GAO
May 2013

United States Government Accountability Office

Report to Congressional Requesters

BUREAU OF
PRISONS
Improvements Needed
in Bureau of Prisons’
Monitoring and
Evaluation of Impact
of Segregated Housing

GAO-13-429

May 2013

BUREAU OF PRISONS

Highlights of GAO-13-429 a report to
congressional requesters

Improvements Needed in Bureau of Prisons’
Monitoring and Evaluation of Impact of Segregated
Housing

Why GAO Did This Study

What GAO Found

BOP confines about 7 percent of its
217,000 inmates in segregated
housing units for about 23 hours a day.
Inmates are held in SHUs, SMUs, and
ADX. GAO was asked to review BOP’s
segregated housing unit practices. This
report addresses, among other things:
(1) the trends in BOP’s segregated
housing population, (2) the extent to
which BOP centrally monitors how
prisons apply segregated housing
policies, and (3) the extent to which
BOP assessed the impact of
segregated housing on institutional
safety and inmates. GAO analyzed
BOP’s policies for compliance and
analyzed population trends from fiscal
year 2008 through February 2013.
GAO visited six federal prisons
selected for different segregated
housing units and security levels, and
reviewed 61 inmate case files and 45
monitoring reports. The results are not
generalizable, but provide information
on segregated housing units.

The overall number of inmates in the Bureau of Prisons’ (BOP) three main types
of segregated housing units—Special Housing Units (SHU), Special
Management Units (SMU), and Administrative Maximum (ADX)—increased at a
faster rate than the general inmate population. Inmates may be placed in SHUs
for administrative reasons, such as pending transfer to another prison, and for
disciplinary reasons, such as violating prison rules; SMUs, a four-phased
program in which inmates can progress from more to less restrictive conditions;
or ADX, for inmates that require the highest level of security. From fiscal year
2008 through February 2013, the total inmate population in segregated housing
units increased approximately 17 percent—from 10,659 to 12,460 inmates. By
comparison, the total inmate population in BOP facilities increased by about 6
percent during this period.

What GAO Recommends
GAO recommends that BOP (1)
develop ADX-specific monitoring
requirements; (2) develop a plan that
clarifies how BOP will address
documentation concerns GAO
identified, through the new software
program; (3) ensure that any current
study to assess segregated housing
also includes reviews of its impact on
institutional safety; and (4) assess the
impact of long-term segregation. BOP
agreed with these recommendations
and reported it would take actions to
address them.

View GAO-13-429. For more information,
contact David C. Maurer at (202) 5129627 or maurerd@gao.gov.

BOP has a mechanism to centrally monitor segregated housing, but the degree
of monitoring varies by unit type and GAO found incomplete documentation of
monitoring at select prisons. BOP headquarters lacks the same degree of
oversight of ADX-specific conditions of confinement compared with SHUs and
SMUs partly because ADX policies are monitored locally by ADX officials.
Developing specific requirements for ADX could provide BOP with additional
assurance that inmates held at ADX are afforded their minimum conditions of
confinement and procedural protections. According to a selection of monitoring
reports and inmate case files, GAO also identified documentation concerns
related to conditions of confinement and procedural protections, such as
ensuring that inmates received all their meals and exercise as required.
According to BOP officials, in December 2012, all SHUs and SMUs began using
a new software program that could improve the ability to document conditions of
confinement in SHUs and SMUs. However, BOP officials acknowledged the
recently implemented software program may not address all the deficiencies
GAO identified. Since BOP could not provide evidence that it addressed the
documentation deficiencies, GAO cannot determine if it will mitigate the
documentation concerns. BOP expects to complete a review of the new software
program by approximately September 30, 2013, which should help determine the
extent to which the software program addresses documentation deficiencies
GAO identified.
BOP has not assessed the impact of segregated housing on institutional safety
or the impacts of long-term segregation on inmates. In January 2013, BOP
authorized a study of segregated housing; however, it is unclear to what extent
the study will assess the extent to which segregated housing units contribute to
institutional safety. As of January 2013, BOP is considering conducting mental
health case reviews for inmates held in SHUs or ADX for more than 12
continuous months. However, without an assessment of the impact of
segregation on institutional safety or study of the long-term impact of segregated
housing on inmates, BOP cannot determine the extent to which segregated
housing achieves its stated purpose to protect inmates, staff and the general
public.

United States Government Accountability Office

Contents

Letter

1
Background
Segregated Housing Unit Population and Number of Cells Have
Increased since Fiscal Year 2008
BOP’s Monitoring of Segregated Housing Policies Varies by Type of
Unit, and Some Facilities’ Documentation Is Incomplete
BOP Estimates that Segregated Housing Costs More than Housing
Inmates in General Population
BOP Has Not Evaluated the Impact of Segregated Housing Units on
Institutional Safety or the Impacts of Long-Term Segregation on
Inmates
Conclusions
Recommendations for Executive Action
Agency Comments and Our Evaluation

5
14
16
29
33
41
42
42

Appendix I

Objectives, Scope, and Methodology

45

Appendix II

Communications Management Units

52

Appendix III

Location and Length of Stay in BOP Segregated Housing
Units

59

Appendix IV

Comments from the Bureau of Prisons

61

Appendix V

GAO Contact and Staff Acknowledgments

66

Table 1: Results of Our Independent Analysis of 51 Selected Inmate
Case Files: Inmate Review and Procedural Protections
Issues, Fiscal Years 2011-2012
Table 2: BOP Average Daily Inmate Costs per Capita by Security
Level, BOP-wide, for Fiscal Year 2012

27

Tables

Page i

30

GAO-13-429 Segregated Housing Units

Table 3: BOP Estimated Daily Inmate Costs per Capita in Selection
of Institutions and Different Types of Segregated Housing
Units, by Security Level for Fiscal Year 2012
Table 4: Site Visits to BOP Institutions

32
46

Figures
Figure 1: Comparison of Segregated Housing Unit Policies: SHU
and SMU
Figure 2: Comparison of Segregated Housing Unit Policies: ADX
Figure 3: Required Procedures for Special Housing Unit (SHU)Disciplinary Segregation
Figure 4: Proportion of BOP Inmates in General Population and in
Segregated Housing Types, as of February 2013
Figure 5: Trends in Inmate Population by Type of Segregated
Housing Unit, Fiscal Year 2008 through February 2013
Figure 6: Special Housing Unit Cell, U.S. Penitentiary,
Leavenworth, Kansas
Figure 7 : Outdoor Recreation Area, Special Management Unit, U.S.
Penitentiary, Lewisburg, Pennsylvania
Figure 8: Interior of Single-Bunked Cell, U.S. Administrative
Maximum Facility, Florence, Colorado
Figure 9: Outdoor Recreation Areas, U.S. Administrative Maximum
Facility, Florence, Colorado
Figure 10: Common Findings from Our Analysis of 45 Monitoring
Reports
Figure 11: Overview of Communications Management Units
Figure 12: Communications Management Unit Cell, Terre Haute,
Indiana
Figure 13: Communications Management Unit Common Area,
Marion, Illinois
Figure 14: Locations of Segregated Housing Units within BOP
Facilities

Page ii

7
9
11
13
16
20
21
23
24
26
53
56
57
59

GAO-13-429 Segregated Housing Units

Abbreviations
ACA
ADX
BOP
CMU
CPD
DHO
DOJ
FCC
GP
HQ
PDS
IPPA
OJP
ORE
PRD
SAM
SHU
SMU
USP

American Correctional Association
Administrative Maximum
Bureau of Prisons
Communications Management Unit
Correctional Programs Division
Discipline Hearing Officer
Department of Justice
Federal Correctional Complex
General Population
Headquarters
Psychology Data System
Information, Policy, and Public Affairs Division
Office of Justice Programs
Office of Research and Evaluation
Program Review Division
Special Administrative Measure
Special Housing Unit
Special Management Unit
United States Penitentiary

This is a work of the U.S. government and is not subject to copyright protection in the
United States. The published product may be reproduced and distributed in its entirety
without further permission from GAO. However, because this work may contain
copyrighted images or other material, permission from the copyright holder may be
necessary if you wish to reproduce this material separately.

Page iii

GAO-13-429 Segregated Housing Units

United States Government Accountability Office
Washington, DC 20548

May 1, 2013
The Honorable Richard J. Durbin
Chairman
Subcommittee on the Constitution,
Civil Rights and Human Rights
Committee on the Judiciary
United States Senate
The Honorable Elijah E. Cummings
Ranking Member
Committee on Oversight
and Government Reform
House of Representatives
The Honorable Robert C. Scott
Ranking Member
Subcommittee on Crime, Terrorism,
Homeland Security, and Investigations
Committee on the Judiciary
House of Representatives
Since the late 1980s, America’s federal prison population increased by
more than 400 percent, accompanied by the use of certain types of
segregated housing units where prisoners are kept apart from the general
inmate population in at times highly restrictive conditions. The
Department of Justice’s (DOJ) Bureau of Prisons (BOP) is responsible for
the custody and care of approximately 217,000 federal inmates. 1 BOP’s
mission is to confine federal inmates in the controlled, safe, secure,
humane, and cost-efficient environments of prisons and communitybased facilities, and to provide work and other self-improvement
opportunities to assist offenders in becoming law-abiding citizens. BOP
operates several types of segregated housing units to separate inmates
from the general inmate population for different purposes, such as

1

As of February 2013, BOP held the majority of inmates in its custody in 119 BOPoperated federal prisons. BOP held about 41,700, or about 19 percent, of the total BOP
federal inmate population in community confinement (residential reentry and home
confinement) and 15 privately managed prisons. We are focusing only on the 119 BOPoperated facilities for the purposes of this review.

Page 1

GAO-13-429 Segregated Housing Units

administrative detention, disciplinary reasons, gang-related activity or
assaulting staff. Specifically, these units are the (1) Special Housing Units
(SHU), where inmates can be placed for administrative reasons, such as
pending transfer to another prison, and for disciplinary reasons, such as
violating prison rules; (2) Special Management Units (SMU), a four-level
program in which inmates can progress from more restrictive to less
restrictive conditions; and (3) the Administrative Maximum (ADX) facility
in Florence, Colorado, for inmates that require the highest level of
security. As of February 2013, BOP confined approximately 12,460
federal inmates—or about 7 percent of inmates in BOP-operated
facilities—in segregated housing units. According to BOP, these
segregated housing units help ensure institutional safety for inmates and
staff. Approximately 435 individuals in ADX are held in what is commonly
referred to as solitary confinement, or single cells alone, for about 23
hours a day. 2
There is little publicly available information on BOP’s use of segregated
housing units. Given the potential high costs, lack of research on their
effectiveness, and possible long-term detrimental effects on inmates, you
requested that we review BOP’s segregated housing unit practices,
including BOP’s standards, reasons for segregating inmates, and costs.
Specifically, this report addresses the following objectives:
1. What were the trends in BOP’s segregated housing unit population
and number of cells from fiscal year 2008 through February 2013?
2. To what extent does BOP centrally monitor how individual facilities
document and apply policies guiding segregated housing units?
3. To what extent has BOP assessed the costs to operate segregated
housing units and how do the costs to confine an inmate in a
segregated housing unit compare with the costs of confining an
inmate in a general inmate population housing unit?
4. To what extent does BOP assess the impact of segregated housing
on institutional safety and the impacts of long-term segregation on
inmates?
Overall, to address our objectives, we reviewed BOP statutory authority
and policies and procedures related to each type of segregated housing

2

According to BOP officials, BOP does not hold anyone in solitary confinement because
BOP staff interacts with inmates who are held in single cells alone.

Page 2

GAO-13-429 Segregated Housing Units

unit and interviewed BOP management officials responsible for inmate
placement and security, monitoring and program compliance, facility and
financial management, and research. To address our first objective, we
obtained and analyzed BOP’s inmate population and number of cells data
for each type of segregated housing unit for the past 5 fiscal years to the
most recent data available—from fiscal year 2008 through February 2013.
We also compared the total inmate population in BOP-operated facilities
with the total segregated housing unit population data to identify trends in
the segregated housing unit population as a share of the total inmate
population in BOP-operated facilities during this period. We assessed the
reliability of BOP’s segregated housing unit inmate population and cell
data by reviewing relevant documentation, interviewing knowledgeable
agency officials about how they maintain the integrity of their data, and
examining the data for obvious errors and inconsistencies. We found the
segregated housing unit inmate population and cell data were sufficiently
reliable for the purposes of this report. We also interviewed BOP
headquarters officials to discuss reasons for the trends in BOP’s
segregated housing unit inmate population and cells.
To address our second objective, we assessed BOP’s monitoring for
each type of segregated housing unit by reviewing monitoring policies,
guidelines and reports. We analyzed BOP’s segregated housing unit
policies and monitoring guidance and compared them against criteria in
Standards for Internal Control in the Federal Government. We also
analyzed BOP’s policies and procedures pertinent to the monitoring of
individual prisons’ compliance with BOP policies, including those of
BOP’s Program Review Division (PRD), which leads monitoring reviews.
In addition, to observe the conditions of confinement, procedural
protections and inmate placement in segregated housing, we conducted
visits to 6 of BOP’s 119 institutions. We chose these institutions to reflect
a range in the types of segregated housing units and security levels. 3
Although the results of our site visits are not generalizable, they provided
insights about BOP monitoring. Further, to assess the methodology and
system BOP employs to monitor, identify, and address deficiencies at
prisons, we requested a selection of monitoring reports from BOP, which
BOP provided for a variety of facilities. Specifically, we analyzed 45 of
187 PRD monitoring reports from 20 of 98 prisons from fiscal years 2007

3

We selected six federal prisons, three of which were complexes that contained multiple
facilities. Thus, our six selected sites contained 11 BOP facilities, 10 of which we visited.
See appendix I for the specific facilities we visited.

Page 3

GAO-13-429 Segregated Housing Units

to 2011. We also reviewed related follow-up monitoring reports to
determine the extent that these prisons resolved deficiencies. Further, we
independently assessed compliance with segregated housing policies for
selected inmates at 2 institutions we visited. For example, we selected a
random sample of 61 case files from inmates housed in segregated
housing units, including SHUs. Although our selection of case files was
not generalizable to all inmates in all types of segregated housing units, it
provided insights into whether these 2 institutions were following BOP
policy. We also reviewed information related to BOP’s new software
program, that includes the SHU application, and compared it against best
practices for project management and criteria in BOP’s monitoring
documentation policies.
To address the third objective, we reviewed BOP fiscal year 2012
average inmate per capita costs for institutions at each security
classification: high security, medium security, low security, minimum
security, administrative, and Federal Correctional Complex. 4 Further, we
analyzed a BOP estimate of fiscal year 2012 inmate per capita costs that
BOP provided in January 2013. We interviewed knowledgeable BOP
officials to understand their processes for developing these cost data and
estimates, and we found the average inmate per capita costs and
estimated inmate per capita costs data to be sufficiently reliable for the
purposes of this report. We also used BOP’s estimated segregated
housing unit versus general population housing inmate per capita cost
data, combined with fiscal year 2012 BOP inmate population data, to
illustrate the possible costs of housing the inmate population in
segregated housing units compared with the costs of housing these same
inmates in general population housing for fiscal year 2012.
To address the fourth objective, we analyzed BOP’s policies, including
program objectives, for each segregated housing unit and policies
governing the provision of mental health services to inmates in
segregated housing units. We also reviewed BOP lockdown information
from fiscal years 2007 through 2012. In addition, we interviewed senior
BOP officials to discuss the extent that BOP has assessed the impact of
segregated housing on institutional safety and their views on the impact
of long-term segregation on inmates. Further, we identified and reviewed

4

BOP operates Federal Correctional Complexes (FCC), which have different missions and
security levels.

Page 4

GAO-13-429 Segregated Housing Units

actions five states have taken regarding segregated housing units. We
selected these five states because they were involved in addressing
segregated housing unit reform and had taken actions to reduce the
number of inmates in segregated housing units. For each of the five
selected states, we reviewed relevant documents on segregated housing,
including placement policies, and we interviewed corrections officials in
these states to understand the reasons for and impact of segregated
housing unit reforms. We visited correctional facilities in two of the five
states—Kansas and Colorado. While the results from our interviews are
not representative of all prisons, they provided us with perspectives on
state actions. We also analyzed studies and reports that describe,
evaluate, or analyze the impact of segregated housing units on
institutional safety and inmates held in these units. We compared BOP’s
mechanisms for evaluating the impact of segregated housing units on
institutional safety, and the impacts of long-term segregation on inmates,
with BOP’s policies and mission statements. Appendix I includes more
details about our scope and methodology.
We conducted this performance audit from January 2012 to April 2013 in
accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain
sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions for our audit objectives. We believe that the
evidence obtained provides a reasonable basis for our findings and
conclusions based on our audit objectives.

Background

BOP operates three main types of segregated housing units: (1) SHUs,
(2) SMUs, and (3) the ADX facility in Florence, Colorado. 5 BOP also
operates Communications Management Units (CMU), where conditions of
confinement are similar to general population and inmates are allowed to
congregate outside their cells for up to 16 hours per day. For information
about CMUs see appendix II. According to BOP policy, all three types of
segregated housing units have the same purpose, which is to separate
inmates from the general inmate population to protect the safety, security,
and orderly operation of BOP facilities, and to protect the public.
However, the specific placement criteria and conditions of confinement

5

BOP also operates other types of segregated housing units, including protective custody
units for inmates requiring protection, sex offender management units for convicted sex
offenders, and a Special Confinement Unit to hold inmates on death row, among others.

Page 5

GAO-13-429 Segregated Housing Units

vary for each type of segregated housing unit. In addition, inmates in
SHUs, SMUs, and ADX are confined to their cells approximately 23 hours
per day. 6 See figures 1 and 2 for a comparison of differences among the
three units.

6

Inmates in segregated housing units may be confined to their cells for fewer or more
hours per day, depending on their unit. For example, inmates in Phase 3 of the ADX StepDown Unit may recreate outside of their cells for three hours per day, or are confined to
their cells for 21 hours per day. Inmates in other types of segregated housing units, such
as SMUs, are permitted five hours of recreation per week, ordinarily on different days, in
which case they would be confined to their cells 24 hours per day on at least two days.
For more information about the number of hours of out-of-cell recreation per segregated
housing unit, see figure 1.

Page 6

GAO-13-429 Segregated Housing Units

Figure 1: Comparison of Segregated Housing Unit Policies: SHU and SMU

a

According to BOP policy, BOP may place inmates in SHU-administrative detention status who are
removed from the general population because they pose a threat to other inmates and staff or the
orderly running of an institution and are (1) under investigation or awaiting a hearing for possibly

Page 7

GAO-13-429 Segregated Housing Units

violating a BOP rule or criminal law, (2) pending transfer to another facility or location, (3) placed in a
SHU for the inmate’s own protection, or (4) in post-disciplinary detention status.
b

BOP lists 91 different types of prohibited acts, which have different punitive measures depending on
their severity, including greatest, high, moderate, and low severity acts. Examples of greatest severity
prohibited acts are killing, attempted or accomplished serious physical assaults and encouraging
others to riot. Examples of high severity prohibited acts are fighting with others and threatening bodily
harm. Examples of moderate severity prohibited acts are indecent exposure, and refusing to obey an
order. Inmates who commit low severity offenses are not eligible to receive disciplinary segregation
as a sanction.
c

BOP provided SHU number of cells data as of fiscal year 2012, and SMU number of cells data as of
November 2012. The population data for both SHUs and SMUs is as of February 2013. Also, the total
SHU population does not include inmates in SHUs within the Florence Administrative Maximum
facility or SHUs within SMUs. These SHU inmates are counted under the ADX and SMU total inmate
populations, respectively. For example, the total SMU population in figure 1 includes inmates in the
SHU within each SMU.
d

SMUs consist of a four-level, 18- to 24-month program. According to BOP policy, an inmate may
progress through the levels depending on his compliance with behavioral expectations, ability to
demonstrate positive “community” interaction skills, and preparedness to function in a general
population setting with inmates from various group affiliations.

Page 8

GAO-13-429 Segregated Housing Units

Figure 2: Comparison of Segregated Housing Unit Policies: ADX

a

The ADX houses BOP inmates who require the tightest controls. The ADX operates five types of
housing unit programs: the Control Unit, a Special Housing Unit, the Special Security Unit, General
Population Units, and the Intermediate Phase (Phase 2) of the Step Down Program. The Transitional
Phase (Phase 3) and Pre-Transfer Phase (Phase 4) of the ADX Step Down Program are physically
located in the United States Penitentiary (USP) Florence (high security) facility.

Page 9

GAO-13-429 Segregated Housing Units

b

If the Designation and Sentence Computation Center (DSCC) staff determine the inmate warrants
consideration for ADX placement, the Chief, DSCC, is to forward the referral packet to the BOP
National Discipline Hearing Administrator, who is to designate a Hearing Administrator who is
experienced in working with and observing inmates and is familiar with the criteria for inmate
placement in different institutions, with an emphasis on ADX.
c

Inmates may be reviewed to be placed in a double-bunked cell, after an initial placement of three
months in a single cell in Phase 4.
d

BOP provided the number of cells data as of fiscal year 2012. The population data are as of
February 2013. The ADX number of cells and population data include the total number of ADX cells
and population physically located in ADX, including the ADX SHU, and the total number of cells and
population in Phases 3 and 4 of the ADX Step Down Unit, physically located at USP Florence. For
example, the total ADX population includes inmates held in the ADX SHU (67 inmates).
e

For more information about BOP policies related to the ADX Control Unit, see BOP program
statement 5212.07 (February 20, 2001), available on BOP’s website.
f

The Executive Panel is composed of the Regional Director of the region where the control unit is
located and the Assistant Director, Correctional Programs Division.

g

Special administrative measures (SAM), which must be authorized by the Attorney General, primarily
limit communication with others, through restricted telephone, correspondence, and visiting privileges.
SAMs based on a substantial risk of death or serious bodily injury may be imposed for a period of up
to 120 days, or, with the approval of the Attorney General, up to one year and may be renewed.
SAMs based on national security threat may be imposed for a period of up to 1 year and may be
renewed. The BOP Director may renew special restrictions within the SAMs if the Attorney General or
federal law enforcement or intelligence agency provides written notification of continued substantial
risk of death or serious bodily injury or threat to national security related to the inmate’s
communications or contacts with other persons. See 28 C.F.R. §§ 501.2, 501.3.

BOP Segregated Housing
Unit Policies

BOP has specific procedural and conditions of confinement policies that
govern each of the three types of segregated housing units. For example,
BOP’s procedural policies govern how determinations are made to place
inmates in each type of segregated housing unit. These determinations
vary based on the level of security and supervision an inmate requires as
well as any prohibited acts committed (e.g., assault against staff or gang
activity). BOP policies require hearings to determine whether an inmate
should be placed in an SMU, SHU-disciplinary segregation, or ADX. 7 In
general, a discipline hearing officer (DHO) not involved in the alleged
infraction presides over the hearing, and inmates have a right to testify
and call witnesses. (See fig. 3 for the required procedures for SHUdisciplinary segregation). In addition, BOP’s procedural policies state that

7

According to BOP, the hearing process is intended to meet the prison disciplinary system
due process requirements established in Wolff v. McDonnell, 418 U.S. 539 (1974). This
includes providing the inmate advance written notice of the alleged violation, and
permitting the inmate to (1) attend the hearing and make a statement, (2) call witnesses,
(3) present documentary evidence, and (4) have staff representation. The inmate is also
provided a written statement of the evidence and reasons for the disciplinary action taken.

Page 10

GAO-13-429 Segregated Housing Units

staff are to periodically review inmates’ status to determine whether they
should remain in SHUs, SMUs, and ADX. 8
Figure 3: Required Procedures for Special Housing Unit (SHU)-Disciplinary Segregation

Further, according to BOP regulations, BOP administers an
Administrative Remedy Program which includes formal procedural
protections and provides all inmates in its custody—including those in
segregated housing units—a method for filing a complaint about their
placement, treatment or conditions while in custody, including placement
in a segregated housing unit. 9 In addition, BOP has specific policies
8

Inmates placed in SHU-administrative detention do not receive a hearing prior to
placement. However, BOP policy requires officials to review the inmates’ records within 3
work days of being placed in SHU-administrative detention, and to review their detention
status within 7 calendar days of their placement and every 30 calendar days thereafter.

9

The Administrative Remedy Program is BOP’s process for filing a complaint, such as
concerns about the appropriateness of placements, or allegation of improper physical or
verbal abuse against facility staff, along with any aspect of an inmate’s confinement and
appeals of disciplinary decisions.

Page 11

GAO-13-429 Segregated Housing Units

governing the conditions of confinement of inmates by segregated
housing unit, such as minimum number of hours of exercise per week,
minimum number of telephone calls and visits, and other privileges.
BOP also requires all of its facilities to be accredited and follow standards
developed by the American Correctional Association (ACA). 10 BOP
policies state that all facilities, security level and housing unit
notwithstanding, must provide the same minimum conditions of
confinement, including clean housing units; nutritionally adequate meals
that meet dietary requirements (such as vegetarian or religious diet);
access to educational, occupational, and leisure time programming; basic
medical and mental health care; and access to a chaplain and basic
religious items according to the inmate’s religious beliefs.
According to BOP officials, BOP does not hold anyone in solitary
confinement because BOP staff frequently visit inmates held in singlebunked cells alone. 11 BOP officials stated BOP staff regularly interact with
inmates during their required monitoring rounds and while providing
meals to inmates. In addition, BOP officials stated that inmates who are in
single cells can interact with other inmates during recreation while in
either the same or separate recreation areas and they are also able to
talk to each other in adjoining cells. However, inmates in these three
types of segregated housing units are subject to more restrictive
conditions of confinement than their counterparts in the general
population.

Population of Segregated
Housing Units

BOP has segregated housing units in prisons located throughout the
country. (See app. III for more information about the location of
segregated housing units). With more inmates held under more restrictive
conditions, often for months or years at a time, segregated housing
represents an important part of BOP’s effort to achieve its goal of

10

ACA’s mission includes the development and promotion of effective standards for the
care, custody, training, and treatment of offenders.
11

BOP holds all inmates in single-bunked cells alone at ADX. Inmates in Phase 4 of the
ADX Step Down program may be double-bunked and are physically located at USP
Florence. As of February 2013, ADX held 435 inmates in single-bunked cells alone. In
addition, as of November 2012, BOP had 360 single-bunked SHU cells and 17 singlebunked SMU cells across its facilities in which inmates are confined to their cells alone for
about 23 hours per day.

Page 12

GAO-13-429 Segregated Housing Units

confining inmates in a safe, secure, and cost-efficient environment. Of all
federal inmates in BOP facilities, about 7 percent are held in segregation
and, as of February 2013, BOP held the majority of segregated inmates—
81 percent, or 10,050 inmates—in SHUs. The second largest population
held in segregation is SMU inmates, who comprise about 16 percent of all
segregated inmates, or about 1,960 inmates. ADX holds 450 inmates,
including 15 inmates in the ADX Step Down Units at the high security
United States Penitentiary (USP) Florence. See figure 4 for inmate
population by segregated unit type as a percentage of the total inmate
population in BOP facilities as of February 2013.
Figure 4: Proportion of BOP Inmates in General Population and in Segregated
Housing Types, as of February 2013

Page 13

GAO-13-429 Segregated Housing Units

Segregated Housing
Unit Population and
Number of Cells Have
Increased since Fiscal
Year 2008

From fiscal year 2008 through February 2013, the total inmate population
in segregated housing units increased approximately 17 percent—from
10,659 to 12,460 inmates. 12 The total inmate population in segregated
housing units increased since fiscal year 2008, but the trends in inmate
population vary by type of segregated housing unit. By comparison, the
total inmate population in BOP facilities increased by about 6 percent
since fiscal year 2008. In addition, the total number of segregated
housing cells in BOP facilities increased by nearly 16 percent. 13 The main
reason for the increase in segregated inmates was the creation of the
SMU program in fiscal year 2008.
SHUs. From fiscal year 2008 through February 2013, the total SHU
population remained about the same at 10,070 and 10,050,
respectively. 14 BOP generally double-bunks inmates in SHUs; however,
BOP has the capability to hold some SHU inmates in single cells. For
example, as of November 2012, BOP had 6,731 double-bunked SHU
cells and 360 single-bunked SHU cells. BOP officials also stated they

12

The percentage of the total inmate population in segregated housing increased from 6
percent in fiscal year 2008 to 7 percent as of February 2013. The total inmate population
in segregated housing units includes the total inmate population in ADX, all SHUs, and all
SMUs in BOP facilities.
13

BOP uses different units of measurement to calculate capacity for its segregated
housing units. BOP calculates the number of inmates SMUs and ADX can hold in terms of
each unit’s rated capacity. Rated capacity reflects the number of prisoners a facility or unit
was designed to house safely and securely and in which BOP can provide inmates
adequate access to services, necessities for daily living, and programs designed to
support their crime-free return to the community. A facility’s rated capacity excludes
medical and SHU bed space, and BOP does not maintain rated capacity data for SHUs.
Instead, BOP provided SHU capacity in terms of number of cells. For the purposes of
analyzing trends in segregated housing unit capacity over time, we used the number of
cells in SHUs, SMUs, and ADX to have the same unit of measurement for all segregated
housing units. BOP provided number of cells data for fiscal years 2008 through 2012 for
SHUs and ADX and number of cells data for fiscal year 2008 through November 2012 for
SMUs.
14
Based on data that BOP provided, these data do not include the inmate population in
SHUs within ADX or SMUs. The SHU inmate population in these units is included under
the total ADX and SMU populations, respectively. Also, based on our analysis, although
the SHU population increased since fiscal year 2008, there has been a decline recently.
According to BOP officials, the SHU population trend might be following the trend in total
population in BOP facilities, which increased from fiscal years 2008 through 2011, then
declined slightly in fiscal year 2012.

Page 14

GAO-13-429 Segregated Housing Units

may add beds to some SHU cells to accommodate the population at a
given facility. 15
SMUs. As shown in figure 5, from fiscal year 2008 through February
2013, the SMU population increased at a faster rate than SHUs and
ADX—from 144 inmates in fiscal year 2008 to 1,960 inmates as of
February 2013. 16 BOP developed SMU capacity by converting existing
housing units in five BOP facilities to 1,270 total SMU cells, as of
November 2012. By March 2013, BOP closed SMUs in two facilities and
moved those SMU inmates into other SMUs or released them from prison
after serving their sentence.
ADX. From fiscal year 2008 through February 2013, the total ADX inmate
population declined by approximately 5 percent from 475 inmates to 450
inmates. During this period, ADX cells remained stable at 623 cells. 17
According to BOP officials, the ADX population has declined overall since
2008 because of the transfer of inmates out of ADX Step Down to the
general population of another high security prison or because inmates are
being placed in SMUs instead of being placed in ADX. (See fig. 5 for the
trends in population growth for SHUs, SMUs, and ADX from fiscal year
2008 through February 2013).

15

According to BOP officials, many SHUs contain at least one single-bunked cell to hold
inmates on a temporary basis who require being placed alone to maintain institutional
safety. In addition, although the SHU at USP Marion has 50 single-bunked cells, the
warden at USP Marion double-bunked and in some cases triple-bunked the SHU when
several inmates needed to be placed in SHU following a large inmate fight in November
2011.
16

These data include inmates in the SHUs within each SMU.

17

These ADX inmate population and cell data include ADX SHU inmates and cells.

Page 15

GAO-13-429 Segregated Housing Units

Figure 5: Trends in Inmate Population by Type of Segregated Housing Unit, Fiscal Year 2008 through February 2013

BOP’s Monitoring of
Segregated Housing
Policies Varies by
Type of Unit, and
Some Facilities’
Documentation Is
Incomplete
BOP Monitors Compliance
Differently across the
Three Types of Segregated
Housing

BOP Headquarters (HQ) has a mechanism in place to centrally monitor
how prisons implement most segregated housing unit policies, but the
degree of BOP monitoring varies depending on the type of segregated
housing unit. In addition, we identified concerns related to facilities’
documentation of monitoring conditions of confinement and procedural
protections.

Page 16

GAO-13-429 Segregated Housing Units

BOP monitors the extent to which individual prisons implement BOP
policies. BOP’s monitoring includes specific steps to check compliance
with requirements for SHUs and SMUs, but not for ADX. BOP’s Program
Review Division is to perform reviews at least once every 3 years to
ensure compliance with BOP policies. However, BOP can review prisons
more frequently if it identifies performance deficiencies. These follow-ups
can occur at 6-month, 18-month, 2-year or 3-year intervals. 18 These PRD
reviews assess compliance with a variety of BOP policies for inmates in
the general population prison and segregated housing. For example, PRD
assesses compliance with BOP policies on conditions of confinement,
such as whether inmates are given three meals a day, provided exercise
time 5 days a week, and are allowed telephone and other privileges.
Following a review at a facility, PRD issues a program review report,
noting deficiencies and findings at the BOP facility. These PRD
monitoring reviews are done on a prison complex basis, which may
include a variety of housing types, including low, minimum, medium, high
security prisons, and the three types of segregated housing units (e.g.,
SHUs, SMUs, and ADX).
To help HQ ensure that PRD monitoring teams are consistently assessing
the extent that individual prisons are complying with general BOP
correctional program and correctional services policies, BOP provides
training for all program review staff. Additionally, new PRD staff are
provided training and accompany experienced staff before being allowed
to conduct a review independently. 19 This training also covers examiner
independence and how to conduct program reviews. BOP also has a
system designed to address problems identified at the individual prisons,
including PRD follow-up with each prison to assess whether PRD
recommendations were implemented. For example, PRD requires
individual prisons to issue reports within 30 days to explain how they
implemented the PRD recommendations to address problems identified in
18

Generally, PRD conducts 15 program reviews covering 15 different areas of BOP
operations at a minimum of every 3 years but may conduct reviews more frequently if prior
reviews identify overall performance deficiencies.
19
According to BOP officials, BOP provides training for PRD program review staff to
conduct on-site monitoring. For example, on-site monitoring generally includes a team of
an average of about five examiners, depending on the size and security level of the
facility. Before a staff member leads an on-site monitoring visit, he or she is required to
shadow an experienced staff member for about 1 year. BOP also trains all employees in
basic correctional duties and inmate supervision. For example, BOP requires all new
examiners to participate in annual refresher training.

Page 17

GAO-13-429 Segregated Housing Units

program reviews. If PRD determines that the prison response is
insufficient, PRD can request that the prison take corrective actions in a
subsequent follow-up report.
We reviewed 43 PRD follow-up reports and found that PRD concluded
that the facilities generally addressed deficiencies identified in all of the
43 reports. For example, one follow-up report was completed within 30
days and identified steps taken by the prison to address each of the four
problem areas—administrative operations, operational security, inmate
management, and intelligence operations—identified in the PRD report.
To address one of the deficiencies related to improper documentation of
exercise, meals, and supervisor assignments in SHUs, PRD required
additional training for the SHU staff. Following training, the prison
determined that it was in compliance with the relevant requirement,
deficiencies were addressed and PRD closed the recommendation. As
part of PRD’s monitoring process, once the facilities document steps
taken to address deficiencies in their follow-up reports, PRD determines
whether to close the recommendations.
As part of the monitoring process discussed above, PRD also checks
compliance with selected SHU- and SMU-specific policies, but has no
requirement to monitor ADX-specific policies. According to documentation
that BOP provided, we determined that BOP’s monitoring system is
designed to assess whether individual BOP prisons are in compliance
with SHU and SMU procedural policies, such as why an inmate is placed
in segregation, and with the specific conditions of confinement. For
example, BOP’s SHU policy requires that prison staff review the inmate’s
status within 3 days of being placed in administrative detention. To
assess compliance with this SHU policy, BOP monitoring guidance
requires PRD staff to review whether the inmate’s status was reviewed
within 3 days of being placed in administrative detention as required. In
addition, PRD also is to verify that prisons completed their quarterly
audits and operational reviews to ensure that procedural protections for
inmates have been followed and that inmates are housed according to
BOP policies. 20 However, as discussed below, BOP does not have

20
Operational reviews conducted by individual facilities provide a status update on all
areas identified by PRD in the previous program review, including operational issues, such
as human resources, financial management, as well as BOP programs in the facility, such
as SHUs. Individual facilities perform operational reviews about 10 to 22 months between
BOP program reviews, depending on the rating the facility received.

Page 18

GAO-13-429 Segregated Housing Units

requirements in place to monitor similar compliance for ADX-specific
policies. BOP’s monitoring policies for each type of segregated housing
unit are described below.
SHU. BOP policies require that PRD monitor SHU policies and review
documentation of 10 percent of inmates held in SHUs in each facility.
BOP policies also require PRD to select 10 inmate files from those held in
SHU disciplinary segregation for a review of procedural protections and
disciplinary procedures. Further, BOP requires PRD to monitor SHU
specific policies that cover additional requirements to monitor conditions
of confinement and procedural protections. 21 BOP incorporates ACA
monitoring standards as part of its SHU policy. See figure 6 for a
photographic example of a SHU cell, which PRD is required to monitor to
ensure the prison provides conditions of confinement for inmates held in
SHUs.

21

BOP’s Correctional Services Program Review Guidelines requires that PRD review 23
SHU specific policies.

Page 19

GAO-13-429 Segregated Housing Units

Figure 6: Special Housing Unit Cell, U.S. Penitentiary, Leavenworth, Kansas

Page 20

GAO-13-429 Segregated Housing Units

SMU. According to BOP policy, PRD is required to monitor a prison’s
compliance with SMU-specific policies, including those SMU-specific
policies that require prisons to provide specific conditions of confinement
and procedural protections. PRD reviews are required to check
compliance with nine SMU-specific policies such as providing inmates
with 5 hours of recreation per week; an opportunity to shower a minimum
of three times per week; and access to visits, correspondence, and
medical and mental health care. According to BOP officials, BOP
incorporates ACA monitoring standards as part of its SMU policy. BOP
also requires PRD to review 25 SMU inmate case files that cover
conditions of confinement for SMU inmates. See figure 7 for a
photographic example of a SMU recreation area, which PRD is required
to monitor to ensure the prison provides conditions of confinement for
inmates held in SMUs.
Figure 7 : Outdoor Recreation Area, Special Management Unit, U.S. Penitentiary,
Lewisburg, Pennsylvania

Page 21

GAO-13-429 Segregated Housing Units

ADX. ADX inmates are included in any PRD program review that covers
the entire Florence prison complex. While PRD has some oversight over
ADX, PRD does not monitor ADX to the same degree that it monitors
SHUs and SMUs. According to BOP officials, except for inmates held in
ADX-SHUs, PRD is not required to monitor ADX-specific conditions of
confinement–such as exercise, telephone, and visitation–as they do for
SHUs or SMUs. For example, PRD reviews do not check for compliance
with ADX-specific policies, such as whether inmates are afforded a
minimum of 7 hours of recreation per week or the minimum of one 15minute phone call per month in the Control Unit.
The ADX-specific policies for recreation, telephone calls, and visits
allowed vary in each of the three ADX housing units: the Control Unit, the
Special Security Unit, and the Step Down Units. (See fig. 2). According to
BOP officials, PRD does not have monitoring requirements for ADXspecific policies because BOP management has not identified ADX as a
high-risk area that needed specific monitoring requirements due to other
oversight mechanisms. For example, BOP HQ reviews the referral and
placement of all inmates in ADX, including a review of each inmate
placed in the Control Unit every 60 to 90 days to determine the inmate’s
readiness for release from the unit. BOP officials also told us that ADXspecific policies are monitored locally by ADX officials.
However, conditions of confinement in ADX housing units are generally
more restrictive than those in SHUs and SMUs. For example, unlike
SHUs and SMUs, nearly all inmates in ADX are confined to single cells
alone for about 23 hours per day. 22 Also, although BOP HQ has
mechanisms to monitor some procedural protections, and ADX officials
locally monitor ADX-specific policies, BOP HQ lacks oversight over the
extent to which ADX staff are in compliance with many ADX-specific
requirements related to conditions of confinement and procedural
protections to the same degree that it has for SHUs and SMUs. According
to PRD officials, PRD does not assess the extent to which ADX provides
conditions of confinement or procedural protections as required under
ADX policy and program statements because it is not required to do so.
As a result, PRD cannot report to BOP management on the extent of

22

Inmates in Phase 3 of the Special Security Unit and Phases 3 and 4 of the ADX Step
Down Unit may be allowed additional recreation time and interaction with others, and
inmates in Phase 4 of the ADX Step Down Unit may be double-bunked. See figure 2 for a
comparison of conditions of confinement by ADX program unit.

Page 22

GAO-13-429 Segregated Housing Units

compliance with these ADX-specific requirements. With such oversight,
BOP headquarters would have additional assurance that inmates held in
BOP’s most restrictive facility are afforded their minimum conditions of
confinement and procedural protections. See figures 8 and 9 for
examples of a cell in the ADX housing unit and recreation areas, which
PRD is required to monitor to some extent to ensure the prison provides
conditions of confinement for inmates held in ADX.
Figure 8: Interior of Single-Bunked Cell, U.S. Administrative Maximum Facility,
Florence, Colorado

Page 23

GAO-13-429 Segregated Housing Units

Figure 9: Outdoor Recreation Areas, U.S. Administrative Maximum Facility,
Florence, Colorado

Standards for Internal Controls in the Federal Government states that an
effective internal control environment is a key method to help agency
managers achieve program objectives. The standards state, among other
things, that monitoring activities are an integral part of an entity’s
planning, implementing, reviewing, and accountability for stewardship of
government resources and achieving effective results. Specific
requirements for PRD to monitor ADX-specific policies to the same
degree that these requirements exist for SHUs and SMUs could help
provide BOP HQ additional assurance that ADX officials are following
BOP policies to hold inmates in a humane manner, in its highest security,
most restrictive facility. The Acting Assistant Director of PRD agreed that
developing such requirements would be useful to help ensure these
policies are followed.

Page 24

GAO-13-429 Segregated Housing Units

Selected Facilities’
Documentation Problems
Related to Conditions of
Confinement and Inmate
Placement Raise Concerns
about the Monitoring of
Inmates Placed in
Segregation

BOP has a mechanism in place to centrally monitor how prisons
implement most segregated housing unit policies. However, given a
selection of PRD monitoring reports from 20 prisons and our independent
analysis of inmate case files at two federal prisons, we identified concerns
related to how facilities are documenting that inmates received their
conditions of confinement and procedural protections, which are
described below.
PRD monitoring reports. We reviewed 45 PRD monitoring reports from
20 prisons that assessed compliance at general population units and
SHUs and SMUs. PRD identified deficiencies in 38 of these reports,
including documentation concerns in 30 reports. 23 As part of our review,
we found PRD monitoring reports identified deficiencies, such as missing
SHU forms, or incomplete documentation that inmates held in
segregation for at least 22 hours per day received all their meals and
exercise as required. 24 For example, segregated inmates in SHUs and
SMUs are entitled the opportunity to have 1 hour of exercise per day but
the documentation at these prisons did not clearly indicate that these
standards were always observed.
According to our review of 45 PRD reports from 20 prisons, we found that
BOP rated 15 prisons as generally compliant with both BOP policies and
policies specific to SHUs and SMUs. 25 However, while BOP found that
these prisons were generally in compliance with segregated housing unit
policies, most of these prisons had some deficiencies. For example, our
analysis of the PRD reports found that, in 38 of the 45 reviews, PRD
identified deficiencies such as missing documentation, monitoring rounds
not being consistently conducted, or inmate review policies not fully
implemented. (See fig. 10 for common deficiencies.)

23
From fiscal years 2007 through 2011, PRD conducted 187 correctional services
program reviews at 98 BOP prisons.
24

For example, some monitoring reports state SHU-specific deficiencies related to missing
BP-A292 forms, which document meals, recreation, and other conditions of confinement.

25
BOP rated these 15 prisons as generally compliant, with ratings such as acceptable,
good or superior. For the remaining 6 prisons, BOP did not provide similar ratings
because they were based on SHU program review observation reports, or part of
quarterly, summary program reports covering several facilities and programs. According to
BOP policy, program review reports for individual facilities are often assigned an overall
rating ranging from superior, good, acceptable, or deficient to at risk. None of the selected
reports rated individual facilities as deficient or at risk.

Page 25

GAO-13-429 Segregated Housing Units

Figure 10: Common Findings from Our Analysis of 45 Monitoring Reports

Notes: Seven reports contained no deficiencies.
a

Cleanliness refers to the living, sanitary conditions of cells where inmates are held.

b

Documentation covers several issues, such as whether medical staff were signing Special Housing
Unit forms daily as required, inmate files adequately processed and documented inmates’ conditions
of confinement, and information was appropriately keyed into the BOP inmate management data
system, SENTRY, and other databases.
c

Monitoring refers to monitoring of inmate status and segregated housing unit policies, such as SHU
monitoring rounds conducted on every shift, or every 30-minute period, and 10 percent of inmate calls
monitored in the past 12 months.
d

Policy refers to monitoring review policies, investigation, inmate classification and program policies.

e

Procedural protections covers compliance with procedural policies, including whether disciplinary
sanctions were appropriately implemented, informal resolutions in place, and incident reports
expunged for appropriate reasons.
f

Security protocol refers to investigative operational policies, ensuring that key equipment is regularly
checked and hot lists are available to alert staff of inmates who pose a security threat.

g

Timeliness refers to conducting monitoring and operational reviews in a timely fashion, and ensuring
that staff investigations are forwarded to BOP HQ within 120 days as required.
h

Training refers to staff training requirements.

To assess how PRD staff conducted monitoring at prisons, we observed
PRD conducting reviews at one prison complex that included two medium
and high security BOP facilities with SHUs. For example, we found that
PRD staff (1) performed monitoring rounds at SHUs, (2) reviewed log
books, and (3) reviewed inmate files, to determine if the facilities followed
the required procedural protections steps. Given our observations, we
concluded that PRD staff monitored these facilities’ compliance with BOP
policies, as called for in PRD’s monitoring guidelines.
Independent analysis of inmate case files. We also conducted an
independent analysis of BOP compliance with SHU-specific policies at

Page 26

GAO-13-429 Segregated Housing Units

three facilities. Specifically, we reviewed a total of 51 segregated housing
files for inmates held in administrative detention and disciplinary-SHU for
fiscal years 2011 and 2012 at three facilities. 26 We found that these three
facilities were generally complying with BOP policies related to inmate
placement and ensuring procedural protections for inmates placed in
SHU-disciplinary segregation, in light of our review of these selected
files. 27 For example, 42 out of 51 inmate case files we analyzed provided
reasons for inmate placement in SHUs, as required by BOP policies.
However, of the 35 case files we reviewed for inmates held in
administrative-SHU – in which we reviewed conditions of confinement,
monitoring, and procedural protections –only 4 files consistently
documented that the inmates were afforded their rights to recreation and
procedural protections. For example, these 4 files consistently
documented that these inmates in SHUs received 1 hour of exercise a
day, 5 days per week, and that the inmates’ status in segregation was
consistently reviewed within 7 days of being placed in the SHU, as well as
meals and recreation, as required by BOP policy. The remaining 31 of the
35 files did not consistently document that the inmates were afforded
these rights. (See table 1.)
Table 1: Results of Our Independent Analysis of 51 Selected Inmate Case Files: Inmate Review and Procedural Protections
Issues, Fiscal Years 2011-2012
Number of inmate case
files in compliance

Type of segregated housing policy

Description

Documentation provided to inmate
about placement

Inmate was provided a copy of administration detention order
when placed into special housing units (SHU); or inmate
received discipline hearing officer (DHO) report that explained
reasons for placement in disciplinary segregation in SHU.

42 out of 51a

BOP regularly monitors inmate status,
conditions of confinement, and
procedural protections

Inmate in SHU received recreation 5 hours per week; inmate
status in SHUs reviewed on a regular basis (e.g., every 3 days
and 7 days).

4 out of 35b
(subset)

Procedural protections provided in
disciplinary segregation

Inmate provided a hearing process and advised of right to appeal 16 out of 17 c
(subset)
the decision
Source: GAO analysis of BOP (2011-2012) inmate files.

26

In addition to 51 inmate SHU case files, we reviewed 10 CMU case files, which are
described in appendices I and II.

27
Out of a total of 51 inmate case files, we reviewed 35 inmate files related to monitoring
of inmate status, conditions of confinement, and procedural protections in administrative
segregation; and 17 inmate files related to procedural protections in disciplinary
segregation. In fiscal year 2012, the total segregated housing inmate population at these
three facilities was 405.

Page 27

GAO-13-429 Segregated Housing Units

a

One inmate in the sample file was randomly selected in both administrative-SHU and disciplinarySHU.
b

According to a selected sample of 35 SHU inmates, which we selected from the sample size of 10
percent of SHU inmates that PRD inspectors use when conducting monitoring reviews. For details,
please see appendix I.
c

Given a selected sample of 17 disciplinary –SHU inmates and their hearing packets, which we
selected largely based on the sample size that PRD inspectors use when conducting monitoring
reviews. For details, please see appendix I.

Given (1) our review of 45 BOP monitoring reports and (2) our
independent analysis of 51 selected inmate case files at three facilities,
we found that that the facilities did not consistently document conditions
of confinement and procedural protections as required under BOP policy
guidelines. For example, 38 out of the 45 reports identified deficiencies
such as missing documentation, monitoring rounds not being consistently
conducted, or inmate review policies not fully implemented. In our
independent analysis of 51 segregated housing unit case files, we
reviewed 35 files focused on determining if BOP regularly monitors
inmates’ status, conditions of confinement, and procedural protections,
and found documentation-related concerns in 31 out of 35 files.
While our selection of reports and site visits cannot be generalized to all
BOP facilities, the extent of documentation concerns indicates a potential
weakness with facilities’ compliance with BOP policies. Without proper
documentation of inmates’ rights and conditions of confinement, neither
we nor BOP HQ can determine whether facility staff have evidence that
facilities complied with policies to grant inmates exercise, meals, and
other rights, as required. In January 2013, BOP officials agreed with our
finding that BOP monitoring reports regularly identified problems with
documentation. BOP officials said that they believed these were
documentation problems caused by correctional officers forgetting to
document the logs, and not instances where inmates were not getting
their food, exercise, and procedural protections granted under BOP
guidelines. They noted that inmates can use the formal grievance
process, called the Administrative Remedy process, if they believe they
have not been granted these rights.
According to BOP officials, in December 2012, BOP began using a new
software program, called the SHU application in all SHUs and SMUs.
BOP officials told us that this new software program could improve the
documentation of the conditions of confinement in SHUs and SMUs, but
acknowledged it may not address all the deficiencies that we identified.
Because this new software was recently implemented, and BOP did not
provide evidence to the extent that it addressed the documentation

Page 28

GAO-13-429 Segregated Housing Units

deficiencies, we cannot determine if it will mitigate the documentation
concerns. In addition, BOP does not have a plan that provides the
specific objectives of the software program, how it will address the
documentation deficiencies, or specific steps BOP will use to verify that
the software will resolve the documentation problems we identified.
According to best practices in project management, the establishment of
clear, achievable objectives can help ensure successful project
completion. 28 A plan that clarifies the objectives and goals of the new
software program and the extent to which they will address
documentation issues we identified, along with time frames and
milestones, could help provide BOP additional assurance that inmates in
these facilities are being treated in accordance with BOP guidance.

BOP Estimates that
Segregated Housing
Costs More than
Housing Inmates in
General Population

BOP does not regularly track or calculate the cost of housing inmates in
segregated housing units. BOP computes costs by facility or complex,
and does not separate or differentiate the costs for segregated housing
units, such as SHUs, SMUs, and ADX that may be within the complex.
For example, Federal Correctional Complex (FCC) Florence in Florence,
Colorado, contains four different facilities, including ADX, one high
security, one medium security, and one minimum security facility, as well
as different types of housing units within most facilities. 29 Specifically,
within the high security facility, there is a SHU and a SMU. According to
BOP officials, segregated housing unit costs are not separated because
most of the costs to operate a facility or complex apply to inmates housed
in all housing units within the facility or complex. 30 BOP officials further
reported that inmates in a segregated housing unit within a facility share
the same costs under the facility’s total obligations, such as utilities, food
services, health services, and facility maintenance, among other things.
BOP officials also stated that BOP aggregates the cost data for an entire

28

Project Management Institute, The Standard for Program Management © (2006).

29
BOP Federal Correctional Complexes include several institutions with different missions
and security levels located in proximity to one another.
30
BOP applies support costs to an institution’s daily inmate per capita costs based on the
percentage of overall support cost to total BOP obligations. Support costs are institutionrelated expenses that are paid by BOP HQ due to centralized billing procedures, such as
phone charges, workers compensation, payroll processing, information technology
support, and costs of Regional and Central Office and Training Center staff. The inmate
daily per capita rate does not include any one time non-routine costs or construction and
renovation costs.

Page 29

GAO-13-429 Segregated Housing Units

facility or complex to reduce paperwork and streamline operations. BOP
also computes an overall average daily inmate per capita cost by security
level for each fiscal year. See table 2 for BOP’s computation of average
daily inmate per capita costs by security level for fiscal year 2012.
Table 2: BOP Average Daily Inmate Costs per Capita by Security Level, BOP-wide,
for Fiscal Year 2012
BOP security level

Actual average daily inmate costs per capita by
security level, BOP-wide

Minimum security

$59.27

Low security

$74.22

Medium security

$72.91

High security

$93.02

Source: BOP (2011-2012).

Note: Using two databases, an accounting system known as the Financial Management Information
System and the population management system known as SENTRY, BOP calculates the inmate daily
per capita costs by dividing the total obligations under the Salaries and Expenses account by total
inmate days for the entire institution, including general population and segregated housing units.
Total inmate days equal the average inmate daily population multiplied by the number of calendar
days for the fiscal year. Inmate days are obtained via the SENTRY Population Management System.
BOP prepares a report reflecting the overall average and annual daily inmate per capita costs and the
average annual and daily inmate per capita costs by security level for each fiscal year. BOP obtains
the average annual and daily per capita costs by security level by consolidating the information for
each institution within that security level to arrive at an “average” inmate per capita cost for that
security level.

BOP officials stated that segregated housing units are more costly than
general prison population housing units because segregated housing
units require more resources—specifically staff— to operate and
maintain. According to BOP officials, the staff-to-inmate ratio in
segregated housing is significantly higher than in the general prison
population, which makes segregated housing units more expensive to
operate. For example, at one high security facility we visited, we
estimated there was an average of 41 inmates to one correctional officer
in the SHU during a 24-hour period. This contrasts to an inmate-tocorrectional-officer ratio of about 124:1 in general population housing
units in the same facility during a 24-hour period. 31 BOP officials at

31

Based on staffing ratios that BOP officials provided, we estimated the average number
of correctional officers assigned to a segregated housing unit. For example, the number of
correctional officers assigned to a SHU varies by shift (e.g., morning, daytime, and
evening shifts), and we estimated the average number of correctional officers assigned to
the SHU in a 24-hour period, based on the different staffing ratios in each shift.

Page 30

GAO-13-429 Segregated Housing Units

facilities we visited stated that ADX, SMUs, and SHUs require more staff
than general population housing because most of the inmates are
confined to their cells for approximately 22 to 24 hours per day. As a
result, they are dependent on the correctional officers for many of the
activities that those in the general inmate population do for themselves.
For example, at least two correctional officers are needed to escort SHU
and SMU inmates to showers and to recreation cells. Some high security
inmates at SMUs require a three-officer escort each time they leave the
cell. Staff are required to bring meals to inmates in their cells in SHUs,
SMUs and ADX three times each day. In addition, staff are also required
to provide laundry services, daily medical visits, and weekly
psychological, educational, and religious visits to inmates in their cells in
SHUs, SMUs and ADX. In contrast, inmates in general population units
can generally access services in other areas of the facility freely, and
therefore can perform these activities without assistance from correctional
officers.
On January 31, 2013, BOP budget officials provided a snapshot estimate
that compares the daily inmate per capita costs in fiscal year 2012 at
ADX, a sample SMU, a SHU at a sample medium security facility, and a
SHU at a sample high security facility. For example, BOP estimates the
daily inmate per capita costs at ADX are $216.12 compared with $85.74
at the rest of the Florence complex. According to BOP estimates, the
inmate per capita costs at the sample SMU facility are $119.71, which are
higher than per capita costs in general population in BOP’s sample high
security facility, which are $69.41.(see table 3). For its estimates of the
costs to operate SHUs, BOP selected Federal Correctional Institution
(FCI) Beckley for a sample medium security facility and U.S. Penitentiary
(USP) Lee for a sample high security facility. According to a senior BOP
official, BOP did not select these facilities because of costs but because
these facilities are a “typical” medium security and high security facility.
The estimated daily costs per inmate at these two sample facilities in
table 3 are lower and not directly comparable to the system-wide average
daily costs per inmate for medium and high security facilities, as shown in
table 2. Please see appendix I for a description of how BOP calculated its
estimated costs.

Page 31

GAO-13-429 Segregated Housing Units

Table 3: BOP Estimated Daily Inmate Costs per Capita in Selection of Institutions and Different Types of Segregated Housing
Units, by Security Level for Fiscal Year 2012
Estimated daily costs per inmate at sample BOP
facilities
General
population
a
units

BOP sample institution and security level

Total facility, including
Segregated
general population and
housing units segregated housing units

Sample Medium Security, Federal Correctional Institution (FCI) Beckley

$57.41

$78.21 (SHU)

$58.74

Sample High Security facility, U.S. Penitentiary (USP) Lee

$69.41

$93.04 (SHU)

$72.39

n/a

$119.71 (SMU)

$97.51b

c

$216.12 (ADX)

$105.25

Sample Special Management Unit (SMU) facility, (USP) Lewisburg
Federal Correctional Complex (FCC), Florence, including the
Administrative Maximum facility (ADX)

$85.74

Source: BOP estimates.

Notes: BOP provided estimates of costs, which are defined as salary and non salary obligations.
a

These costs exclude the staffing costs for segregated housing units within each facility. BOP
selected FCI Beckley for the sample medium security facility and USP Lee for the sample high
security facility. According to a senior BOP official, BOP did not select these facilities based on costs
but because they considered them to be a “typical” medium security facility and a high security
facility.
b

These costs include the costs of USP Lewisburg, which is an entirely SMU facility, and a minimum
security prison camp.
c

BOP’s estimate of the daily costs per inmate for the Florence FCC excludes the staffing costs for
ADX and includes the staffing costs for the rest of the complex—the general population in the
medium security, high security and camp facilities, the USP Florence SMU, and the SHUs within the
medium and high security facilities.

According to these cost estimates that BOP provided, we estimated that
the total cost of housing 1,987 inmates in SMUs in fiscal year 2012 was
$87 million. If these inmates were housed in a sample BOP medium or
high security facility, the total cost would have been about $42 million and
$50 million, respectively. Also, given BOP estimates, we calculated that
the total cost to house 435 inmates in ADX in fiscal year 2012 was about
$34 million. If these inmates were housed in a medium security or high
security facility, the total costs would have been about $9 million and $11
million, respectively. Moreover, the estimated costs of housing 5,318 SHU
inmates at the cost estimated by BOP for the sample medium security
facility, FCI Beckley, would be $152 million, which is more expensive than
housing inmates in medium security general population housing units
which would cost an estimated $112 million. Similarly, the estimated cost
of housing 2,701 SHU inmates at the cost estimated by BOP for the
sample high security facility, USP Lee, would be $92 million, compared
with housing inmates in high security general population housing units,
which costs an estimated $69 million.

Page 32

GAO-13-429 Segregated Housing Units

According to BOP officials, the use of SMUs can reduce BOP costs. The
officials said that SMUs resulted in reduced assault rates and a reduction
in the number of facility lockdowns. 32 Senior BOP budget officials noted
that there are significant financial costs associated with keeping disruptive
inmates in the general prison population who can cause a serious
incident and lead to costly lockdowns. For example, according to BOP
data, from fiscal years 2007 through 2011, lockdowns and disturbances
led to losses totaling about $23 million. These officials explained that,
during a lockdown, a facility has to use its entire staff to perform security
and custodial duties at the expense of other duties.

BOP Has Not
Evaluated the Impact
of Segregated
Housing Units on
Institutional Safety or
the Impacts of LongTerm Segregation on
Inmates
BOP Has Not Assessed the
Extent to Which
Segregated Housing
Impacts Institutional
Safety

BOP has not assessed the extent to which all three types of segregated
housing units—SHUs, SMUs, and ADX— impact institutional safety for
inmates and staff. 33 Although BOP has not completed an evaluation of the
impact of segregation, BOP senior management and prison officials told
us that they believed segregated housing units were effective in helping
to maintain institutional safety. According to BOP officials, SMUs helped

32
According to BOP, a lockdown is an emergency security practice in which inmates are
locked in their cells and movement is restricted (e.g., dining, showering, recreating, and
programming outside of cells is halted) for immediate or long-term control of a crisis or to
prevent a crisis situation. In addition, BOP officials noted the number of lockdowns
declined during a time frame when the total BOP inmate population rose.
33

BOP’s Office of Research and Evaluation (ORE), which conducts research and
evaluations of BOP programs, reported in March 2012 that a SMU study was underway.
However, as of February 2013, BOP officials could not confirm when the SMU study
would be completed.

Page 33

GAO-13-429 Segregated Housing Units

reduce assault rates BOP-wide and reduced the number of lockdowns
due to conflict and violence from 149 in fiscal year 2008 to 118 in fiscal
year 2010, during a period when the overall inmate population increased.
BOP, however, could not provide documentation to support that these
reductions resulted from the use of SMUs.
Although state prison systems may not be directly comparable to BOP,
there may be relevant information from efforts states have taken to
reduce the number of inmates held in segregation. 34 Five states we
reviewed have reduced their reliance on segregation—Colorado, Kansas,
Maine, Mississippi, and Ohio—prompted, according to state officials, by
litigation and state budget cuts, among other reasons. These states
worked with external stakeholders, such as classification experts and
correctional practitioners, to evaluate reasons why inmates were placed
in segregation and implemented reforms that reduced the number of
inmates placed in segregated housing units. 35 After implementing
segregated housing unit reforms that reduced the numbers of inmates
held in segregation, officials from all five states we spoke with reported
little or no adverse impact on institutional safety. While these states have
not completed formal assessments of the impact of their segregated
housing reforms, officials from all five states told us there had been no
increase in violence after they moved inmates from segregated housing
to less restrictive housing. In addition, Mississippi and Colorado reported
cost savings from closing segregated housing units and reducing the
administrative segregation population. For example, Colorado closed a
high security facility in 2012, which state officials reported led to cost
savings of nearly $5 million in fiscal year 2012 and $2.2 million in fiscal
year 2013. According to Colorado officials, segregation reform efforts
helped lead to the closure of this high security facility. In Mississippi,
reforms in segregation also led to the closure of a supermax facility in

34
For example, both state departments of corrections and BOP are required to house,
clothe, and feed inmates in a safe and secure setting, but BOP noted the federal
correctional system and states are subject to different sentencing laws, which affect the
types of inmates and types of segregation each system manages.
35
State administrative segregation units, some of which are referred to as supermax
facilities, are designed to hold the most dangerous inmates. Supermax facilities are
designed to separate violent or disruptive inmates from general prison population and
generally require confining inmates in a single cell up to 23 hours per day, with minimal
contact with staff or other inmates. Some states have implemented segregated housing
reforms earlier than others. For example, Ohio initiated segregated housing reforms about
a decade ago.

Page 34

GAO-13-429 Segregated Housing Units

early 2010, which Mississippi Department of Corrections officials reported
saved the state nearly $6 million annually.
All five states changed their criteria for placing inmates in segregated
housing, which helped them reduce their segregated inmate populations.
Of the five states, three—Colorado, Mississippi, and Ohio—reviewed and
changed the classification for placing inmates in administrative SHUs and
two—Kansas and Maine—established new or modified the criteria for
placement of inmates in SMUs. For example, in 2007, Mississippi found
that approximately 800 inmates (or 80 percent) did not meet its revised
criteria for placement in administrative segregation. Before reforms,
inmates would generally be transferred directly from admittance to
administrative segregation without consideration of the inmate’s offense
and would generally remain in segregation without regular review of the
inmate’s status irrespective of whether the inmate had committed any
serious misconduct. After implementing reforms, Mississippi adopted new
criteria that stated inmates could be held in administrative segregation
only if they committed serious infractions, were active high-level members
of a gang, or had prior escapes or escape attempts from a secure facility.
According to Mississippi officials, this reform did not lead to an increase in
violence, assault rates, or serious incidents.
In 2011, after a study with external stakeholders that reviewed and
recommended changes to Colorado’s administrative segregation
operations, Colorado revised its policies for placement of inmates in
segregated housing. Subsequent to the external study’s completion,
Colorado began reviewing all offenders that had been in administrative
segregation for longer than 12 months and found that nearly 37 percent
or about 321 inmates in administrative segregation could be moved to
close custody general population. 36 After Colorado revised its
classification criteria and increased oversight of the inmate review
process, the number of inmates held in segregation decreased from 60
per month in 2011 to approximately 20 to 30 per month in 2012.
According to Colorado state officials, these reforms did not lead to an
increase in violence.

36

Custody levels in Colorado refer to level of supervision and are identified as minimum,
minimum restricted, medium, and close.

Page 35

GAO-13-429 Segregated Housing Units

In addition, in 2011, Maine’s Department of Corrections reformed its
inmate placement policies for SMUs. After changing the criteria and
classification for holding inmates in SMUs, Maine significantly reduced
the number of inmates in its 132-cell SMU, by closing a 50-cell section of
its supermax SMU. Inmates removed from the SMU were reintegrated
into a less restrictive, general population setting, and according to
officials, there was no increase in incidents of violence.
While the policies and procedures for segregated housing vary between
states and BOP, and their experiences may not be directly comparable,
there may be lessons for BOP in the states’ experiences reducing their
reliance on segregated housing. According to BOP officials, BOP
generally uses larger states, such as California, Texas, or New York, for
comparison, and that the five states included in our report may not be
comparable with BOP. BOP officials also told us, in response to the
findings from these states, that BOP has more comprehensive
classification criteria, reviews, and procedural protections than the states.
As a result, they indicated that BOP might not have the same reductions
in costs and inmates in SHUs found at the state level. However, without
an assessment of the impact of segregated housing, BOP cannot
determine the extent to which placement of inmates in segregation
contributes to institutional safety and security. Such an assessment is
also important to inform DOJ and congressional decision making about
the extent to which segregation meets BOP’s key programmatic goals for
institutional safety. Our past work and the experience of leading
organizations have demonstrated that measuring and evaluating
performance allows organizations to track progress they are making
toward intended results—including goals, objectives, and targets they
expect to achieve—and gives managers critical information on which to
base decisions for improving their programs. 37

37

For example, see GAO, Information Sharing, DHS Could Better Define How it Plans to
Meet Its State and Local Mission and Improve Performance Accountability, GAO-11-223
(Washington, D.C.: Dec. 16, 2010), Managing for Results: Enhancing Agency Use of
Performance Information for Management Decision Making, GAO-05-927 (Washington,
D.C.: Sept. 9, 2005); Program Evaluation: Studies Helped Agencies Measure or Explain
Program Performance, GAO/GGD-00-204 (Washington, D.C.: Sept. 29, 2000); Agency
Performance Plans: Examples of Practices That Can Improve Usefulness to
Decisionmakers, GAO/GGD/AIMD-99-69 (Washington, D.C.: Feb. 26, 1999); and
Managing for Results: Strengthening Regulatory Agencies’ Performance Management
Practices, GAO/GGD-00-10 (Washington, D.C.: Oct. 28, 1999).

Page 36

GAO-13-429 Segregated Housing Units

Given that BOP maintains data on assault, violence, and lockdown rates
across all prison facilities, BOP senior officials reported that evaluating
the relationship between assault rates and segregation might help them
evaluate the impact of segregated housing. An assessment of the
effectiveness of segregation, including consideration of practices across
local and state correctional systems, could better position BOP to
understand the extent to which different types of segregated housing
units meet BOP mission goals to ensure institutional safety for inmates
and staff.
On January 31, 2013, BOP officials told us that the BOP Director had
authorized the solicitation of an independent review of segregated
housing and, once a contract is awarded, they expect the study to be
completed during fiscal year 2014. BOP officials explained that the
study—with the objective of identifying improvement in BOP’s practice
and policy—is to review segregated housing, including identifying best
practices across the correctional spectrum, such as inmate management,
and mental health, among other areas. According to BOP, the statement
of work for this solicitation requires the recipient to provide an
assessment of the use and practices of segregated housing units in BOP.
However, it is unclear to what extent the review will assess the extent that
segregated housing units contribute to the safety and security of inmates
and staff and ensures that BOP meets its mission goals.

Page 37

GAO-13-429 Segregated Housing Units

BOP Conducts Regular
Assessments of the Mental
Health of Inmates in
Segregated Housing, but
Has Not Conducted an
Evaluation of Impacts of
Long-Term Segregation on
Inmates

BOP psychologists are required to provide an initial intake screening of
each inmate within 30 days of the inmate’s arrival in a BOP facility.
Moreover, BOP requires that psychological staff visit inmates in
segregated housing on a weekly basis and provide psychological
assessments after 30 consecutive days in the SHUs, SMUs, and ADX
Control and Special Security Units. 38 According to BOP’s Psychology
Services Branch Administrator, these weekly visits and psychological
assessments provide staff an opportunity to intervene when and if they
find that an inmate is having difficulty in segregation. BOP also has a
suicide prevention program, which includes training for all staff and
additional supplemental training for staff working in segregation. In
addition, inmates receive information on suicide prevention upon their
arrival at an institution and the availability of mental health services while
in segregated housing. BOP also develops “hot list” memos that are
posted in SHUs to help inform staff of inmates who may have specific
mental health concerns or suicidal tendencies.
While BOP conducts regular assessments of mental health of inmates,
BOP has not evaluated the impact of long-term segregation on inmates.
BOP’s Office of Research and Evaluation (ORE) officials said they have
not studied the impact of long-term segregation on inmates because of
competing priorities related to studying impacts of prisoner reentry, drug
treatment, and recidivism. 39 In addition, BOP officials explained that there
are methodological concerns related to finding an appropriate control
group of inmates to compare with inmates held in segregation. We
recognize the methodological limitations; however, a 2010 Colorado study
that was funded by DOJ identified a comparison group of inmates in order
to evaluate the psychological impact of segregation.

38

For example, the 30-day psychological assessment is to include an interview with the
inmate, assessment of each inmate’s adjustment to his or her surroundings, and any
threat the inmate poses to self, staff, and other inmates. We currently have work under
way for the House Committee on Oversight and Government Reform and the House
Judiciary Committee, which is reviewing the extent to which BOP monitors and assesses
the cost and quality of inmate mental health.

39
BOP’s Office of Research and Evaluation (ORE), which reported that BOP is in the
early stages of a study dedicated to evaluating the impact of SMUs on offenders. BOP
does not yet have an estimated completion date for the study.

Page 38

GAO-13-429 Segregated Housing Units

BOP officials, including psychologists, at four of the six facilities we visited
reported little or no adverse impact of segregation on inmates. 40 Some of
these psychologists and BOP HQ officials cited the 2010 DOJ-funded
study of the psychological impacts of solitary confinement in the Colorado
state prison system. 41 This study showed that segregated housing of up
to 1 year may not have greater negative psychological impacts than non
segregated housing on inmates. While the DOJ-funded study did not
assess inmates in BOP facilities, BOP management officials told us this
study shows that segregation has little or no adverse long-term impact on
inmates. BOP’s Psychology Services Branch Administrator explained that
the impact is dependent on each individual inmate. For example, she told
us that a small number of inmates with mental disorders, such as
schizophrenia, actively seek placement in segregation, and some appear
to function reasonably well in this environment.
We reviewed several studies on the impact of segregated housing on
inmate mental health, and several suggest that long-term segregation or
solitary confinement can cause significant adverse impacts. See appendix
I for information about criteria used to select studies in our review. These
reports describe possible adverse impacts of segregation, including
exacerbation or recurrence of preexisting illnesses, illusions,
oversensitivity to stimuli, and irrational anger, among other symptoms,
although it is unclear how applicable the conditions studied are to BOP
segregated housing. Other reports addressed the possible effect of
segregation on other outcomes, such as recidivism or new convictions
after release from prison. 42 Few reports, however, incorporate a
comparison between inmates in segregation versus inmates not in

40

The psychologist at one facility reported that segregation could adversely impact an
inmate’s mental health.
41

Maureen L. O’Keefe et al, Colorado: One Year Longitudinal Study of the Psychological
Effects of Administrative Segregation, National Institute of Justice Report, U.S.
Department of Justice, Washington, D.C. (2010).
42
For an example of some studies we reviewed, see S. Grassian, “Psychiatric Effects of
Solitary Confinement, “ Journal of Law & Policy, vol. 22, no. 325 (2006); D. Lovell, L. C.
Johnson and K. C. Cain, “Recidivism of Supermax Prisoners in Washington State,” Crime
& Delinquency, vol. 53, no. 4 (2007); D. P. Mears and W. D. Bales, “Supermax
Incarceration and Recidivism,” Criminology, vol. 47, no. 4 (2009); H. A. Miller and G. R.
Young, “ Prison Segregation: Administrative Detention Remedy or Mental Health
Problem,” Criminal Behaviour and Mental Health, vol. 7:85-94 (1997); and P. Smith, “The
Effects of Solitary Confinement on Prison Inmates: A Brief History and Review of the
Literature,” Crime and Justice, vol. 34:441-568 (2006).

Page 39

GAO-13-429 Segregated Housing Units

segregation, limiting the ability to draw conclusions about the impact of
segregation. A comparison of inmates held in segregation with those in
general population would be important for understanding the extent to
which any adverse psychological impacts are unique to long-term
segregation.
While most BOP officials told us there was little or no clear evidence of
mental health impacts from long-term segregation, BOP’s Psychology
Services Manual explicitly acknowledges the potential mental health risks
of inmates placed in long-term segregation. Specifically, it states that
BOP “recognizes that extended periods of confinement in Administrative
Detention or Disciplinary Segregation Status may have an adverse effect
on the overall mental status of some individuals.” In addition, according to
BOP’s mission statement, BOP protects society by confining offenders in
prisons that are, among other things, safe and humane. In our prior work,
we reported that DOJ stresses the importance of evidence-based
knowledge in achieving its mission. Specifically, DOJ’s Office of Justice
Programs (OJP) supports DOJ’s mission by sponsoring research to
provide objective, independent, evidence-based knowledge to meet the
challenges of crime and justice, such as the 2010 Colorado state prison
system study. In addition, BOP’s ORE is responsible for conducting
research and evaluation of BOP programs, but ORE has not conducted
studies on the impact of long-term segregation on inmates. Further,
according to generally accepted government auditing standards,
managers should evaluate programs to provide external accountability for
the use of public resources to understand the extent to which the program
is fulfilling its objectives. 43
To help BOP HQ assess inmates placed in segregation, BOP maintains a
psychology data system (PDS) that is used to document all mental health
screenings and staff visits by psychologists and treatment specialists, and
a Bureau Electronic Medical Record (BEMR) that documents all staff
visits by physicians and medication provided. Given that BOP’s PDS and
BEMR systems maintain data on the mental health of inmates and BOP’s
Psychological Services Manual states there may be potential adverse
effects from long-term segregation, a study that uses existing information
to assess the impact of segregation on inmates would better position

43

See GAO, Designing Evaluations: 2012 Revision, GAO-12-208G (Washington, D.C.:
January 2012).

Page 40

GAO-13-429 Segregated Housing Units

BOP to understand the effects of segregation, including any related to
inmates’ mental health. BOP’s Psychology Services Branch Administrator
agreed that such a study would be useful. As of January 2013, BOP
announced that the bureau is considering the development of procedures
for conducting individualized mental health case reviews of inmates held
in long-term segregation, i.e., inmates housed in SHUs or the ADX
Control Unit for more than 12 continuous months and inmates who fail to
progress through the SMU or ADX General Population Step Down
phases in a timely manner. These reviews would be conducted at BOP
HQ, and if the review found any concerns, the reviewers would contact
prison staff to discuss strategies to reduce or eliminate the identified
mental health concerns. However, the proposal is still under
consideration, has not yet been implemented across all prison facilities,
and we cannot determine the extent to which this proposal will
systematically assess the long-term impact of segregated housing on
inmates.

Conclusions

Over the past 5 years, the number of BOP inmates in segregated housing
has grown at a faster rate than the general inmate population. With more
inmates held under more restrictive conditions, often for months or years
at a time, segregated housing represents an important part of BOP’s
effort to achieve its primary goal of confining inmates in a safe, secure,
and cost-efficient environment. While BOP has a mechanism to centrally
monitor many of its segregated housing unit policies, BOP does not
centrally monitor the policies specific to its most restrictive segregated
prison, the ADX facility. As a result, BOP has less assurance that ADX
staff consistently follows ADX-specific policies to the same degree that
these requirements are followed for SHUs and SMUs. We also found that
prison officials were not consistently documenting that inmates’ conditions
of confinement, such as food and exercise privileges, were being met.
BOP has taken initial steps toward addressing these documentation
issues by implementing new software that may help track the monitoring
of SHUs and SMUs. However, BOP has not developed a plan to clarify
the objectives and goals of the new software program, with time frames
and milestones that explain the extent to which it will address
documentation issues we identified.
BOP officials believe that segregated housing helps maintain institutional
safety. Given BOP’s increased reliance on segregated housing and the
higher costs associated with its use, it is notable that BOP has not studied
the impact of segregated housing on inmates, staff, and institutional
safety. As BOP considers options for conducting a study of segregated

Page 41

GAO-13-429 Segregated Housing Units

housing, BOP may want to consider lessons learned from some state
initiatives that reduced the number of inmates held in segregation without
significant, adverse impacts on violence or assault rates. In addition,
BOP’s own policies recognize that long-term segregation may have a
detrimental effect on inmates. While BOP does regularly check the mental
health of inmates in segregated housing, BOP has not conducted an
assessment of the long-term impact of segregation on inmates.

Recommendations for
Executive Action

To improve BOP’s ability to centrally oversee the implementation of
segregated housing policies, we recommend that the Director of the
Bureau of Prisons take the following two actions:
(1) develop ADX-specific monitoring requirements and
(2) develop a plan that clarifies the objectives and goals of the new
software program, with time frames and milestones, and other means,
that explains the extent to which the software program will address
documentation concerns we identified.
To ensure that BOP’s use of segregated housing furthers BOP’s goal to
confine inmates in a humane manner and contributes to institutional
safety without having a detrimental impact on inmates held there for long
periods of time, we recommend that the Director of the Bureau of Prisons
take the following two actions:
(1) ensure that any current study to assess segregated housing units also
includes an assessment of the extent that segregated housing contributes
to institutional safety, and consider key practices that include local and
state efforts to reduce reliance on and the number of inmates held in
segregated housing and
(2) assess the impact of long-term segregation on inmates in SHUs,
SMUs, and ADX.

Agency Comments
and Our Evaluation

We provided a draft of this report to DOJ for its review and comment.
BOP provided written comments on this draft, which are reproduced in full
in appendix IV. BOP concurred with all of our recommendations. BOP
also provided technical comments on the report on April 19, 2012, which
we incorporated as appropriate.

Page 42

GAO-13-429 Segregated Housing Units

BOP concurred with the first recommendation that BOP develop ADXspecific monitoring requirements. BOP stated that it will conduct a
Management Assessment to identify aspects of the Control Unit at ADX
that are vulnerable to violations of policy. BOP further noted that it would
develop guidelines, as appropriate, to be incorporated into the program
review guidelines. If fully implemented across all ADX housing units,
BOP’s planned actions will address the intent of this recommendation.
BOP concurred with the second recommendation that BOP develop a
plan with timeframes and milestones, to explain the extent the software
program will address documentation concerns. BOP stated that the goal
of the new software program is to help ensure compliance with
requirements to maintain accurate and complete records on conditions
and events in segregated housing units. BOP indicated that they will
conduct a program review by September 30, 2013 to determine if the
SHU documentation deficiencies have been reduced. If fully
implemented, BOP’s planned actions will address the intent of this
recommendation.
BOP concurred with the third recommendation that BOP ensure any
current study to assess segregated housing units also includes an
assessment of the extent that segregated housing contributes to
institutional safety. BOP stated that the current scope of work for the
Special Housing Review and Assessment will include an assessment of
how segregated housing units contribute to institutional safety. BOP
further noted that the scope of work will include consideration of key
practices of local and state correctional systems. If fully implemented,
BOP’s planned actions will address the intent of this recommendation.
BOP concurred with the fourth recommendation that BOP assess the
impact of long-term segregation on inmates in SHUs, SMUs, and ADX.
BOP stated that the assessment of mental health of inmates is consistent
with its public safety mission. BOP stated that BOP will develop and
distribute an expanded mental health screening tool for psychology staff,
which will help conduct a longitudinal assessment of: (1) inmates housed
in SHUs or the ADX Control Unit for more than 12 continuous months;
and (2) those inmates who fail to progress through the SMU or ADX
General Population Step Down phases in a timely manner. In addition,
BOP stated that its review of segregated housing units will include an
evaluation of inmate mental health history and a review of BOP’s mental
health assessment process. If fully implemented, BOP’s planned actions
will address the intent of this recommendation.

Page 43

GAO-13-429 Segregated Housing Units

As agreed with your offices, unless you publicly announce the contents of
this report earlier, we plan no further distribution until 30 days from the
report date. At that time, we will send copies to the Attorney General,
Director of the Bureau of Prisons, selected congressional committees,
and other interested parties. In addition, the report will be available at no
charge on the GAO website at http://www.gao.gov.
Should you or your staff have any questions concerning this report,
please contact David Maurer at (202) 512-9627 or by email at
maurerd@gao.gov. Contact points from our Offices of Congressional
Relations and Public Affairs may be found on the last page of this report.
Key contributors to this report are listed in appendix V.

David C. Maurer
Director, Homeland Security
and Justice Issues

Page 44

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology
Appendix I: Objectives, Scope, and
Methodology

Our objectives for this report were to address the following questions:
1.What were the trends in the Bureau of Prison’s (BOP) segregated
housing unit population and number of cells from fiscal year 2008 through
February 2013?
2.To what extent does BOP centrally monitor how individual facilities
document and apply policies guiding segregated housing units?
3.To what extent has BOP assessed the costs to operate segregated
housing units and how do the costs to confine an inmate in a segregated
housing unit compare with the costs of confining an inmate in a general
inmate population housing unit?
4.To what extent does BOP assess the impact of segregated housing on
institutional safety and the impacts of long-term segregation on inmates?
Overall, to address our questions, we analyzed BOP’s statutory authority
and policies and procedures (e.g., BOP’s inmate placement, procedural
protections, and general conditions of confinement for segregated
housing units—Special Housing Units (SHU), Special Management Units
(SMU), and the Florence Administrative Maximum facility (ADX)—and
Communications Management Units (CMU)). BOP considers CMUs as
self-contained general population housing units. However, since CMU
inmates are separated from general inmate population and have
restrictive conditions, such as 100 percent of their communications
monitored and noncontact visits, we include CMUs within the scope of our
review, as described in appendix II.
To address the first question, we obtained and analyzed BOP’s number of
cells and inmate population data for each type of segregated housing unit
and the CMUs. We focused our data analysis on the period of fiscal year
2008 through February 2013 or the past five fiscal years to the most
recent data available. 1 We assessed the reliability of the inmate
population and number of cells data by (1) participating in an electronic
demonstration of the SENTRY database that BOP uses to generate

1

BOP provided population data at the end of each fiscal year for fiscal years 2008, 2009,
2010, 2011, 2012, and as of February 2013. BOP also provided number of cells data from
fiscal years 2008 through 2012 for SHUs, ADX and CMUs, and from fiscal year 2008
through November 2012 for SMUs.

Page 45

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

required inmate population, (2) reviewing existing information about the
data and the system that produced them, (3) examining the data for
obvious errors and inconsistencies, and (4) interviewing BOP officials
knowledgeable about the data. We determined that the required data
elements were sufficiently reliable for the purposes of this report.
To address the second question, we analyzed BOP’s policies and
procedures pertinent to the monitoring of individual prisons’ compliance
with segregated housing unit policies. To observe the conditions of
confinement, procedural protections, and inmate placement in segregated
housing, we conducted visits to 6 of 119 BOP federal institutions. We
chose these institutions because of different types of segregated housing
units and varying security levels they contain. As shown in table 4, the six
prisons we visited cover the three main types of segregated housing
units—SHUs, SMUs, and ADX—as well as CMUs.
Table 4: Site Visits to BOP Institutions

Institution name

Types of segregated
housing units Included

Security level of the
individual facilities
we visited

Federal Correctional Complex
(FCC) Allenwood Complex

SHUs, SMUs

High security, medium
security

U.S. Penitentiary (USP) Lewisburg

SHUs, SMUs

High security

FCC Florence

SHUs, SMUs, ADX

Administrative
maximum, high
security, medium
security

U.S. Penitentiary (USP)
Leavenworth

SHUs

Medium security

USP Marion

SHUs, CMUs

Medium security

FCC Terre Haute

SHUs, CMUs

High security, medium
security

Source: GAO analysis of BOP information.

During the site visits, we interviewed institutional management officials
and toured the prison to observe inmate housing, recreational areas, food
service, and educational and vocational programming. We also
interviewed officials from BOP’s Program Review Division (PRD), which
leads monitoring reviews, and officials from BOP’s Correctional Programs
Division (CPD), which has primary responsibility for inmate placement
and procedural policies at segregated housing units. Because we did not
visit all BOP facilities and did not randomly select the facilities we visited,
our results are not generalizable to all BOP facilities. However, we

Page 46

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

selected the sites to provide perspectives on different types of segregated
housing units and varying security levels, which were useful in
understanding population trends, BOP monitoring of conditions of
confinement and procedural policies, cost, and the impact of segregated
housing.
Further, for our second question, we assessed BOP’s monitoring for each
type of segregated housing unit by reviewing monitoring policies,
guidelines, and reports. We analyzed BOP’s segregated housing unit
policies and monitoring guidance and compared them against criteria in
Standards for Internal Control in the Federal Government. We also
assessed the methodology and system BOP employs to monitor, identify,
and address deficiencies at prisons; we reviewed 45 of 187 PRD
monitoring reports from 20 of 98 facilities that PRD monitored during the
period from fiscal years 2007 to 2011. We requested a selection of PRD
correctional services monitoring reports, which BOP provided for a variety
of facilities during this time period. In addition, we requested monitoring
reports for the facilities we visited for our site visits. We also reviewed 43
follow-up monitoring reports related to the 45 monitoring reports to
determine the extent that prisons resolved deficiencies identified in the
monitoring reports.
We reviewed these PRD monitoring reports to summarize common
findings and deficiencies relevant to our engagement related to
cleanliness, conditions of confinement, documentation, procedural
protection, monitoring, policy, security protocols, timeliness, and training.
We developed a methodology for selecting these areas to assess the
extent that BOP monitored conditions of confinement, procedural policies,
and other key issues identified in the monitoring reports. One analyst
reviewed each report and highlighted any common findings and
deficiencies noted in the report. A second analyst independently verified
the findings and deficiencies identified. We also interviewed PRD officials
responsible for doing on-site monitoring, and interviewed senior BOP
officials who are responsible for developing monitoring policy guidance to
understand the degree and methodology of monitoring used.
To provide an independent analysis of BOP compliance with segregated
housing unit policies at selected prisons, we developed a data collection
instrument (DCI) according to BOP’s monitoring policies, and guidance
and questions. Our DCI is similar to questions used during PRD periodic
on-site monitoring reviews of segregated housing unit policies at SHUs,
SMUs, and general prison policies at CMUs. We selected two of the six
institutions we visited—FCC Terre Haute and USP Marion. At each

Page 47

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

institution, we selected a random sample of case files from fiscal years
2011 to 2012, of inmates currently housed in segregated housing units—
including SHU-administrative detention, SHU-disciplinary segregation,
and CMUs—totaling 61 files. These 61 inmate case files include 51 SHU
inmate case files, and 10 CMU inmate case files. We selected the inmate
case files from SHUs using the same sample size BOP PRD inspectors
use when conducting correctional services monitoring reviews of SHUs.
For example, according to BOP PRD monitoring guidance for correctional
services reviews of SHUs, PRD inspectors are to review documentation
of 10 percent of inmates currently in SHU to determine whether the
inmates are afforded specific conditions of confinement, inmates’
placement and status in SHU are regularly reviewed, and other SHU
policies are followed. Accordingly, we selected the case files of 10
percent of inmates in SHUs in the two institutions for our analysis.
According to PRD monitoring guidance for the review of disciplinary-SHU,
PRD inspectors are to review 10 disciplinary hearing packets. For our
review, we selected 17 disciplinary inmate case files and hearing packets
because we were interested in understanding the extent to which BOP
provided procedural protections for inmates held in disciplinary-SHU. We
randomly selected the inmate case files from both SHUs and CMUs from
a roster of inmates in each SHU or CMU at the time of our visit. Although
our selection of files was not generalizable to all inmates in all types of
segregated housing units, it provided insights into whether these
institutions were following BOP policy. We used the DCIs to extract
information relevant to BOP’s monitoring policies, inmate placement,
conditions of confinement and procedural protections for inmates held in
SHU-administrative detention, SHU-disciplinary segregation, and CMUs.
One analyst summarize information from the inmate case file, and a
second analyst verified the DCI information collected. A third analyst
reviewed and summarized information collected from the DCIs. In
addition, we observed PRD staff conduct on-site monitoring of SHUs and
CMU at two facilities.
We also reviewed information and documentation received related to
BOP’s new software program, that includes the SHU application,
compared against best practices for project management and criteria in
BOP’s monitoring documentation policies. For example, we reviewed
implementation dates and plans, training materials used across BOP
facilities, and analyzed BOP monitoring policies, and interviewed PRD
officials to understand to what extent the new SHU application addresses
any documentation concerns we identified during our engagement.

Page 48

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

To address the third question, we reviewed BOP fiscal year 2012 average
inmate per capita costs for prisons at each major security level: high
security, medium security, low security, and minimum security levels.
These inmate per capita costs cover all costs associated with the day-today operation of the entire institution, including health services, uniform,
food, programming, and contractual services and equipment costs related
to each prison. According to BOP, the inmate daily per capita costs are
calculated as total obligations as reported in BOP’s Salaries and
Expenses appropriations account divided by total inmate days. Further, in
January 2013, BOP provided a snapshot estimate of fiscal year 2012
inmate per capita costs broken out by segregated housing versus general
population housing at four institutions: (1) USP Lewisburg, a SMU facility;
(2) FCC Florence, which includes ADX Florence; (3) a sample medium
security facility (FCI Beckley); and (4) a sample high security facility (USP
Lee), which both include SHUs. We interviewed BOP officials from the
Administration Division, who have responsibility over financial and facility
management, about their processes for developing the estimates.
According to senior BOP officials, BOP selected these facilities because
they considered them “typical” medium security and high security
facilities. We found BOP’s segregated housing versus general population
housing inmate per capita cost data to be sufficiently reliable for the
purposes of presenting an overview of possible costs. For illustration
purposes, we also used BOP’s estimated segregated housing versus
general population housing inmate per capita cost data, combined with
BOP inmate population data, to estimate the costs of housing the number
of inmates in ADX, all SMUs, and all SHUs, BOP-wide, as of fiscal year
2012 compared with the costs to house these same amount of inmates in
general population housing units for fiscal year 2012. For example, to
estimate the total costs of housing the total SMU inmate population in
SMUs, BOP-wide, for fiscal year 2012, we multiplied BOP’s estimated
daily inmate per capita costs for USP Lewisburg SMU by the total SMU
population times 366 days, or the number of calendar days in 2012. To
estimate the costs of housing this same number of SMU inmates in
general population housing in a medium security or high security facility,
we multiplied the total SMU population, BOP-wide, by BOP’s estimated
daily inmate per capita costs for the sample medium facility, FCI Beckley,
times 366 days, and estimated daily inmate per capita costs for the
sample high security facility, USP Lee, times 366 days, respectively.
To address the fourth question, we reviewed BOP’s policies, including
program objectives, for each segregated housing unit and policies
governing the provision of mental health services to inmates in
segregated housing units. We also reviewed BOP lockdown data from

Page 49

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

fiscal year 2008 through fiscal year 2012. We also interviewed officials
from BOP’s Correctional Programs Division (CPD), which also includes
the Psychology Services Branch that is responsible for mental health
services. We also interviewed officials from BOP’s Office of Research and
Evaluation (ORE), who produce reports and research corrections-related
topics. During these interviews, we discussed the lack of BOP studies
that assess the impact of segregated housing units on institutional safety
and inmates and staff, and their views on the impact of long-term
segregation, including their views on the impact of segregation on
inmates, including those with mental illness. We also discussed the
impacts of segregation with officials from the Council of Prison Locals, the
union that represents all nonmanagement staff working in BOP facilities.
To identify actions states have taken regarding segregated housing that
may be relevant to BOP, we reviewed actions taken by five states—
Colorado, Kansas, Maine, Mississippi, and Ohio. We selected these five
states because they (1) were involved in addressing segregated housing
reform and (2) had taken actions to reduce the number of inmates held in
segregation. For each of the five selected states, we reviewed relevant
documents on segregated housing, and in four states we reviewed
placement policies. For four of the five selected states, we reviewed
relevant reports on their segregated housing unit conditions for context.
While conducting site visits to BOP prisons in Kansas and Colorado, we
also visited state correctional facilities in those two states. We interviewed
corrections officials at these facilities and the other states regarding
reasons for reducing the segregated housing unit population and any
reported impact of the segregated housing unit reforms on institutional
safety. While the reports and results from our interviews are not
representative, they provided us with perspectives on state actions to
reduce segregated housing
There are dissimilarities between federal and state prison systems—
legally and structurally, to name a few––that limit the comparability
between federal and state correctional systems. We are unable to
generalize about the types of actions other states have taken to reform
segregated housing policies and reduce the number of inmates held in
segregation and any effects. Nevertheless, the information we obtained
through these visits provided examples of state responses to reforming
segregation and reducing inmates housed in segregated housing units.
We also discussed with BOP officials the state actions we identified.
Further, to identify the universe of reports and studies that describe,
evaluate, or analyze the impact of segregated housing, including any

Page 50

GAO-13-429 Segregated Housing Units

Appendix I: Objectives, Scope, and
Methodology

long-term impacts associated with mental illness, we used a multistaged
process. First, we (1) conducted key word searches of criminal justice,
legal, and social science research databases; (2) searched academic,
nongovernment and stakeholder interest group-related Web sites, such
as those of Vera, American Civil Liberties Union (ACLU), and Urban
Institute, (3) reviewed bibliographies, published summaries, metaanalyses, and prior GAO reports on segregated housing; and (4) asked
academic corrections experts to identify evaluations. Our literature search
identified over 150 documents, which included articles, opinion pieces,
published reports, and studies related to segregated housing. We further
identified studies that compared inmates in segregated housing with
inmates in the general population. We reviewed these reports and studies
to gain a broader understanding of the potential impacts of segregated
housing and of the extent and quality of research available on the subject.
We compared BOP’s mechanisms for evaluating the impact of
segregated housing units on institutional safety, or the impacts of longterm segregation on inmates, with BOP’s policies and mission
statements.
We conducted this performance audit from January 2012 to April 2013 in
accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain
sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions given our audit objectives. We believe that the
evidence obtained provides a reasonable basis for our findings and
conclusions for our audit objectives.

Page 51

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units
Appendix II: Communications Management
Units

BOP established CMUs in 2006 and 2008, in two institutions to house
inmates who require increased monitoring of their communications with
the public to protect the safety, security, and orderly operation of BOP
facilities and the public. 1 Inmates in CMUs have 100 percent of their
communications monitored by BOP officials and are allowed only
noncontact visits with family and friends. According to each prison’s
institution supplement guidelines, CMUs are self-contained general
population housing units in which inmates reside; eat; and participate in
all educational, religious, visiting, unit management, and work
programming in the unit, similar to general population inmates. 2 From
fiscal year 2008 to February 2013, the total CMU population increased
from 64 inmates to 81 inmates. 3 See figure 11 for an overview of CMUs.

1

BOP operates CMUs in two medium security BOP facilities. BOP established a CMU at
FCI Terre Haute in 2006 and a second CMU in USP Marion in 2008. As part of this
review, we visited both CMUs.

2

According to BOP officials, BOP does not have a national policy governing CMUs.
Rather, the prisons containing the CMUs have developed institution supplements, or local
guidelines and procedures that govern the CMU inmate review process and conditions of
confinement. The CMU institution supplements for FCI Terre Haute and USP Marion are
generally the same.

3

BOP has had 113 CMU inmate cells since fiscal year 2008. The CMU population and
cell totals include the CMU SHUs.

Page 52

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

Figure 11: Overview of Communications Management Units

a

BOP promulgated a proposed rule in April 2010 to establish specific parameters for CMU operations
and put inmates and the public on notice of CMUs; the comment period closed in June 2010, and a
final rule has not been published. Under the proposed rule, BOP’s Assistant Director of the
Correctional Programs Division would become the authorizing official for CMU designations.
Communications Management Units. 75 Fed. Reg. 17,324 (proposed Apr. 6, 2010) (to be codified at
28 C.F.R. pt. 540, subpt. J).
b

The total CMU inmate population and number of cells includes SHUs within the CMUs.

Placement Criteria and
Procedural Policies

According to a BOP memorandum, BOP places inmates in CMUs for
several reasons, including conviction, conduct or involvement related to

Page 53

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

international or domestic terrorism, and commission of prohibited activity
related to misuse or abuse of approved communication methods while
incarcerated, or for other reasons. 4
Inmates referred to CMUs do not receive a hearing prior to placement in
CMUs. 5 According to the prison’s institution supplement guidelines, an
inmate assigned to a CMU is to receive a notice of transfer to the CMU
within 5 days of arrival in the unit, including reasons for placement and
notice of the right to appeal the transfer through the administrative
remedy process. At the institution, prison officials are to review the CMU
inmate’s status every 6 months, according to BOP’s national policy that
applies to all inmates in BOP custody. The guidelines also call for prison
officials to regularly review an inmate’s readiness to be transferred out of
a CMU by examining a number of factors, including programming needs
and if the original reasons for CMU placement still exist. After conducting
the review, prison officials may recommend to the warden that an inmate
be transferred out of the CMU.

General Conditions of
Confinement

All CMU inmates are segregated from the general population in selfcontained housing units to regulate and monitor their communications
with persons in the community. However, they are allowed to congregate
outside their cells, but within these self-contained housing units, for 15 to
16 hours per day like inmates in the general population. 6 Inmates in
CMUs require 100 percent live monitoring of their telephone calls and
social visits, and a review of their incoming and outgoing social mail. 7 All
4

According to BOP’s Memorandum for Continued CMU Designation, dated October 14,
2009, BOP also refers inmates to CMUs for the following reasons: attempt or propensity to
contact victims of the inmate’s current offense of conviction, and/or conviction or conduct
indicating a propensity to coordinate illegal activity through communication with persons in
the community, or evidence of a potential threat to the safety and security of prison
facilities or the public, as a result of the inmate’s unmonitored communication with persons
in the community.

5

As of February 2013, there is litigation pending in federal court addressing the issue of
whether the lack of a hearing prior to placement in a CMU adequately protects inmates’
constitutional rights. See Aref v. Holder, No. 10-0539 (D.D.C. filed Apr. 1, 2010).

6

According to the CMU institution supplements, inmates in the USP Marion CMU will
generally be housed in single-bunked cells, and inmates in the FCI Terre Haute CMU will
generally be housed in double-bunked cells.

7

Legal and special mail (e.g., attorney, federal courts) can be sealed and delivered to unit
management.

Page 54

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

telephone calls and social visits are also recorded, and they must occur in
English only, unless the call is previously scheduled and conducted
through simultaneous translation monitoring. Other than increased
communications monitoring, BOP officials stated that conditions of
confinement in these units are the same as conditions of confinement for
inmates in other medium security general population housing units. This
includes (1) access to medical and mental health services; (2) meals that
meet inmate dietary requirements served in common dining areas; (3)
access to recreation and leisure in a common area daily up to 16 hours
per day, including table games, television in the common areas, and
some aerobic exercise equipment; (4) religious service opportunities; and
(5) access to law library services.
Also, like general population housing, each CMU contains a SHU
dedicated to housing inmates in need of being placed in SHUadministrative detention or SHU-disciplinary segregation status. See
figures 12 and 13 for photographs of a CMU.

Page 55

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

Figure 12: Communications Management Unit Cell, Terre Haute, Indiana

Page 56

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

Figure 13: Communications Management Unit Common Area, Marion, Illinois

BOP Monitoring of CMUs

As previously discussed, BOP headquarters has a mechanism in place to
centrally monitor how prisons implement most housing unit policies, but
the degree of monitoring varies depending upon the type of housing. In
addition, we reviewed PRD monitoring reports, assessed how PRD
conducted monitoring at one of the two prisons with CMUs, and
conducted an independent analysis of BOP compliance at these two
prisons.
At one of the two prisons with CMUs we visited, we observed that PRD
checked compliance with general prison policies, as well as SHU-specific
policies, but PRD does not have requirements to monitor CMU-specific
policies. CMU inmate files may be included in any PRD program review
that covers the entire prison complex. According to BOP officials,
although not required, BOP may randomly select some CMU inmate files

Page 57

GAO-13-429 Segregated Housing Units

Appendix II: Communications Management
Units

as part of the prison complex during periodic PRD reviews. 8 However,
PRD does not have requirements to monitor CMU-specific policies found
in the institution supplement guidelines. According to BOP officials,
additional monitoring for CMUs is not required because they do not have
the same kinds of restrictive conditions of confinement that are the
subject of SHU- and SMU-specific monitoring steps.
As part of our review of PRD monitoring reports, we found that 8 of the 45
monitoring reports covered these two prisons with CMUs. PRD found that
these prisons were in general compliance with BOP policies, and none of
these PRD monitoring reports identified any findings or deficiencies
specific to the CMUs. To assess how PRD staff conducted monitoring at
one of these prisons, we observed PRD conduct reviews at the CMUs in
accordance with PRD guidelines. In light of our observations, we found
that PRD staff (1) performed monitoring rounds at CMUs, (2) reviewed log
books, and (3) reviewed inmate files, to determine if the prisons followed
the required procedural protections steps. In addition, we also conducted
an independent analysis of BOP compliance with CMU-specific policies at
the two prisons with CMUs. Specifically, we reviewed a total of 10 files for
inmates held in CMUs for fiscal years 2011 and 2012 at these two
facilities. We found that all 10 inmate case files we analyzed provided
reasons for inmate placement in CMUs, as required by BOP institution
supplements. However, similar to the documentation problems we noted
in the body of the report, we found documentation deficiencies during our
review of the CMU files. For example, 2 out of the 10 inmate case files we
reviewed did not include documentation that unit team staff regularly
monitored the inmate’s CMU status every 6 months and ensured that
inmates were afforded their rights to programming activities. Without
complete documentation, BOP headquarters cannot be assured that
inmates in CMUs are receiving the procedural protections and conditions
of confinement to which they are entitled, as stated in BOP policy and
institution supplements.

8

According to BOP officials, BOP considers CMUs as general population housing, and
thus does not require separate oversight like SHUs or SMUs.

Page 58

GAO-13-429 Segregated Housing Units

Appendix III: Location and Length of Stay in
BOP Segregated Housing Units
Appendix III: Location and Length of Stay in
BOP Segregated Housing Units

Location of Segregated
Housing Units

BOP has segregated housing units in prisons located throughout the
country. For example, BOP has SHUs in 109 out of its 119 facilities.
Three facilities have SMUs. See figure 14 for a map of the locations of
each type of segregated housing unit.

Figure 14: Locations of Segregated Housing Units within BOP Facilities

Length of Stay

According to BOP, the length of stay inmates serve in segregated
housing units varies, and BOP does not track an inmate’s total length of
stay or establish a maximum length of stay for inmates in any type of
segregated housing unit. An inmate’s length of stay in segregated
housing varies depending on the inmate’s program needs and status,

Page 59

GAO-13-429 Segregated Housing Units

Appendix III: Location and Length of Stay in
BOP Segregated Housing Units

reason for placement, and behavior while in the unit. BOP policy provides
the expected length of stay for some segregated housing units. For
example, according to BOP officials, placement of inmates in SHUs is
intended to be temporary. Inmates may be sanctioned to 1 to 18 months
in a SHU for disciplinary reasons, given the severity of infraction. Also,
BOP policy states inmates placed in SMUs, the ADX Step Down Units,
and ADX Special Security Unit may participate in structured, phased
programs where they can progress or “step down” to general population
after approximately 18 to 36 months if they maintain good behavior.
However, according to BOP officials, an inmate may remain in any of the
segregated housing units if the inmate continues to be disruptive or BOP
officials determine through the review process that the inmate’s original
reason for placement still exists.

Page 60

GAO-13-429 Segregated Housing Units

Appendix IV: Comments from the Bureau of
Prisons
Appendix IV: Comments from the Bureau of
Prisons

Page 61

GAO-13-429 Segregated Housing Units

Appendix IV: Comments from the Bureau of
Prisons

Page 62

GAO-13-429 Segregated Housing Units

Appendix IV: Comments from the Bureau of
Prisons

Page 63

GAO-13-429 Segregated Housing Units

Appendix IV: Comments from the Bureau of
Prisons

Page 64

GAO-13-429 Segregated Housing Units

Appendix IV: Comments from the Bureau of
Prisons

Page 65

GAO-13-429 Segregated Housing Units

Appendix V: GAO Contact and Staff
Acknowledgments
Appendix V: GAO Contact and Staff
Acknowledgments

GAO Contact

David C. Maurer, (202) 512-9627 or maurerd@gao.gov

Staff
Acknowledgments

In addition to the contact named above, Ned George, Assistant Director;
Pedro Almoguera; Lori Achman; Carla Brown; Jennifer Bryant; Frances
Cook; Michele Fejfar; Eric Hauswirth; Lara Miklozek; Linda Miller; Jessica
Orr, Meghan Squires; Helene Toiv; and Yee Wong made key
contributions to this report.

(441036)

Page 66

GAO-13-429 Segregated Housing Units

GAO’s Mission

The Government Accountability Office, the audit, evaluation, and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and
policies; and provides analyses, recommendations, and other assistance
to help Congress make informed oversight, policy, and funding decisions.
GAO’s commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of
GAO Reports and
Testimony

The fastest and easiest way to obtain copies of GAO documents at no
cost is through GAO’s website (http://www.gao.gov). Each weekday
afternoon, GAO posts on its website newly released reports, testimony,
and correspondence. To have GAO e-mail you a list of newly posted
products, go to http://www.gao.gov and select “E-mail Updates.”

Order by Phone

The price of each GAO publication reflects GAO’s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO’s website,
http://www.gao.gov/ordering.htm.
Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537.
Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional information.

Connect with GAO

Connect with GAO on Facebook, Flickr, Twitter, and YouTube.
Subscribe to our RSS Feeds or E-mail Updates. Listen to our Podcasts.
Visit GAO on the web at www.gao.gov.

To Report Fraud,
Waste, and Abuse in
Federal Programs

Contact:
Website: http://www.gao.gov/fraudnet/fraudnet.htm
E-mail: fraudnet@gao.gov
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional
Relations

Katherine Siggerud, Managing Director, siggerudk@gao.gov, (202) 5124400, U.S. Government Accountability Office, 441 G Street NW, Room
7125, Washington, DC 20548

Public Affairs

Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, DC 20548

Please Print on Recycled Paper.