Solitary Confinement

What Georgia Runs, and What the International Standard Calls It

The international benchmark is precise. Rule 44 of the UN Standard Minimum Rules for the Treatment of Prisoners — the Nelson Mandela Rules — defines solitary confinement as “the confinement of prisoners for 22 hours or more a day without meaningful human contact,” and defines prolonged solitary confinement as solitary confinement in excess of 15 consecutive days; Rule 43 prohibits indefinite and prolonged solitary confinement in all circumstances (#1476, #6619, #1450, #1477). In 2011, UN Special Rapporteur on Torture Juan E. Méndez concluded that “any imposition of solitary confinement beyond 15 days constitutes torture or cruel, inhuman or degrading treatment or punishment” (#6620, #1479); in February 2020 Special Rapporteur Nils Melzer reaffirmed that “prolonged or indefinite isolation … may amount to torture” (#6621, #1480). The Mandela Rules are persuasive authority in the United States, not enforceable law (#19329).

Georgia does not use the word. GDC runs a “Tier Segregation Management System,” implemented in August 2013 and described in the agency's own Facilities Division fact sheet as a stratification pathway for moving an offender from administrative segregation to lower security levels (#6593, #20860). The vocabulary is Tier I (short-term disciplinary, protective-custody and transient housing), Tier II (long-term administrative segregation) and Tier III (the Special Management Unit at Georgia Diagnostic and Classification Prison), alongside generic Administrative Segregation under SOP 209.06 and a juvenile variant, RHA-JOAS, under SOP 209.11 (#19247, #19252, #19260, #19265). A GDC press release states the SMU is one of 10 GDC facilities with a Tier Program (#20885), and the agency's Close Security Facility Fact Sheet names Hancock, Hays, Macon, Smith, Telfair, Valdosta and Ware as designated Tier I & II facilities (#6595). Tier I, however, is authorized at every facility type in the system — state prisons, county correctional institutions, private prisons, intensive treatment facilities, transitional centers, probation detention centers and probation boot camps (#19247, #19363).

The scale was set at implementation. GDC's FY2014 Annual Fiscal Report states the department “developed and implemented the Tier Segregation Management System at all Level V facilities” and that, “as part of this system, 2,721 beds were converted to Tier beds” (#19297) — the only system-wide quantification of the Tier system GDC has ever published, and a bed count rather than a population count. The confinement floor is written into the policies. Tier I guarantees five hours per week of out-of-cell exercise, at least one hour a day on five days (#19270); Tier II guarantees “a minimum of five (5) hours per week,” with no daily distribution specified at all (#19269); generic Administrative Segregation guarantees one hour a day, five days a week (#19271). Five hours out of 168 yields 163 hours in cell, or 23.3 hours a day, with zero guaranteed out-of-cell time on the two non-exercise days (#19272). That arithmetic is GPS's, derived from GDC's stated minimums; it describes a policy floor, not what any individual experiences.

For population, the best independent figure comes from the 2019 ASCA-Liman Time-in-Cell survey, which reports Georgia's restrictive-housing population at 2,147 — 4.9 percent of the 44,073 people in reporting Georgia facilities, against a weighted average of 5.0 percent across eleven responding jurisdictions (#20817, #20819). The most recent figure comes from a GDC Tier Management Bed Report obtained through an open records request by the Southern Center for Human Rights: as of October 2025, more than 2,100 people were housed in Georgia's “tiers,” including almost 1,000 in the long-term tiers (#20801, #20869). A single undated GDC bed report from early October 2026 gives Tier II as 8 prisons, 884 beds and 804 occupied; Tier III as one location, 160 beds and 128 occupied; and the three programs together — Tier II, Tier III and the STEP step-downs — as 1,163 beds with 1,028 occupied (#20762, #20763, #20765).

Whether any of this is solitary confinement in the Mandela Rules sense turns on cell occupancy, and GDC does not answer the question. Rule 44 has two elements: 22 or more hours a day in cell, and the absence of meaningful human contact. GPS can establish the first by arithmetic from GDC's own minimums. It cannot establish the second from the SOPs, and does not assume it. SOP 209.09 guarantees single occupancy in Tier III, SOP 209.55 guarantees it in Tier III STEP, SOP 209.45 permits double occupancy in Tier II STEP, and SOP 209.08 says nothing about occupancy in Tier II at all; SOP 209.06 treats single-celling in administrative segregation as an enumerated exception requiring assignment by medical, mental-health or classification staff plus evaluation within 24 or 48 hours (#19278, #19331, #19279, #19404). The implication of that drafting is that multiple occupancy is the ordinary condition. Where a Tier II cell holds two people, the 23.3 hours a day are served with a cellmate present — still 23 hours locked in a cell, but not “without meaningful human contact” in the Rule 44 sense. Two conditions of confinement are therefore running inside the same program under the same rules, and the policy distinguishes neither, guarantees neither, and reports neither (#19510).

The Special Management Unit: 78 Percent Held Past Two Years

The SMU is the one part of the Tier system GDC quantifies, because it is carried as a standalone facility with its own facility code in the monthly Inmate Statistical Profile series. It has a stated capacity of 192 single-bunked beds in six cellblocks (#19302, #19484, #6600), and as of September 1, 2026 it held 154 people; across 75 monthly snapshots from July 2020 to September 2026 the population ranged from 149 to 195 (#19304). That series implies the unit has run below its bed capacity throughout. SCHR reported a steeper decline after the 2019 reforms, from 180 in October 2017 to about 100 (#20867).

What the unit was in 2017 is what the litigation record documents. The Gumm court adopted findings, based on Dr. Craig Haney's expert report, that SMU residents were confined in cells “smaller than the average parking space” — approximately 6 feet by 9 feet — with as little as five hours per week out of cell, no outside light, no congregate religious or educational programming, and frequently locked food-port flaps (#19326, #6598, #1511). People in the SMU were confined to isolation cells for nearly 24 hours a day on average as of 2017, unable even to see out of a window, and a number of inmates were confined literally 24 hours a day for months at a time (#6604). In the most restrictive cellblocks, prisoners had no books or personal property and could not leave their cells for a minimum of 90 days on arrival (#1512).

The durations are the finding that distinguishes the SMU from anything most states operate. As of July 2017, 78 percent of SMU prisoners — 141 of 182 — had been held in isolation more than two years (#1448, #7286). Forty-four percent (80 of 182) had been held more than four years (#1516); 26 percent (47 of 182) more than five (#1517). SCHR, drawing on the Gumm expert record, put “about 20 percent” of SMU inmates at six or more years and the average stay at three to four years (#6609, #6610). The individual cases are in the court record: Timothy Gumm was held continuously in the SMU for seven and a half years, with 14 separate recommendations over four years that he be transferred out (#1518, #19326, #6612); Johnny Mack Brown for nine years (#1519); Robert Watkins for eight to ten (#1520, #6613); Daniel Barfield for eight (#6611). People were released directly from the SMU to the community at sentence expiration with no transitional programming (#6677, #19326).

Dr. Craig Haney inspected the unit in October 2017 and described it as “one of the harshest and most draconian” facilities he had seen “in decades of conducting evaluations,” with prisoners “among the most psychologically traumatized persons [he] ha[d] ever assessed in this context,” concluding that “some of the inmates' psychological harm … may be irreversible and even fatal” (#6624, #1510, #20412). His report documented a cell block full of inmates with serious mental illness, a man locked for months inside a pitch-black cell, and another man, naked and psychotic, whose cell was covered in blood (#6643). At that inspection the SMU held approximately 180 people, of whom 70 — 39 percent — were designated mentally ill, and Haney's opinion was that it was “dangerous” to house mentally ill people there (#6615, #6606, #1449, #20413, #20414). The 2017 record also includes two suicides in the SMU (#6603), and the plaintiffs in the later contempt litigation cited the suicides of two SMU inmates as evidence that people with serious mental illness remained in the unit after the conditions caused them to decompensate and engage in acts of self-harm (#20411).

The conditions documented in the 2024 contempt proceedings were no less severe. GDC officials placed people in “strip cells” on arrival at the SMU, taking their clothing and leaving them naked or near-naked for hours or days (#1526). One prisoner testified to a cell whose broken toilet was filled with feces and urine from prior occupants, where he was forced to urinate in a cup and pour it into the sink, had no mattress and no clothing, and was held in freezing temperatures; the GDC attorney did not refute the testimony (#1528). Six prisoners testified about being denied showers, out-of-cell time, programming, cell cleanout, and access to kiosks and book carts (#1529). Under the Mandela Rules' 15-day threshold, even the unit's reformed parameters sit far outside the international standard: the Tier III program has a stated minimum of 390 days (60 + 60 + 90 + 90 + 90) and an outer limit of 24 months — 26 and roughly 49 times the 15-day threshold respectively, before the qualifications discussed below (#19260, #19329, #19330). And the SMU is the only part of the system where GDC has been required to guarantee single occupancy, four hours a day out of cell Monday through Friday, general-population-standard food, weekly library access, tablet access, and written notification to a Regional Director whenever out-of-cell time is cancelled for three consecutive days (#19281, #19282).

Tier II: The Larger Program With the Fewer Protections

SOP 209.08, “Administrative Segregation – Tier II,” effective April 11, 2016, governs Georgia's long-term administrative segregation program and has not been revised since — nearly three years before the Gumm settlement (#19324). Its own introduction states that the program “is an offender management process and is not a punishment measure,” a characterization the published summary repeats (#19291). Three features of the policy sit in tension with that framing.

First, it sets no maximum duration. “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 to determine whether to keep them there — indefinite administrative segregation as a matter of written policy (#19252). Second, the discipline SOP routes people into it: SOP 209.01 caps disciplinary isolation at 30 consecutive days in prisons and 14 days in probation detention centers and integrated treatment facilities, and provides that when isolation time exceeds 30 days because it is necessary to control and encourage behavioral change, a recommendation for placement in the Tier II program is warranted (#19292, #19352). The effect written into the policy is that Tier II is the mechanism for continuing confinement once the disciplinary-isolation limit has been reached. Third, the admission criteria are broad enough to be satisfied without any adjudication. Criterion 1 covers an offender “noted as a threat to the safe and secure operation of the Facility,” which “may include, but is not limited to” documented security-threat-group activity, notoriety of crimes, high supervision requirements, and offenders who “have either been threatened with bodily harm or threatened others with bodily harm” — the last of which describes a victim (#19253). Criterion 5 is failure in or refusal to participate in Tier I (#19487); criterion 10 is “assaultive histories” (#19253). The asymmetry between admission and release is written into the policy: a facility can place someone in Tier II on its own authority, including immediately in an emergency, but release requires a statewide Criminal Investigations Unit clearance and the personal decision of the Director of Field Operations (#19256). The 90-day review that continues confinement is appealed to the Warden — the same official who approves placement (#19258). Section IV.D.1.l allows the Warden or designee to modify every condition and privilege in the policy on a finding of undefined “extenuating operational circumstances,” with no time limit, no documentation requirement, no notice to the prisoner, and no reporting obligation to Central Office (#19276).

Duration figures for Tier II appear in GDC's own PREA facility audits but nowhere in the SOP. The Hays State Prison Cycle 2 audit describes Tier II as “long-term segregation/270 days”; Macon's Cycle 2 and Cycle 4 audits describe it as running “from nine months to indefinite” and “from 270 days to an indefinite date”; Georgia State Prison's Cycle 2 audit records that Tier II offenders “may take up to two years to progress” through the phases (#19494, #19496, #19497, #19478, #19403). Georgia told the ASCA-Liman survey that it had “Tier II step down units for offenders on phase 3+, who have been in restrictive housing for 270+ days” (#20814). None of these figures has any basis in SOP 209.08, which states no duration at all (#19403).

On mental health, the effect written into the policy architecture is that prisoners with the higher mental-health classifications are routed away from the reformed unit and remain eligible for the unreformed one. SOP 209.09 bars both MH Level III and MH Level IV offenders from Tier III and directs them to a Specialized Mental Health Treatment Unit. SOP 209.08 excludes only MH Level IV; MH Level III may be placed in Tier II on a mental-health recommendation (#19254). GDC's own mental-health staffing policy, SOP 508.01, classifies “Specialized Mental Health Treatment Units (SMHTUs/TIER II, etc.)” among the highest-need units in the system, alongside death row, the SMU and acute care, and assigns them a ratio of at least one mental health counselor per 20 offenders (#19288). The SCHR complaint alleged that over 70 percent of the approximately 300 people in Georgia State Prison's Tier II program had serious mental illness (#20887, #6614) — an allegation in a complaint, not a court finding, at a facility that closed on February 19, 2022 (#19369).

Tier II is also the larger program by a wide margin, and it is nearly full. On the October 2026 snapshot, Tier II comprised 884 beds across eight prisons against the SMU's 160, and four of the eight were at or within one bed of full: Hays at 151 of 151, Telfair at 117 of 117, Ware at 138 of 138, and Hancock at 82 of 83; Macon was at 119 of 121, and Valdosta had 49 of its 129 beds empty (#20762, #20750, #20754, #20756, #20749, #20751, #20755, #20766). Hays alone holds the largest restrictive-housing program outside the SMU — 151 Tier II beds plus 59 STEP beds, 210 in all, every one occupied (#20767). The protections won in Gumm attach only to Tier III: the 24-month ceiling, the mental-health exclusion, the 60-day and 90-day out-of-cell evaluations, the four-member statewide retention panel and the personal approval requirement all appear in SOP 209.09 and none of them appears in SOP 209.08 (#19324, #19261). The program with the least protection is the larger one.

What Isolation Does: The Clinical Record and the Deaths

The clinical literature on solitary confinement is unusually old and unusually consistent. Dr. Stuart Grassian identified “SHU syndrome” in a 1983 article in the American Journal of Psychiatry based on clinical observations of 14 men in long-term solitary in Massachusetts, describing hypersensitivity to external stimuli, affective disturbances, difficulties with thinking, concentration and memory, disturbances of thought control and problems with impulse control; in severe cases he described “florid delirium — a confusional psychosis with intense agitation, fearfulness, and disorganization” (#6623, #7277, #1456). He later wrote that solitary conditions are “strikingly toxic to mental functioning,” that even psychologically resilient inmates suffer “severe psychological pain,” and that the harm may produce “prolonged or permanent psychiatric disability” (#1455, #1457). Dr. Craig Haney's 2003 study documented symptom prevalence among prisoners in solitary: 91 percent reported 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 review published in Crime and Justice in 2018 found the research on psychological harm from solitary has been “strikingly consistent since the early nineteenth century” across methodologies, populations and time periods (#1454).

The quantitative syntheses point the same way. A meta-analysis of solitary confinement and psychological symptoms pooled five studies to a standardized mean difference of 0.45 (95% CI 0.29–0.61, p < 0.001) (#20586). A systematic review covering 13 studies and 382,440 inmates, 23 percent of whom had been exposed to solitary, found higher-quality evidence that solitary was associated with increases in adverse psychological effects, self-harm and mortality (#20587), and analysis showed the deterioration went beyond that attributable to incarceration generally or to prior mental illness (#20588). A meta-analysis published in PLOS One in June 2025 synthesized data from 171,300 inmates and found significantly greater psychological distress, more psychiatric symptoms including self-harm and thought disorders, and greater need for mental health services and hospitalization among prisoners in disciplinary confinement (#1458). A 2024 Washington State study of 106 randomly sampled prisoners in long-term solitary found clinically significant depression, anxiety and guilt in 50 percent of participants, with qualitative themes of social isolation, loss of identity and sensory hypersensitivity (#1459, #1560). Hagan et al. (2018) found that among 119 recently released individuals, 28 percent screened positive for PTSD, rising to 43 percent among those with solitary-confinement exposure against 16 percent among those without (#7278, #7332). Research has also linked any time in solitary to premature death after release — all-cause hazard ratio 1.24, suicide 1.78, homicide 1.54, and opioid overdose within the first two weeks 2.27 (#20524, #1466) — with a North Carolina cohort showing adjusted hazard ratios of 1.81 for suicide and 1.61 for homicide within a year of release after more than 14 consecutive days in restrictive housing (#20526).

Self-harm concentrates in isolation. Kaba et al., analyzing 244,699 incarcerations in the New York City jail system, found that although only 7.3 percent of admissions involved any solitary confinement, 53.3 percent of acts of self-harm and 45.0 percent of potentially fatal self-harm occurred in that group; after controlling for length of stay, serious mental illness and demographics, individuals punished with solitary were 6.9 times more likely to commit self-harm (#6625, #6626). Across sources, people in solitary make up approximately 6–8 percent of incarcerated populations and account for approximately half of suicides in correctional settings (#1447, #5922, #6631, #1467). In Georgia, the DOJ's October 2024 findings report states that segregation “can cause severe psychological damage, especially when it involves near-complete isolation and sensory deprivation, or when the segregation extends for a prolonged period of time,” and recounts a transgender woman placed in isolation after filing PREA complaints who was denied a request to be moved and died by suicide in the isolation unit the next day (#20405, #20406, #20460). The same report found that “GDC fails to control violence even in its segregated housing units” (#6904). GDC's own 2023 internal audits found severe lapses in staff and supervisor rounds in segregation units in at least nine prisons; at one medium-security men's prison the audit found long gaps between checks, some days with no checks at all, and that all check sheets reviewed by the auditor were incomplete (#2621).

People have died in these units. GPS's death registry records the following. Miguel Angel Duran died March 1, 2026, at Central State Prison; the registry lists the cause as suicide, age 44, and the underlying record places the death in segregation. Justin Waymon Hollingsworth died June 26, 2025, at Rogers State Prison; the registry lists the cause as suicide, age 43. Sheqweetta Vaughan died July 9, 2025, at Arrendale State Prison; GPS's registry lists her cause as Unknown/Pending — GPS does not know it. Hallie Marie Reed died May 5, 2024, at Arrendale State Prison; the registry lists the cause as homicide, age 23. Sherry Elaine Joyce died April 27, 2024, at Arrendale State Prison; the registry lists the cause as homicide, age 61. Christina Marie Buttery died December 21, 2022, at Pulaski State Prison; the registry lists the cause as Unknown/Pending, age 34. James Wheeler died October 29, 2017, at Wilcox State Prison at age 39, and Demitri Carter died October 29, 2017, at Phillips State Prison at age 25; both are coded as suicide in the Department's own 2015–2019 cause-of-death data, obtained by the UCLA Law Behind Bars Data Project through an open records request. Those two causes are GDC's administrative coding, not autopsy findings — no pathologist examined the body for that determination. GPS's underlying record describes Reed's and Joyce's deaths as occurring in the same Lee Arrendale mental health unit, eight days apart (#5088).

Other documented deaths sit alongside them. DOJ documented that in February 2023 an incarcerated person was found dead in his restrictive-housing cell at Calhoun State Prison, wrapped in mattress padding, after no one had entered the cell for two days; staff had shut off his water supply and closed the chow flap, and the cause of death was dehydration with renal failure (#6902, #124, #6655). Jenna Mitchell, a transgender woman held in solitary at Valdosta State Prison, died by suicide on December 6, 2017, after her mother reported suicide threats to the warden; her family settled for $2.2 million on December 6, 2021 (#6774). The estates of James Wheeler and Demitri Carter settled for $750,000 and $700,000 respectively, both in 2021 (#6798, #6799).

The Counting Problem: What GDC Publishes and What It Withholds

Georgia publishes no count of the people in Tier segregation. A systematic review of the GDC report corpus GPS holds — 110 non-monthly GDC reports and all 34 distribution sections of the monthly Inmate Statistical Profile series — found restrictive-housing language in five years of annual fiscal reports, a reduction percentage and a delta in two, and never a population (#19306). The monthly profile series publishes 34 distributions and none reports segregation, restrictive housing or Tier status; “institution_type” offers only county CI, county jail, private prison, probation detention center, RSAT center, state prison and transitional center, and “supervision” offers only close, medium and minimum (#19306). GPS's own open-records archive contains no production in which the phrase “restrictive housing” appears, and none in which “Tier II” appears (#19306). GDC has not publicly disclosed suicide and self-harm incidents by tier and facility, mental-health classification breakdowns of the Tier II/III population, or length-of-stay distributions for restrictive housing (#6669). In GPS's own mortality data, the segregation flag is populated on only 1 of 67 recorded suicides, so GPS cannot speak to the solitary-suicide association from its own records at all (#8408). Georgia also did not answer the 2021 ASCA-Liman survey; the comparison table lists its restrictive-housing count as “0 (No 2021 Survey Response)” against a custodial population of 46,315 (#20818).

What can be counted is counted by accident. Because the SMU is a standalone facility, its population appears in the monthly facility table: 154 people on September 1, 2026 (#19304). Tier I and Tier II sit inside ordinary prisons and are absorbed invisibly into those prisons' facility totals (#19304). The one alternative source is GDC's PREA facility audits, which often state segregation cell counts. Summing the most recent audit per facility that states a figure gives 3,401 segregation cells across 30 facilities — 3,182 across 28 state-operated facilities — and that is a floor, not a total (#19398, #19504). On a single-cycle basis, the 23 state facilities with a structured Cycle 2 figure total 2,967 cells (#19505). The facility figures range from 409 at Georgia Diagnostic and Classification to 4 at the Long Unit (#19438, #19462), and they include large numbers at medium-security and private facilities — 197 at Calhoun, whose Cycle 2 audit gives its security level as “4/Medium,” and 170 at the private Coffee Correctional Facility (#19442, #19444). Other single-facility counts include 336 at Smith, 280 at Telfair, 252 at Hays, 223 at Hancock, 192 at Macon, 167 at Baldwin, 161 at Valdosta and 125 at Ware (#19439, #19440, #19441, #19437, #19443, #19445, #19446, #19447). Every state prison in the corpus that reports a number reports segregation cells (#19511), and double-bunked segregation and Tier housing are not exceptional in Georgia — they are the ordinary described condition at most of the facilities that describe the condition at all (#19528, #19395).

Three caveats belong on those numbers, and GPS states them rather than resolving them. First, they are cell counts, not bed counts and not population counts. The audits themselves say many of these cells hold two people: Lee State Prison states “16 cells with 32 beds,” Wilcox “72 double bunked segregation cells,” Montgomery “eight (8) two-man segregation cells,” Macon “48 double occupancy cells,” Telfair “40 single cells and 40 double-bunked cells,” and Smith “48 double bunked cells housing 96 offenders per dorm” (#19398, #19526, #19428, #19469). The bed count is materially higher than 3,401, and the number of people is a separate unknown; GPS has interpolated neither. Second, the total mixes vintages spanning Cycle 2 (2017–2019) through Cycle 4 (2021–2025), and the most recent audits of five of the nine Tier prisons dropped the segregation-cell field altogether (#19398). Third, it counts all segregation, not only Tier program housing; the Tier units are a subset (#19398). It must not be presented as a trend against GDC's FY2014 statement that 2,721 beds were converted to Tier beds — a bed count at implementation across all Level V facilities, a decade earlier, and a different measure (#19398, #19297). For a sense of scale at one prison, GDC's FY2018 report records that automated segregation-rounds technology installed at Valdosta State Prison captured “all tier units (178 Cells)” (#19301) — one facility's Tier units approaching the size of the entire SMU.

The agency's own disclosures about reduction are shaped the same way. GDC's FY2018 Annual Fiscal Report states that “since January 2017, the number of offenders in extended restrictive housing has decreased by 41 percent (671),” achieved partly through STEP, Specialized Protective Custody housing, Specialized Mental Health Treatment Units, administrative reviews, changes in the discipline process and policy restructuring (#19298, #20780). The FY2019 report states that “Facilities Division has reduced the Extended Restrictive Housing population by 40% (675) since January 2017” (#19299). Both statements report a delta with no endpoints: GDC states how many people left and what percentage that was, and never states the population before or after. The two are also not reconcilable to a single baseline — the same January 2017 starting point yields 671 people at 41 percent in one report and 675 people at 40 percent in the next, and the FY2018 report's own Commissioner letter attributes the 41 percent reduction to FY2018 alone while its body attributes it to the eighteen months since January 2017 (#19299, #19298). GPS has back-solved approximately 1,640 and 1,690 from those pairs, and flags those numbers explicitly as GPS arithmetic on inconsistent agency statements that must never be cited as a GDC-published population figure (#19299).

Two structural facts make the gap closable, and GPS has identified a current SMU population count — with duration, mental-health diagnosis and demographic breakdown — as a priority open-records target (#1551). SOP 209.07 and SOP 209.08 both require per-cell, per-bed accountability logs recording cell number, bed number, offender name and number, race, and in and out times (#19296). And GDC maintains a single statewide Tier Segregation Manager whose function is to place people within the Tier system, who by necessity holds or can produce a statewide view of Tier placements (#19257).

Reform: What the Courts Won, What Other States Did, and What Georgia Left Untouched

Gumm v. Ford, the class action over the SMU at Georgia Diagnostic and Classification Prison, produced the most detailed set of solitary-confinement reforms in Georgia's history. The January 2019 settlement agreement required a minimum of three hours out-of-cell time in common areas plus one hour of outdoor recreation; within six months, two hours per week of computer time or educational classes and GED programs; prison-issued tablets in cells at all times; access to programming, mental-health evaluations and books; food servings consistent with general-population standards; a maximum of 24 months in the SMU except in narrowly defined circumstances; and committee review for transfer six to twelve months before release (#1523). The May 7, 2019 settlement order imposed the general 24-month limit, mandatory 60-day or 90-day reviews with out-of-cell mental-health evaluations by a licensed mental-health professional before assignment and at each review, a bar on housing Mental Health Level III or above prisoners in Tier III/SMU, and $425,000 in attorney's fees, with an initial three-year term subsequently extended (#6641). GDC codified the terms into SOP 209.09 alone: a 13-month (390-day) minimum program across five wings and five phases, a 24-month ceiling immediately qualified by six retention criteria, quarterly review of anyone held past 24 months by a four-member statewide panel of the Deputy Director, Field Operations, the Statewide Mental Health Director, the Statewide Medical Director and a member of the Office of Legal Services, and personal approval by the Director, Field Operations or the Assistant Commissioner for Facilities for any retention beyond 24 months (#19260, #19261, #19330).

Compliance failed anyway. In April 2024, Chief Judge Marc T. Treadwell found GDC in contempt, found the department had “no desire or intention” to comply with the December 2018 settlement, ordered an independent monitor, and imposed $2,500-per-day fines accumulating at $75,000 every 30 days for six months, plus additional attorney's fees (#6803, #20404, #20779). The court's April 19, 2024 order recounted falsified therapy records, including documenting that an inmate attended treatment sessions after he was already dead (#5952). The fee order credits the court-appointed monitor Gary Mohr — and the daily contempt fines — with the defendants having “either fixed or were in the process of fixing most problems in the SMU” after “years of defiance and prevarication” (#20900). The DOJ's October 2024 findings report notes both the consent decree and the April 2024 contempt finding (#20404, #20779), and recommends that GDC “[c]onduct a review of restrictive housing unit practices and remedy all noncompliance with GDC SOP 209.06 Administrative Segregation … and applicable legal standards including PREA” (#20788).

The structural limit of that litigation is scope. Every operative term is written in SMU and Tier III vocabulary; the class arose from conditions in a single facility; the agreement was codified into SOP 209.09; and SOP 209.08, which governs Tier II, contains no duration cap, no mental-health exclusion, no panel, no mental-health evaluation at placement, and no out-of-cell minimum beyond five hours per week (#19324). The programme with the least protection is the larger one, and it has gone unrevised since 2016. The 2025 and 2026 amendments to the Tier III SOP show which direction the policy has moved at the margins: the escape criterion for Tier III admission was loosened from two criteria requiring violence or multiple escapes within three years to a single criterion, “Escaped or attempted escape from a secure correctional facility or law enforcement custody,” with no violence condition and no look-back (#20856), and the mental-health evaluation for Tier III placement may now be performed by telehealth (#20857). In the legislature, House Bill 714 — which did not pass — would have prohibited consecutive periods in restrictive housing and capped it at no more than 90 days per year unless the offender is a danger to himself, herself, others, or the security of the facility (#20873, #20816). The 2024 Georgia Senate Study Committee recommended converting all existing facilities and constructing all new facilities to single-cell housing (#7583) and heard testimony defining solitary confinement as 22 or more hours a day in cell (#2952) — but a search of the Senate Press Office release announcing the committee's final report for the terms “segregation,” “solitary confinement” and “restrictive housing” returns no match (#20901).

Other states have moved further, and the outcomes are documented. New York's HALT Solitary Confinement Act, signed April 2021 and effective April 2022, limits solitary to 15 consecutive days, defines solitary as 17 or more hours a day in a cell, bars the practice for people 21 and under, 55 and older, pregnant or postpartum individuals, people with disabilities and people with serious mental illness, requires a minimum of four hours of out-of-cell programming daily, and mandates evidentiary hearings for all placements (#1492, #1493, #1494). Connecticut and Nevada enacted 15-day maximums in June 2023, aligning with the Mandela Rules; New Jersey's Isolated Confinement Restriction Act set a 20 consecutive day maximum in 2020 (#1498, #1499). Seven states enacted solitary-confinement legislation in 2021 alone — Arkansas, Colorado, Connecticut, Kentucky, New York, North Carolina and Tennessee (#1491) — and Colorado, Delaware, North Dakota and Vermont reported that they no longer housed anyone under their restrictive-housing definitions (#1500). At the federal level, the First Step Act (2018) codified a ban on juvenile solitary confinement and remains the only enacted federal legislation directly addressing the practice; the End Solitary Confinement Act, introduced in 2023 and reintroduced in 2025, has not passed (#1502, #1501).

North Dakota reduced solitary confinement by 74.28 percent between 2016 and 2020 through a partnership with Amend at UCSF, including significant reductions for people with serious mental illness (#7761, #7659, #4685), and Oregon reported reductions of 55.7 percent to 73.9 percent (#7762). Colorado reduced its administrative segregation population from approximately 1,500 people — 7 percent of the prison population — in 2011 to under 200 by 2017 and then largely eliminated it; following the ban in two mental-health prisons, assaults, forced cell entries and the use of heavy restraints declined by 40 percent (#6665, #6666). Virginia reduced its restrictive-housing population from 5 percent in 2016 to 4.1 percent in 2017 and 3 percent in 2018 (#20828), and the proportion of people in its Step-Down Program with a mental-health diagnosis dropped from 47 percent in February 2016 to 32 percent in July 2017 (#20829). Oregon's randomized comparison of its Step Up Program against the Intensive Management Unit found 10.0 percent of the SUP group maintained an MH-R mental-health classification against 53.9 percent in the IMU (p < .001) (#20825). South Carolina's Restoring Promise randomized trial produced an 83 percent reduction in restrictive-housing stays during the first year of participation (#4676, #5026), and Haney testified that Mississippi's reduction of supermax administrative segregation produced an overall reduction in misconduct and violence system-wide (#19005).

The legal road in Georgia is open, not closed. The Eleventh Circuit, which governs Georgia, Alabama and Florida, has not issued a definitive ruling on the constitutional limits of solitary confinement under the Eighth Amendment (#1503), and the Supreme Court has never ruled directly on whether solitary confinement is constitutional (#1584, #1508). In Hope v. Harris, the Court denied certiorari in 2023 in a case involving a man held in continuous solitary for 27 years in a 54-square-foot cell who received one personal phone call since 1994 (#1507, #1549, #1548). In Porter v. Pennsylvania DOC, the Third Circuit found that 33 years of solitary can demonstrate cruel and unusual punishment, and in Williams v. Secretary Pennsylvania DOC the Third Circuit held that “someone with a known preexisting serious mental illness has a constitutional right not to be held in prolonged solitary confinement without penological justification,” denying qualified immunity (#1504, #6636). Madrid v. Gomez established in 1995 that conditions at Pelican Bay SHU violated the Eighth Amendment as applied to prisoners with mental illness — “the mental equivalent of putting an asthmatic in a place with little air to breathe” — while declining to hold long-term SHU confinement of mentally healthy prisoners unconstitutional (#6635, #6876, #1557). Under current doctrine, solitary is not per se unconstitutional, but becomes so when duration, conditions and individual vulnerability combine to produce the unnecessary and wanton infliction of pain (#1583).