Hailu v. Commissioner of Social Security

UNITED STATES DISTRICT COURT WESTERN DISTRICT OF NEW YORK DEREJE H.,1 Plaintiff, v. 19-CV-6514-LJV DECISION & ORDER COMMISSIONER OF SOCIAL SECURITY, Defendant. On July 9, 2019, the plaintiff, Dereje H. (“Dereje”), brought this action under the Social Security Act. He seeks review of the determination by the Commissioner of Social Security (“Commissioner”) that he was not disabled.2 Docket Item 1. On March 22, 2020, Dereje moved for judgment on the pleadings, Docket Item 13; on August 4, 2020, the Commissioner responded and cross-moved for judgment on the pleadings, Docket Item 19; and on August 25, 2020, Dereje replied, Docket Item 20. 1 To protect the privacy interests of social security litigants while maintaining public access to judicial records, this Court will identify any non-government party in cases filed under 42 U.S.C. § 405 (g) only by first name and last initial. Standing Order, Identification of Non-government Parties in Social Security Opinions (W.D.N.Y. Nov. 18, 2020). 2 Dereje applied for both Social Security Income (“SSI”) and Disability Insurance Benefits (“DIB”). One category of persons eligible for DIB includes any adult with a disability who, based on quarters of qualifying work, meets the Social Security Act’s insured-status requirements. See 42 U.S.C. § 423 (c); see also Arnone v. Bowen, 882 F.2d 34, 37-38 (2d Cir. 1989). SSI, on the other hand, is paid to a person with a disability who also demonstrates financial need. 42 U.S.C. § 1382 (a). A qualified individual may receive both DIB and SSI, and the Social Security Administration uses the same five-step evaluation process to determine eligibility for both programs. See 20 C.F.R §§ 404.1520(a)(4) (concerning DIB); 416.920(a)(4) (concerning SSI). For the reasons stated below, this Court grants Dereje’s motion in part and denies the Commissioner’s cross-motion.3 STANDARD OF REVIEW “The scope of review of a disability determination . . . involves two levels of inquiry.” Johnson v. Bowen, 817 F.2d 983, 985 (2d Cir. 1987). The court “must first decide whether [the Commissioner] applied the correct legal principles in making the determination.” Id. This includes ensuring “that the claimant has had a full hearing under the . . . regulations and in accordance with the beneficent purposes of the Social Security Act.” Moran v. Astrue, 569 F.3d 108, 112 (2d Cir. 2009) (quoting Cruz v. Sullivan, 912 F.2d 8, 11 (2d Cir. 1990)). Then, the court “decide[s] whether the determination is supported by ‘substantial evidence.’” Johnson, 817 F.2d at 985 (quoting 42 U.S.C. § 405 (g)). “Substantial evidence” means “more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Richardson v. Perales, 402 U.S. 389, 401 (1971) (quoting Consol. Edison Co. v. NLRB, 305 U.S. 197, 229 (1938)). “Where there is a reasonable basis for doubt whether the ALJ applied correct legal principles, application of the substantial evidence standard to uphold a finding of no disability creates an unacceptable risk that a claimant will be deprived of the right to have her disability determination made according to correct legal principles.” Johnson, 817 F.2d at 986 . 3 This Court assumes familiarity with the underlying facts, the procedural history, and the ALJ’s decision and will refer only to the facts necessary to explain its decision. DISCUSSION Dereje argues that the Commissioner erred in three ways.4 Docket Item 13-1. First, he argues that the finding by the Administrative Law Judge (“ALJ”) that Dereje medically improved beginning on February 24, 2017, is unsupported by substantial evidence. Docket Item 13-1 at 15. Second, he argues that the ALJ failed to develop the record. Id. at 21. And third, he argues that the Appeals Council improperly failed to consider new and material evidence. Id. at 23. This Court agrees that the ALJ erred in finding medical improvement as of February 24, 2017, and, because that error was to Dereje’s prejudice, remands the matter to the Commissioner. “Once a claimant establishes the existence of a disabling condition, . . . [the] claimant is entitled to a presumption that the classification will not change unless the condition, governing statutes, or regulations change.” Carbone v. Astrue, 2010 WL 3398960 , at *12 (E.D.N.Y. Aug. 26, 2010). Under the medical-improvement standard, the Commissioner may terminate a recipient’s benefits when there is “substantial evidence that the individual’s condition has improved to the point that he or she is no longer disabled.” De Leon v. Sec’y of Health & Human Servs., 734 F.2d 930, 936 (2d Cir. 1984). Before an ALJ may find a medical improvement, there must be “substantial evidence that the recipient’s condition has improved in a manner relevant to the recipient’s ability to work, and that the recipient can now engage in substantial gainful 4 Neither Dereje nor the Commissioner contest the ALJ’s finding that Dereje was disabled from April 3, 2013, to February 23, 2017. See Docket Items 13-1, 19-1. This Court therefore will not disturb that finding, and this decision and order relates only to the ALJ’s finding of a medical improvement beginning on February 24, 2017. activity.” Rodriguez v. Saul, 2020 WL 5200691 , at *2 (W.D.N.Y. Sept. 1, 2020) (citing Daif v. Astrue, 2008 WL 2622930 , at *5 (E.D.N.Y. July 1, 2008); Williams v. Barnhart, 2002 WL 618605 , at *4 (S.D.N.Y. Apr. 18, 2002) (“After a declaration of disability entitling the claimant to DIB and SSI benefits, benefits can be terminated based on a finding that the relevant impairment has ceased, no longer exists[,] or is not disabling.”)); see also 20 C.F.R. §§ 404.1594 (a), 416.994(b). Under Social Security regulations, a medical improvement is “any decrease in the medical severity of [a claimant’s] impairment[s,] which was present at the time of the most recent favorable medical decision that [a claimant was] disabled or continued to be disabled.” 20 C.F.R. §§ 404.1594 (b)(1), 416.994(b)(1)(i). “A determination that there has been a decrease in medical severity [of a claimant’s impairments] must be based on improvement in the symptoms, signs, and/or laboratory findings associated with [the] impairments.” 20 C.F.R. § 404.1594 (b)(1) (internal marks omitted); see also 20 C.F.R. § 416.994 (b)(1)(i). To make this determination, therefore, the ALJ “must compare the current medical severity of the impairment to the medical severity of that impairment at the time of the most recent favorable medical decision.” Veino v. Barnhart, 312 F.3d 578, 586-87 (2d Cir. 2002) (citation omitted) (internal marks omitted). In closed-period disability cases5 like this one, “the most recent favorable medical decision for 5 “A closed period of disability refers to when a claimant is found to be disabled for a finite period of time[,] which started and stopped prior to the date of the administrative decision granting disability status.” Carbone, 2010 WL 3398960 , at *13, n.12 (internal marks and citations omitted). Although the medical-improvement standard generally “is applied at a continuing disability review regarding a prior adjudication,” see id. at *13 , district courts within the Second Circuit—including in the Western District of New York—as well as several circuit courts have held that the medical-improvement standard is appropriately used in closed-period cases, see Torres v. Colvin, 2018 WL 3301437 , at *3 (W.D.N.Y. July 5, comparison purposes is the disability onset date.” McDonagh v. Acting Comm'r of Soc. Sec., 2017 WL 9286987 , at *10 (S.D.N.Y. Nov. 27, 2017), report and recommendation adopted, 2018 WL 2089340 (May 2, 2018). Stated more simply, the ALJ compares the current severity of the claimant’s condition with the severity at the time when the claimant was last found to be disabled. See id. Once an ALJ finds a medical improvement, the ALJ then must determine whether the medical improvement relates to the claimant’s ability to work. 20 C.F.R. §§ 404.1594 (a), 416.994(b). A medical improvement relates to a claimant’s ability to work when there has been an “increase in [the claimant’s] functional capacity to do basic work activities.”6 20 C.F.R. §§ 404.1594 (b)(3), 416.994(b)(1)(iii). To make this determination, the ALJ first must craft a new RFC “based on the current severity of the [claimant’s] impairment[s], . . . and compare the new RFC with the RFC before the putative medical improvements.” Shepard v. Apfel, 184 F.3d 1196, 1201 (10th Cir. 1999) (citing 20 C.F.R. §§ 404.1594 (c)(2)). “The ALJ may find medical improvement related to an ability to do work only if an increase in the current RFC is based on objective medical evidence.” Id. (citing 20 C.F.R. §§ 404.1594 (c)(2)). 2018) (citing cases). The ALJ here purported to apply this standard, and the parties do not contest that the medical-improvement standard applies here. 6 Basic work activities are defined as “the abilities and aptitudes necessary to do most jobs[, including] exertional abilities such as walking, standing, pushing, pulling, reaching and carrying, and non[-]exertional abilities and aptitudes such as seeing, hearing, speaking, remembering, using judgment, dealing with changes and dealing with both supervisors and fellow workers.” 20 C.F.R. §§ 404.1594 (b)(4), 416.994(b)(1)(iv). I. MEDICAL IMPROVEMENT The ALJ found that Dereje was disabled from April 3, 2013, to February 23, 2017, but improved medically as of February 24, 2017. Docket Item 7-2 at 31, 33. To support this finding, the ALJ relied primarily on treatment notes dated February 23, 2017, from Thomas McElligott, M.D. See id. at 33 . [Dereje] has a normal mood and affect. His speech is normal and behavior is normal. He is not actively hallucinating. Thought content is not paranoid. Cognition and memory are not impaired. He expresses no homicidal and no suicidal ideation. Patient is pleasant and cooperative today. He is attentive. Docket Item 7-10 at 108. The ALJ also cited treatment notes dated February 15, 2018, for the proposition that Dereje’s depression had “resolved” and that he was “nonsuicidal[, and] non-homicidal and had no hallucinations.” Docket Item 7-2 at 33 (citing Docket Item 7-10 at 200). And she cited treatment notes dated May 24, 2018, noting that Dereje’s “Risperdal was decreased due to concerns of gynecomasia.” Id. (citing Docket Item 7-10 at 287). But none of the treatment notes cited by the ALJ support a finding that Dereje had medically improved as of February 24, 2017. First, as a matter of simple logic, the February 2018 treatment notes cannot support a finding that Dereje’s condition had improved a full year earlier—in February 2017. The February 2018 notes do not refer to February 24, 2017, nor do they address how Dereje’s mental health symptoms improved or changed over time. See Docket Item 7-10 at 200. For that reason, the February 2018 notes are not evidence of the “symptoms, signs, [or] laboratory findings associated with [Dereje’s] impairments” on February 24, 2017, and they therefore cannot support a “determination that there ha[d] been a decrease in medical severity” on that date.7 See 20 C.F.R. § 404.1594 (b)(1). In fact, the February 2018 notes indicate that Dereje’s depression may have resolved on August 3, 2017, months after the ALJ found a medical improvement, but that his psychosis remained unresolved as of February 15, 2018. See Docket Item 7-10 at 200 (describing Dereje’s “[p]sychosis” as lasting from “9/8/2016 – [p]resent” and his “[d]epression” from “3/4/2016 - 8/3/2017”). Thus, while the February 2018 notes might support a finding that Dereje experienced some “decrease in [the] medical severity” of his impairments by February 2018—or perhaps by August 3, 2017, when his depression resolved—they do not support a finding that he experienced such a decrease by February 24, 2017. See 20 C.F.R. § 404.1594 (b)(1). And they may not support even the conclusion that he improved by February 2018 in a meaningful way given that Dereje’s psychosis continued. What is more, there is nothing in Dr. McElligott’s notes from February 23, 2017, that support a finding that Dereje’s mental impairments had medically improved by then. On February 23, 2017, Dr. McElligott described Dereje as having a “normal mood[,] affect . . . and behavior”; “not actively hallucinating”; and being “cooperative.” Docket Item 7-10 at 108. But Dr. McElligott made almost identical psychological findings both before and after this date. For example, four months prior, on October 17, 2016, Dr. McElligott described Dereje as presenting with “a normal mood and affect[,]” and having “normal” behavior, judgment, and thought content. Id. at 71 . Similarly, on April 20, 7 For this same reason, the remark that Dereje’s Risperdal was decreased in May 2018, see Docket Item 7-10 at 287, also cannot support a finding that he medically improved on February 24, 2017. 2017, Dr. McElligott observed that Dereje “ha[d] a normal mood and affect. His speech [was] normal and behavior [was] normal[, and his t]hought content [was] normal.” Id. at 117 . On June 22, 2017, Dr. McElligott wrote that Dereje “ha[d] a normal mood and affect” and “normal” behavior, judgment, and thought content. Id. at 128 . And on February 15, 2018, Dr. McElligott again described Dereje’s behavior and thought content as “normal”; he also observed that Dereje “exhibit[ed] a depressed mood” that day. Id. at 176 . Therefore, nothing in Dr. McElligott’s notes from February 23, 2017, show a medical improvement. On the contrary, Dr. McElligott completed a medical source statement that same day in which he opined Dereje had been “totally disabled since 2013,”8 indicating that Dr. McElligott considered Dereje’s conditions still to be quite severe. See Docket Item 7-7 at 189. Because the Commissioner bears the burden to 8 Although an opinion on the ultimate issue—that a claimant is “disabled”—is not binding on the Commissioner, see 20 C.F.R. § 404.1527 (d)(1), the ALJ still is required to analyze what is behind such an opinion or to recontact the provider to request a translation of the opinion into language that fits the Social Security context, see Snell v. Apfel, 177 F.3d 128, 133 (2d Cir. 1999) (the fact that the “ultimate finding of whether a claimant is disabled” is “reserved to the [C]ommissioner . . . means that the Social Security Administration considers the data that physicians provide but draws its own conclusions”) (emphasis added). And because Dr. McElligott was one of Dereje’s treating physicians, his opinion was entitled to controlling weight so long as it was “well-supported [sic] by medically acceptable clinical and laboratory diagnostic techniques and is not inconsistent with the other substantial evidence in [the claimant’s] case record.” 20 C.F.R. § 404.1527 (c)(2). Before giving less-than-controlling weight to Dr. McElligott’s opinion, the ALJ was required to “explicitly consider, inter alia: (1) the frequency, length, nature, and extent of treatment; (2) the amount of medical evidence supporting the opinion; (3) the consistency of the opinion with the remaining medical evidence; and[] (4) whether the physician is a specialist.” Greek v. Colvin, 802 F.3d 370, 375 (2d Cir. 2015) (quotations and alterations omitted). The ALJ did not appear to do so, however. See Docket Item 7-2 at 29. On remand, the ALJ should weigh Dr. McElligott’s—and the other treating physicians’—opinions in accordance with this rule. show that Dereje had medically improved, Dr. McElligott’s notes—which at most show the absence of a worsening problem—are insufficient. See Carbone, 2010 WL 3398960 , at *12 (explaining that “the medical-improvement standard shifts the burden of proof to the Commissioner”). This error was particularly significant given that Dereje had long-standing relationships with mental health professionals whose notes and opinions offered a more complete picture of his mental health. A review of their progress notes suggests that Dereje may have improved at some point during the course of his treatment. But there is no evidence to support a finding that he improved by February 24, 2017. When Dereje began mental health treatment at the Genesee Mental Health Center on March 17, 2016, his therapist rated his risk for acute suicide, chronic suicide, and chronic violence as “[m]oderate.” Docket Item 7-8 at 7-8. A few months later, Dereje still was having paranoid thoughts and other delusions related to his ex-wife. See id. at 278 (“she [has] vodoo [sic] on me, my sister told me”). By December 2016, Dereje felt that he was improving and began attending Bible study classes. See Docket Item 7-9 at 84. In May 2017, he began to think about looking for a job after he graduated from his recovery program. Id. at 131 . On June 2, 2017, Dereje’s risk for suicide and violence dropped to “[l]ow,” id. at 158 , and on June 23, 2017, Dereje reported becoming more social, applying for a job, and taking more walks, id. at 179 . On August 30, 2017, his therapist described him as “in maintenance,” id. at 278 , and by September 2017, Dereje graduated from his recovery program, id. at 231 . On October 20, 2017, Dereje’s therapist opined that he had no mental functional limitations and could work a 40-hour work week. Docket Item 7-10 at 28-29. But despite the improvement, Dereje still struggled with his mental health for much of this time. As explained above, Dereje remained a “[m]oderate” risk for suicide and violence until June 2, 2017. Compare Docket Item 7-8 at 7-8, with Docket Item 7-9 at 158. His depression did not resolve until August 3, 2017, and his psychosis remained unresolved at least through February 2018. See Docket Item 7-10 at 200. Similarly, when Dereje began treatment in March 2016, his LOCUS score9 was 20, in the range of Level 4—the third-most severe category—indicating the need for Medically Monitored Non-Residential Services; his LOCUS score did not suggest that he was in the improved category of “Recovery Maintenance & Health Management” until November 28, 2017. See Docket Item 7-9 at 280, 294 (LOCUS score improving from Level 3 to Level 1 on November 28, 2017, “[d]ue to increased social supports and decrease[d] symptoms”); see also Docket Item 7-8 at 200 (LOCUS score improving from Level 4 to Level 3 on June 7, 2016, due to increased independence and community engagement). The point is not that Dereje never improved; indeed, he may well have improved by October 20, 2017, when his therapist thought that he was capable of working 40 hours a week. See Docket Item 7-10 at 28-29. But there simply is no evidence in the 9 A LOCUS score, also known as “Level of Care Utilization System of Psychiatric and Addiction Services,” is a criteria-assessment tool used to gauge the level of mental health and addiction care that an individual needs. The LOCUS provides for six levels of care ranging from Level 1 (score of 10-13) to the Level 6 (score of 28 or more). The higher the LOCUS score, the more mental health services an individual requires. See LOCUS/CALOCUS, American Association of Community Psychiatrists, https://sites.google.com/view/aacp123/resources/locus (last visited Feb. 9, 2021); see also Level of Care Utilization System for Psychiatric and Addiction Services, Adult Version 2010, American Association of Community Psychiatrists (Mar. 20, 2009), available for download at http://www.communitypsychiatry.org. record tied to February 24, 2017, or supporting the conclusion that Dereje had medically improved as of that day. Stated another way, the most that the record suggests is that Dereje medically improved at some point during his treatment—probably months after February 24, 2017. See, e.g., Docket Item 7-9 at 158 (June 2, 2017: risk of suicide and violence dropped to “[l]ow”); Docket Item 7-10 at 28-29 (October 20, 2017: therapist opined that Dereje had no mental functional limitations); Docket Item 7-9 at 280 (November 28, 2017: LOCUS score improved to Level 1). The ALJ’s error probably prejudiced Dereje: there was a period of time after February 24, 2017, the ALJ’s date of improvement, when Dereje likely was eligible for benefits. Remand therefore is appropriate for the ALJ to reconsider, and to explain in light of the objective medical evidence, the date by which Dereje may have medically improved. See 20 C.F.R. § 404.1594 (b)(1). II. RELATED TO THE ABILITY TO WORK The ALJ also found that Dereje’s medical improvement was related to his ability to work because, as of February 24, 2017, “he was able to concentrate and attend for greater periods.” Docket Item 7-2 at 34. The ALJ specifically found that beginning on February 24, 2017, Dereje had the mental RFC to perform “simple routine repetitive tasks[ and] work[] independently with only occasional instruction.” Id. He “only occasional[ly could] interact[] with co-workers; [he could not] work[] tandem [or] work with the public[;] and he would be off task up to 10% of the workday.” Id. These limitations are quite specific, but no physician or other medical provider opined about any of them. Instead, the ALJ grounded the mental RFC in treatment notes beginning in May 2017, three months after Dereje’s disability supposedly ended. See id. (reasoning that “by May 2017, the record indicates that [Dereje] began to discuss in therapy that he would start driving a taxi again when he was finished with the Restart program”). Indeed, the ALJ does not cite any evidence before May 2017 to support her finding that Dereje’s functional capacity had increased by February 24, 2017. See id. What is more, the two opinions that the ALJ discussed and weighed in the context of the new mental RFC were from May 30, 2018, and July 24, 2018, and neither shed any light on Dereje’s functional capacity on February 24, 2017. See id. at 35 (citing Docket Item 7-10 at 26 (opinion of H. Bronstein, M.D., dated May 30, 2018)); id. (citing Docket Item 7-10 at 345 (opinion of Daniel Hoefling dated July 24, 2018)). Neither opinion therefore could support the mental RFC that the ALJ found—at least not as of February 24, 2017. See id.; see also Shepard, 184 F.3d at 1201 (explaining that to find that a medical improvement is related to a claimant’s ability to work, the ALJ first must craft a new RFC “based on the current severity of the [claimant’s] impairment[s], . . . and compare the new RFC with the RFC before the putative medical improvements”). And to the extent that the ALJ relied implicitly on Dr. McElligott’s February 23, 2017 treatment notes when crafting the new mental RFC, for the reasons stated above those notes alone cannot support a finding that Dereje’s functional capacity had increased by then. See supra (explaining that Dr. McElligott’s notes did not show a medical improvement as of that date). In sum, to find a medical improvement related to Dereje’s ability to work, the ALJ was required to base Dereje’s new mental RFC “on objective medical evidence.” See Shepard, 184 F.3d at 1201 . But the ALJ did not do that. Instead, she included specific limitations of her own, not based on any medical provider’s opinion. For that reason, the ALJ erred by relying on her own lay judgment to evaluate the medical records and formulate the RFC.10 That error prejudiced Dereje. For example, the increased mental RFC limited Dereje to 10% time off task, Docket Item 7-2 at 34, and according to the Vocational Expert, anything more than that would preclude substantial gainful employment, see id. at 73 (opining that employers will tolerate “nothing more than 10% of the day [off task] exclusive of regular breaks” before terminating an employee). The finding that Dereje would be off task only 10% of the time therefore was necessary for the ALJ to conclude that Dereje was no longer disabled as of February 24, 2017. But how the ALJ concluded Dereje would be no more than 10% off task after February 24, 2017, is anyone’s guess. Even worse, not only did the ALJ fail to cite any evidence before May 2017 to support the new mental RFC, but that RFC conflicts with the medical evidence from around that time. On February 6, 2017, for example, Dereje’s psychiatrist, M. Saleem Islamil, M.D., categorized Dereje’s concentration and attention as only “[f]air.” Docket Item 7-9 at 107. But the RFC required Dereje to stay on task at least 90% of the workday—with the expectation that he still would complete any work missed while off 10 It otherwise is “unclear to the Court how the ALJ, who is not a medical professional, came up with this highly specific RFC determination.” Perkins v. Berryhill, 2018 WL 3372964 , at * 3 (W.D.N.Y. July 11, 2018); see also Balsamo v. Chater, 142 F.3d 75, 81 (2d Cir. 1998) (“In the absence of a medical opinion to support [an] ALJ’s finding as to [a claimant’s] ability to perform [a certain level of] work, it is well-settled that the ALJ cannot arbitrarily substitute his own judgment for competent medical opinion.”); Jermyn v. Colvin, 2015 WL 1298997 , at *19 (E.D.N.Y. Mar. 23, 2015) (“[N]one of these medical sources assessed [the p]laintiff’s functional capacity or limitations, and therefore provide no support for the ALJ’s RFC determination.”). task—and “work[] independently with only occasional instruction.” See Docket Item 7-2 at 34 (RFC limiting Dereje to being off task 10% of the workday); see also id. at 73 (Vocational Expert testifying that an employer would tolerate an individual being off task 10% of the day “with the caveat that the expected amount of work is completed, even with the time off task”). Likewise, in February 2017, Dereje still had a “[m]oderate” risk for suicide and violence. See, e.g., Docket Item 7-9 at 85 (progress notes dated 12/5/2016: “[m]oderate” risk for acute and chronic suicide and chronic violence); id. at 158 (progress notes dated June 2, 2017: risk of suicide and violence dropped to “[l]ow”). But the RFC would require Dereje to “occasional[ly] interact[] with co-workers.” See Docket Item 7-2 at 34. And the ALJ did not even attempt to reconcile any of those inconsistencies. Along the same lines, on March 31, 2016, Dereje’s therapist, Nicole Smith, MHC, opined that Dereje was “[m]oderately [l]imited [and u]nable to function 10-25% of the time” in the following functional-limitation categories: “[d]emonstrates the capacity to follow, understand[,] and remember simple instructions and directions”; “[d]emonstrates the capacity to perform simple and complex tasks independently”; “[d]emonstrates the capacity to maintain attention and concentration for [illegible] tasks”; “[d]emonstrates the capacity to regularly attend to a routine and maintain a schedule”; “[d]emonstrates the capacity to maintain basic standards of hygiene and grooming”; and “[d]emonstrates the capacity to perform low stress and simple tasks.” Docket Item 7-10 at 37. The ALJ afforded Smith’s opinion “some weight,” see Docket Item 7-2 at 29, and incorporated those limitations into the RFC rendering Dereje disabled from April 3, 2013, to February 23, 2017, id. at 26 (RFC included being “off task in excess of 20% of the workday”); id. at 30 (ALJ explaining that “I have included a limit of 20% for attention and concentration in the [RFC] assessment based in part on Ms. Smith’s statement”). The ALJ did not explain how those limitations suddenly vanished on February 24, 2017. And it was not until October 20, 2017, that Nicole Lambert, LMHC—another therapist who Dereje began seeing on February 19, 2017—opined that Dereje had “[n]o evidence of limitation” in each of the above categories and that, as of October 20, 2017, Dereje was able to work up to 40 hours per week. Docket Item 7-10 at 28-29. Simply put, it is not clear to the Court how the ALJ was able to glean from the bare medical data that, as of February 24, 2017, Dereje’s mental RFC increased to the point that he was able to work. That is a second, independent error requiring remand. CONCLUSION This Court therefore remands the matter for reconsideration of Dereje’s medical improvement. Specifically, the ALJ should reconsider, and explain in light of the objective medical evidence, the date by which Dereje had medically improved. See 20 C.F.R. § 404.1594 (b)(1) (“A determination that there has been a decrease in medical severity [of a claimant’s impairments] must be based on improvement in the symptoms, signs, and/or laboratory findings associated with [the] impairments.”) (internal marks omitted). If it is unclear to the ALJ when that date is—which well may be the case—the ALJ should contact Dereje’s providers for an opinion about that date. See Rosa v. Callahan, 168 F.3d 72, 79 (2d Cir. 1999) (“[W]here there are deficiencies in the record, an ALJ is under an affirmative obligation to develop a claimant’s medical history ‘even when the claimant is represented by counsel.’”). If, based on the medical evidence, the ALJ finds a medical improvement, she then must craft an RFC based on objective medical evidence, not on her own lay opinion, to decide whether that medical improvement is sufficient for Dereje to return to work. See 20 C.F.R. §§ 404.1594 (c)(2), 416.994(b)(2)(iii). And if there is insufficient evidence in the record as to Dereje’s function limitations on that date, then the ALJ should fill that gap by contacting Dereje’s medical providers. See Rosa, 168 F.3d at 79 . The Court “will not reach the remaining issues raised by [Dereje] because they may be affected by the ALJ’s treatment of this case on remand.” Watkins v. Barnhart, 350 F.3d 1297, 1299 (10th Cir. 2003); see also Bonet ex rel. T.B. v. Colvin, 2015 WL 729707 , at *7 (N.D.N.Y. Feb. 18, 2015). The Commissioner’s motion for judgment on the pleadings, Docket Item 19, is DENIED, and Dereje’s motion for judgment on the pleadings, Docket Item 13, is GRANTED in part and DENIED in part. The decision of the Commissioner is VACATED, and the matter is REMANDED for further administrative proceedings consistent with this decision. SO ORDERED. Dated: February 24, 2021 Buffalo, New York /s/ Lawrence J. Vilardo LAWRENCE J. VILARDO UNITED STATES DISTRICT JUDGE