The plaintiff-appellant, Mr. Childress, applied to the Social Security Administration for disability benefits in 2008, when he was 35. Turned down in 2010 after a hearing before an administrative law judge, he appealed to the district court, which initially remanded the case to the Social Security Administration for reevaluation of the medical opinions in the record and reconsideration of the plaintiffs credibility. The upshot was a second hearing, held in 2013, before the same administrative law judge, who again ruled that Childress was not disabled; and this time the district court affirmed, precipitating his appeal to us. He complains that the administrative law judge again failed to give sufficient weight to the opinions of his treating physicians and did not accurately assess his credibility or his capacity to work.
The administrative law judge ruled in the second round that Childress could perform only limited sedentary work (and nothing more strenuous), because he is capable of standing .for only 25 to 30 minutes at a time and of walking 1 to 2 blocks at a time, adding up to a maximum of 2 hours a day of either walking or standing; he is capable of sitting 45 to 60 minutes at a time, for a total of 6 hours a day; and he is capable of carrying out workplace instructions — provided he was given them no more than 30 days earlier. The administrative law judge also ruled that he must avoid strenuous work, dangerous machinery, unprotected heights, and exposure to extreme heat.
The medical evidence presented by Chil-dress’s treating physicians was extensive. A cardiologist who had been treating Chil-dress for years, Dr. Theodore Addai, reported that Childress suffers from congestive heart failure, cardiomyopathy (another disease that diminishes cardiac performance), severe asthma, COPD (chronic obstructive pulmonary disease, actually one or more of a set of distinct diseases, all of which however are debilitating, progressive, and potentially fatal lung diseases), occasional chest pain, obesity (he weighs 350 pounds yet is only 69 inches — five feet nine inches — tall), hypertension, and dysp-nea (difficult or uncomfortable breathing, resulting in shortness of breath). He was prescribed a number of medications: Ad-vair, Benazepril, Coreg, Diovan, Lanoxin, Lasix, Norvasc, Proventil, and Spiriva. We are not told whether any of them have side effects that are harmful or that affect work capacity, either in general or with specific respect to Childress.
The percentage of blood pumped out of the ventricles with each contraction of the heart (i.e., each heartbeat) is called the “ejection fraction” (EF). The EF number helps a health-care provider determine whether a patient has heart disease. A normal heart has an EF of 50 to 75 percent in the left ventricle. (The right ventricle can have a lower EF without being abnormal, because it pumps blood only to the lungs, whereas the left ventricle pumps blood to the rest of the body.)
*791By 2010, Childress’s left-ventricle EF had fallen to 35 to 40 percent (though later in the year it rose to a normal 66 percent). ■The cardiologist estimated that in an eight-hour workday Childress would be able to stand or walk for no more than one hour and to sit for no more than two hours'.
The ups and downs continued, but the downs predominated. A stress test in 2011 showed “poor exercise tolerance for his age” (though he was still in his 30s). He had shortness of breath even at home, fatigue, tingling sensations, swelling in his feet and ankles, tightness in his chest, flashes of light in his vision, tingling all over, and continued diagnoses of serious heart problems. The following year his ejection fraction fell to a dangerously low 20 to 25 percent, requiring implantation of a cardiac defibrillator. The following year his ejection fraction rose to 30 to 35 percent, still abnormally low. Other heart problems that he was diagnosed with included diastolic dysfunction, severely dilated left ventricular chamber size, severely dilated left and right atriums, and mild aortic valve insufficiency.
Nor is the ejection fraction the only evidence that Childress’s left ventricle is impaired. If the left ventricle is wider than 6 centimeters in its diastolic state (that is, when it is expanded and full of blood), this indicates a severe thinning of the heart muscle. See 20 C.F.R. Part 404, Subpart P, App. 1, 4.02(A)(1). Childress’s heart measurements have consistently revealed that his left ventricle is more than 6 centimeters wide. This was actually noted by the administrative law judge at Childress’s first hearing. Yet at the second hearing, the one now under review, she did not mention it, thus overlooking an important fact supportive of his claim to be disabled.
Another doctor, Kari Cataldo, evaluated and treated Childress beginning in 2008. Her diagnoses were similar to Dr. Addai’s, but with the addition of diagnosing crackles in the lungs, bronchial markings, a hernia, acute bronchitis, depression, a systolic heart murmur, increased lung markings, increased dyspnea and chest pain, wheezing and coarse breath sounds, edema (swelling) of the ankles, headaches, and pain and swelling in the legs. In 2009 another doctor appeared on the scene, Patrick Hartman, who diagnosed acute bronchitis and hypertension, coarse breath sounds, shortness of breath, coughing, fatigue, decreased exercise tolerance, severe fatigue, congestive heart failure, chronic obstructive pulmonary disease, obesity, etc., and an ejection fraction,- still subnormal, of 35 to 40 percent. Childress also had painful cysts on his legs, which Dr. Hartman treated.
Dr. Hartman estimated that Childress would be able to sit for 3 hours total and stand or walk for 1 hour total in an 8-hour workday, though he would also have to lie down for an hour every 1 to 2 hours during the workday — which would reduce his workday. If he lay down only every 2 hours, say at 11 a.m. and 1 p.m. and 3 p.m., and his workday was 9 a.m. to 5 p.m., that would reduce the actual working part of his “working day” to 5 hours (9 to 11 a.m., 12 to 1 p.m., 2 to 3 p.m., and 4 to 5 p.m.), rendering him unemployable. Dr. Hartman also thought it likely that Childress’s medical symptoms would worsen in a work environment, and he concluded that Chil-dress is totally disabled from gainful employment.
Two nonexamining state agency physicians looked at a severely incomplete set of the plaintiffs medical records, and concluded without any real evidence that he can walk about 6 hours a day and sit about 6 hours a day. The second such physician summarily agreed with the first, after re*792viewing treatment records that ended at the end of January 2009.
Childress testified at length at his hearing before the administrative law judge. He described the bad effect of his obesity on his breathing (he is constantly short of breath), his joints, and his ability to get up and sit down. He testified that he was unable to lose weight through either dieting or exercise, despite trying — in part because his medications, which include steroids, have caused him to gain thirty to forty pounds. He leaves his home infrequently, sleeps a good deal during the day, has trouble showering and dressing, does very little work around the house, does not cook or do his laundry, and gets help from his parents in cleaning the house. His testimony was not contradicted. The picture it painted is not that of a person capable of full-time work even of a sedentary character, especially since Childress had not attended college; and although when younger he had had occasional jobs, they were as a cook and as a satellite antenna installer, which are not sedentary jobs.
Although the testimony of Drs. Addai and Hartman, if believed, established that Childress was totally disabled from gainful employment, the administrative law judge disbelieved them on the ground that them testimony was “not consistent with the medical record as a whole, or with their own progress notes.” Yet she gave little weight to the evidence presented by the nontreating physicians — and rightly so, as their evidence was essentially worthless since they had not had access to Chil-dress’s full (abundantly full) medical record. Her decision rejecting Childress’s claim of total disability rested on her belief that the two treating physicians had given evidence concerning Childress’s activities that contradicted the record: that Dr. Ad-dai had recommended that Childress walk 30 minutes a day, five to seven days a week, to help with his cardiac functioning — and he did walk, almost daily. He had worked as a part-time cook for three months ending in November 2009. And despite his disabilities he lives alone and cares for himself.
These are feeble points. Ability to walk 30 minutes a day doesn’t contradict Chil-dress’s need for sleeping (or at least lying down) at work 3 hours a day, which would surely disable him from gainful, full-time employment, as the vocational expert agreed. Furthermore the Social Security Administration defines gainful sedentary employment as comprising at least 6 hours of sitting and 2 hours of standing or walking — not 30 minutes. The administrative law judge seems not to have realized that Childress’s treating physicians considered all his problems in combination when assessing his ability to stand or sit for long periods of time. That is the correct approach. Engstrand v. Colvin, 788 F.3d 655, 661 (7th Cir. 2015).
Childress lives by himself but receives lots of assistance in his home from his parents and does very little himself besides sleeping. As for working as a part-time cook for three months, the precise nature of the job and the reasons for his being employed for so short a period were never explored in the hearing. We have emphasized that “working sporadically or performing household chores are not inconsistent with being unable to engage in substantial gainful activity.” Engstrand v. Colvin, supra, 788 F.3d at 661. See also Bjornson v. Astrue, 671 F.3d 640, 647 (7th Cir. 2012), noting that failing to recognize the difference between performing activities of daily living with flexibility (and often with help from family and friends) and performing to the standards required by an employer “is a recurrent, and deplorable, feature of opinions by administra*793tive law judges in social security disability cases.”
The administrative law judge was critical of the fact that two of Childress’s visits to doctors’ offices recorded symptoms after he’d run out of his medications, when one would expect his symptoms to worsen. But there is no evidence that he failed to take any of his medications after April 2009, and yet despite that his symptoms were worse by March 2010. The medications, numerous and formidable as they were, were unable to improve his perilous health.
Very strangely the administrative law judge thought that the fact that, as mentioned by the treating physicians, Chil-dress has been given a Class III rating on the New York Heart Association Functional Classification scale showed he can do full-time sedentary work. Not so. A Class III rating is indicative of “marked limitation in physical activity,” such that “less than ordinary activity causes fatigue, palpitation, or dyspnea.” See American Heart Association, “Classes of Heart Failure,” www.heart.org/HEABTORG/Conditions/ HeartFailure/AboutHeartFailure/Classes-of-Heart-Failure_UCM_306328_Article.jsp (visited Jan. 3, 2017).
Another curious observation made by the administrative law judge was that Chil-dress does not have the “usual objective signs of severe pain ... such as abnormal weight loss or muscular atrophy.” No, he does not have those particular objective signs of severe pain; but does the administrative law judge really believe that an obese person with very serious heart and lung problems will not have severe pain? Nor is severity the only pain dimension worth considering; to be plagued by constant pain can be torture even if the individual spasm is not severe. Similarly, it was wrong of the administrative law judge to fault Childress for not taking strong opioids for pain, when his medical conditions did not require them and, as is now well known, opioids can be very dangerous.
She said “the doctor concluded [that Childress had] possible sleep apnea, and advised referral for a sleep study. This is not a medically determinable impairment.” What could she have meant? Sleep apnea is a serious, medically determinable condition, which if not diagnosed and treated can kill.
Between them, Drs. Addai and Hartman treated Childress for a total of about 30 months, the treatment including numerous medical tests. There is no suggestion that the two doctors are not reputable physicians specializing in the specific health problems that afflict Childress. Their testimony and records were entitled to significant weight, see 20 C.F.R. § 404.1527(c); Meuser v. Colvin, 838 F.3d 905, 912 (7th Cir. 2016) (per curiam), yet the administrative law judge gave their testimony and records virtually no weight, as a result leaving the record of the hearing almost blank. Her conclusion that Childress can sit for at least 6 hours a day at work and stand and/or walk another 2 hours a day, while sleeping 3 hours a day at work and missing 3 or more days of work per month, and keep to this schedule week after week, and continue employed until his sleeping on the job and absenteeism is noticed by the employer is — absurd. The vocational expert admitted that an employee who misses three or more days of work a month is unemployable. The administrative law judge never explained why she thought this limitation did not apply to Childress.
A final botch by the administrative law judge was her remark that if Childress was really suffering he would have stopped smoking cigarettes earlier; he stopped only four days prior to the second hearing. But there is no evidence of whether on balance *794his smoking relieved or increased his pain, or, most important, given the gravity of his medical condition, whether stopping smoking would improve his health significantly — as we explained in Shramek v. Apfel, 226 F.3d 809, 812-13 (7th Cir. 2000). Nor is there evidence of how much he was smoking before he stopped.
The judgment of the district court is reversed and the case remanded to that court with directions to remand the case to the Social Security Administration.