Plaintiff began receiving supplemental security disability benefits in 1976. In December 1977, she was informed that her disability had ceased. This determination was reversed on March 31, 1978 by an administrative law judge (Judge Ohanian) after a hearing, and plaintiff’s benefits continued. In June 1979, she was informed that her disability had ceased in April 1979. After a hearing on August 20,1979, another ALJ (Judge Varni) upheld this decision. When the Appeals Council declined to review the decision of the ALJ, it became the final decision of the Secretary. In due course, adopting the conclusions and findings of a federal magistrate, the district court affirmed, and this appeal followed.

Claimant argues that the decision was not based on substantial evidence, that various of her procedural rights were abused, and that the determination resulted from illegal use of social security regulations as a substitute for an individualized inquiry into her disability.

Claimant’s attack on the use of 20 C.F.R. §§ 416.902 et seq. as the basis for a disability determination is not well taken. However, because the decision of the Secretary is not supported by substantial evidence, we reverse.

I. BACKGROUND

Claimant underwent a lumbar laminecto-my in 1973. On the basis of a report from Dr. Tsuneo Hirabayashi relating to a November 28, 1977 consultative examination, the Social Security Administration determined that claimant’s disability had ceased in November of 1977. In his decision overturning this determination, Administrative Law Judge Ohanian observed:

He went on to say:

Judge Ohanian’s resolution of these questions was expressed in his findings that:

However, Judge Ohanian also noted that 20 C.F.R. § 404.1507 provides that a claimant who “willfully fails to follow prescribed treatment cannot by virtue of such failure be found to be under a disability,” and observed that “the claimant . . . should not expect the Administration to continue payment of benefits, if she willfully fails to follow prescribed treatment to alleviate her impairment.” In this connection, Judge Ohanian found that her impairment “could reasonably be regarded as partially remediable and possibly substantially remediable,” and that “[consequently, she should take reasonable action to treat and alleviate her symptoms and impairments.” He also found that her impairments should be reevaluated “at some appropriate future date.”

Thereafter, claimant began undergoing treatment in the office of Dr. Morrie Brandman, a specialist in family practice, and subsequently from Dr. William P. Brac-ciodieta, a neurologist. She also availed herself of the vocational services of the California Department of Rehabilitation.

In June 1979, the Administration informed claimant that her disability was considered to have ended in April 1979, and that her benefits would cease at the end of June 1979. In August, Administrative Law Judge Varni upheld this determination. His decision was based on medical reports from Dr. Brandman and Dr. Bracciodieta, and on the results of electromyographic and radiological studies undertaken at Dr. Brac-ciodieta’s request. It was further based on his assessment of the claimant’s testimony at the hearing as to subjective symptoms of pain, and his assessment of the testimony of Mr. Rene Garcia, a program supervisor with the California Department of Rehabilitation, who testified on claimant’s behalf.

II. THE MEDICAL REPORTS

The central issue in this appeal is whether the various medical reports furnish an adequate basis for the findings and decision of the tribunals below that claimant’s disability has ceased. We turn first, therefore, to an examination of those reports.

a. The Brandman Reports

Judge Varni’s review of the Brandman reports was as follows:

Study of his entire decision suggests quite clearly that Judge Varni relied in part on the notations in the Brandman reports to the effect that the claimant “feels fine” or is “feeling not too bad” to discount her own testimony that she was, and had been, in almost constant, severe pain. The other information he abstracts from the Brand-man reports does not seem to have any direct bearing on whether she was still suffering from her earlier lumbar problems, nor whether she was willfully failing to seek effective treatment for them. In light of this reliance, examination of the Brand-man reports themselves is revealing. First of all, the entries relating to blood pressure, weight, and pulse are clearly in a different handwriting from the remainder of the notations. They all come at the beginning of the notations for each day’s visit. All of the damning comments (“feels fine,” etc.) occur in these initial notations. These facts suggest (although they do not prove) that these notes were written not by Dr. Brand-man or Dr. Jones, but by a nurse or other office assistant. That possibility raises obvious questions as to the weight these comments ought to be given in assessing the claimant’s disability. Those questions are nowhere discussed, or even mentioned, by the ALJ, the magistrate, or the district judge. We decline to speculate whether we would be obliged to uphold an explicit finding that these remarks represented the clinical impressions of a treating physician. In the absence of such a finding, however, we are not persuaded that the remarks furnish any significant evidence of the claimant’s medical condition. Secondly, although Judge Varni relates only that Dr. Brand-man treated claimant for hypertension, the notes clearly indicate a diagnosis of “Post Lumbar Disc Pathology.” These records simply do not contain any evidence that we can detect that the claimant’s condition had improved at all.

b. The First Bracciodieta Report

This report, dated 4/24/79, is described by Judge Varni as follows:

Left out is Dr. Bracciodieta’s “diagnostic impression”: “Chronic low back pain syndrome with probable chronic, non-progressive radiculopathy.” Also unmentioned is the fact that this report indicates no change, as far as we can tell, from the neurological study performed by Dr. Hira-bayashi in November 1977, in which, aside from slightly diminished sensations in the medial half of the left foot, and a slight weakness in motor power on the right side, the results were all normal. That report, of course, was already determined by Judge Ohanian in the 1978 hearing to be consistent with the continuing existence of claimant’s disability.

c. The Electromyographic and X-Ray Studies

Judge Varni also relied on the EMG and X-ray studies ordered by Dr. Bracciodieta. Judge Varni correctly relates that the myo-grams were not “positive for the existence of any neuropathology.” Left out, however, is this statement of the physician who performed the tests:

Judge Varni also correctly reports that the X-rays disclosed only “mild degenerative and post-surgical changes.” Again, however, this represents no change from the clinical status of the claimant at the time of the prior hearing.

The account above comprehends all of the medical evidence relied on below to support the Secretary’s determination that the claimant’s disability has ceased. We are of the opinion that it does not comprise substantial evidence of any change in the claimant’s condition since the 1978 hearing and determination of disability. Under the controlling principles of law, as we explain below, that fact is dispositive.

III. APPLICABLE LEGAL STANDARDS

When the Administration has determined that a claimant’s disability has ceased, the burden of proof to establish otherwise lies with the claimant. Gonzalez v. Harris, 631 F.2d 143, 145 (9th Cir. 1980). Furthermore, the burden “is a continuing one. It does not cease or shift after an initial ruling of disability has been had.” Id. In an appropriate case, however, a pri- or ruling of disability can give rise to a presumption that the disability still exists. “Once evidence has been presented which supports a finding that a given condition exists, it is presumed in the absence of proof to the contrary that the condition has remained unchanged.” Rivas v. Weinber-ger, 475 F.2d 255, 258 (5th Cir. 1973). A presumption, of course, does not affect the ultimate burden of proof. It does, however, impose “on the party against whom it is directed the burden of going forward with evidence to rebut or meet the presumption.” Fed.R.Evid. 301.

The Secretary’s own regulations provide that the decision of an ALJ on a disability question becomes binding “on all parties to the hearing” if none of the specified avenues of review are exercised by the claimant or the Secretary. 20 C.F.R. § 416.1455 (1981). In the present case, neither the Secretary nor the claimant saw fit to disturb the 1978 determination by an ALJ that the claimant was disabled at that time. The question before us now, therefore, is whether that determination gave rise to a presumption at the time of the 1979 hearing that the claimant was still disabled — a presumption which the Secretary was required to “meet or rebut” with evidence that her condition had improved in the interim. We answer that question in the affirmative. We are unable to discern any reason why the familiar principle that a condition, once proved to exist, is presumed to continue to exist, should not be applied when disability benefits are at stake. The Fifth Circuit has expressly approved such application. Rivas v. Weinberger, supra. Accord, Prevette v. Richardson, 316 F.Supp. 144, 146 (D.S.C.1970). In Prevette, the district court reversed the Secretary’s determination denying disability benefits where there had been a prior determination of disability, and where neither of the two subsequent medical reports indicated recovery by the claimant. The case before us is almost identical.

As we have said, when a claimant is entitled to the benefit of a presumption that her disability still exists, the burden is still on her to prove her case. All the presumption does is impose on the Secretary a burden to come forward with evidence that her condition has changed. Whether that burden has been met is a judgment to be made initially by the Secretary, and that judgment cannot be overturned on appeal if it meets the “substantial evidence” standard. See, e.g., Gonzalez v. Harris, supra, 631 F.2d at 145 . But where, as here, there is essentially no evidence to support a conclusion that the claimant’s condition has changed, the substantial evidence standard has not been met. Accordingly, we REVERSE.

. Although Dr. Brandman’s name and signature appear on the Medical Report Form filed on May 23, 1979 with the Administration, the substantive portions of the form simply refer to the attached medical records. Those records bear the signature of, and seem to have been compiled by, a Dr. R. Fleming Jones, apparently a physician working in Dr. Brandman’s office. Nevertheless, the parties to this action, the ALJ, the magistrate, and the district judge, all refer to Dr. Brandman as the treating physician and the author of the reports. . See footnote 1, supra. . The notes for 8/14/78 include the following: Diag: (1) Hypertension (2) Post-Lumbar Disc Pathology Those for 8/22/78 include: Diag: (1) Hypertension (2) Post-Lumbar Disc Syndrome (3) Chronic Tension Those for 9/8/78 include: Diag: (1) Hypertension (2) Lumbar Disc Syndrome (3) Chronic Tension [illegible] . Judge Varni’s observations were as follows: At the request of Dr. Bracciodieta electro-myographic studies of the lower extremities and associated paravertebral musculature were performed together with a nerve conduction [sic] studies of both lower extremities. The nerve conduction studies of both lower extremities were within normal limits and not demonstrative of any neuropatholo-gy. The electromyogram was also interpreted as being a normal study so far as could be determined considering the claimant’s prior lumbar laminectomy. In any event neither electrodiagnostic study was positive for the existence of any neuropathology. X-rays of the lumbosacral spine taken on April 25, 1979 were interpreted by a radiologist as showing evidence of a partial laminectomy on the right at the L5 level. There was said to be a very slight disc space narrowing at the L4-5 level otherwise the stature, alignment, and disc spaces were unremarkable. No significant degenerative hypertrophic changes were apparently found by the radiologist. Flexion and extension views demonstrated appropriate motion at all levels within normal limits. The impression was mild degenerative and post-surgical changes as previously noted.