Solitary Confinement

What Georgia's isolation system actually is

Georgia does not call most of its isolation solitary confinement. It calls it the Tier Segregation Management System, implemented in August 2013 (#6593, #20860). In FY2014 the Department of Corrections reported that it had "developed and implemented the Tier Segregation Management System at all Level V facilities" and that "2,721 beds were converted to Tier beds" (#19297). That is the only system-wide quantification of the Tier system GDC has ever published, and it is a bed count, not a population count. It establishes that the system was built at a scale of thousands of beds, not hundreds.

The international benchmark is the UN Standard Minimum Rules for the Treatment of Prisoners. Rule 44 defines solitary confinement as confinement "for 22 hours or more a day without meaningful human contact," and defines prolonged solitary confinement as more than 15 consecutive days; Rule 43 prohibits prolonged and indefinite solitary in all circumstances (#1476, #6619, #1450). Georgia's own written out-of-cell minimums exceed the 22-hour element by arithmetic. Tier I guarantees five hours of exercise per week across five days (#19270). Tier II guarantees "a minimum of five (5) hours per week of exercise" with no daily distribution specified at all (#19269). Five hours out of a 168-hour week leaves 163 hours in cell — 23.3 hours a day (#19272). Tier III, the unit that was litigated, guarantees four hours out of cell Monday through Friday, which averages 21.1 hours a day but reaches 24 hours on Saturdays and Sundays, when no out-of-cell time is guaranteed (#19272).

Rule 44 has a second element — the absence of meaningful human contact — and Georgia's policy determines whether it is met by which unit a person is assigned to. Tier III requires single occupancy in every phase (#19278). Tier III's step-down program requires it too; Tier II's step-down program expressly permits double occupancy (#19284). SOP 209.08, which governs Tier II, says nothing about occupancy at all, and the generic administrative segregation policy treats a single cell as an enumerated exception requiring assignment by medical, mental health or classification staff (#19404, #19279). GPS's own synthesis puts the consequence plainly: whether a person in the Tier system is held alone or with a cellmate, "and therefore whether a 23-hour-a-day confinement is solitary confinement within the Mandela Rules definition or not, is determined by which unit they are assigned to — and GDC's Tier II policy neither specifies nor reports it" (#19510).

The three programs differ sharply in what they promise. Tier I is short-term segregation, capped at 30 days except that the cap is disapplied entirely for voluntary and involuntary protective custody (#19249), and it is authorized at every facility type in the system — state prisons, county correctional institutions, private prisons, transitional centers, probation detention centers and boot camps (#19247). Tier II is long-term administrative segregation, defined in policy as a "stratification plan" that "is not a punishment measure," with no maximum duration written into the SOP (#19252, #19291). Tier III is the Special Management Unit at the Georgia Diagnostic and Classification Prison: a five-phase program with a stated minimum of 390 days (60 + 60 + 90 + 90 + 90) and a 24-month presumptive limit subject to a six-criterion override (#19260).

Scale is where the public record thins out. Summing the most recent PREA audit figure for each facility that states one yields 3,401 segregation cells across 30 facilities — a floor, and a cell count rather than a bed count, since many of those cells hold two people (#19398). As of October 2, 2026, the eight facilities reporting Tier II beds held 884 of them: Hays 151 (#20750), Ware 138 (#20756), Valdosta 129 (#20755), Macon 121 (#20751), Telfair 117 (#20754), Smith 111 (#20753), Hancock 83 (#20749) and McRae Women's Facility 34 (#20752). Four of those prisons were at or within one bed of full, and Macon was at 119 of 121 (#20766). Against that, the Special Management Unit reported 160 Tier III beds (#20757). Only the Special Management Unit's population is actually published, and only as a side effect of the SMU being carried as a standalone facility: 154 people on September 1, 2026, ranging between 149 and 195 across 75 monthly snapshots since July 2020, against 192 beds (#19304, #19302).

The Special Management Unit: the unit that was litigated

The Special Management Unit at the Georgia Diagnostic and Classification Prison in Jackson is the state's most restrictive solitary confinement facility, where, in the Southern Center for Human Rights' description, prisoners "spend years wholly isolated in a small concrete room" (#21926). GDC describes it as six cellblocks of single-bunked cells (#6600). Cells measured roughly six feet by nine — about 54 square feet, smaller than an average parking space — with solid metal doors and only a small glass window (#1511, #6598). In the most restrictive cellblocks, prisoners were held 22 to 24 hours a day alone, without books or personal property, barred from leaving the cell for a minimum of 90 days on arrival (#1512). The unit was characterized by a constant din of yelling and banging, a permeating stench of feces, and dampness and mildew from in-cell showers (#1514).

The duration data is the core of the record. As of July 2017, 141 of 182 people in the SMU — 78 percent — had been held more than two years (#1448); 80, or 44 percent, more than four years (#1516); 47, or 26 percent, more than five years (#1517). SCHR's 2017 letter to GDC, drawing on the litigation record, put the average stay at three to four years and said about 20 percent had been held six years or more (#6610, #6609). Individual cases ran longer: Timothy Gumm, the lead plaintiff, was held continuously for seven and a half years despite 14 separate transfer recommendations over four years (#1518, #6612); Johnny Mack Brown for nine years (#1519); Robert Watkins for eight to ten (#1520). Thirty-nine percent of the unit's population had a diagnosed mental illness (#1449), and at the 2017 inspection 70 of roughly 180 people were designated mentally ill (#6606).

Dr. Craig Haney inspected the unit in October 2017 and called it "one of the harshest and most draconian" facilities he had seen "in decades of conducting evaluations," describing the prisoners as "among the most psychologically traumatized persons [he] ha[d] ever assessed in this context" and concluding that "some of the inmates' psychological harm … may be irreversible and even fatal" (#6624, #1510, #19326). His report described "a cell block full of inmates with serious mental illness; a man who had been locked for months inside a pitch-black cell; and another man, naked and psychotic, whose cell was covered in blood" (#6643), and described E Wing as "bedlam-like, as chaotic and out-of-control as any such unit I have seen in decades of conducting such evaluations" (#21840).

Gumm filed a handwritten pro se complaint in 2015 after five years in isolation, following a failed escape attempt in 2010 (#1521, #25482). The settlement that followed — approved in 2019 — imposed a general 24-month limit on SMU confinement, mandatory 60-day or 90-day reviews with out-of-cell mental health evaluations by a licensed provider, a bar on holding anyone at Mental Health Level III or above in Tier III, minimum out-of-cell time and programming, documentation requirements, and $425,000 in attorney's fees (#6641, #1523, #26228). In April 2024 the court overseeing that agreement held GDC in contempt for failing to comply and imposed monetary sanctions, described in one account as a $2,500-a-day fine (#20404, #21547, #21953). The contempt order records that extremely low staffing at the unit made taking people to the outdoor pen impossible, and Chief U.S. District Judge Marc Treadwell described SMU officials as "obstinately disobedient," making "no real effort to affect change at all" (#21869, #21866). The court also found that SMU officials could produce no documentation of the mental health screening or counseling the settlement had promised (#21871).

Tier II: the larger program the settlement never reached

The protections won in Gumm attached to the unit that was litigated. SOP 209.08, governing Tier II, has not been revised since April 11, 2016 — nearly three years before the settlement — and contains none of it: no duration cap, no mental health exclusion, no review panel, no mental health evaluation at placement, and no out-of-cell minimum beyond five hours per week (#19324, #19252). The programme with the least protection is the larger one, operating at eight facilities and holding more than five times the SMU's published population.

Tier II is indefinite as a matter of written policy. Offenders who "cannot return to general population because they pose a serious threat to the safety and security of the institutional operation shall remain in Phase 3," reviewed every 90 days (#19252). Facility audits written by GDC's own PREA reviewers describe the duration differently again: Hays State Prison documents Tier II as "long-term segregation/270 days" (#19403), Macon State Prison as "from nine months to indefinite" and later "from 270 days to an indefinite date" (#19418), and Georgia State Prison as taking "up to two years to progress … through the required phases" (#19478). None of those durations appears in SOP 209.08, which states no duration at all.

Admission is broad. The first eligibility criterion is that "the offender is noted as a threat to the safe and secure operation of the Facility," which "may include, but is not limited to" gang documentation, notoriety of crimes, high supervision needs, and "offenders who have either been threatened with bodily harm or threatened others with bodily harm" — language that reaches the person who was threatened (#19253). Refusal to participate in Tier I is itself a ground for escalation to Tier II (#19487). None of the first, fifth or tenth criteria requires a disciplinary adjudication (#19253).

Conditions are thin and adjustable. Medical rounds are required three times per week, expressly excluding weekends and holidays (#19336), a standard that conflicts with the health services policy requiring daily rounds in restrictive housing (#19337). A person placed in Tier II has their tablet confiscated until Phase 3, after which it may be used only for education and programming, and video visits are limited to one per calendar month within a six-hour window on two named weekdays (#19285). One clause lets the Warden modify every condition in the SOP — hygiene, food, exercise, telephone, visitation, counselling — on a finding of "extenuating operational circumstances," undefined, with no time limit, no documentation requirement and no notice to the prisoner (#19276). Tier III has no equivalent; it requires written justification in the electronic logbook and written notice to the North Regional Director whenever out-of-cell time is cancelled for three or more consecutive days (#19282).

Occupancy is the sharpest difference. Tier II housing is documented as double-bunked across the state in GDC's own audit language: Hancock State Prison's G-1 unit "hous[es] a maximum of 96 inmates, in double occupancy cells" (#19572); Smith State Prison's dormitories comprise "48 double bunked cells housing 96 offenders per dorm" (#19526); Georgia State Prison's G Building held "up to 218 offenders in double bunk cells" while its K Building held 100 in "single cell lockdown cells" (#19474, #19525). Two occupancy modes run inside the same program at the same prison, and the policy distinguishes neither.

Exit is not a facility decision. A facility can place someone in Tier II on its own authority, including immediately in an emergency, but release runs through the Unit Manager, the Warden and Regional Director, a Criminal Investigations Unit security-threat-group clearance, and finally "the Director of Field Operations who will make the final decision" (#19256). The appeal of the 90-day review that continues confinement goes only to the Warden — the same official who approved the placement in the first instance (#19258). And in 2025 GDC's own auditors found that people "assigned to the Tier II Program" and other lockdown units "are not being afforded a minimum of five (5) hours per week of exercise outside the cells," that "[e]xercise is not being offered by policy," and verified the finding against door sheets (#21525). One close-security prison's Tier units carried a stated capacity of 303 beds — larger than the entire Tier III Special Management Unit (#19393, #19523).

What isolation does, and what Georgia measures about it

The clinical literature on solitary confinement is old, consistent, and unusually blunt. Dr. Stuart Grassian identified "SHU syndrome" in 1983 as a "major, clinically distinguishable psychiatric syndrome" — hypersensitivity to external stimuli, affective disturbance, difficulties with thinking, concentration and memory, disturbances of thought control and problems with impulse control — based on 14 prisoners in long-term solitary in Massachusetts (#6623, #7277). In severe cases he described "florid delirium — a confusional psychosis with intense agitation, fearfulness, and disorganization" (#1456), and found that even psychologically resilient prisoners suffer "severe psychological pain," with harm that may result in "prolonged or permanent psychiatric disability" (#1457). A 2018 review in Crime and Justice found the research on psychological harm from solitary "strikingly consistent since the early nineteenth century" across methodologies, populations and time periods (#1454).

Dr. Craig Haney's 2003 study documented symptom prevalence among prisoners held in solitary: 91 percent reporting anxiety, 86 percent oversensitivity to stimuli, 83 percent social withdrawal, 77 percent chronic depression, 70 percent an impending nervous breakdown, and 68 percent heart palpitations (#1460, #1461, #1462, #1463, #1464, #1465). A meta-analysis pooling five studies found a standardized mean difference of 0.45 (95% CI 0.29–0.61) on mental health scales (#20586); a systematic review covering 13 studies and 382,440 people found solitary associated with increases in adverse psychological effects, self-harm and mortality, and specifically that the effect appears to be beyond that of general incarceration or prior mental illness (#20587, #20588). A 2025 PLOS One meta-analysis synthesizing 171,300 people found significantly greater psychological distress, more psychiatric symptoms including self-harm and thought disorders, and greater need for mental health services among prisoners in disciplinary confinement (#1458). A 2024 Washington State study of 106 people in long-term solitary found clinically significant depression, anxiety and guilt in 50 percent (#1459).

Self-harm and suicide are where the research becomes an accounting problem. People in solitary confinement are roughly 6 to 8 percent of the incarcerated population and account for approximately half of prison suicides (#1447, #6631, #5922). In a study of 244,699 New York City jail admissions, 7.3 percent of admissions involved any solitary confinement, but 53.3 percent of self-harm acts and 45.0 percent of potentially fatal self-harm occurred in that group; after controlling for length of stay, serious mental illness and demographics, people punished with solitary were 6.9 times more likely to commit self-harm (#6625, #6626). A separate cohort found people ever assigned to solitary 3.2 times as likely to commit an act of self-harm per 1,000 days (#20615), and people with mental illness in solitary approximately 7 times more likely to self-harm than those in the general population (#1467). Among 119 recently released people, 28 percent screened positive for PTSD symptoms, rising to 43 percent among those with solitary exposure (43 percent of participants) against 16 percent among those without (#7278, #7332). Post-release mortality follows the same line: any time in restrictive housing was associated with a hazard ratio of 1.24 for death in the first year after release, 1.78 for suicide, 1.54 for homicide and 2.27 for opioid overdose in the first two weeks (#20524, #20526).

Professional bodies have treated this as settled. The American Psychiatric Association's December 2012 position statement (retained 2017) says prolonged segregation of prisoners with serious mental illness "with rare exceptions, should be avoided," defining prolonged as longer than three to four weeks (#6628, #6887) — a standard that, on the SMU duration data, almost no one in Georgia's SMU could meet. The American Public Health Association called solitary a public health crisis and called for an end to long-term use (#6629); the National Commission on Correctional Health Care recommends prohibiting it for seriously mentally ill, juvenile and pregnant people and limiting it to under 15 days for everyone else (#6630). The U.S. Department of Justice states that segregation can cause severe psychological damage, especially with near-complete isolation and sensory deprivation or over prolonged periods (#20405, #25955). Retired SMU warden Jose Morales described the trajectory from the inside: "from the day that they arrive, to a year later, they decompensate. They're normal mentally, but then they start talking to themselves. They start to hallucinate" (#21868).

The dead

Two men died by suicide in the Special Management Unit at Georgia Diagnostic and Classification Prison in 2017 (#6603). Nearly half of those assigned to the unit had documented mental disorders requiring treatment (#26227). The court in the SMU litigation recorded the plaintiffs' claim that people with serious mental illness remained in the unit even after the Tier III conditions caused them to decompensate and engage in self-harm (#20411). In one documented federal case, a transgender woman who filed Prison Rape Elimination Act complaints and asked to be moved from isolation had her request rejected and died by suicide in the isolation unit the next day (#25948, #20406). Jenna Mitchell died by suicide in solitary confinement at Valdosta State Prison on December 6, 2017, after her mother reported suicide threats to the warden; her family settled for $2.2 million in 2021 (#6774).

In February 2023 a man was found dead in his restrictive-housing cell at Calhoun State Prison, leaning against the door and wrapped in mattress padding. No one had entered the cell for two days; staff had shut off his water supply and closed the chow flap. The cause was dehydration with renal failure (#6902, #6655, #22479). GDC's own internal audits in 2023 found severe lapses in staff and supervisor rounds in segregation units in at least nine prisons, including long gaps between checks, days with no checks at all, and check sheets that were uniformly incomplete (#2621). A retired SMU warden described the arithmetic: "We were required to conduct 30-minute checks at a minimum, and if we had anyone on suicide watch, they would have to be checked every 15 minutes … When you have a short staff of six people out of a 27-man post, you can't do it" (#21924).

GPS's death registry records the following people, each of whom the evidence above places in isolation, segregation or a restrictive-housing cell, or whose death the record ties to conditions in one. Miguel Angel Duran died March 1, 2026, at Central State Prison; GPS's registry lists the cause as Suicide, age 44 (#6908). Justin Waymon Hollingsworth died June 26, 2025, at Rogers State Prison; the registry lists the cause as Suicide, age 43 (#6909). Sheqweetta Vaughan died July 9, 2025, at Arrendale State Prison; the registry lists the cause as Unknown/Pending, age 32 — she was found in a segregation cell in the 90s Fahrenheit, and the Georgia Bureau of Investigation could not determine cause or manner of death (#6894). Ricardo Castavio Daughtry died June 10, 2024, at the Special Management Unit; the registry lists the cause as Unknown/Pending, age 41. Daughtry was the lead plaintiff in the SMU class action (#21904). Dominique Cornelius Cole died June 8, 2025, at Wilcox State Prison; the registry lists the cause as Homicide, age 37. Carlito Omar Soldiew-Acosta died June 22, 2024, at Dooly State Prison; the registry lists the cause as Homicide, age 38 — he was found in his isolation cell and had been dead for at least 24 hours, if not longer (#25413). Bedarius Jarkeez Clark died August 21, 2021, at Baldwin State Prison; the registry lists the cause as Homicide, age 26 — he was found unresponsive in the prison's segregation unit (#25143). Thomas Henry Giles died October 28, 2020, at Augusta State Medical Prison; the registry lists the cause as Homicide, age 31. Christina Marie Buttery died December 21, 2022, at Pulaski State Prison; the registry lists the cause as Unknown/Pending, age 34. Melissa Christine Thacker died July 25, 2021, at Pulaski State Prison; the registry lists the cause as Suicide, age 27 (#24865).

Two further deaths come from GDC's own cause coding in the 2015–2019 series, obtained by the UCLA Law Behind Bars Data Project through an open records request. Those causes are the Department's coding, not autopsy findings — no pathologist examined the body for that determination. James Wheeler died October 29, 2017, at Wilcox State Prison; GDC's coding lists the cause as Suicide, age 39. He had a history of self-harm and was placed in solitary before he hanged himself (#6798). Demitri Carter died October 29, 2017, at Phillips State Prison; GDC's coding lists the cause as Suicide, age 25, after multiple prior attempts (#6799).

The legal standard Georgia has not been held to

There is no Supreme Court ruling directly on whether solitary confinement is constitutional (#1584). The most significant statement remains In re Medley in 1890, where the Court recognized solitary's severe psychological effects — prisoners falling into a "semi-fatuous condition," becoming "violently insane," or committing suicide (#1508). In 2023 the Court denied certiorari in Hope v. Harris, involving a man held in continuous solitary for 27 years, 22 to 24 hours a day in a 54-square-foot cell, who had received one personal phone call since 1994 (#1507, #1548, #1549). Justice Kennedy's concurrence in Davis v. Ayala — "years on end of near-total isolation exact a terrible price" — is a signal, not a holding (#6632).

The operative doctrine is that solitary is not per se unconstitutional but becomes so when duration, conditions and individual vulnerability combine to inflict unnecessary and wanton pain or to deprive identifiable human needs (#1583). Madrid v. Gomez in 1995 held that conditions at Pelican Bay's SHU imposed cruel and unusual punishment on mentally ill prisoners, likening prolonged solitary for the mentally ill to "the mental equivalent of putting an asthmatic in a place with little air to breathe," and ordered them removed (#6635, #6876). The Third Circuit found in 2020 that 33 years of solitary can demonstrate cruel and unusual punishment (#1504), and in 2024 held in Williams v. Secretary that "[i]t was clearly established that someone with a known preexisting serious mental illness has a constitutional right not to be held in prolonged solitary confinement without penological justification," rejecting qualified immunity (#6636). The Sixth Circuit reached a similar result in Finley v. Huss in 2024 (#1585).

Georgia sits in the Eleventh Circuit, which has not issued a definitive ruling on the constitutional limits of solitary confinement under the Eighth Amendment (#1503). The Georgia cases that exist are narrow: Canupp v. Paul, where the court had "little trouble" finding an Eighth Amendment violation for 14 days in an isolation cell with no running water (#21554), and Turner v. Warden, where a cold strip-cell claim from the Special Management Unit failed (#20001). The Supreme Court's one direct intervention on segregation conditions in a southern prison system was Hutto v. Finney, which upheld a 30-day cap on punitive isolation as part of a comprehensive remedy and affirmed fee awards against the Arkansas Department of Correction (#26379, #26378).

The due-process angle runs the other way from the Eighth Amendment one. Wilkinson v. Austin in 2005 found that near-total solitary, indefinite duration and parole disqualification together created a liberty interest requiring procedural protections (#19327). Indefinite duration is present in Tier II, which has no outer limit, and absent from Tier III, which has a 24-month ceiling — so, on Wilkinson's own reasoning, the procedural argument is stronger for the program the litigation never touched (#19327). GDC's courts have instead produced a consent decree: since 2019 the Department has been subject to one covering SMU conditions, and the April 2024 contempt finding was issued under it (#20404, #21547).

The Department of Justice's October 2024 findings report adds the federal government's own conclusions about the wider system: that GDC "fails to control violence even in its segregated housing units and exposes incarcerated persons to an unreasonable risk of harm due to its inappropriate use of segregated housing" (#6904); that people at Georgia Diagnostic spent weeks in segregation without out-of-cell recreation (#21777); that people at Wilcox in segregation received no required services, programs or outside recreation (#22663); and that queer and transgender prisoners reported being placed in solitary after reporting sexual assault or other violence, or because they were in mental health crisis — making solitary a punitive response to victimization (#1536). The report recommends that GDC "[c]onduct a review of restrictive housing unit practices and remedy all noncompliance" with its own administrative segregation policy and applicable legal standards including PREA (#20877).

What Georgia does not publish, and what has worked elsewhere

The central evidentiary problem is that Georgia does not publish a Tier population. A systematic review of 110 non-monthly GDC reports and all 34 distributions in the monthly Inmate Statistical Profile series found no published count of the Tier or restrictive-housing population in any of them; the phrase "restrictive housing" is absent from all 22 GDC open-records productions in GPS's archive, and "Tier II" from all 261 (#19306). The only reason a Tier III figure exists is that the SMU is carried as a facility row (#19304). GDC has also not disclosed suicide and self-harm incidents by tier and facility, mental health classification breakdowns of the Tier II and Tier III populations, or length-of-stay distributions for restrictive housing (#6669), and it stopped publishing cause-of-death data after February 2024 (#6893). Georgia submitted no response at all to the 2021 ASCA-Liman survey of time-in-cell (#20818).

Where GDC has published numbers, they do not reconcile. The FY2018 Annual Fiscal Report states that "[s]ince January 2017, the number of offenders in extended restrictive housing has decreased by 41 percent (671)" (#19298). The FY2019 report states that the Extended Restrictive Housing population was reduced "by 40% (675) since January 2017" (#19299). Two different deltas, two different percentages, one stated baseline, and never a population at either end. Back-solving a baseline from those pairs yields roughly 1,640 or 1,690 depending on which statement you use; those are GPS arithmetic on inconsistent agency statements and must never be cited as a GDC-published figure (#19299). The shape of the disclosure — a delta with no endpoints — is itself the finding.

Several named records would answer the question the published data cannot. GDC's Tier I and Tier II policies require per-cell, per-bed accountability logs recording who is housed in each Tier cell and bed, updated to the minute (#19296). Board of Corrections Rule 125-3-2-.10(f) prohibits confinement in an isolation cell beyond 30 days without the review and approval of a Division Director (#19350), and a Central Office approval record exists for every such confinement — counting those approvals for a defined period would give a floor on long-stay isolation independent of any Tier figure (#19351). Tier III also generates a specific written record whenever out-of-cell time is cancelled for three or more consecutive days (#19282); no equivalent exists for Tier I, Tier II or the juvenile program.

The comparative record is not ambiguous. North Dakota, working with Amend at UCSF from 2015, reduced solitary confinement by more than 74 percent, with significant decreases in restrictive housing placements for people with serious mental illness (#4685, #7659, #7761). Colorado cut its administrative segregation population from approximately 1,500 people — about 7 percent of its prison population — in 2011 to under 200 by 2017, then largely eliminated it (#6665); after banning solitary in two mental health prisons, assaults, forced cell entries and the use of heavy restraints declined by 40 percent (#6666). Oregon achieved reductions of 55.7 to 73.9 percent (#7762). Haney testified that Mississippi, which drastically reduced supermax administrative segregation by transferring people to mainline prisons, experienced an overall reduction in misconduct and violence system-wide (#19005). New York's HALT Act limits solitary to 15 consecutive days and bars it entirely for people 21 or younger, 55 or older, pregnant or postpartum, disabled, or with serious mental illness (#1492, #1493); Connecticut and Nevada enacted 15-day maximums in 2023 (#1498); New Jersey set a 20-day maximum in 2020 (#1499); and Colorado, Delaware, North Dakota and Vermont reported holding no one under restrictive-housing definitions at all (#1500). Against those benchmarks, Georgia's Tier II program has no outer limit in policy (#19252), and its own auditors found in 2025 that the five-hour weekly exercise floor was not being delivered (#21525).