This case presents for review a January 31,1990, decision of the Board of Veterans’ Appeals (BVA or Board) denying service connection for post-traumatic osteoarthritis of the knees, hips, spine, left ankle, and shoulders. We reverse the Board’s decision as it pertains to the right shoulder because we find the decision to be “arbitrary and capricious, an abuse of discretion, or otherwise not in accordance with law”. 38 U.S.C. § 7261(a)(3)(A) (formerly 4061(a)(3)(A)). As to the knees, hips, spine, left ankle, and left shoulder, we vacate the decision and remand the case for readjudi-cation and a statement of the Board’s findings and conclusions and the “reasons or bases” for those findings and conclusions. 38 U.S.C. § 7104(d)(1) (formerly 4004(d)(1)); See Gilbert v. Derwinski, 1 Vet.App. 49 (1990).
I. BACKGROUND
The appellant, a veteran of the Second World War, had active duty, including service as a turret gunner on a B-17, in the United States Army Air Corps from December 13, 1942, to October 29, 1945. On February 10, 1944, while on a bombing mission over Germany, his plane was shot down by enemy anti-aircraft and fighters. He was captured and held as a prisoner of war of the German government for 14V2 months, until April 26, 1945, when he was returned to U.S. military control.
In early 1947 he filed a claim for disability compensation benefits from the Veterans’ Administration (now the Department of Veterans Affairs) (VA) for arthritis and other disabilities. In support of his claim, he submitted a January 21, 1947, medical statement from Dr. Stewart, a private physician, who stated that he had been treating the veteran for two months for “acute arthritis”. R. at 15. A January 23, 1947, VA exam revealed “[n]o orthopedic abnormality at this time.” R. at 19. In April 1947 the claim was denied by a VA Regional Office (RO) rating decision. The rating board noted that the VA exam “failfed] to find chronic residual Arthritis from the acute condition reported by Dr. Stewart.” R. at 23.
In May 1947 the veteran submitted a statement describing a forced parachute jump from 13,000 feet at the time his plane was shot down, the cold conditions of the POW camps, and a forced march he endured from February to April 1945. In January 1948 a second statement from Dr. Stewart was submitted. Dr. Stewart stated that the veteran had complained of pain in the hands, feet, and knees in 1947. He also stated that the veteran “clearly had an acute arthritis” of the knees in January 1947, and since then he “has been able to continue regularly at work. However, for periods of several days his knees are moderately painful.” R. at 28. In 1949 three statements from fellow POWs were submitted on behalf of the veteran. All three recounted that they were exposed to severe weather and endured a poor diet while prisoners of war. They also stated that they withstood a forced march of 600 to 700 miles from February to April of 1945 and that during that time they marched and slept in the open in cold and snow. Another VA examination was performed in March 1949. Complaints of pain in the knees and the right shoulder were noted. The diagnosis was “[ojrganic orthopedic disease not found.” R. at 35. In April 1949 the VARO awarded the veteran service connection (rated at 10 percent) for a condition related to his POW experiences, but unrelated to arthritis. R. at 39.
Another VA examination was conducted in early 1951. Complaints of stiffness and weakness of the left knee were noted. Physical examination found “no swelling, deformity or tenderness of any major joint.” R. at 47. On January 19, 1951, the arthritis claim was again denied by VARO rating decision. R. at 49.
In 1974 the veteran was hospitalized at a private facility and diagnosed with “os*443teoarthritis, left knee.” R. at 51. In 1978, and again in 1982, the claim was denied by the RO. On March 15, 1983, the BVA issued a decision denying service connection for arthritis. The BVA found that the “[m]ultiple joint arthritis ... is not a result of the veteran’s experiences as a prisoner of war.” R. at 60.
In January 1984 a former-prisoner-of-war protocol examination was conducted by VA, producing a diagnosis of “osteoarthritis generalized] 10-yrs plus”. R. at 67. It was also noted that the veteran had had a total left knee replacement in 1983. The record indicates that in 1984 his right knee was also replaced.
In June 1987 another VA physical examination was conducted by Dr. Crowder, a VA physician. He noted that there were “minimal findings of osteoarthritis with no significant swelling of any of the joints at the present time.” R. at 85. He also noted that the “major problem” complained of by the veteran was the right shoulder and that the veteran also had complained of “problems” with the lower back, the ankles, the feet, the wrists, and the hands. Ibid. In conjunction with the physical examination, a radiological examination of the shoulders, the lumbar spine, the ankles, the feet, the hands, and the wrists was performed by Dr. Frere, a VA physician. He recorded the following opinion: “[Mjoderate to moderately severe arthritis of the wrists”, “moderate post-traumatic arthritis of the right ankle”, “minimal osteoarthritis of the left ankle”, “moderate degenerative changes and arthritis of the lower lumbar spophyseal joints and sacroiliac joints”, “moderate arthritis of the right shoulder”, and “minimal arthritis of the left shoulder.” R. at 90. Also, as to the right shoulder the examiner stated: “I would presume the patient has had previous shoulder trauma, perhaps even dislocations and this is a residual osteoarthritis.” R. at 89. As to the wrists, the examiner stated: “I presume this is post-traumatic though it could represent a[n] erosive osteoarthritis. Looks more post-traumatic than erosive.” R. at 90. After receiving Dr. Frere’s opinion, Dr. Crowder made a diagnosis of “[d]e-generative arthritis” involving “primarily” both knees, involving “probably” the lumbar spine and both ankles, and involving the shoulders, particularly on the right. R. at 91.
On February 8, 1988, the BVA issued a decision based upon a de novo review of the appellant’s claim as a result of the enactment of Public Law No. 99-576, § 108(a)(2), 100 Stat. 3248, 3252 (1986), which added “post-traumatic osteoarthritis” to the list of former POW presumptive disorders. The BVA granted service connection for post-traumatic osteoarthritis of the wrists and the right ankle (service medical records show that the veteran had sprained his right ankle in service in 1943), both of which were “presumed to be related to service.” R. at 97. The BVA found that “arthritis involving other joints, including the knees, hips, shoulders, lumbar spine or left ankle, is not shown to be of service origin or to be traumatic in nature.” Ibid.
In 1988, following the BVA’s denial of the claim, the appellant submitted a 1976 statement of Dr. Conley, a private physician, noting that the veteran had a “long standing history of degenerative arthritis involving the knees, occasionally the shoulders.” R. at 102. He also submitted a November 8, 1988, statement of Dr. Jef-fries, a private physician, regarding the arthritis of the knees. Dr. Jeffries noted that he had treated the veteran’s knee problems since 1982 when he had found “much more severe arthritis than one would expect in a 59 year old man.” R. at 107. He stated that “[wjhile it is impossible to abs[o]lutely prove or disprove the chain of events leading to the severe arthritis which Mr. Bailey has sustained and which ultimately lead to bilateral total knee joint replacements, it is my opinion that the trauma which he received during World War II was substantially involved in this process in a causal manner.” R. at 108. He also made the following observations: (1) “the negative impact on the knee [caused by parachute jumping] was definitely real and substantial and I believe within the reasonable degree of medical certainty is causally related to later knee *444arthritis”, (2) “[t]he very severe conditions found in prisoner of war camps as well as the forced march that Mr. Bailey endured would have been very traumatic to his knees and I believe within a reasonable degree of medical certainty would have been causally related to his subsequent knee arthritis”, and (3) Dr. Stewart’s 1948 statement “would certainly substantiate that Mr. Bailey was having problems very soon after his liberation at the end of World War II and in general go along with the picture of early post traumatic arthritis.” R. at 107.
In December 1988 a VA examination was conducted by Dr. Crowder again. Dr. Crowder noted that the veteran “has had arthritis or pain and bursitis perhaps in his shoulders. Has had injections in them. At the present they are doing relatively well.” R. at 117. The diagnosis was: “Generalized osteoarthritis with particular involvement of the wrists, knees and ankles, and also the lumbosacral spine, etc. NOTE: As was said by Dr. Jeffries, it is impossible to say whe[ jther this is traumatic in origin or not. I suppose theoretically with his parachute jumps in service, some of the etiology could have been traumatic although we have no specific instance to relate to. Even the forced marches theoretically may have been etiologic.” R. at 126.
On February 3, 1989, the RO rating board issued a decision denying the claim, finding that “[ojutside of mere speculation, we cannot find any basis on which to associate the arthritis of the knees, hips, shoulders, lumbar spine, and left ankle with the experience of his military service.” R. at 128.
On January 31, 1990, the BVA issued a decision denying the claim. The BVA stated that the newly submitted evidence “suggests that arthritis of multiple joints, other than the wrists and right ankle, is of degenerative etiology, which is associated with the aging process. Medical opinion to the contrary notwithstanding, we must conclude that, as the Board did previously, the veteran does not have traumatic arthritis in his knees, hips, spine, left ankle and shoulders.” Jack L. Bailey, BVA 90-02581, at 6 (Jan. 31, 1990). A timely appeal to this Court followed pursuant to 38 U.S.C. §§ 7252(a) (formerly § 4052(a)) and 7266 (formerly § 4066).
II. ANALYSIS
A.
The first issue we address is whether, as required by 38 U.S.C. § 5108 (formerly 3008), new and material evidence was submitted to reopen the claim in 1988 with respect to the shoulders, knees, hips, spine, and left ankle. We hold that new and material evidence was submitted to reopen the claim. In Colvin v. Derwinski, 1 Vet.App. 171, 174-75 (1991), this Court held “that to justify a reopening on the basis of new and material evidence, there must be a reasonable possibility that the new evidence, when viewed in the context of all the evidence, both old and new, would change the outcome.” See also Godwin v. Derwinski, 1 Vet.App. 419, 424 (1991). In the present case, the new evidence included the 1976 statement of Dr. Conley noting a “longstanding history of degenerative arthritis involving the knees, occasionally the shoulders”; the 1988 statement of Dr. Jeffries providing an opinion that the trauma the veteran received in WWII was “substantially involved” in causing the veteran’s arthritis of the knees; and the 1988 VA examination conducted by Dr. Crowder which noted that the veteran “has had arthritis or pain and bursitis perhaps in his shoulders”, “crepitance in the right shoulder”, “generalized osteoarthritis” of the wrists, knees, ankles, and lumbosacral spine, and the theoretical possibility that the generalized osteoarthritis may be traumatic in origin. R. at 102, 107-8, 117, and 126. In light of the BVA’s finding in its February 1988 decision that the arthritis of the knees, hips, shoulders, lumbar spine, and left ankle was not shown to be traumatic in nature, the above evidence, specifically raising the possibility that the arthritis was traumatic in nature, created a reasonable possibility that when the new evidence was viewed in the context of all the evidence there would be a change in out*445come such that the arthritis of those joints would be found to be traumatic in nature.
Because new and material evidence was submitted to reopen the claim, the BVA was required under 38 U.S.C. § 5108 to “review the disposition of the claim” with respect to the arthritis of the knees, hips, spine, left ankle, and shoulders. In reviewing the disposition of a claim, the BVA must look “at the new and material evidence in the context of the other evidence of record and not in isolation.” Jones v. Derwinski, 1 Vet.App. 210, 215 (1991). However, that was not the standard of review applied by the BVA here. Rather, it examined only the “newly submitted evidence” and found that it did “not establish that the arthritis of the knees, hips, spine, left ankle, and shoulders is traumatic in nature or otherwise related to the veteran’s active service.” Bailey, BVA 90-02581, at 6, 7. A remand on this basis would be required under our precedents. See, e.g., Jones, at 215; Manio v. Derwinski, 1 Vet.App. 140, 146-47 (1991).
However, in our review of the BVA’s January 31, 1990, finding that the arthritis of the knees, shoulders, hips, left ankle, and spine was not traumatic in nature, we also find reversible error in part.
B.
38 U.S.C. § 1112(b)(12) (formerly § 312(b)(12)) provides in pertinent part:
[I]n the case of a veteran who is a former prisoner of war and who was detained or interned for not less than thirty days, the disease of ... post-traumatic osteoarthritis, ... which became manifest to a degree of 10 percent or more after active military, naval, or air service shall be considered to have been incurred in or aggravated by such service, notwithstanding that there is no record of such disease during the period of service.
38 U.S.C. § 1112(b)(12). The statute does not require that the post-traumatic osteoarthritis become manifested within a certain period of time after separation from service in order for the presumption of service connection to arise. Nor is there any requirement that there be a finding that the trauma occurred in service. See Godwin, 1 Vet.App. at 426. The presumption arises, as long as a post-traumatic osteoarthritis becomes manifest, to a degree of 10 percent, at any time after separation from service. The presumption of service connection is rebutted under 38 U.S.C. § 1113 (formerly § 313) only where affirmative evidence to the contrary, evidence establishing intercurrent injury or disease, or willful misconduct by the veteran, is present. 38 U.S.C. § 1113.
In proposing the traumatic arthritis presumption in 1986, the Senate Committee on Veterans’ Affairs made special note of the “disagreement as to the adequacy of current medical science to distinguish between arthritis resulting from earlier trauma and arthritis which is the result of other causes or which normally occurs during the aging process.” S.Rep. No. 444, 99th Cong., 2d Sess. 30, reprinted in 1986 U.S.Code Cong. & Admin.News 5468, 5469, 5480.
C.
We hold that the Board’s conclusion in its 1990 decision that “the veteran does not have traumatic arthritis in his [right] shoulder[]”, Bailey, BVA 90-02581, at 6, must be reversed because it was reached in a manner that was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law” under 38 U.S.C. § 7261(a)(3)(A). In reviewing the claim of post-traumatic arthritis of the right shoulder, the Board, as noted in part A, above, was required by 38 U.S.C. § 5108 to review all of the evidence of record.
The evidence of record as to the right shoulder includes: complaints as early as 1949 of pain in the right shoulder; the 1987 VA examination report of Dr. Crowder which noted that the “major problem” complained of by the veteran was the right shoulder; the 1987 opinion of Dr. Frere of “moderate arthritis of the right shoulder” and his statement that “I would presume the patient has had previous shoulder trauma, perhaps even dislocations and this is a residual osteoarthritis”; the 1976 state*446ment (submitted in 1988) of Dr. Conley that the veteran “had a longstanding history of degenerative arthritis involving the knees, occasionally the shoulders”; and the 1988 VA examination report of Dr. Crowder which noted that the veteran “has had arthritis or pain and bursitis perhaps in his shoulders”. R. at 35, 85, 89-90, 102, 117.
The record also indicates that in 1988 the Board awarded service connection for post-traumatic arthritis of the wrists based on “evidence of record indicating that the arthritic disease of these joints is post-traumatic in nature.” R. at 97. That evidence included arthritic-type complaints of the wrists in 1947 and 1948 and Dr. Prere’s 1987 opinion that the veteran had “moderate to moderately severe arthritis of the wrists” and his statement that “I presume this is post-traumatic arthritis though it could represent a[n] erosive osteoarthritis. Looks more post-traumatic than erosive.”
However, the Board in its 1990 decision denied the claim for post-traumatic arthritis of the right shoulder despite the very similar evidence, including the complaints of right shoulder pain in 1947 and Dr. Frere’s opinion that the veteran had “moderate arthritis of the right shoulder” and “has had previous shoulder trauma, perhaps even dislocations and this is a residual osteoarthritis.” Whether the Board was arbitrary and capricious in the process it employed to find in 1988 that the arthritis of the wrists was post-traumatic in nature and that the arthritis of the right shoulder was not is not an issue that is before us here. However, in reviewing the reopened claim in 1990, the Board was obligated to review the old evidence in the context of the new evidence, which indicated that the right shoulder arthritis was, even as of 1976, “longstanding”. In light of the evidence of record and the previous findings of the Board, the Board’s conclusion in its 1990 decision, in the face of the almost identical evidence to that in the record as to the wrists, that the right shoulder arthritis was due to the aging process rather than trauma was the result of an arbitrary and capricious decision-making process under 38 U.S.C. § 7261(a)(3)(A).
We note that at oral argument VA counsel argued that the section 1112(b)(12) presumption was rebutted as to the right shoulder under section 1113 by a lack of post-service complaints as to the right shoulder, whereas, the presumption had not been rebutted as to the wrists because of the 1947 clinical records indicating “arthritic type complaints”. Without deciding whether rebuttal under section 1113 could be based on such a lack of post-service complaints, we point out that the record indicates that the veteran did indeed complain of “some pain in the right shoulder” as early as 1949 at a VA examination. R. at 34. Thus, the presumption as to the right shoulder could not have been rebutted by a lack of evidence of post-service complaints.
D.
As to the veteran’s claim to entitlement to service connection for arthritis of the knees, hips, spine, left ankle, and left shoulder, we remand the claim to the BVA for readjudication and a statement of “reasons or bases” for the Board’s findings and conclusions under 38 U.S.C. § 7104(d)(1). In denying the claim as to those joints (and the right shoulder), the Board made the following statement: “Medical opinion to the contrary notwithstanding, we must conclude that, as the Board did previously, the veteran does not have traumatic arthritis of the knees, hips, spine, left ankle and shoulders.” Bailey, BVA 90-02581, at 6.
In Gilbert v. Derwinski, 1 Vet.App. 49 (1990), we stated:
[T]he Board must identify those findings it deems crucial to its decision and account for the evidence which it finds to be persuasive or unpersuasive. These decisions must contain clear analysis and succinct but complete explanations. A bare conclusory statement, without both supporting analysis and explanation, is neither helpful to the veteran, nor ‘clear enough to permit effective judicial review’, nor in compliance with the statutory requirements.
Id. at 57 (citation omitted). See also Moore v. Derwinski, 1 Vet.App. 401, 404 (1991) and cases cited there.
*447The above quoted statement from the BYA decision is exactly the type of “bare conclusory statement” which Gilbert held to be inadequate. See also Sammarco v. Derwinski, 1 Vet.App. 111, 113-14 (1991); Ohland v. Derwinski, 1 Vet.App. 147, 149-50 (1991). The Board provided no explanation as to what “medical opinion to the contrary” it was referring, nor did it explain why such medical opinion was unpersuasive. Further, the Board provided no explanation for its statement that the “evidence suggests that arthritis of multiple joints, other than the wrists and right ankle, is of degenerative etiology, which is associated with the aging process.” Bailey, BVA 90-02581, at 6. The Board did not indicate what evidence suggested that the arthritis was associated with the aging process, nor did it explain why that evidence was more persuasive than the “medical opinion to the contrary”.
We note in particular that the Board provided no analysis of Dr. Jeffries’ opinion that trauma received by the veteran during World War II was “substantially involved ... in a causal manner” in the severe arthritis of the knees which ultimately lead to the total knee replacements. R. at 108. Nor did the Board provide any explanation as to what bearing, if any, the evidence that the veteran had had, as early as 1976, a “longstanding history of degenerative arthritis of the knees” (R. at 102), that he had been seen for “acute arthritis” of the knees as early as 1948 (R. at 35), and that in 1982 he had had “much more severe arthritis than one would expect in a 59 year old man” (R. at 107) had on its totally unsupported conclusion that that arthritis was due to the aging process.
■ We note further that in Colvin, 1 Vet.App. at 175, the Court held that the BVA must support its decision with “a medical basis other than the panel’s own unsubstantiated [medical] opinion.” Although that may not be the explanation for the reference to the aging process by the Board here, the requirements of Colvin must be followed on remand and, if the medical evidence of record is insufficient, the BVA is always free to supplement the record by seeking an advisory opinion, ordering a medical examination or providing medical authority, from recognized medical treatises or otherwise, that clearly supports its ultimate conclusions. See Colvin, at 175; Murphy v. Derwinski, 1 Vet.App. 78, 81 (1990).
At oral argument, VA counsel argued that Dr. Jeffries and Dr. Crowder, unlike the RO and the BVA, did not have before them the complete record, and, therefore, their opinions as to the possible causes of the arthritis should be given less weight. That may or may not be a valid reason for finding those opinions unpersuasive; however, the Board did not articulate it as a reason or basis for denying the claim. Thus, it does not alleviate the need for a remand for a statement of the reasons or bases for the Board’s findings and conclusions.
In addition, the Board provided no explanation or support for its statement that “the facts in this case” do not raise a reasonable doubt. See 38 U.S.C. § 5107(b) (formerly 3007(b)); see also 38 U.S.C. § 1154(b) (formerly § 354(b)) (reasonable doubt specifically applicable to combat veterans); 38 C.F.R. § 3.102 (last sentence) (1990) (to the same effect as the statute). In readjudicating this claim on remand, the Board should take into account in determining whether the veteran is entitled to the benefit-of-the-doubt the statement quoted above from the Senate Report that there is disagreement as to the adequacy of current medical science to distinguish between arthritis resulting from trauma and arthritis resulting from the aging process or other causes.
Finally, we note that if on remand the veteran is successful with respect to his claim for entitlement to service connection for post-traumatic osteoarthritis of the knees, the veteran, based on Dr. Jeffries’ opinion of the status of the arthritis in 1982 prior to the two knee-replacement procedures, might have a claim for discretionary reimbursement from VA under 38 U.S.C. § 503(a) (substantially similar to former *448§ 210(c)(2)) (administrative error), for any expenses he incurred for those procedures that were performed in private facilities.
III. CONCLUSION
Accordingly, we reverse the January 31, 1990, finding of the Board that the veteran’s right-shoulder arthritis was not traumatic in nature because, based on the evidence of record and the previous findings of the Board, we hold that that finding was reached in a manner that was “arbitrary and capricious, an abuse of discretion, or otherwise not in accordance with the law.” under 38 U.S.C. § 7261(a)(3)(A). The right-shoulder claim is remanded for a determination of the degree of disability of the veteran’s right shoulder post-traumatic osteoarthritis, including a determination of whether the right-shoulder arthritis had manifested itself to the degree of 10 percent for purposes of presumptive service connection under 38 U.S.C. § 1112(b)(12). As to the claim for entitlement to service connection for arthritis of the knees, hips, spine, left ankle, and left shoulder, we vacate and remand the decision for readjudi-cation in accordance with this opinion.
REVERSED IN PART AND VACATED AND REMANDED IN PART.