of onset or manner of affliction, and answered “yes” to the question “Are you satisfied this is an occupational disease or injury?” In an “Attending Physician’s Statement” dated September 13, 1984, Dr. Mardirosian further answered, in response to a question whether the patient’s condition arose out of his employment, “Yes, due to lifting.”
During December 1984, Dr. Mardirosian referred Syverson to Dr. Marie Olestad, another orthopedic surgeon on the clinic staff. Dr. Olestad noted that Syverson suffered “occasional sharp pain with lifting,” and that “the history is one of a wear and tear arthritis in the AC joints.”
Relying on selected segments of Dr. Mardirosian’s testimony, the Bureau asserts that “the evidence does no more than speculate as to a possible causal relationship between any lifting and the arthritic injury.” The doctor testified on several occasions that it was only “possible,” and that he could not “hundred percent affirm,” that Syverson’s heavy lifting at work was related to his arthritic condition. The doctor also testified that he researched the problem and that “I have not been able to find an article really dealing specifically with this [AC] joint here that would produce degenerative changes so rapidly due to strain because of the work itself.... Now I presume it’s possible.”2
However, Dr. Mardirosian testified more fully, as follows:
“Q [By Mr. Buchanan] Apparently by that time, in the number of times that you had seen Mr. Syverson, it was a concern in your mind, was it not, that he avoid liftings because of those shoulders?
“A [By Dr. Mardirosian] Right.
“Q And the earlier history was, of course, that he had expressed to you that this discomfort apparently presented itself when lifting?
“A That’s my original note as he indicates.
“Q In the latter part of September of ’84, you were advising to avoid lifting, were you not, Doctor?
“A Right.
“Q That was to protect the irritation or aggravation in those shoulders?
“A Protect any strain that might cause.
* # * # # *
“Q So various sites and components, all of which were associated with shoulder pain in this patient; is that correct?
“A Yes.
“Q And apparently a history of employment that required lifting, is that correct?
“A The lifting would have certainly produced a strain on the joints that might have been abnormal as well.
* * * * * *
“Q If I may at the risk of restating, is it your opinion to a reasonable medical certainty that Mr. Syverson’s use of those joints in handling appliances and furniture had to be taken into consideration in discussing this shoulder problem?
“A Well, I cannot see any other way to tell you that, how his condition had become worse otherwise_” [Emphasis added.]
While Dr. Mardirosian may have been unable to medically pinpoint the precise origin of Syverson’s arthritic shoulder condition, we believe that his testimony clearly demonstrates that he considered Sy-verson’s employment to be a substantial contributing factor in the development of *692his arthritis. Considering the whole of Dr. Mardirosian’s testimony together with the other evidence in the record, we conclude that a reasoning mind could not have reasonably determined, as did the Bureau, that the greater weight of the evidence is that Syverson’s arthritic condition had no causal relationship with his employment or that it was not an injury arising out of and in the course of his employment.
Accordingly, the district court’s judgment reversing the Bureau’s dismissal of Syverson’s claim for benefits is affirmed. The case is remanded to the Bureau for a determination of the benefits to which Sy-verson is entitled.
LEVINE and GIERKE, JJ., concur.. The term "fairly traceable to the employment" is defined as only a disease which:
"a. Arises under conditions wherein it is apparent to the rational mind upon consideration of all the circumstances that there is a direct causal connection between the conditions under which the work is performed and the disease;
"b. Can be seen to have followed as a natural incident of the work as a result of the exposure occasioned by the nature of the employment;
“c. Can be fairly traced to the employment;” § 65-01-02(12), N.D.C.C.
Section 65-01-02(7)(a), N.D.C.C., further provides:
“Ordinary diseases of life to which the general public outside of the employment is exposed shall not be compensable except where the disease follows as an incident to, and in its inception is caused by a hazard to which an employee is subjected in the course of his employment. The disease must be incidental to the character of the business and not independent of the relation of employer and employee. The disease includes impairment and effects from radiation fairly traceable to the employment. It need not have been foreseen or expected, but after it is contracted, it must appear to have had its origin in a risk connected with the employment and to have flowed from that source as a rational consequence.”
. The Bureau also relies on an "Office Memo” prepared by its medical doctor. However, the memo indicates that further information was needed to review the claim and suggests only a "possible dismissal.”