MANION, Circuit Judge.
Louquetta O’Connor-Spinner, who is 47, suffers from depression and several physical impairments. Several times since 2001 she has applied for Disability Insurance Benefits and Supplemental Security Income, and six years ago we invalidated the Social Security Administration’s denial of her 2004 request for benefits.
O’Connor-Spinner v. Astrue,
627 F.3d 614 (7th Cir. 2010). We concluded that the assigned administrative law judge had committed two errors relating to O’Connor-Spinner’s depression. First, the ALJ had not asked a testifying vocational expert to assess how O’Connor-Spinner’s employment prospects would be affected by her moderate limitation on concentration, persistence, and pace. And, second, the ALJ had ignored a psychologist’s opinion that O’Connor-Spin-ner also faces a moderate limitation on her ability to accept instructions from, and respond appropriately to, supervisors. We instructed the Agency to remedy these mistakes, but instead of complying with this simple directive, a different ALJ contradicted his colleague and declared that O’Connor-Spinner’s depression is not, and never has been,. a severe impairment. O’Connor-Spinner again has sought judicial review, and she argues that the medical evidence contradicts this assertion. We agree, and once more we must remand this case to the Agency for further proceedings.
I. Background
As we did in our previous opinion, we omit discussion of O’Connor-Spinner’s physical impairments because the parties agree that those afflictions — degenerative disk disease, bilateral carpal tunnel syndrome, sleep apnea, “restrictive lung disease,” and obesity — are severe and would themselves limit O’Connor-Spinner to sedentary work. The effect of O’Connor-Spin-ner’s depression on her ability to maintain employment is, as before, the sole issue on appeal.
Evidence of O’Connor-Spinner’s depression predates the December 2003 onset date that she alleged when applying for benefits in 2004. Doctors treating her physical ailments had described O’Connor-Spinner as appearing depressed, noted a history of depression, and catalogued prescriptions for antidepressants. O’Connor-
Spinner had visited a community mental-health center three times during 2002. Progress notes document her reports of recurring agitation, impulsivity, fatigue, and crying spells, as well as two or three “explosive episodes” weekly involving violent behavior and memory blackouts. O’Connor-Spinner also had recounted suicide attempts and difficulty managing anger (the latter prompting a court to grant a neighbor’s request for a restraining order). Clinicians had diagnosed depression and an “adjustment disorder with mixed disturbance of emotions and conduct.” This type of adjustment disorder is marked by “clinically significant emotional or behavioral symptoms” of both depression and conduct disturbances such as “violation of the rights of others or of major age-appropriate social norms and rules.” Am. Psyohi-ATRio Ass’n, Diagnostic & Statistioal Manual of Mental DISORDERS 679-80 (4th ed. text revision 2000). O’Connor-Spinner was prescribed the antidepressants Elavil and Prozac. During these visits her Global Assessment of Functioning
had been scored at 50, 52, and 55; a score of 50 indicates serious symptoms or “serious impairment in social, occupational, or school functioning,” while scores between 51 and 60 indicate moderate functional difficulty or symptoms.
Id.
at 34;
see also Price v. Colvin,
794 F.3d 836, 839 (7th Cir. 2015). Even so, a state-agency psychologist who reviewed O’Connor-Spinner’s medical records in October 2002 (in connection with her 2001 application for benefits) had opined that she didn’t have a severe mental impairment at that time.
No mention of psychological counseling appears in O’Connor-Spinner’s medical records for the three years following that psychologist’s review, but medical providers treating her physical ailments in 2004 and 2005 alluded to her history of depression and noted current prescriptions for antidepressants. And in May 2004 a different state-agency psychologist, Dr. Kamla Paul, examined O’Connor-Spinner and specifically diagnosed a depressive disorder. Dr. Paul noted many of the same symptoms documented in 2002, including crying fits and violent outbursts — in one of the latter, O’Connor-Spinner had threatened her husband with a butcher knife. Although Dr. Paul rated as adequate many of O’Connor-Spinner’s abilities, including abstract thinking and long-term memory, the doctor gauged her general knowledge and short-term memory as poor and remarked that “she gets confused.” Dr. Paul noted that O’Connor-Spinner was dys-phoric and showed a flat affect, and assigned her a GAF of 55.
Later that same month, a third state-agency psychologist, Dr. Donna Unversaw, reviewed the medical file. She opined that O’Connor-Spinner’s depression caused a moderate limitation on concentration, persistence, and pace, although she asserted (without explanation) that this limitation would not prevent O’Connor-Spinner from performing moderately complex tasks. Dr. Unversaw also .opined that O’Connor-Spin-ner had moderate difficulty accepting instructions from supervisors and responding appropriately to their criticisms.
At O’Connor-Spinner’s hearing before the first ALJ in January 2006, she testified that she thinks frequently of suicide, sleeps for days at a time, rarely leaves home, and skips social activities that she once enjoyed. A vocational expert, re
sponding to a “hypothetical” from the ALJ listing O’Connor-Spinner’s physical limitations but not her limitation on concentration, persistence, and pace or her difficulty relating to supervisors, testified that she could perform several jobs existing in significant numbers in the national economy. In a written decision issued in June 2006, the ALJ concluded at Step 2 of the 5-step analysis,
see
20 C.F.R. §§ 404.1520(a)(4)(ii), 416.920(a)(4)(ii), that O’Connor-Spinner’s depression was severe (as were her physical ailments). The ALJ found that she was not disabled, however, reasoning at Step 5 that she retained the ability to perform the simple, sedentary jobs identified by the vocational expert. The Appeals Council declined review.
O’Connor-Spinner sought judicial review,
see
42 U.S.C. § 405(g), and though a district judge upheld the ALJ’s decision, we agreed with O’Connor-Spinner that the ALJ should have included her moderate limitation on concentration, persistence, and pace among the limitations given to the vocational expert.
See O’Connor-Spinner,
627 F.3d at 618-21. In reaching this conclusion, we specifically rejected the Commissioner’s argument that the ALJ necessarily had incorporated this limitation by restricting O’Connor-Spinner to simple, repetitive’ tasks.
Id.
at 621. We also concluded that the ALJ should have addressed Dr. Unversaw’s opinion that O’Connor-Spinner suffered from a moderate limitation on her ability to accept instructions and criticism from supervisors since, as Social Security Administration regulations make clear, “even a moderate limitation on responding appropriately to supervisors may undermine seriously a claimant’s ability to work.”
Id.
(citing 20 C.F.R. § 404.1545(c); SSR 85-15). We directed the Agency to remedy these two shortcomings.
Our opinion reversing the district court was issued in November 2010, but for reasons not disclosed in the record, the Social Security Administration did not act on that decision until April 2012. By then more than six years had passed since O’Connor-Spinner’s hearing before the ALJ, and she supplemented the record with evidence accumulated during that period. Records show that in March 2006 — after the hearing but three months before the ALJ’s decision — she had returned to the community mental-health center seeking help with symptoms of depression and anxiety. This time she was diagnosed with “major depression, recurrent severe.” A diagnosis of “major depression” means,
by definition,
that an individual’s “symptoms cause clinically significant distress or impairment in social, occupational, or other important areas of functioning.” Am. Psychiatrio Ass’n,
supra
at 356. O’Connor-Spinner’s GAF score once again was recorded as 50, indicating serious symptoms or functional impairments. The diagnosis of major depression is repeated in records from 2007 and 2008. So are O’Connor-Spinner’s self-reports of suicidal thoughts. One therapist documented an incident in which O’Con-nor-Spinner, in a fit of anger, had discarded all of her medications — prompting her husband to remove his guns from their home “for safety.”
Dr. Robert Kurzhals, a fourth state-agency psychologist, had examined O’Con-nor-Spinner in December 2009 (after she again applied for benefits while her previous appeal was under consideration in this court). O’Connor-Spinner had reported feeling depressed most days and said that about eight times a month she stays home because of her depression. Dr. Kurzhals had observed that O’Connor-Spinner was marginally groomed, slow-speaking, depressed, and “perhaps mildly impaired” in concentration and attention. He also had noted without comment that, when, tested, she could recall only five digits forward and three digits backward (this same re-
suit prompted the earlier examining physician, Dr. Paul, to conclude that O’Connor-Spinner’s short-term memory was “poor”). When asked to count down from one hundred by sevens (a common technique, called “serial sevens,” for assessing concentration and memory), O’Connor-Spin-ner counted on her fingers and slowly responded, “92, 84, 77, 70, 63, 56, 48.” Dr. Kurzhals had written that she “presented as a depressed and somewhat irritable individual,” diagnosed O’Connor-Spinner as bipolar (based on her self-reported history of suffering from the disorder), and assigned her a GAF score of 51.
Also in December 2009, yet another state-agency psychologist, Dr. William Shipley, had reviewed O’Connor-Spinner’s medical records and opined, without an examination, that her mental impairment was not severe. Dr. Shipley checked a box on a form to indicate that O’Connor-Spin-ner suffered from bipolar disorder (based solely on the opinion of Dr. Kurzhals), but he failed to reconcile this conclusion with, or even acknowledge, the diagnoses of depression and, later, major depression from her treatment providers. Neither did he comment on the explicit diagnosis of a “depressive disorder” given by Dr. Paul, the first of the two state-agency psychologists who
examined
O’Connor-Spinner. Dr. Shipley ticked off evidence that, in his view, proved her “social functioning and concentration ... to be intact”: O’Connor-Spinner still was capable ... of attending to her own personal hygiene (although that ability was “marginal,” according to Dr. Kurzhals); she could shop in stores, manage money, and drive; and she enjoyed watching TV. He then questioned O’Con-nor-Spinner’s credibility, and concluded that she was only mildly limited in the three categories used in evaluating the effects of mental illness: activities of daily living; social functioning; and concentration, persistence, and pace. Thus, Dr. Ship-ley had said, it was “reasonable to conclude that the claimant’s condition is not severely limiting at this time.”
Apart from contradicting the existing evidence, Dr. Shipley’s assessment also stands in stark contrast to the conclusions of medical providers who continued to treat O’Connor-Spinner. In March 2010, a few months after Dr. Shipley offered his assessment, a nurse practitioner specializing in pain management had examined O’Connor-Spinner. The nurse practitioner had noted significant signs of depression, remarked on the limited efficacy of O’Con-nor-Spinner’s prescribed antidepressants, and recorded her revelation that more often than not she thinks she’d be better off dead. O’Connor-Spinner continued taking those medications, however, and later in 2010 a clinician at the community mental-health center had written that she “presented as depressed and her affect was constricted,” that her cognition was impaired, and that she “struggled” with both serial sevens and fours. The clinician opined that the effectiveness of O’Connor-Spinner’s medications should be evaluated, repeated the diagnosis of “major depressive disorder, recurrent,” and assigned her a GAF score of 55.
About a year later, near the end of 2011, O’Connor-Spinner told a psychiatrist at the community mental-health center that her medications were working well and her depression had improved. Although the doctor wrote that O’Connor-Spinner seemed happier and more stable, her GAF score remained at 55.
Meanwhile, our simple directive to remedy the first ALJ’s missteps in addressing O’Connor-Spinner’s mental limitations was never carried out. Instead the Agency reassigned the case to a different ALJ, who in June 2012 reopened the evidentiary hearing. The new ALJ posed to a different vocational expert — the only witness — a “hypothetical” that, according to the ALJ,
was “based on the previous ALJ’s finding.” True, the new ALJ’s hypothetical incorporates all of O’Connor-Spinner’s
'physical
restrictions, but it does not mention her limitation on concentration, persistence, and pace, nor does it say anything about difficulties in responding appropriately to supervisors. After the vocational expert had identified several jobs that could be performed by someone with the listed physical restrictions, the ALJ asked if the vocational expert’s opinion would be different if “due to limitations in concentration, persistence, or pace ... this person cannot do any work involving strict production quotas or fast paced.” The vocational expert replied that his opinion would not change because, he explained, he already had excluded jobs which require “constant handling and fingering.” The ALJ then asked if the same jobs would be available to a person who would be “off task” for 10% of the workday. That degree of distraction, the vocational expert replied, would be tolerable for unskilled work. But, he added, at more than 15% the person would be “very unlikely” to maintain employment. O’Connor-Spinner’s attorney then asked if a “moderate” limitation on the “ability to accept instructions and respond appropriately to criticism from supervisors” would further limit the jobs available to someone with O’Connor-Spin-ner’s physical limitations. Counsel defined “moderate” as 15% to 20% below normal, prompting the vocational expert to reply that this limitation would undermine the ability to maintain employment and, if persistent, “would likely cause significant problems, meaning to retain the jobs.”
In a written decision, the new ALJ found that O’Connor-Spinner had not engaged in substantial gainful activity since 2003 (Step 1); that her degenerative disc disease, bilateral carpal tunnel syndrome, sleep apnea, lung disease, and obesity (but
not
her major depression) constituted severe impairments (Step 2); that none of these impairments, alone or in combination, met or equaled a listed impairment (Step 3); that she had the residual functional capacity to perform sedentary work with several physical restrictions (Step 4); and that given her age, education, work experience, and residual functional capacity O’Connor-Spinner could perform jobs which exist in significant numbers and thus wasn’t disabled (Step 5). Although the ALJ had identified the salient issue on remand as “the claimant’s depression and what, if any, work-related limitations it causes,” the bulk of his analysis concerns his
rejection
at Step 2 of depression as a severe impairment.
That analysis acknowledges our previous opinion, but instead of complying with our instructions, the new ALJ asserted that “both errors cited by the court in its review of the case were based on the convoluted assessment by Dr. Unversaw ... that is internally inconsistent, virtually incoherent, and totally unsupported.” The ALJ gave three reasons for excoriating Dr. Unversaw: (1) on a worksheet she had described O’Connor-Spinner’s difficulty with concentration, persistence, and pace as mild but then, in assessing her mental residual functional capacity, concluded that O’Connor-Spinner was moderately limited; (2) the psychologist’s opinion that O’Con-nor-Spinner had difficulty relating to supervisors was communicated by marking a form instead of through narrative; and (3) Dr. Unversaw had not, in the ALJ’s view, supported her findings adequately.
Dr. Unversaw’s evaluation, the ALJ insisted, “stands alone in suggesting severe depression with two moderate limitations.” The ALJ acknowledged that (shortly after her January 2006 hearing) O’Connor-Spin-ner had been diagnosed (with major depression, but he did not mention that the precise diagnosis was “major depression, recurrent severe.” Neither did he explore
the import of the specific diagnosis. The ALJ referred only in passing to Dr. Kur-zhals’s most recent examination of O’Con-nor-Spinner for the state agency, without discussing any of the psychologist’s statements supporting O’Connor-Spinner’s claim of disability. Instead the ALJ credited the opinion issued a few days later by Dr. Shipley, who decided
without examining
O’Connor-Spinner that her depression was not a severe impairment.
In accepting that opinion, the ALJ minimized the import of a decade of records from the community mental-health center. Those records document treatment for, and recurrent symptoms of, O’Connor-Spinner’s major depressive disorder, but the ALJ focused on her latest visit at the end of 2011 when she had reported improvement. And, while blinding himself to documentation of the symptoms she did experience, the ALJ stated that O’Connor-Spinner’s treatment records “provide little or no information about any problems with social functioning or concentration, persistence, and pace” (this despite the fact that clinicians at the mental-health center were tasked with treating O’Connor-Spinner’s depression, not evaluating her ability to work). The ALJ also weighed against O’Connor-Spinner her purported “lack of mental health treatment” between 2002 and 2006, when, in fact, throughout that time she was taking antidepressants that only a medical provider- treating her depression would have prescribed. Moreover, the new ALJ rejected outright the consistent reports from different practitioners— including both of the examining psychologists hired by the Agency — that O’Con-nor-Spinner had GAF scores ranging from 50 to 55. The ALJ brushed aside those scores, which signal moderate to severe functional limitations, as “subjective” and having “little value.”
The ALJ then observed that physicians treating O’Connor-Spinner’s
physical
ailments had not mentioned the specific
mental
limitations underlying our previous opinion:
Furthermore, in all of the voluminous documents that address physical impairments, there is scarcely even a hint of significant problems with social functioning or concentration, persistence and pace. Indeed, indications of mental issues are conspicuously absent from those records. It is reasonable to assume that if the claimant were rude and discourteous to her doctors, if she reacted inappropriately to their “criticism” (such as repeated admonitions to lose weight), or if she were having a difficult time concentrating, it would have been observed and mentioned somewhere in the hundreds of pages of medical records.
The ALJ gave as an example that nurses attending O’Connor-Spinner during an overnight hospital stay had not documented any “mental aberrations,” which must mean, the ALJ reasoned, that her depression was not severe.
The ALJ then concluded:
Therefore, in response to the court’s remand order, the undersigned finds that the claimant’s depression has never been a severe impairment. Accordingly, she does not have — and never did have — a moderate limitation in concentration, persistence, or pace that would require corresponding work-related limitations in her residual functional capacity. The undersigned also rejects Dr. Unversaw’s opinion that she has a moderate limitation in social functioning that impacts her ability to accept instructions and respond appropriately to criticism from supervisors.
The district court upheld the ALJ’s decision, so the ALJ’s ruling is the final decision of the Commissioner of Social Security.
See Liskowitz v. Astrue,
559 F.3d 736, 739 (7th Cir. 2009).
II. Discussion
On appeal O’Connor-Spinner contends that the ALJ mischaracterized her depression as not severe and, consequently, failed to “capture the full scope of her emotional limitations” in framing questions to the vocational expert. The Commissioner, while saying that O’Connor-Spinner simply complains about a handful of “notations” in the medical record which the ALJ “overlooked,” actually understands the central issue in this case to be the ALJ’s rejection at Step 2 of depression as a severe impairment. The Commissioner devotes her brief to arguing that the ALJ considered all evidence “he found important” and was not required to comment on each piece of evidence cited by O’Connor-Spinner. Yet never does the Commissioner explicitly argue that the ALJ’s Step 2 finding, if erroneous, was harmless or that the ALJ actually followed the instructions we gave in returning this case to the Agency.
Had the new ALJ followed our narrow instructions to address the shortcomings in his predecessor’s decision, this appeal would be simple (if necessary at all). But instead the ALJ reevaluated O’Connor-Spinner’s condition himself and, despite new evidence that
strengthens
the earlier finding that she suffers from severe depression, found the opposite.
Having reviewed the record, we cannot agree with the Commissioner’s assertion at oral argument that a “mountain of evidence” supports the ALJ’s rejection of depression as a severe impairment. The Step 2 determination is “a
de minimis
screening for groundless claims” intended to exclude slight abnormalities that only minimally impact a claimant’s basic activities.
Thomas v. Colvin,
826 F.3d 953, No. 15-2390, 2016 WL 3439015, at *5 (7th Cir. June 22, 2016). But here the ALJ decided that “major depression, recurrent
severe”
isn’t a severe impairment based on the opinions of two state-agency psychologists who did not even examine, let alone treat, O’Connor-Spinner. That determination is not supported by substantial evidence and, indeed, strikes us as nonsensical given that the diagnosis, by definition, reflects a practitioner’s assessment that the patient suffers from “clinically significant distress or impairment in social, occupational, or other important areas of functioning.” Am. Psy-ohiatriC Ass’n,
supra
at 679-80. We have not found a published opinion from any circuit in which an ALJ declared that major depression was not a severe impairment, although two unpublished decisions soundly reject this assertion,
see Magwood v. Comm’r,
417 Fed.Appx. 130 (3d Cir. 2008);
Wick v. Barnhart,
173 Fed.Appx. 597 (9th Cir. 2006). Rather than relying on the guidance of professionals and evidence from O’Connor-Spinner’s treating sources, the ALJ “played doctor” by substituting his opinion for their medical judgment.
The ALJ’s decision is replete with examples of cherry-picking evidence supporting his finding while ignoring contradictory information.
See Bates v. Colvin,
736 F.3d 1093, 1099 (7th Cir. 2013) (reversing because ALJ had ignored line of evidence supporting claim of serious mental-health issues);
Scrogham v. Colvin,
765 F.3d 685, 698-700 (7th Cir. 2014) (reversing because ALJ had disregarded evidence undermining finding that claimant wasn’t disabled). The ALJ said nothing about progress notes from treatment providers that document serious symptoms (i.e., suicidal thoughts, angry and violept behavior, and staying in bed on multiple days per month), yet he singled out the one record documenting a period when O’Connor-Spinner’s symptoms had waned.
See Larson v. Astrue,
615 F.3d 744, 751 (7th Cir. 2010) (observing that “symptoms that ‘wax and wane’ are not inconsistent with a diagnosis of recurrent, major depression”);
Bauer v. Astrue,
532 F.3d 606, 609 (7th Cir. 2008) (noting that person whose
chronic mental illness is treated with heavy drugs likely will experience better and worse days). The ALJ inexplicably drew a negative inference from the fact that doctors treating O’Connor-Spinner’s
physical
ailments had not commented on her concentration, memory, or social functioning. The ALJ seized on Dr. Kurzhals’s observation that O’Connor-Spinner did not appear manic, yet he ignored the psychologist’s further observations that O’Connor-Spinner was depressed and irritable and had performed poorly on memory tests during the examination. What is more, the ALJ berated Dr. Unversaw’s findings as “virtually incoherent” because of what appears to be a typo. Overall the ALJ was wrong to conclude, despite the plethora of evidence to the contrary, that O’Connor-Spinner’s depression was not severe.
Because the new ALJ eliminated depression at Step 2, he did not take into account
any
effects which the disorder might have on O’Connor-Spinner’s ability to maintain employment. The ALJ thus failed to “build an accurate and logical bridge between the evidence of mental impairments” and his ultimate conclusion that O’Connor-Spinner remains capable of performing sedentary work.
See Yurt v. Colvin,
758 F.3d 850, 858-59 (7th Cir. 2014);
O’Connor-Spinner,
627 F.3d at 619—21. Had the ALJ not excluded depression at Step 2, he would have been required to fully explore the restrictions caused by O’Connor-Spinner’s depression.
See Villano v. Astrue,
556 F.3d 558, 563 (7th Cir. 2009). Given the considerable additional evidence introduced after our remand to the Agency, including the diagnosis of “major depression, recurrent severe,” it is inconceivable that O’Connor-Spinner’s depression was
less
likely than before to impede her ability to work. Greater restrictions are more probable, but, at the very least, the ALJ was compelled — as we said before — to account for O’Connor-Spinner’s limitation on concentration, persistence, and pace and also to address— not ignore — Dr. Unversaw’s opinion that she is moderately limited in responding appropriately to supervisors.
To her credit the Commissioner does not defend the new ALJ’s assertion that he accomplished the objective of our remand even while explicitly declining to follow our instructions. According to the ALJ, even if he
had found a moderate limitation in concentration, persistence, or pace and added a restriction to work without strict production quotas or fast pace, the [vocational expert] testified - it would not affect the jobs he named, because the inability to perform constant handling and fingering had already eliminated most fast-paced jobs.
The ALJ specifically had asked the vocational expert whether, because of “limitations in concentration, persistence, or pace ... this person cannot do any work involving strict production quotas or fast pace.” Yet the Commissioner has not cited, nor have we found, any authority supporting the ALJ’s speculation that eliminating jobs with strict production quotas or a fast pace may serve as a proxy for including, as part of the claimant’s mental residual functional capacity, a moderate limitation on concentration, persistence, and pace. Perhaps the vocational expert would have substantiated that speculation if asked, but, as before, the ALJ did not ask.
That the ALJ did not ask is especially troubling given the vocational expert’s concession that O’Connor-Spinner would be “very unlikely” to maintain employment if she was off task 15% or more of the time, or if she had difficulty responding appropriately to supervisors. Agency regulations, as far as we can tell, do not quantify what is meant by a “moderate” restriction, but the regulations do instruct ALJs to
rate the degree of limitation on a 5-point scale of none, mild, moderate, marked, and extreme.
See
20 C.F.R. § 404.1520a. If a moderate impairment on maintaining concentration, persistence, and pace equates to being off task at least 15% of the time then, according to the vocational expert, O’Connor-Spinner is essentially unemployable. This testimony at least calls into question the ALJ’s assertion that eliminating jobs which require strict production quotas or a fast pace sufficiently accounts for a moderate limitation on concentration, persistence, and pace.
III. Conclusion
Ten years
ago the first ALJ to address O’Connor-Spinner’s impairments committed errors that prompted us to return the case to the Agency with simple instructions. And six years after our remand the matter remains unresolved because the new ALJ failed to comply with our instructions. Accordingly, we VACATE the judgment of the district court and REMAND the case to the Agency for further consideration. On remand the ALJ should determine what limitations are caused by O’Connor-Spinner’s “major depression, recurrent severe” and fully explore with a vocational expert the effect of those limitations on O’Connor-Spinner’s ability to engage in competitive employment.