Wayne Winans appeals the denial of disability insurance benefits contending there is not substantial evidence to support the finding of no disability. We agree and reverse.

FACTS

On August 19, 1983, Winans applied for disability insurance benefits under Title II of the Social Security Act (“the Act”). His application alleged disability since June 18, 1982 due to narcolepsy. Narcolepsy is a rare syndrome of “recurrent attacks of sleep, sudden loss of muscle tone (cataplexy), hypnagogic hallucinations and sleep paralysis.” The Merck Manual of Diagnosis and Therapy 1412 (13th ed. 1977). Cataplexy, from which Winans also suffers, is “a condition in which there are abrupt attacks of muscular weakness and hypotonia triggered by an emotional stimulus such as mirth, anger, fear, or surprise. It is often associated with narcolepsy.” Dorland’s Illustrated Medical Dictionary 228 (26th ed. 1985).

Winans’ application was denied initially and upon reconsideration. Winans requested a hearing before an Administrative Law Judge (“AU”). On March 11, 1985, the AU found that Winans was not disabled. The AU concluded from medical records and testimony that Winans “does not have an impairment or combination of impairments listed in, or medically equal to one listed in Appendix 1, Subpart P, Regulations No. 4.” Excerpt of Record (“ER”) p. 8. This became the final decision of the Secretary when the Appeals Council declined to review the AU’s decision.

Having exhausted his administrative remedies, Winans brought suit in federal district court to obtain judicial review of the Secretary’s final decision. The district court affirmed the Secretary’s decision that Winans was not entitled to disability benefits under the Act. This appeal followed.

STANDARD OF REVIEW

“The role of a court in reviewing the Secretary’s decision is a limited one.” Allen v. Heckler, 749 F.2d 577, 579 (9th Cir. 1985). This court will set aside a denial of benefits only if “the Secretary’s findings are based upon legal error or are not supported by substantial evidence in the record as a whole.” Taylor v. Heckler, 765 F.2d 872, 875 (9th Cir.1985) (citation and empha sis omitted); see Allen, 749 F.2d at 579 (citation omitted).

Substantial evidence means “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Richardson v. Perales, 402 U.S. 389, 401 , 91 S.Ct. 1420, 1427 , 28 L.Ed.2d 842 (1971) (citation omitted). This court must consider the record as a whole, weighing both the evidence that supports and the evidence that detracts from the Secretary’s conclusion. Jones v. Heckler, 760 F.2d 993, 995 (9th Cir.1985). However, we “may not reweigh the evidence, substitute [our] own judgment for the Secretary’s, or give vent to feelings of compassion.” Bowman v. Heckler, 706 F.2d 564, 566 (5th Cir.1983). At the same time, this court must do more than merely rubber-stamp the ALJ’s decision. Garfield v. Schweiker, 732 F.2d 605, 609-610 (7th Cir. 1984).

EVIDENTIARY RECORD

At the time of his administrative hearing, Winans was 56 years old, with a high school education and technical training as an electrician. Winans testified that at the time of the hearing, he had at least three attacks daily which caused an abnormal and almost irresistible urge to sleep. Win-ans usually has a two to three minute warning before an attack and he can either lie down or try to fight it off. If he tries to fight it off, he eventually falls asleep anyway; and when fighting it, Winans has both alterations of awareness and loss of consciousness. Additionally, he is very susceptible to cataplexy attack in which there is a brief loss of muscle control of his head and upper body. Winans also testified that he suffers from related disorders: sleep paralysis, hypnagogic hallucinations, depression and tachycardia.

Winans testified that the narcoleptic symptoms began in 1955. Winans was first diagnosed as having narcolepsy in 1961 when he was evaluated by the Navy Regional Medical Center in San Diego. He began using Dexadrine, a stimulant, but due to an irregular heartbeat (tachycardia) and other developments, he is unable to take Dexadrine in a quantity sufficient to suppress the narcolepsy. The narcolepsy progressed to the point where, according to Winans, he was unable to work for over two years beginning in 1975.

In 1977, Winans underwent evaluation at the Stanford Sleep Disorders Clinic. Merrill M. Mitler, Ph.D., Administrative Director at the Clinic, reported on October 13, 1977, that Winans had a history of excessive daytime sleepiness, cataplexy, sleep paralysis, and hypnagogic hallucinations. The Nocturnal Polysomnogram and a Multiple Sleep Latency test confirmed the diagnosis of narcolepsy without qualification. These tests relied on an occipital electro-encephalogram (EEG), an electro-oculogram (EOG), an electro-myogram (EMG) and an electro-cardiogram (ECG). With the support of the Veterans Administration (“VA”), Winans was able to return to work after rehabilitation in 1978.

Winans began to experience increasing problems in 1981. Cylert was added to his medication, but was reduced because it caused his face and ears to flush. Clinic notes covering treatment from 1981 through August, 1983 indicate continuing medicine adjustment. In July, 1982, Win-ans’ pulse rate increased to 120 beats per minute. In an attempt to decrease his heart rate, Winans was given Dexadrine and Cylert. Several months later, his heart rate was reduced to 100 beats per minute. During this period Winans was sent to the VA hospital in Portland. The VA rated him 100% disabled, but not permanently so. In December, 1982, he was doing better— his blood pressure was 130/88 in the supine position and his pulse was 88.

In January, 1983, Winans noted some chest tightness. Persistent tachycardia was noted in February, 1983. In June, 1983, his tachycardia had improved with a gradual decrease in Cylert, and Winans continued treatment with Altvan, Inderal and Dexadrine. He had ankle swelling and bruising due to a cataplexy attack. His blood pressure elevated to 170/100 and he was prescribed the sedative Valium. An August, 1988 chart note indicated intermittent tachycardia although it was better with decreased Cylert. His narcolepsy was no worse since the decrease in Cylert and he was “actually doing well now.” He was “sleeping a lot — bad days only up 5 hours/days.”

Dr. John A. Melson has been Winans’ treating physician since 1973. In a January 5, 1984 report, he indicated that Win-ans was having between four and five narcoleptic attacks per day which were mixtures of both cataplexy and/or sleep attacks. Dr. Melson noted that according to Winans’ wife, he had been sleeping more than he was awake. Dr. Melson related Winans’ difficulty with medication and tachycardia and observed that Winans, “like many patients, is unable to take Dexadrine in a dosage required to suppress narcoleptic symptoms. In his case, this has never been very effective in blocking cataplexy anyway.” In July, 1984, Dr. Melson concluded that the frequent occurrence of both narcoleptic attacks and cataplexy made Winans unable to engage in substantial gainful activity. Winans is extremely limited in his ability to drive, falls asleep watching television, has given up most activities and spends an inordinate amount of time in bed.

Winans last worked in 1982. He quit working on the recommendation of Dr. Melson. Winans testified that he was injured on the job in April, 1981, left a lot of work undone because of sleeping at work, and was worried about a cataplexy attack while working above ground level.

On May 7, 1985, Dr. Melson reported that, in his medical opinion, Winans has a “degree of disability almost identical to that which would be engendered if his attacks were epileptic in origin rather than narcoleptic.”

The Secretary’s “examining” physicians, Drs. Ford and Reubens, agreed that Win-ans had narcolepsy, but were of the opinion that Winans’ impairment did not meet or equal any listed impairment. It is significant to note that these “examining” physicians did not in fact personally examine Winans, but merely reviewed the documentary evidence of his affliction.

THE ALJ’S OPINION

Despite this uncontradicted evidence of the nature and degree of Winans’ condition, the AU found him not disabled. From the analysis and findings of the AU, it is apparent that the basis of his ruling is that Winans’ narcolepsy did not meet or equal in severity the requirements of the “Listing of Impairments.” The AU held that “the medical evidence establishes that the claimant has severe narcolepsy, but that he does not have an impairment or combination of impairments listed in, or medically equal to one listed in Appendix 1, Subpart P, Regulations No. 4.” ER p. 8.

DISCUSSION

The Act provides that certain individuals who are “under a disability” shall receive disability benefits. 42 U.S.C. § 423 (a)(1)(D). The Act defines disability as the “inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment which ... can be expected to last for a continuous period of not less than 12 months.” 42 U.S.C. § 423 (d)(1)(A). The Act further provides that a claimant will be found disabled only if his impairment(s) “are of such severity that he is not only unable to do his previous work but cannot, considering his age, education and work experience, engage in any other kind of substantial gainful work which exists in the national economy....” 42 U.S.C. § 423 (d)(2)(A).

Since 1978, the Secretary has implemented this definition of disability by applying a five-step sequential inquiry into whether or not an applicant for benefits should be considered “disabled” and thus eligible for benefits. 20 C.F.R. § 404.1520 (1986). Known as sequential evaluation, this system embodies elements of both the medical and vocational requirements of the definition of disability. It consists of a series of five questions asked in a specific order. The sequential evaluation is continued until a question is answered affirmatively or negatively in such a way that a decision can be made that a claimant is either disabled or not disabled. 20 C.F.R. § 404.1520 (b)-(f).

The third question asks whether the claimant’s impairment meets or equals the medical listings contained in 20 C.F.R. § 404 , Subpart P, Appendix 1. If a claimant meets the listing or his condition is the medical equivalency, he is presumptively disabled. 20 C.F.R. § 404.1520 (d). Narcolepsy has been determined to be the statutory equivalent of the listed impairment of epilepsy. Boeltz v. Bowen, 648 F.Supp. 753 (W.D.Pa.1986); Collord v. Heckler, 633 F.Supp. 902 (N.D.Ill.1986); Stout v. Heckler, 579 F.Supp. 237 (D.Idaho 1984).

Winans does not contend that he suffers from a listed impairment. He argues that he suffers an impairment equivalent to the listed impairment of epilepsy. That he does is his treating physician’s opinion. The examining physicians, Drs. Ford and Reuben, selected by the Secretary as required by the Regulations, agree tliat Win-ans has narcolepsy. Their opinions differ only as to the degree of impairment.

This case is not one of contradictory evidence of a physical impairment, but one in which a treating physician’s opinion on the ultimate question of the degree of impairment differs from that of the examining physicians. Dr. Ford’s opinion that Win-ans’ impairment did not meet or equal any listed impairment and Dr. Reuben’s opinion that Winans could perform non-hazardous work, the AU apparently found, were more convincing than Dr. Melson’s opinion that Winans’ degree of impairment is almost identical to that which would be engendered if his attacks were epileptic in origin.

The general rule is that conflicts in the evidence are to be resolved by the Secretary and that his determination must be upheld when the evidence is susceptible to one or more rational interpretations. Sprague v. Bowen, 812 F.2d 1226, 1230 (9th Cir.1987). “But when the conflict is between the opinions of a treating physician and an examining physician it is the rule in this circuit that ‘[i]f the AU wishes to disregard the opinion of the treating physician, he ... must make findings setting forth specific, legitimate reasons for doing so that are based on substantial evidence in the record.’ ” Id. (quoting Murray v. Heckler, 722 F.2d 499, 502 (9th Cir. 1983)). The rationale behind affording the treating physician’s opinion greater weight is that “he is employed to cure and has a greater opportunity to know and observe the patient as an individual.” Id. (citations omitted).

In the case at hand, the AU concluded that Winans’ condition does not meet or equal a listed impairment. This conclusion is not supported by substantial evidence because no specific reasons were given for disregarding Dr. Melson’s opinion to the contrary. We find nothing in the AU’s decision which indicates why Dr. Melson’s medical findings, reports, and opinion were disregarded. Because the AU did not state reasons based on substantial evidence, we reverse the decision to deny benefits. Id. at 1232.

CONCLUSION

The decision whether to remand the case for additional evidence or simply to award benefits is within our discretion. Id. We. are convinced that substantial evidence does not support the Secretary’s decision. We therefore REVERSE and REMAND for payment of benefits.

Winans’ request for attorney’s fees is DENIED. See Albrecht v. Heckler, 765 F.2d 914, 916 (9th Cir.1985).

. More than 50% of the patients with narcolepsy seen at the Clinic are totally disabled (in medical terms) by age 40. Stout v. Heckler, 579 F.Supp. 237 , 238 n. 1 (Idaho 1984). . The Act's requirements regarding insured status and retirement age are not at issue here. See 42 U.S.C. § 423 (a)(1).