Georgia's Tier Segregation System

The ladder and the SOP machinery that builds it

Georgia's Tier Segregation Management System was implemented in August 2013, according to GDC's Tier Management Fact Sheet, consolidating a multi-tier restrictive-housing architecture [#6593, #20803]. The only system-wide quantification GDC has ever published of that build-out is in its FY2014 Annual Fiscal Report: “Developed and implemented the Tier Segregation Management System at all Level V facilities. As part of this system, 2,721 beds were converted to Tier beds” [#19297]. That is a bed figure from implementation in 2014 — not a current capacity, and not a count of people — but it establishes that the system was built at a scale of thousands of beds, and it supplies the meaning of “Level 5 facility” in the Tier II SOP: Level V is the GDC security class at which the Tier system was installed [#19297].

Four stages sit on the ladder, each governed by its own standard operating procedure. Tier I (SOP 209.07, effective 04/30/2015) is defined as the “Program for the involuntary and voluntary placement and management of offenders in short-term segregated housing,” and its applicability section reaches “All State Prisons, County CIs, Private Prisons, Intensive Treatment Facilities, Transitional Centers, Probation Detention Centers, and Probation Boot Camp” [#19247, #19363]. It is not limited to close-security prisons or to a designated subset of facilities. GDC's fact sheet describes Tier I as “the Disciplinary, Protective Custody and Transient Housing” with eight placement grounds [#19491].

Tier II (SOP 209.08, effective 04/11/2016) is titled “Administrative Segregation - Tier II.” The policy defines the program as a “Long Term Administrative Segregation stratification plan” and provides that 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]. It contains no maximum duration [#19252]. The GDC Fact Sheet quoted in PREA audits describes Tier II as “Administrative Segregation” with twelve eligibility criteria; the audits use the two terms interchangeably [#19492, #19486].

Tier III / Special Management Unit (SOP 209.09, effective 04/23/2025) is a “minimum 13-month (390 days) program,” requiring at least 60 days in E-Wing, 60 in F-Wing, 90 in D-Wing, 90 in C-Wing and 90 in B-Wing — arithmetic that sums to exactly 390 [#19260]. Section IV.A.1.f states that offenders “shall not be held in the Tier III Program for more than 24 months,” then immediately qualifies it with six criteria permitting longer retention, including a catch-all for an offender who “poses such an exceptional, credible, and articulable risk … that no facility other than the Tier III Program facility is sufficient to contain the risk” [#19260, #19330]. STEP completes the ladder: SOP 209.45 (Tier II STEP) and SOP 209.55 (Tier III STEP), both effective 04/25/2019, are textually near-identical except that Tier II STEP provides “Cells may be single or double occupancy” where Tier III STEP provides “All cells are single occupancy” [#19283, #19284].

The published record is thinner than the SOPs. The Fact Sheet as quoted in PREA audits describes no Tier III at all — its “Phase I, II and III” are phases within Tier II — and the only Tier III reference in the entire 273-document PREA corpus is the Georgia Diagnostic and Classification Prison Cycle 3 audit [#19392]. No SOP in the corpus lists which facilities are designated Tier facilities: SOP 209.08 says it “applies to designated Georgia Department of Corrections facilities” without naming them, and a keyword search of the 1,432-document GPS SOP corpus returns no designation list [#19264]. SOP 508.01, effective 2023, refers to “designated SMHTU/TIER II facilities” without naming them, which establishes that a current designation document exists inside GDC but is not published [#19370].

What the audits do establish is that named Tier program housing appears in the audits of exactly nine facilities — Georgia State Prison, Smith, Hays, Valdosta, Telfair, Macon, Hancock, Georgia Diagnostic and Classification, and Ware — and all nine are close-security prisons, with not one exception in 273 documents [#19359]. Seven of those (Hancock, Hays, Macon, Smith, Telfair, Valdosta, Ware) are the designated Tier I and II facilities per GDC's 2018–2019 fact sheets [#6595]; that list should be cited as GDC's designation as of 2018–2019, not as the current set, because Georgia State Prison closed in 2022, SOP 209.09 was revised in 2025, and a current designation document is missing [#19366].

GDC's framing of the programme travels with the evidence. SOP 209.08 Section I states: “This program is an offender management process and is not a punishment measure,” and SOP 209.06's published summary says “segregation is not a form of abuse or punishment” [#19291]. Set against that, SOP 209.01 caps disciplinary isolation at 30 days in prisons and 14 days in probation detention centres and integrated treatment facilities, and provides that where isolation “exceeds thirty (30) days due to being necessary to control and encourage behavioral change, a recommendation for placement of the offender in the Tier II program is warranted” [#19292, #19352]. SOP 209.06 also lists, among its own grounds for administrative segregation, “Serving sanctions for a violation of Facility/Center rules” [#19293].

Conditions: the policy floor, and the occupancy question it leaves open

The out-of-cell minimums are the clearest numbers in the system, and they are stated as guarantees. Tier I: “Exercise periods shall be available for a minimum of five (5) hours per week, with at least one (1) hour per day on five days,” leaving two days a week on which the policy guarantees no out-of-cell time at all [#19270]. Tier II: “Offenders will be offered a minimum of five (5) hours per week of exercise,” with no daily distribution specified, so the five hours may lawfully be delivered in any pattern across the week [#19269]. Tier II STEP: “a minimum of four (4) hours out of cell time per day, seven (7) days per week” [#19283]. Tier III: “All offenders shall receive a minimum of four (4) hours out of cell time per day, Monday through Friday” [#19281].

Run the arithmetic on a 168-hour week, and the policy floors diverge sharply. Tier I, Tier II, generic administrative segregation and the juvenile RHA-JOAS programme all produce 163 hours in cell per week, or 23.3 hours per day, with a 24-hour floor on the days the exercise entitlement does not cover [#19272]. Tier II STEP and Tier III STEP produce 20.0 hours per day [#19272]. Tier III averages 21.1 hours per day on weekdays and reaches 24 hours on Saturdays and Sundays, when no out-of-cell time is guaranteed [#19272]. The step-downs therefore guarantee 5.6 times more out-of-cell time than Tier II itself, and Tier III — the stage GDC labels most restrictive — guarantees four times more than Tier II [#19273].

That matters because of how the UN Standard Minimum Rules for the Treatment of Prisoners define the term. Mandela Rule 44 has two elements: confinement for 22 hours or more a day, and the absence of meaningful human contact [#19331]. GPS can establish the first element for Tier I and Tier II by arithmetic from GDC's own minimums. The second element turns entirely on cell occupancy, and on that question GDC's Tier II policy says nothing. SOP 209.08 imposes no single-occupancy requirement, while 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 — drafting whose necessary implication is that multiple occupancy is the ordinary condition [#19279, #19404]. SOP 209.08's own religious-practice clause is drafted on the same assumption: offenders “may individually pray or individually engage in religious practices in their assigned cells or by their assigned beds,” where the alternative “or by their assigned beds” only has work to do if some people are assigned a bed within a shared space [#19277].

Two conditions of confinement are accordingly being run inside the same programme under the same rules — 23 hours a day alone in a cell, and 23 hours a day locked in a cell with another person — and GDC's Tier II policy distinguishes neither, guarantees neither, and reports neither [#19510]. The facility-level record shows both. Smith State Prison's Cycle 4 audit, with onsite dates from 2025-02-03, describes dorms of “48 double bunked cells housing 96 offenders per dorm,” one for Tier 1 and one for Tier 2 [#19427, #19526]. Georgia State Prison's Cycle 2 audit documents both modes side by side: Tier II at G Building double-bunked for up to 218 people, Tier II at K Building “100 inmates in single cell lockdown cells” [#19474, #19475, #19525]. At Hays, Tier II units F1, F2, G1 and G2 each held 32 one-man cells while Tier I units D2 and E held 64 in two-man cells [#19470, #19471]. Hancock's answers across cycles state that “G-1 is what is known as a ‘Tier 2’ unit … housing a maximum of 96 inmates, in double occupancy cells located on two tiers” [#19572]. Telfair's E-1 and E-2 held 80 and 70 in two-man cells; Ware's J-Building held 100 Tier 1 offenders double-bunked; Dooly's J2 held 96 in double bunks [#19430, #19429, #19426].

Other conditions follow the same pattern of written floors subject to facility discretion. Tier II cells “must be equipped and furnished in a manner consistent with cells designated as Tier II cells” — a circular standard, where Tier III and both STEP programmes instead require furnishing consistent with the general population [#19275]. A single clause, Section IV.D.1.l, lets the Warden modify every Tier II condition — hygiene, food, exercise, telephone, visitation, counselling — on a finding of “extenuating operational circumstances,” an undefined term, with no time limit, no documentation requirement and no notice to Central Office [#19276]. The tablet is confiscated on Tier II placement and returned only at Phase III, and video visitation is limited to one visit per calendar month [#19285]. Medical rounds are the subject of an unresolved conflict: the health services SOP 507.04.33 (2022) sets daily rounds in restrictive housing, while Tier II's SOP 209.08 requires them three times per week “excluding weekends and holidays,” and Tier I's SOP 209.07 says the same [#19334, #19336, #19337]. Board of Corrections Rule 125-4-4-.08 requires three weekly medical checks [#19354]. GDC has not harmonised the two standards, and the older, less protective one is written into the policy that governs the Tier unit itself [#19337]. The staff inspection schedule compounds it: of the officials required to inspect Tier housing, only the Institutional Duty Officer must attend on weekends and holidays [#19340, #19341, #19342]. The 2024 Georgia Senate Study Committee recommended converting all existing facilities and building all new ones as single-cell facilities [#7583, #2975]; the Tier II SOP, unchanged since 2016, does not.

Placement, review and release: where the policy is loose and the exit is tight

Tier I has eight placement grounds, and only one presupposes a pending disciplinary charge. Ground A is that “the offender is noted as a threat to the safe and secure operation of the Facility”; ground H is that the offender “has not been classified due to their arrival at a Facility,” meaning a person can be held in segregated housing purely because they have just arrived and have not yet been classified [#19248]. The 30-day cap on Tier I is “the only outer time limit anywhere in the adult Tier architecture, and it is disapplied for protective custody” — including involuntary protective custody imposed by staff [#19249]. Double bunking in protective custody requires the Warden's approval, capped at 72 hours, with written Central Office approval beyond that [#19251].

Tier II's admission criteria include: a noted threat to safe and secure facility operations, with illustrative categories covering documented STG activity, “notoriety of crimes” and offenders who “have either been threatened with bodily harm or threatened others with bodily harm”; failure in Tier I or refusal to participate; excessive destruction of state property; and “Transfer from Georgia Diagnostic and Classification Prison,” which is a standalone eligibility ground for long-term administrative segregation [#6596, #19487, #19488]. Criterion 1 is open-ended on its face — “noted as a threat,” “may include, but is not limited to,” no evidentiary standard and no requirement of a disciplinary conviction, and it expressly covers a person who has been threatened, i.e. a victim rather than an aggressor [#19253].

The procedural texture is lopsided. Placement is decided at facility level: the Classification Committee holds an Administrative Segregation hearing within 96 hours of assignment, using a body made up of the Tier II Program Unit Manager, the OIC, the assigned counselor and, as applicable, the assigned mental health counselor, with the Warden reviewing for approval within seven business days, and the Warden able to authorise immediate assignment in an emergency [#19256, #19490]. Getting out is a Central Office decision: the Unit Manager recommends, the Warden and Regional Director review, the Criminal Investigations Unit STG Coordinator investigates and reports within ten business days, and the file goes to the Director of Field Operations, who “will make the final decision,” after which the statewide Tier Segregation Manager determines placement [#19256]. A facility can put a person into Tier II on its own authority, including immediately; it cannot take a person out [#19256].

The appeal structure carries the same asymmetry. The initial placement appeal goes to the Director of Field Operations [#19256]. The appeal against the 90-day review that continues confinement goes only to the Warden — the same Warden/Designee who, under Section IV.C.4, approves Tier II placement in the first instance [#19258]. Every subsequent 90-day decision to keep a person in indefinite administrative segregation is therefore reviewed only by the official who authorised the confinement, with no route to Central Office [#19258]. Exit is also gated on programming: the offender must be actively participating in, and where appropriate must complete, the O.U.T. cognitive behavioural program before release, and is then assigned to Tier I for up to thirty days of reorientation before general population [#19259]. The SOP contains no provision for what happens if the O.U.T. program is not offered at the facility [#19259].

Duration figures for Tier II appear nowhere in SOP 209.08 and only in facility audits: Hays Cycle 2 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 “up to two years” for progression through the phases [#19494, #19496, #19497, #19478, #19403]. None of these figures has any basis in the SOP [#19403]. At the Tier III end, retention past 24 months requires quarterly review using Attachment 12 by a four-member statewide panel — the Director of 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 Commissioner or Assistant Commissioner for Facilities [#19261]. An offender within 12 or 6 months of maximum release date must be considered for Tier III STEP [#20783].

Two further features shape accountability. GDC's grievance SOP 227.02 prohibits retaliation for filing a grievance, but makes involuntary assignment to administrative segregation non-grievable, routing the challenge into the classification appeal that terminates with the Warden; and it provides that “a Grievance must not be placed in the Offender's file or referred to in SCRIBE case notes” — where SCRIBE case notes are among the records the Unit Manager reviews to decide release from Tier II [#19319, #19320]. And the reason no Tier population figure is published is not that the data does not exist: GDC's own annual reports state that the “TIER 2 SCRIBE module went live March 2019 to effectively track offenders in Extended Restrictive Housing,” and that FY2020 added to the SCRIBE Warden's Dashboard, for Tier II offenders, “the number of days an offender has spent in restrictive housing Tier II” [#19300]. GDC also employs a single statewide Tier Segregation Manager whose function is to place individuals within the Tier system, a position that necessarily holds or can produce a statewide view [#19257]. What is missing is a rule requiring anyone to compile it: a review of the 209 series found no SOP naming any Tier roster, periodic report or Central Office aggregate submission, and the term “Central Office” does not appear in SOP 209.08 at all [#19308].

The scale GDC publishes, and the arithmetic it makes impossible

Georgia does publish a Tier III population, but only as a side effect. The Special Management Unit at Jackson appears as a facility row in the monthly Inmate Statistical Profile because it is a standalone facility with its own GDC facility code; as of 2026-09-01 it held 154 people, and across 75 monthly snapshots from 2020-07-01 to 2026-09-01 the population ranged from 149 to 195 [#19304]. Against the SMU's 192 beds, that implies the unit has run below bed capacity throughout the published series [#19304, #19302, #19484]. It is one facility, it is not the statewide Tier population, and it says nothing about Tier I or Tier II, whose units sit inside ordinary prisons and are absorbed invisibly into those prisons' facility totals [#19304]. In GDC's October 2026 bed table the SMU is carried at 160 Tier III beds, 128 occupied, 32 vacant, 80.00 percent occupied [#20757, #20763].

The rest of the picture only became visible when GDC produced a per-facility bed table in October 2026 — the first per-facility count of Georgia's Tier II population GPS holds from the agency [#20768]. It lists Tier II at 8 prisons with 884 beds and 804 occupied (91.0 percent), STEP at 4 prisons with 119 beds and 96 occupied, and Tier III at one location with 160 beds and 128 occupied; all three programmes together come to 1,163 beds and 1,028 occupied [#20762, #20764, #20765]. The per-facility Tier II figures are Hancock 83 beds (82 occupied), Hays 151 (151), Macon 121 (119), McRae Women's Facility 34 (28), Smith 111 (89), Telfair 117 (117), Valdosta 129 (80) and Ware 138 (138) [#20749, #20750, #20751, #20752, #20753, #20754, #20755, #20756]. Four of the eight were at or within one bed of full [#20766]. As of October 2025, more than 2,100 people were housed in Georgia's “tiers,” including almost 1,000 in the long-term tiers, where stays can last for months or even years [#20801].

The table has to be read with its own limits attached. It covers Tier II, Tier III and STEP only — GPS's request excluded Tier I, disciplinary isolation and protective custody, so 1,163 beds is not a total of Georgia's segregation beds and must not be compared directly with the 2,721 beds GDC's FY2014 report says were converted to Tier beds at implementation, or with audit-reported segregation-cell counts, which include Tier I [#20771]. It carries no as-of date: the PDF was created 2026-10-02 at 17:29 EDT, and the counts cannot be tied to a specific reporting period [#20839]. It does not split the 119 STEP beds between Tier II STEP and Tier III STEP [#20854, #20776]. Occupied beds are a point-in-time figure, not admissions, lengths of stay, or the number of people who passed through Tier II in a year [#20773], and the production contains no count of people as distinct from occupied beds, no time series and no lengths of stay [#20777]. McRae Women's Facility is the only women's facility on the Tier II list [#20769].

The one published population trend is both small and abandoned. The SMU appears as a single state-prison facility row in each of the 22 GDC Profile of Inmate Admissions and 22 Profile of Inmate Releases editions; in FY2025 the SMU recorded 5 admissions, 0.06 percent of state-prison admissions — a figure that counts people admitted to GDC custody directly at the SMU, not transfers into the Tier III programme, which is how most people reach it [#19305]. GDC reported reducing its extended restrictive housing population by 41 percent (671 people) since January 2017 in FY2018, and by 40 percent (675 people) since the same January 2017 baseline in FY2019 — two statements that cannot be reconciled to a single baseline, and both shaped as a delta with no endpoints [#19298, #19299]. Back-solving a baseline from those percentages yields roughly 1,640 and roughly 1,690; those are GPS arithmetic performed on two inconsistent agency statements and must never be cited as a GDC-published population figure [#19299]. From FY2021 onward the subject disappears from the annual reports as a programme narrative; the last year in which GDC published any figure about the size or trend of its restrictive-housing population was FY2019 [#19303].

Against all of that, a systematic review of 110 non-monthly GDC reports and all 34 distributions of the monthly Inmate Statistical Profile series found no published count of the Tier or restrictive-housing population [#19306, #6608]. The “supervision” distribution offers only Close, Medium and Minimum; there is no housing-status field [#19306]. Because SOP 209.09, SOP 209.45 and SOP 209.55 all classify Tier III, Tier II STEP and Tier III STEP offenders as Close Security, the Tier population is contained within GDC's published Close Security count but cannot be isolated from it — making the Tier population arithmetically unrecoverable rather than merely unpublished [#19307, #19262]. Georgia submitted no response at all to the 2021 ASCA-Liman survey, whose comparison table lists its custodial population at 46,315 with “0 (No 2021 Survey Response)” for the restrictive-housing count [#20818] — even though Georgia had told the same survey that it had “Tier II step down units for offenders on phase 3+, who have been in restrictive housing for 270+ days” [#20814].

Per-facility scale, visible only in the PREA audits

All 273 GDC-published PREA facility audit reports in the GPS index were scanned for programme-sense Tier language and for segregation-housing counts, with each figure attributed to the audit that states it [#19514]. Summing the most recent audit per facility that states a count gives 3,401 segregation cells across 30 facilities, or 3,182 across 28 when restricted to state-operated facilities; on a single-cycle basis the 23 state facilities with a structured Cycle 2 Pre-Audit Questionnaire figure total 2,967, rising to 3,078 across 25 when Wilcox and Montgomery are added from narrative [#19398, #19504, #19505, #19506]. The largest stated counts are Georgia Diagnostic and Classification at 409 (Cycle 2), Smith 336 (Cycle 2), Telfair 280 (Cycle 2), Hays 252 (Cycle 2), Hancock 223 (Cycle 4 and again Cycle 3), Calhoun 197 (Cycle 2), Macon 192 (Cycle 2), Coffee 170 (Cycle 4), Wilcox 168 (Cycle 3), Baldwin 167 (Cycle 2), Valdosta 161 (Cycle 2), Ware 125 (Cycle 2) and Phillips 100 (Cycle 3) [#19438, #19439, #19440, #19441, #19437, #19442, #19443, #19444, #19433, #19445, #19446, #19447, #19436]. At the other end sit Montgomery 16, Burruss 10, Emanuel Women's 7, Whitworth Women's 7, West Central 5 and the Long Unit 4 [#19456, #19458, #19459, #19460, #19461, #19462].

Four caveats govern every one of those numbers. First, this is a cell count: not a bed count and not a population count. The audits are explicit that these are different quantities — 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” [#19435, #19433, #19434, #19432]. The bed count is materially higher than 3,401 and the number of people is a separate unknown that GPS has not interpolated and should not [#19398]. Second, it is a floor, not a total: at least six state facilities with published audits contribute zero because no audit of them ever states a figure — Arrendale, Augusta State Medical Prison, Walker, Helms Facility, Emanuel Unit and McRae Women's Facility [#19415]. Third, it mixes vintages: the figures span Cycle 2 (2017–2019), Cycle 3 (2019–2021) and Cycle 4 (2021–2025), and the most recent audits of five of the nine Tier prisons — Georgia Diagnostic and Classification, Hays, Smith, Macon and Wilcox — dropped the segregation-cell field, so those facilities contribute older numbers [#19508, #19416]. Fourth, it counts all segregation, not only Tier programme housing; Tier units are a subset [#19507].

The Tier label also does not track where segregation capacity actually sits. The programme designation appears only at close-security prisons, without exception, but segregation and administrative-segregation housing is documented at medium-security prisons throughout the system, sometimes at larger scale: Calhoun State Prison, whose Cycle 2 audit gives its security level as “4/Medium,” reports 197 segregation cells — more than close-security Baldwin (167), Ware (125) or Phillips (75) [#19399, #19442, #19530]. Wilcox, described in its own audits as “housing medium security, general population inmates,” reports 168 in Cycle 3, up from 24 double-bunked in Cycle 1 and 72 in Cycle 2 [#19467, #19433]. Every state prison in the corpus that reports a number reports segregation cells, close security or medium [#19511].

Three per-facility Tier figures stand apart. GDC's FY2018 Annual Fiscal Report records that Guard 1 Technology installed at Valdosta State Prison “captures automated segregation rounds in all tier units (178 Cells)” — a cell count for one prison's Tier units alone, approaching the size of the entire Special Management Unit [#19301]. Hays State Prison's Cycle 2 audit itemises 303 stated Tier beds across Tier I and Tier II units (D2 64, E 64, F 64, G 64, H 47), a stated capacity larger than the entire Tier III SMU [#19393, #19523]. Georgia State Prison's Cycle 2 audit itemises 424 stated Tier beds (E-South 106, G Building 218, K Building 100), more than twice the SMU's 192, with 13 Tier Program dorms recorded in its Cycle 1 audit [#19394, #19524, #19477]. Georgia State Prison closed on 2022-02-19, so those figures are historical [#19369]. The only facility-level Tier II population figure in any source GPS holds is an allegation: the Southern Center for Human Rights' September 2021 class action complaint alleged approximately 300 people in Georgia State Prison's Tier II programme, over 70 percent of them experiencing serious mental illness — an allegation, not a finding, describing a facility that has since closed [#19367, #6614]. Set against the SMU's published range of 149 to 195, that would make one prison's Tier II programme roughly twice the size of the entire Tier III programme; that comparison is flagged as derived, and GPS cannot and must not extrapolate from one facility to a system total [#19368].

The audit corpus is the best per-facility segregation dataset GDC publishes, and it is internally inconsistent [#19513]. The defects are recorded because they are defects in GDC's own published record. Georgia State Prison's “Cycle 3” audit is the Cycle 2 audit republished — both 181 pages, extracted text identical after whitespace normalisation at 546,523 characters with zero differences, both carrying a 2019-03-22 report date, so its segregation figure must not be counted twice [#19408, #19400]. Telfair State Prison's Cycle 4 audit contradicts itself: “The PCM indicated the facility does not have administrative segregation cells” in one section and “the facility head reported having 360 administrative segregation beds” in the same report's interviews section [#19400]. Pre-Audit Questionnaire segregation fields contradict the same audit's narrative in at least three cases, including Georgia State Prison Cycle 2's questionnaire figure of 12 segregation cells against a narrative itemising 106 Tier I cells, 218 Tier II beds and 100 Tier II single lockdown cells [#19411, #19400]. Label defects leave the unit of measure undetermined in Washington, Dooly and Pulaski [#19413]. A single sentence of boilerplate about delayed PREA education appears in roughly 60 audits, including audits of facilities that demonstrably have no Tier housing, badly inflating any naive keyword search [#19515]. The word “tier” in a PREA audit usually means the upper or lower gallery of a cellblock, not the Tier programme [#19401]. Cycle 1 audits are two-to-four-page summaries with no Physical Plant table [#19414]. And Georgia State Prison Cycle 2's security-level field reads “Minimum” while the same report's narrative states “This facility is a close security institution” [#19405]. Any published GPS figure drawn from this corpus must name the facility, the cycle and the field, and must not silently reconcile a questionnaire figure against a narrative one [#19513].

On women, the record is thinner still. All seven facilities GDC designates as Tier I and II are men's prisons, and neither of GDC's two close-security-classified women's facilities — McRae Women's Facility and Arrendale — is on the designation list, though the Tier SOPs are written in gender-neutral terms and SOP 209.07 applies Tier I system-wide, so women in GDC custody are subject to Tier I segregation [#19364]. Whether GDC operates a Tier II programme for women, and if not, what long-term segregation mechanism applies to women, is answered nowhere in the SOP corpus or published fact sheets; the record that would answer it is the current facility designation document [#19365]. McRae Women's Facility has no PREA audit in the corpus under that name [#19374]. Of the twelve currently active close-security state prisons, eight have documented Tier programme housing — the seven designated facilities plus Georgia Diagnostic and Classification [#19361].

Litigation, the standards gap, and the dead

The one part of the Tier system that has been litigated is the Special Management Unit at Georgia Diagnostic and Classification Prison. Gumm v. Ford, No. 5:15-cv-00041-MTT-CHW in the Middle District of Georgia before Chief Judge Marc T. Treadwell, began as a handwritten pro se complaint filed 2015-02-12 [#19321]. The court adopted findings, based on Dr. Craig Haney's expert report, that SMU residents were confined in cells “smaller than the average parking space,” 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; at the 2017 inspection the SMU held approximately 180 people, of whom 70 were designated mentally ill, and Timothy Gumm himself had been held continuously for seven and a half years despite fourteen transfer recommendations over four years [#19326, #6598]. Haney described the unit as “one of the harshest and most draconian” he had seen “in decades of conducting evaluations,” and the residents as “among the most psychologically traumatized persons [he] ha[d] ever assessed in this context” [#6624].

The May 2019 settlement 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 mental health professional, a prohibition on housing Mental Health Level III or above prisoners in Tier III/SMU, minimum out-of-cell time and programming standards, and $425,000 in attorney's fees [#6641]. In April 2024 the district court found the agency had “no desire or intention” to comply, ordered an independent monitor, imposed fines of $2,500 per day — $75,000 every 30 days for six months — and documented that officials had falsified therapy records, including recording that a man attended treatment sessions after he was already dead [#19322]. In April 2024 the U.S. Department of Justice expanded its CRIPA investigation of Georgia prisons to include restrictive housing, disciplinary practices and special education services; the findings report issued 1 October 2024 addressed only the violence and sexual abuse prongs, and the restrictive-housing findings have not been released [#19311]. What that report did find bears directly on this page: “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” [#6607, #6904].

The scope of the reform is the central fact about the Tier system's standards gap. Every operative settlement term is written in SMU and Tier III vocabulary, the certified class arose from conditions in a single facility, and GDC codified the settlement into SOP 209.09 alone, which now carries the 24-month ceiling, the mental-health exclusion, the 60-day and 90-day evaluations, the quarterly four-member panel and the requirement of personal approval by the Commissioner or Assistant Commissioner for Facilities to hold anyone past 24 months [#19324, #19261]. SOP 209.08, governing Tier II, has none of it and has not been revised since 2016-04-11 — nearly three years before the settlement [#19324]. The programme with the least protection is the larger one. On Wilkinson v. Austin, 545 U.S. 209 (2005), the factor the Court identified as central — indefinite duration — is present in Tier II and absent from Tier III, so on Wilkinson's own reasoning the due-process argument is stronger for the programme the litigation bypassed; that is an argument from a holding, not a holding, and no court has applied Wilkinson to Georgia's Tier II programme [#19327].

Mental-health routing compounds the problem. Tier III excludes both MH Level III and Level IV, referring them to a Specialized Mental Health Treatment Unit; Tier II excludes only Level IV and admits Level III on mental-health recommendation [#19254]. GDC's own mental-health policy places SMHTUs/TIER IIs in the highest-need tier of its staffing hierarchy alongside death row, the SMU and acute care, and assigns them the second-richest counsellor ratio in the system, at least 1.0 FTE per 20 offenders [#19288]. Suicide-risk questioning is monthly in Tier I and weekly in Tier II, phrased as “should be asked,” not “shall” — so a person serving an entire 30-day Tier I term could be asked about suicidal ideation exactly once [#19287]. Against the Mandela Rules' 15-day prolongation threshold, Tier I's 30-day cap is twice the threshold and unbounded for protective custody; Tier II has no cap and a 90-day review cycle that is itself six times the threshold and indefinitely repeatable; Tier III's stated 390-day minimum is 26 times the threshold and its 24-month outer limit approximately 48.7 times it [#19329].

The consequences are documented in GDC's own audit and inspection record as well as in litigation. GDC's 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 and some days with no checks at all, and noted that all check sheets reviewed by the auditor were incomplete [#2621]. At Calhoun State Prison, DOJ documented that a person in restrictive housing was found dead, wrapped in mattress padding; no one had entered his cell for two days, his water supply had been shut off and the chow flap closed, and the cause of death was dehydration with renal failure [#6902, #124]. The 2024 Georgia Senate Study Committee, where the Southern Center for Human Rights' Ms. Hollie defined solitary confinement as being held in a cell for 22 or more hours per day, heard that solitary was used at Georgia State Prison — a special mission mental health facility — until its closure in 2022 [#2952, #7578, #2957]. And in the SMU litigation, the plaintiffs claimed individuals with serious mental illness remained in the SMU even after the extreme conditions of the Tier III program caused them to decompensate and engage in acts of self-harm, citing the suicides of two SMU inmates [#20411].

Two people named in GPS's death registry died in this system's restrictive and mental-health housing in the spring of 2024. Sherry Elaine Joyce died April 27, 2024, at Arrendale State Prison; GPS's registry lists the cause as Homicide and her age as 61. Hallie Marie Reed died May 5, 2024, at Arrendale State Prison; GPS's registry lists the cause as Homicide and her age as 23. Both deaths occurred in the same Lee Arrendale mental health unit eight days apart, according to the record before the 2024 Senate Study Committee, which also records that Reed had requested protective custody after Joyce's death and was denied [#5088].