OPINION
GARTH, Circuit Judge.
Elizabeth Ramirez appeals from an Order of the United States District Court for the Eastern District of Pennsylvania affirming the decision of the Commissioner of Social Security. The Commissioner had denied Ramirez’s claims for Supplemental Security Income (“SSI”) under Title XVI of the Social Security Act (the “Act”). 42 U.S.C. §§ 1381 — 1383(f). On appeal, Ramirez contends, as she did in the District Court, that the hypothetical question posed by the Administrative Law Judge (“ALJ”) to the vocational expert failed to accurately convey all of Ramirez’s limitations, and that the Commissioner’s ensuing decision is, therefore, not supported by substantial evidence. The District Court affirmed the ALJ’s decision and Ramirez timely appealed to this Court. For the reasons set. forth below, we vacate the District Court’s order and remand for further proceedings before the ALJ.
I.
A.
Early Procedural History
Ramirez is a 47-year-old divorced mother of two children with no significant work experience. In August 1994, she applied for disability insurance benefits and SSI under the Act, claiming that she was disabled by asthma, bad nerves and a thyroid condition. Six months later, she began receiving mental health treatment for an anxiety disorder.
After the Commissioner denied Ramirez’s application initially and on reconsideration, she requested a hearing before an ALJ. At Ramirez’s request, the ALJ dismissed Ramirez’s claim for disability benefits. The ALJ denied Ramirez’s re
maining claim for SSI, finding that Ramirez was not disabled by her physical or mental impairments. The SSA Appeals Council subsequently vacated the ALJ’s decision and remanded with instructions to explain certain findings and take additional evidence on Ramirez’s mental impairments.
B.
Ramirez’s Mental Health Treatment/Evaluation
Ramirez first sought mental health treatment in February 1995 from Dr. H.H. Park, who diagnosed her with generalized anxiety disorder with depression. Over the next twenty-two months, Dr. Park prescribed various medications with varying degrees of success.
In December 1996, Ramirez stopped seeing Dr. Park. She did not resume mental health treatment until September 1998. At the request of her attorney, however, she underwent a comprehensive psychological evaluation by Dr. Craig Weiss in April 1997. Dr. Weiss concluded that Ramirez had an “Anxiety Disorder ... with significant symptoms of depression, social phobia, obsessive-compulsive, and mood ineon-gruent hallucinations.” (Administrative Record (“A.R.”) at 303.) With respect to “functional limitations,” Dr. Weiss opined that Ramirez (1) had moderate restriction in activities of daily living, (2) had marked to extreme difficulties in maintaining social functioning, (3) experienced frequent deficiencies of concentration, and (4) continually experienced episodes of deterioration.
Almost two years earlier, in 1995, Dr. Louis Poloni, a state agency psychologist, had completed a Psychiatric Review Technique Form (“PRTF”) on which he had also assessed Ramirez’s mental impairments in four broad areas of mental functioning. Dr. Poloni concluded that Ramirez: (1) had no restriction in activities of daily living, (2) had slight difficulties in maintaining social functioning, (3) often experienced deficiencies of concentration, persistence, or pace, and (4) never experienced episodes of deterioration or decom-pensation in work-like settings. Based on those findings, Dr. Poloni concluded that Ramirez had a severe anxiety disorder, but that Ramirez’s condition did not meet or equal any of the mental impairments deemed by the SSA to be presumptively disabling. Consequently, Dr. Poloni proceeded to complete a Mental Residual Functional Capacity (“MRFC”) form, which is meant to assess a claimant’s ability to perform either the claimant’s previous work or other work in the national economy. Dr. Poloni determined that Ramirez could perform simple, routine unskilled work.
Ramirez’s mental functioning was also in 1998 assessed by Dr. Herman Rudnick, a Board certified psychiatrist. Dr. Rudnick concluded that Ramirez suffered from anxiety-related and personality disorders. As to the four broad areas of mental functioning, Dr. Rudnick found that Ramirez (1) had only moderately limited daily activities, (2) had moderately limited social functioning, (3) often experienced deficiencies of concentration, persistence, or pace, (4) and did not experience any episodes of deterioration or decompensation. Like Dr. Poloni before him, Dr. Rudnick found that Ramirez’s mental impairments did not meet or equal the criteria of a listed impairment. As to Ramirez’s residual functional capacity, Dr. Rudnick opined that Ramirez could not perform complex or complicated work and would need to be able to contact her home from work, but that there was no need to limit Ramirez’s interaction with the public or with coworkers.
C.
Procedural History After Remand
Following the remand by the SSA Appeals Council, the ALJ held a second hear
ing. The ALJ considered, among other things, all of the evidence described above concerning Ramirez’s mental impairments. Near the conclusion of the second hearing on remand, the ALJ posed the following hypothetical question to vocational expert Julie Stratton:
I will begin by asking you to assume that we’re talking about an individual of Ms. Ramirez’s age, education and prior work history. And I’d like you to further assume that this individual’s capable of performing a range of sedentary work. The work should be performed in a well ventilated facility, with no exposure to dust, fumes, pets, animals, chemicals, or temperature extremes. The work should provide for occasional breaks, for the individual use of an inhaler or pump. The work should involve simple one to two step tasks. The work should not require the individual during the course of performing the work to travel outside of the workplace. And ... the work setting should provide reasonable opportunity for the individual to make and receive personal phone calls. Within the boundaries of these limitations, ... are there jobs in the regional or national economy that the individual could perform?
(A.R. at 457-58.) It is significant that neither Dr. Weiss’s evaluation nor his conclusions found expression in the hypothetical.
The purpose of the hypothetical was to assess Ramirez’s residual functional capacity. Stratton responded that, notwithstanding the limitations contained in the hypothetical, there were several jobs in the local and national economy that the hypothetical claimant could perform, including assembler, hand packer, and production inspector.
The ALJ then issued a written decision in which she determined that Ramirez was not disabled and therefore she again denied Ramirez’s application for SSL In reaching her decision, the ALJ relied heavily on Stratton’s testimony.
The ALJ attached to her written decision a completed PRTF, as she was required to do under the then-existing Social Security Regulations. In a section of the PRTF entitled, “Functional Limitation and Degree of Limitation,” the ALJ noted that Ramirez “often” experienced “deficiencies of concentration, persistence, or pace resulting in a failure to complete tasks in a timely manner (in work settings or elsewhere).” (Appendix at 93.) The ALJ’s hypothetical, however, had not mentioned this particular limitation.
Ramirez appealed the ALJ’s decision, but the Appeals Council declined to grant review. Ramirez then challenged the ALJ’s ruling in the United States District Court for the District of Eastern Pennsylvania, where Ramirez and the Commissioner filed cross-motions for summary judgment. The District Court referred the cross-motions to a Magistrate Judge for a Report and Recommendation.
Although Ramirez made a number of claims, one of her primary arguments was that the ALJ had failed to include in her hypothetical the finding she had made on the PRTF that Ramirez often suffered deficiencies in concentration, persistence, or pace. The Magistrate Judge rejected all of Ramirez’s arguments except for the PRTF argument, finding that it was “not clear whether a limitation in concentration, persistence or pace within the hypothetical ... would have changed the vocational expert’s response.” The Magistrate Judge recommended that the District Court remand to the ALJ to allow for further testimony by a vocational expert.
The Commissioner objected only to the Magistrate Judge’s finding that the ALJ’s hypothetical was deficient. The District Court adopted those portions of the Magis
trate’s Report to which no objection was raised, but disapproved that portion of the Magistrate Judge’s Report which found the ALJ’s hypothetical to be defective. The District Court explained that “the standards articulated by the Third Circuit do not mandate that the ALJ articulate verbatim to the vocational expert the findings recorded on the PRTF form” and proceeded to find that the limitations the ALJ had included in her hypothetical “accurately reflect[ed] the evidence contained in the record.” (Appendix at 20, 23.) Accordingly, the District Court entered summary judgment for the Commissioner.
Ramirez now appeals from the District Court’s Order. Ramirez contends that (a) the ALJ’s hypothetical did not adequately incorporate the PRTF finding concerning Ramirez’s deficiencies in concentration, persistence, or pace; and (b) the ALJ improperly made adverse credibility findings.
II.
We have jurisdiction to hear this appeal pursuant to 28 U.S.C. § 1291. We may reverse the District Court’s grant of summary judgment to the Commissioner only if the ALJ’s findings were not supported by “substantial evidence.”
Burns v. Barnhart,
312 F.3d 113, 118 (3d Cir.2002). If, however, an ALJ poses a hypothetical question to a vocational expert that fails to reflect “all of a claimant’s impairments that are supported by the record[,] ... it cannot be considered substantial evidence.”
Chrupcala v. Heckler,
829 F.2d 1269, 1276 (3d Cir.1987).
III.
We consider first Ramirez’s argument that the ALJ’s hypothetical was deficient.
A.
Administrative Framework
Under the Act, a person who has a “disability” is entitled to SSI payments from the Social Security Administration (“SSA”). 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 result in death or which has lasted or can be expected to last for a continuous period of not less than 12 months.” 42 U.S.C. § 423(d)(1)(A). The definition is qualified, however, as follows:
An individual shall be determined to be under a disability only if his physical or mental impairment or impairment 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). A person seeking SSI may allege that they have disabling physical impairments, mental impairments, or both.
Acting pursuant to its rulemaking authority, the SSA has promulgated regulations establishing a five-step sequential evaluation process to determine if a claimant has a disability. 20 C.F.R. § 404.1520 (2003). At step one, the SSA will find that a claimant is not disabled unless he demonstrates that he is not working at a “substantial gainful activity.” 20 C.F.R. §§ 404.1520(b), 416.920(b). At step two, the SSA will find no disability unless the claimant shows that he has a “severe impairment,” defined as “any impairment or combination of impairments which significantly limits [the claimant’s] physical or mental ability to do basic work activities.” 20 C.F.R. §§ 404.1520(c), 416.920(c). If the claimant successfully demonstrates that he has a severe impairment, the SSA determines at step three whether the impairment is on a list of impairments presumed severe enough by the SSA to render one dis
abled; if so, the claimant qualifies. 20 C.F.R. §§ 404.1520(d), 416.920(d). If, however, the claimant’s impairment is not on the list, the inquiry proceeds to step four and the SSA assesses whether the claimant has the “residual functional capacity” to perform his previous work. Unless he shows that he cannot, he is determined not to be disabled.
If the claimant survives step four, the fifth step requires the SSA to consider “vocational factors” (the claimant’s age, education, and past work experience) and to determine whether the claimant is capable of performing other jobs existing in significant numbers in the national economy. 20 C.F.R. §§ 404.1520(f), 404.1560(c), 416.920(f), 416.960(c). Unlike some of the earlier stages in the evaluation process, the burden of proof at step five is on the agency. Here, Ramirez was found able to perform other jobs.
B.
Evaluating Mental Impairments
In 1985, the SSA issued revised regulations to evaluate individuals who claimed to be disabled as a result of a mental impairment. 20 C.F.R. § 416.920a (1999). The revised regulations implemented a new technique that required the SSA to evaluate a claimant’s mental impairments in four broad areas of functioning: (1) activities of daily living, (2) social functioning, (3) concentration, persistence, or pace, and (4) deterioration or decompensation in work or work-like settings. 20 C.F.R. § 416.920a (1999). The third functional area-which is of the most concern in this appeal-was rated on a five point scale of never, seldom, often, frequent, and constant. 20 C.F.R. § 416.920a(b)(3) (1999).
Under the revised regulations, the adjudicator applied the new technique by completing a form known as the Psychiatric Review Technique Form (PRTF). 20 C.F.R. § 416.920a(d) (1999). Based on the adjudicator’s findings on the PRTF with respect to these four areas of functioning, he would determine at step two of the sequential evaluation process if the claimant had a “severe mental impairment.” 20 C.F.R. § 416.920a(c)(l) (1999). If so, the adjudicator would proceed to the third step and determine if the claimant’s impairment met or equaled one of the impairments found on the list of impairments presumed severe enough to render a person disabled. 20 C.F.R. § 416.920a(c)(2) (1999).
In 1996, the SSA issued Social Security Ruling 96-8p “[t]o state the [SSAJ’s policies and policy interpretations regarding the assessment of residual functional capacity (RFC) in initial claims for disability benefits under [the Act].” Ruling 96-8p discussed the PRTF and the role it plays in the five-step analysis:
The psychiatric review technique described in 20 CFR 404.1520a and 416.920a and summarized on the Psychiatric Review Technique Form (PRTF) requires adjudicators to assess an individual’s limitations and restrictions from a mental impairment(s) in categories identified in the “paragraph B” and “paragraph C” criteria of the adult mental disorders listings. The adjudicator must remember that the limitations identified in the “paragraph B” and “paragraph C” criteria are not an RFC assessment but are used to rate the severity of mental impairment(s) at steps 2 and 3 of the sequential evaluation process. The mental RFC assessment used at steps 4 and 5 of the sequential evaluation process requires a more detailed assessment by itemizing
various functions contained in the broad categories found in paragraphs B and C of the adult mental disorders listings in 12.00 of the Listing of Impairments, and summarized on the PRTF.
SSR 96-8p (July 2,1996).
C.
The ALJ’s Hypothetical Did Not Adequately Convey Ramirez’s Limitations
As we have previously noted, the ALJ who reviewed Ramirez’s application noted on the PRTF that Ramirez “often” suffers from “deficiencies of concentration, persistence, or pace resulting in a failure to complete tasks in a timely manner (in work settings or elsewhere).” (Appendix at 93.) Ramirez argues on appeal that the ALJ erred by not including this limitation in the hypothetical that she posed to the vocational expert. As we explain below, we agree with Ramirez and hold that the hypothetical did not accurately convey all of Ramirez’s impairments, and the limitations they cause, and therefore the ALJ’s decision is not supported by substantial evidence.
We have not previously decided the precise issue of whether certain findings on a PRTF must be included in an ALJ’s hypothetical. We have, however, stated in the clearest of terms that an ALJ’s hypothetical must include all of a claimant’s impairments. For example, in
Chrupcala v. Heckler,
829 F.2d 1269, 1276 (3d Cir.1987), the claimant argued that the vocational expert’s opinion was deficient because it failed to account for all of the claimant’s impairments. We agreed, noting that the ALJ’s hypothetical question “did not reflect the fact of constant and severe pain which [the claimant] testified to and which we have explained was supported by objective medical findings in the record.”
Id.
We explained that “[a] hypothetical question must reflect
all of
a claimant’s impairments that are supported by the record; otherwise the question is deficient and the expert’s answer to it cannot be considered substantial evidence.”
Id.
(citing
Podedworny v. Harris,
745 F.2d 210 (3d Cir.1984) and
Wallace v. Secretary,
722 F.2d 1150 (3d Cir.1983)) (emphasis added).
We recently reaffirmed this principle in
Burns v. Barnhart,
312 F.3d 113, 122 (3d Cir.2002). There, the claimant argued that the vocational expert’s testimony did not provide substantial evidence because the ALJ’s hypothetical questions regarding the claimant’s residual functional capacity failed to incorporate the claimant’s borderline intellectual functioning. The Commissioner argued there, as it does here, that the ALJ’s use of “simple repetitive one, two-step tasks” in the hypothetical was sufficiently descriptive to encompass the findings concerning the claimant’s limited intellectual functioning. We disagreed, however, explaining that the reference to simple tasks did not “specifically convey” the claimant’s intellectual limitations and that “greater specificity” was required.
Id.
at 123.
Although we have not previously held whether findings on a PRTF about a claimant’s concentration, persistence, or pace must be included in an ALJ’s hypothetical, some of our sister Circuits have dealt with this issue. For example, in
Howard v. Massanari,
255 F.3d 577 (8th Cir.2001), the claimant argued on appeal that the ALJ had failed to convey in his hypothetical the finding that the claimant
often suffered from deficiencies in concentration, persistence, or pace.
Id.
at 581. The ALJ had, however, asked the vocational expert to assume that the claimant would be capable of performing simple, routine, repetitive tasks.
Id.
The Eighth Circuit Court of Appeals held that the hypothetical “adequately capture[d]” the claimant’s deficiencies, in part because the state psychologist who had made the finding also prepared a “functional capacity assessment” in which he concluded that the claimant could “sustain sufficient concentration and attention to perform at least simple, repetitive, and routine cognitive activity without severe restriction of function.”
Id.
at 582.
But see Newton v. Chater,
92 F.3d 688 (8th Cir.1996),
infra.
The Sixth Circuit Court of Appeals reached a similar outcome in
Smith v. Halter,
307 F.3d 377 (6th Cir.2001). In
Smith,
the ALJ also found that the claimant often suffered from deficiencies in concentration, persistence, or pace. In his hypothetical, the ALJ instructed the vocational expert to assume the claimant had mental impairments limiting him to jobs that were routine and low stress, and did not involve intense interpersonal confrontations, high quotas, unprotected heights, or operation of dangerous machinery.
Id.
at 378. On appeal, the Sixth Circuit concluded that the ALJ “went beyond” the simple findings included in the PRTF and determined that the claimant’s concentration problems were minimal or negligible, and then “translated [the claimant’s] condition into the only concrete restrictions available to him ... and duly incorporated them into his hypothetical.... ”
Id.
at 379.
Other Courts of Appeal (or in one case a different panel of the same Circuit) have been less forgiving. For example, in
Kasarsky v. Barnhart,
335 F.3d 539 (7th Cir.2003), the ALJ noted on the PRTF that the claimant frequently suffered from deficiencies in concentration, persistence, or pace. In his hypothetical, the ALJ mentioned that the claimant’s borderline intelligence seriously limited, but did not preclude him from, understanding, remembering, and carrying out detailed instructions.
Id.
at 544. On appeal, the Seventh Circuit stated that it saw “nothing in the hypothetical that takes into account the ALJ’s own earlier observation ... that [the claimant] suffered from
frequent
deficiencies of concentration, persistence, or pace.”
Id.
The court acknowledged that there might be an explanation for the omission, but it explained that it had “no way of knowing that.”
Id.
Accordingly, it remanded the case for further proceedings.
In
Newton v. Chater,
92 F.3d 688 (8th Cir.1996), the court found that a hypothetical was defective because it failed to adequately convey the claimant’s deficiencies in concentration, persistence, or pace. The ALJ’s hypothetical had described a person with a minimal ability to read and write, a borderline range of intelligence, a ninth or tenth grade education, an inability to perform highly skilled or technical work, a capacity for simple jobs, and a demonstrated ability to control his drinking problem.
Id.
at 694. The Commissioner argued that the deficiencies of concentration, persistence or pace did not have to be included in the hypothetical because the ALJ had limited the hypothetical to simple jobs and two medical professionals had testified that the claimant’s deficiencies did not significantly limit his abilities to follow short and simple instructions and make simple work-related decisions.
Id.
at 695. The court disagreed, noting that the vocational expert had testified on cross-examination that the claimant’s concentration and persistence deficiencies related to basic work habits needed to maintain employment. Thus,
the court remanded for further proceedings.
Turning to the instant appeal, the ALJ asked vocational expert Stratton to assume a hypothetical individual with Ramirez’s background and the following limitations: sedentary work in a well-ventilated environment, with no exposure to dust, fumes, pets, animals, chemicals, or temperature extremes; occasional breaks necessary for the use of an inhaler or pump; no more than simple one or two-step tasks; no travel outside the workplace; and a reasonable opportunity to receive and make personal telephone calls.
We are not satisfied that these limitations take into account the ALJ’s own observation (both in her opinion and in the PRTF) that Ramirez
often
suffered from deficiencies in concentration, persistence, or pace. The first several limitations that the ALJ included in her hypothetical pertain to Ramirez’s physical impairments and therefore have no bearing on her mental impairments. The only limitations that relate to Ramirez’s mental impairments are the limitations to simple tasks, the restriction on travel, and the phone privileges.
These limitations do not adequately convey all of Ramirez’s limitations. The Commissioner contends that the limitation to one to two step tasks is sufficient, but we agree with the Magistrate Judge that a “a requirement that a job be limited to one to two step tasks, as was stated in the hypothetical relied upon by the ALJ, does not adequately encompass a finding that [Ramirez] ‘often’ has ‘deficiencies in concentration, persistence, or pace,’ as was noted by the ALJ both in her decision and on the PRTF attached to the decision.” (Appendix at 72.) Most importantly, this limitation does not take into account deficiencies in pace. Many employers require a certain output level from their employees over a given amount of time, and an individual with deficiencies in pace might be able to perform simple tasks, but not over an extended period of time. If Ramirez often suffers deficiencies in pace and this had been included in the hypothetical, vocational expert Stratton may have changed her answer as to whether there were jobs in the local or national economy that Ramirez could perform. In fact, the vocational expert testified that each of the jobs suitable for Ramirez (assembler, packer, and inspector) would have daily production quotas and that Ramirez would have to maintain a certain degree of pace to maintain those jobs.
This omission from the hypothetical runs afoul of our directive in
Chrupcala
that a “hypothetical question posed to a vocational expert ‘must reflect
all
of a claimant’s impairments,”
Chrupcala,
829 F.2d at 1276, as well as our statement in
Bums
that “great specificity” is required when an ALJ incorporates a claimant’s mental or
physical limitations into a hypothetical.
Bums,
312 F.3d at 122. Indeed, the SSA’s own ruling requires a “more detailed assessment” of the claimant’s mental limitations at step five of the disability analysis.
See
SSR 96-8p (July 2,1996).
Of course, there may be a valid explanation for this omission from the ALJ’s hypothetical. For example, the ALJ may have concluded that the deficiency in pace was so minimal or negligible that, even though Ramirez “often” suffered from this deficiency, it would not limit her ability to perform simple tasks under a production quota. The record, however, would seem to suggest otherwise. At the second hearing, Dr. Rudnick — upon whose testimony the ALJ relied — was asked the following question: “What happens to [Ramirez’s] ability to handle pace, for example, in a work situation, where there’s a certain amount of work that has to be done in an eight hours or two hours or whatever segment?” (A.R. at 451.) Although the ALJ briefly interceded before Dr. Rudnick could answer, Dr. Rudick eventually replied that Ramirez’s ability to maintain a full-time job depended primarily on “the proximity to where her children would be” because Ramirez’s anxiety-disorder is in large part attributable to her “need to feel that she has to be reasonably protective of her children.” While this might lead a neutral observer to conclude that Ramirez’s deficiencies in pace could be overcome by finding a job close to her children, the ALJ did not include this limitation in her hypothetical. Instead, the ALJ provided only for a reasonable number of personal phone calls. If this accommodation would not remedy Ramirez’s deficiency in concentration and pace, the vocational expert might have given a different answer to the hypothetical.
Relying on Social Security Ruling 96-8p, which we reproduced in part earlier in this opinion, the Commissioner contends that the “PRTF does not document specific functional limitations for RFC purposes, bur rather assesses functional loss from a claimant’s mental impairments only with respect to broad areas of functioning.” In other words, the Commissioner argues that the PRTF findings are relevant only in steps two and three of the sequential evaluation process, before any assessment of a claimant’s residual functional capacity is made.
We cannot concur in the Commissioner’s evaluation of the PRTF findings. While SSR 96-8p does state that the PRTF findings are “not an RFC assessment” and that step four requires a “more detailed assessment,” it does not follow that the findings on the PRTF play no role in steps four and five, and SSR 96-8p contains no such prohibition.
In conclusion, we hold that the ALJ’s hypothetical did not adequately capture and recite all of Ramirez’s mental impairments and the limitations caused by those impairments. In reaching that holding, one factor we cannot ignore is that the burden shifts to the Commissioner at step five to prove that the claimant can perform a job that exists in the national economy.
See Bums,
312 F.3d at 119 (“At the final step-step five-the burden shifts to the Commissioner to show that the claimant can perform ‘other work.’ ”).
IV.
We have considered Ramirez’s remaining arguments and, after reviewing the record, are persuaded that they are without merit.
V.
For the foregoing reasons, we will vacate the Order of the District Court and remand to the District Court so that it can in turn remand to the Commissioner for
further proceedings consistent with this Opinion.