ORIGINAL ARTICLE

Torture vs Other Cruel, Inhuman,
and Degrading Treatment
Is the Distinction Real or Apparent?
Metin Bas¸og˘lu, MD, PhD; Maria Livanou, PhD; Cvetana Crnobaric´, MD

Context: After the reports of human rights abuses by the

US military in Guantanamo Bay, Iraq, and Afghanistan,
questions have been raised as to whether certain detention and interrogation procedures amount to torture.
Objective: To examine the distinction between vari-

ous forms of ill treatment and torture during captivity
in terms of their relative psychological impact.
Design and Setting: A cross-sectional survey was conducted with a population-based sample of survivors of
torture from Sarajevo in Bosnia and Herzegovina, Banja
Luka in Republica Srpska, Rijeka in Croatia, and Belgrade in Serbia.
Participants: A total of 279 survivors of torture accessed

through linkage sampling in the community (Banja Luka,
Sarajevo, and Rijeka) and among the members of 2 associations for war veterans and prisoners of war (Belgrade).
Main Outcome Measures: Scores on the Semi-

structured Interview for Survivors of War, Exposure to
Torture Scale, Structured Clinical Interview for DSMIV, and Clinician-Administered PTSD (posttraumatic
stress disorder) Scale for DSM-IV.

W

Results: Psychological manipulations, humiliating treatment, exposure to aversive environmental conditions, and
forced stress positions showed considerable overlap with
physical torture stressors in terms of associated distress
and uncontrollability. In regression analyses, physical torture did not significantly relate to posttraumatic stress
disorder (odds ratio, 1.41, 95% confidence interval, 0.892.25) or depression (odds ratio, 1.41, 95% confidence interval, 0.71-2.78). The traumatic stress impact of torture (physical or nonphysical torture and ill treatment)
seemed to be determined by perceived uncontrollability
and distress associated with the stressors.
Conclusions: Ill treatment during captivity, such as psychological manipulations, humiliating treatment, and
forced stress positions, does not seem to be substantially different from physical torture in terms of the severity of mental suffering they cause, the underlying
mechanism of traumatic stress, and their long-term psychological outcome. Thus, these procedures do amount
to torture, thereby lending support to their prohibition
by international law.

Arch Gen Psychiatry. 2007;64:277-285

IDELY ACCEPTED

definitions of torture, such as that
provided by the
Convention Against
Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment,1 refer to
torture as “severe pain or suffering, whether
physical or mental,” inflicted on a person

See also page 275
Author Affiliations: Section of
Trauma Studies, Institute of
Psychiatry, King’s College,
University of London, London,
England (Drs Bas¸og˘lu and
Livanou); and Department of
Psychiatry, Clinical Hospital
Zvezdara, Belgrade, Serbia
(Dr Crnobaric´).

for particular purposes. After reports2 of human rights abuses by the US military in
Guantanamo Bay, Iraq, and Afghanistan,
a US Defense Department working group
report3 on detainee interrogations and a
US Justice Department memorandum4 on
US torture policy argued for a fairly narrow definition of torture that excludes

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
277

mental pain and suffering caused by various acts that do not cause severe physical
pain. According to this definition, various interrogation and detention procedures, such as blindfolding, hooding,
forced nudity, isolation, forced standing,
rope bondage, deprivation (of sleep, light,
water, food, or medical care), and psychological manipulations designed to break a
person’s resistance (eg, humiliating treatment or other acts designed to create fear,
terror, or helplessness in the detainee), do
not constitute torture. This report also
stated that proof of “severe mental pain or
suffering” associated with torture requires proof of “prolonged mental harm,”
such as the development of posttraumatic stress disorder (PTSD). The implications of such a narrow definition of tor-

WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

ture have raised serious concerns in the human rights
community.
The problems in defining torture arise in part from
lack of sufficient knowledge on 3 issues: (1) the severity
of mental suffering associated with particular stressors
during detention or captivity, (2) the psychological mechanisms by which these stressors exert their traumatic
impact, and (3) their long-term psychological effects. Various forms of ill treatment, whether physical or psychological, might share the same mechanisms of traumatization as torture and therefore lead to similar psychological
outcomes.5 To our knowledge, no study has yet examined these issues, partly owing to the difficulties in conducting research in this area.
A recent study6 of war survivors in former Yugoslavia, 279 of whom had a history of torture, provided a valuable opportunity to examine some of these issues. Using
the structured Exposure to Torture Scale, we obtained
information on the degree of perceived distress and loss
of control associated with 46 stressor events commonly
reported by survivors of torture. We tested the following hypotheses: (1) forms of torture that involve severe
physical pain are associated with more distress and uncontrollability than all other stressors that do not involve severe physical pain and (2) exposure to physical
forms of torture is more likely to be associated with PTSD
and depression than stressors that do not involve severe
physical pain. Because substantial evidence7-12 during the
past 30 years suggests that unpredictability and uncontrollability of stressors play a role in the development of
anxiety and fear, we also tested the hypothesis that perceived distress and uncontrollability of the torture stressors, rather than mere exposure to them, would be associated with a greater likelihood of PTSD and depression
(hypothesis 3).
METHODS

STUDY DESIGN AND SAMPLING
Because the details of the method are presented elsewhere,6 they
will be summarized herein. The study involved 1358 survivors of war trauma from Banja Luka, Sarajevo, Rijeka, and Belgrade. “Target sampling”13 was used to ensure adequate representation of 5 survivor groups of interest (combat veterans,
torture survivors, refugees, internally displaced people, and survivors of the North Atlantic Treaty Organization bombardment of Belgrade) and sufficient numbers of cases of PTSD in
the sample to test the study hypotheses. We attempted to minimize sampling bias in the targeted groups as much as possible
by using linkage sampling.14 This method involved tracing and
contacting survivors in the community through “key informants” (project staff and their acquaintances, contacts in various nongovernment organizations, and the study participants). To minimize sampling bias with respect to psychological
status, the key informants were asked to make a list of their
friends or acquaintances who had an experience of a particular index stressor, disregarding any available information on
their psychological status. These survivors were then contacted and invited to participate in the study. Once the interview was completed, each survivor was asked to list all friends
or acquaintances with a similar trauma experience. This process continued until the targeted sample size for a particular
stressor of interest was achieved.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
278

The present study is based on a subsample of 279 survivors
of torture (102 from Belgrade, 58 from Rijeka, 52 from Sarajevo, and 67 from Banja Luka). The Belgrade sample included
mainly former army conscripts recruited from 2 associations
for war veterans and prisoners of war in Belgrade. The Banja
Luka and Sarajevo samples included former soldiers and civilian ex-detainees recruited from the community. The Rijeka
sample consisted mostly of men who had been captured in Vukovar and sent to collective camps. The inclusion criteria were
experience of torture, age 18 to 65 years, literacy, absence of
past or present psychotic illness, and willingness to give written consent for participation in the study.

MEASURES
The measures included the Semi-structured Interview for Survivors of War,6 the Structured Clinical Interview for DSM-IV
(version 2),15 and the Clinician-Administered PTSD Scale for
DSM-IV.16 The Semi-structured Interview for Survivors of War,
modified from an earlier version for survivors of torture,17 included an Exposure to War Stressors Scale (54 war-related stressors) and an Exposure to Torture Scale that elicited information on 46 different forms of torture and related stressors. Each
stressor event was rated as absent or present and also for associated distress (0=not at all distressing, 1=slightly distressing, 2=moderately distressing, 3=fairly distressing, and 4=extremely distressing) and loss of control (0=completely in control
and 4=not at all in control/entirely helpless). The perceived
distress rating reflected anxiety, fear, discomfort, or any other
distressing emotion experienced during the event. The perceived control rating was based on detailed information about
the behavioral and cognitive coping strategies (eg, acts designed to protect oneself from life-threatening or distressing
events, cognitive dissociation, distraction, and distressreducing beliefs/thoughts/interpretations relating to the event)
used in avoiding a particular stressor event or lessening the pain
or distress during the event. A Global Distress Rating and a
Global Sense of Control Rating were also used to assess the survivors’ overall perceived distress or loss of control during the
torture. These ratings were demonstrated to have sufficient validity in a previous study.18 The Clinician-Administered PTSD
Scale ratings of social and occupational disability (0 = no
adverse impact and 4 = extreme impact, little or no social/
occupational functioning) were also used as measures of posttorture psychiatric status. Because these ratings were highly intercorrelated, they were summed to derive a measure of social/
occupational disability.
The interviews were conducted between March 11, 2000, and
July 30, 2002, by 21 psychiatrists and psychologists from 4 study
sites. They were standardized with a senior psychiatrist (C.C.)
who had received extensive training from the main author (M.B.)
for this purpose. The concordance rates in assessments, detailed in a previous study,6 were satisfactory. Written informed
consent was obtained for all study procedures. The study was
approved by the research ethics committee of the Institute of Psychiatry, King’s College, University of London.

STATISTICS
The assessment of PTSD was conducted in relation to the most
distressing event reported by the participants. Because some
survivors reported a war-related traumatic event other than torture as their most distressing experience, a diagnosis of PTSD
in their case related to that event and not to their torture. These
cases were included in all analyses involving the Exposure to
Torture Scale but were excluded from those involving the diagnosis of PTSD. Between-group comparisons involved ␹2 tests
or 1-way analyses of variance. Hierarchical logistic regression

WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

analyses examined the factors related to current PTSD and depression. A multiple regression analysis (sequential entry) was
conducted to examine the predictors of social/occupational disability. The data were analyzed by one of us (M.B.) using a software program (SPSS Version 12; SPSS Inc, Chicago, Ill).
RESULTS

SAMPLE CHARACTERISTICS
Two hundred forty-one participants (86.4%) were men
and 192 (68.8%) were married. The mean (SD) participant age was 44.4 (10.2) years. Fifty-one participants
(18.3%) were Bosniaks, 56 (20.1%) were Croats, 165
(59.1%) were Serbs, and 7 (2.5%) were of mixed or other
ethnic origin. Thirty-nine participants (14.0%) had a primary school education, 168 (60.2%) had a secondary
school education, 35 (12.5%) had a high school education, and 37 (13.3%) had a university or postgraduate
education. One hundred ninety-one survivors (68.5%)
had combat experience, 141 (50.5%) had refugee experience, 140 (50.2%) had internal displacement experience, 192 (68.8%) had detention camp experience, and
178 (63.8%) had prisoner-of-war experience. Participants reported a mean (SD) of 19.0 (5.9) war-related
stressors and 19.3 (7.5) torture-related stressors. The mean
(SD) time since last torture was 96.3 (24.6) months.
Excluding the 49 survivors whose PTSD assessments
were conducted in relation to war stressors other than
torture, 174 (75.7%) of the 230 survivors had lifetime
PTSD and 128 (55.7%) had current PTSD (according to
the Clinician-Administered PTSD Scale), 39 (17.0%) had
a current major depressive episode, 40 (17.4%) had a past
major depressive episode, and 34 (14.8%) had at least 1
anxiety disorder other than PTSD. Of the 128 survivors
with PTSD, 38 (29.7%) also had major depression (16.5%
of the whole sample).
PERCEIVED DISTRESS AND LOSS OF CONTROL
ASSOCIATED WITH STRESSORS
Table 1 provides the percentages of participants rating

the stressors as fairly/extremely distressing and slightly/
not controllable at all and the mean distress and control
ratings. The stressors were grouped under 7 categories to
facilitate comparison. The physical torture category included stressors that involved physical pain. The other categories included mostly stressors that are regarded in the
US Justice Department memorandum4 as acts that do not
constitute torture because they do not involve infliction of
severe physical pain (hereafter referred to as nonphysical
stressors). Although sexual torture also has a strong humiliation component, it was taken as a separate category
to allow comparison with the other stressors.
More than 80% of the survivors rated 30 stressors as
fairly to extremely distressing, which included, in addition to physical torture, various psychological manipulations, humiliating treatment, and forced stress positions. The mean distress ratings for physical torture
stressors ranged from 3.2 to 3.8. The mean distress ratings for 16 (48.5%) of the 33 stressors from the other categories were in the same range. Sham executions, wit(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
279

nessing torture of close ones, threats of rape, fondling of
genitals, and isolation were associated with at least as
much if not more distress than some of the physical torture stressors. There was thus substantial overlapping between physical torture and other stressors in terms of associated distress. The control ratings also showed a similar
pattern.
High distress ratings for 30 stressors did not necessarily mean that the distress scale lacked sufficient discriminatory power. This finding reflected the fact that
certain stressors occur concurrently in a torture setting,
reinforcing the effects of each other, as will be discussed
later. The distress ratings showed sufficient variability
across different types of stressors. For example, the mean
distress ratings for physical torture stressors were higher
than those for some stressors in the deprivation of basic
needs category (eg, deprivation of medical care, prevention of personal hygiene, food deprivation, denial of privacy, and infested surroundings). Another finding that
supports the validity of the Exposure to Torture Scale is
that control ratings were generally lower than distress ratings, suggesting that the 2 ratings measured different constructs. This finding makes theoretical sense in that survivors often can exercise some degree of control over the
stressors, despite experiencing high levels of distress.5
To facilitate comparison across the stressor categories, the distress ratings for the items in each category
were averaged to obtain a single distress rating in relation to each stressor category. The same procedure was
also conducted with the control ratings. The Figure shows
the mean distress and control ratings for each stressor
category. The mean distress ratings were by and large similar across stressor categories.
RELATIONSHIP BETWEEN
CUMULATIVE IMPACT OF STRESSORS
AND PSYCHOLOGICAL OUTCOME
The concurrent occurrence of stressors in a torture setting often leads to highly correlated variables in assessment, which makes it difficult to disentangle the effects
of individual stressors. Therefore, we examined the relative cumulative impact of physical vs nonphysical stressors. Distinguishing survivors who had only physical torture from those who had only nonphysical torture was,
however, not possible because most participants reported at least 1 physical torture event and all reported
at least 1 nonphysical torture stressor. The most commonly reported physical torture was beating (87.8%).
Therefore, we divided the sample into 3 subgroups defined by those who reported nonphysical torture only
(n = 20) (group 1), nonphysical torture plus beating
(n=44) (group 2), and nonphysical torture plus at least
1 form of physical torture other than or in addition to
beating (n=166) (group 3). In group 3, all but 5 survivors had an experience of beating and at least 1 other form
of physical torture. Thus, this distinction contrasted nonphysical torture with “low-intensity” physical torture (eg,
only beating) and “high-intensity” physical torture (eg,
all physical torture stressors, including beating).
One survivor who reported rape but no other physical torture was included in the physical torture category
WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

Table 1. Perceived Distress and Uncontrollability Associated With Stressors
Participants, No. (%)
Rating, Mean (SD)
Stressor
Sexual torture
Rape
Fondling of genitals
Sexual advances
Physical torture
Palestinian hanging*
Suffocation/asphyxiation
Electric torture
Falaqa†
Burning of parts of body
Forced extraction of teeth
Stretching of the body
Beating
Hanging by hands or feet
Needles under toenails or fingernails
Beating over the ears with cupped hands
Pulling/dragging/lifting by hair
Psychological manipulations
Sham executions
Witnessing torture of close ones
Threats of rape
Threats against family
Witnessing torture of others
Threats of death
Threats of further torture
Fluctuation of interrogator’s attitude
Humiliating treatment
Throwing of urine/feces at detainee
Stripping naked
Excrement in food
Verbal abuse
Mockery/humiliation
Exposure to forced stress positions
Rope bondage
Forced standing with weights on parts of the body
Forced standing
Restriction of movement
Exposure to sensory discomfort
Exposure to extreme heat or cold
Exposure to bright light
Cold showers
Exposure to loud music
Deprivation of basic needs
Isolation
Prevention of urination/defecation
Blindfolding
Sleep deprivation
Water deprivation
Deprivation of medical care
Prevention of personal hygiene
Food deprivation
Denial of privacy
Infested surroundings

Frequency

Fairly/Extremely
Distressing

Slightly/Not at All
in Control

Distress

Control

16 (5.7)
30 (10.8)
43 (15.4)

15 (93.8)
29 (96.6)
37 (86.1)

13 (81.3)
23 (76.7)
31 (72.1)

3.9 (0.5)
3.7 (0.5)
3.4 (0.8)

3.4 (0.8)
3.2 (0.9)
2.9 (1.2)

4 (1.4)
27 (9.7)
23 (8.2)
47 (16.8)
20 (7.2)
11 (3.9)
20 (7.2)
245 (87.8)
19 (6.8)
5 (1.8)
112 (40.1)
164 (58.8)

4 (100.0)
26 (96.2)
22 (95.6)
45 (95.7)
19 (95.0)
10 (90.9)
17 (85.0)
229 (93.9)
17 (89.5)
4 (80.0)
103 (92.0)
135 (82.3)

2 (50.0)
24 (88.9)
18 (78.3)
31 (66.0)
13 (65.0)
10 (90.9)
14 (70.0)
121 (49.6)
12 (63.2)
4 (80.0)
61 (54.5)
76 (46.4)

3.8 (0.5)
3.8 (0.5)
3.7 (0.6)
3.6 (0.5)
3.6 (0.6)
3.6 (0.7)
3.5 (1.1)
3.5 (0.7)
3.5 (0.6)
3.4 (0.9)
3.4 (0.8)
3.2 (0.8)

2.8 (1.0)
3.5 (0.7)
3.2 (1.2)
2.9 (1.1)
2.8 (1.2)
3.5 (0.7)
3.0 (1.1)
2.5 (1.1)
2.8 (1.1)
2.8 (1.1)
2.6 (1.1)
2.3 (1.2)

86 (30.9)
81 (29.0)
43 (15.4)
143 (51.3)
202 (72.4)
252 (90.3)
211 (75.6)
179 (64.2)

82 (95.4)
80 (98.7)
41 (95.4)
126 (88.1)
185 (91.6)
214 (85.0)
173 (82.0)
125 (70.2)

57 (66.3)
48 (59.3)
30 (69.8)
83 (58.1)
95 (47.0)
134 (53.2)
98 (46.4)
61 (34.2)

3.7 (0.7)
3.6 (0.5)
3.6 (0.6)
3.4 (0.9)
3.4 (0.7)
3.3 (0.9)
3.2 (0.8)
2.9 (1.0)

2.8 (1.3)
2.7 (1.1)
2.8 (1.2)
2.5 (1.1)
2.4 (1.1)
2.5 (1.2)
2.4 (1.1)
1.9 (1.3)

35 (12.6)
100 (35.8)
12 (4.3)
264 (94.6)
257 (92.1)

31 (88.5)
80 (80.0)
9 (75.0)
191 (72.6)
189 (73.5)

21 (60.0)
55 (55.0)
8 (66.7)
96 (36.5)
101 (39.3)

3.4 (0.7)
3.2 (1.0)
3.2 (0.8)
3.0 (1.0)
3.0 (1.0)

2.6 (1.4)
2.5 (1.3)
2.9 (1.2)
2.0 (1.2)
2.0 (1.2)

133 (47.7)
30 (10.8)
211 (75.6)
261 (93.5)

116 (87.2)
24 (80.0)
160 (75.8)
185 (70.9)

64 (48.1)
11 (36.7)
72 (34.1)
75 (28.8)

3.2 (0.8)
3.1 (1.0)
3.0 (0.9)
2.8 (1.0)

2.4 (1.2)
2.1 (1.2)
2.1 (1.1)
1.9 (1.2)

107 (38.4)
63 (22.6)
82 (29.4)
89 (31.9)

79 (73.8)
45 (72.6)
58 (70.7)
54 (60.7)

41 (38.4)
25 (40.3)
33 (40.3)
25 (28.1)

3.0 (0.9)
2.9 (0.9)
2.9 (1.0)
2.8 (0.9)

2.1 (1.2)
2.2 (1.1)
2.1 (1.2)
1.9 (1.1)

91 (32.6)
115 (41.2)
70 (25.2)
194 (69.5)
166 (59.7)
138 (49.5)
216 (77.4)
206 (73.8)
217 (77.8)
188 (67.4)

79 (86.8)
99 (86.1)
59 (84.3)
157 (80.9)
123 (74.1)
95 (68.9)
139 (64.4)
122 (59.2)
129 (59.4)
112 (60.2)

46 (50.6)
55 (47.8)
40 (57.1)
85 (43.8)
62 (37.4)
58 (42.0)
65 (30.1)
56 (27.2)
47 (21.7)
44 (23.7)

3.5 (0.9)
3.3 (0.8)
3.3 (0.9)
3.1 (0.8)
3.0 (1.0)
2.9 (1.1)
2.7 (1.0)
2.7 (0.9)
2.6 (1.1)
2.6 (1.0)

2.4 (1.1)
2.4 (1.2)
2.6 (1.2)
2.2 (1.2)
2.1 (1.3)
2.1 (1.3)
1.9 (1.2)
1.8 (1.2)
1.7 (1.2)
1.7 (1.2)

*Hanging by the wrists tied at the back.
†Beating of the soles of the feet.

(because rape involves an element of physical force),
whereas other stressors of a sexual nature (ie, fondling
of genitals and sexual advances) were grouped together
with the other stressors. Table 2 provides a comparison of the 3 groups in demographic and trauma charac(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
280

teristics and posttrauma outcome. Group 1 included more
women and unmarried survivors than did the other
groups, but the groups were otherwise similar regarding demographic characteristics and time since torture.
Group 3 had more severe overall trauma exposure than
WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

Extremely Distressing/
4
Not at All in Control

Fairly Distressing/
3
Slightly in Control

3.5
(0.8)

3.4
(0.7)

Distress Rating
Control Rating

3.3
(0.6)

3.0
(1.1)
2.4
(1.0)

3.0
(0.9)
2.4
(1.0)

3.0
(0.8)

2.1
(1.2)

Moderately Distressing/
2
in Control

2.8
(0.9)
2.0
(1.1)

2.8
(0.8)
2.0
(1.1)

2.0
(1.0)

Slightly Distressing/
1
Fairly in Control

Not at All Distressing/
0
Completely in Control

Sexual Torture

Physical Torture

Psychological
Manipulations

Humiliating
Treatment

Exposure to Forced
Stress Positions

Exposure to Sensory
Discomfort

Deprivation
of Basic Needs

Figure. Mean (SD) distress and control ratings for the 7 stressor categories.

Table 2. Comparison of Survivors Exposed to Physical and Nonphysical Torture Forms

Characteristic
Sex (male), No. (%)
Marital status (married), No. (%)
Current PTSD, No. (%)
Lifetime PTSD, No. (%)
Current MDE, No. (%)
Age, mean (SD), y
Education, mean (SD)§
Time since torture, mean (SD), mo
War events reported, mean (SD), No.
Torture events reported, mean (SD), No.
Stressors from each event category, mean (SD), No.
Psychological manipulations
Humiliating treatment
Deprivation of basic needs
Exposure to sensory discomfort
Sexual torture (other than rape)
Exposure to forced stress positions
Global Distress Rating, mean (SD)
Global Sense of Control Rating, mean (SD)
Social disability (0-4), mean (SD)
Occupational disability (0-4), mean (SD)
Severity of PTSD symptoms (0-4), mean (SD)
Beck Depression Inventory, mean (SD)

Group 1:
Nonphysical
Torture Only
(n = 20)

Group 2:
Nonphysical
Torture ؉ Beating
(n = 44)

Group 3:
Nonphysical
؉ Physical Torture
(n = 166)

␹2 or
F Statistic

P
Value*

12 (60.0)
8 (40.0)
9 (45.0)
12 (60.0)
3 (15.0)
43.0 (13.1)
2.4 (0.8)
97.0 (30.1)
19.5 (5.2)
11.6 (3.9)

38 (86.4)
35 (79.5)
19 (43.2)
31 (70.5)
3 (6.8)
46.4 (10.2)
2.1 (1.0)
90.6 (24.7)
16.3 (4.6)
13.7 (5.2)

146 (88.0)
113 (68.1)
100 (60.2)
131 (78.9)
33 (19.9)
44.4 (9.8)
2.3 (0.9)
95.6 (24.3)
22.6 (6.3)
22.7 (6.6)

11.1†
9.9†
5.1†
4.3†
4.3†
1.0‡
0.5‡
0.8‡
20.4‡
57.7‡

.01
.01
.08
.12
.12
.38
.59
.45
.001
.001

3.3 (1.6)
1.6 (0.6)
4.6 (2.6)
0.5 (0.8)
0.2 (0.5)
1.4 (0.7)
3.5 (0.6)
2.4 (1.2)
1.2 (1.2)
1.0 (1.2)
1.5 (1.3)
19.8 (15.8)

3.4 (1.7)
1.9 (0.6)
4.2 (2.5)
0.8 (1.0)
0.1 (0.4)
1.7 (0.9)
3.5 (0.6)
2.3 (0.8)
1.0 (0.9)
0.8 (1.0)
1.5 (1.0)
18.8 (12.8)

4.9 (1.6)
2.7 (0.8)
6.6 (2.4)
1.5 (1.3)
0.4 (0.7)
2.6 (0.9)
3.7 (0.5)
2.5 (1.0)
1.3 (1.0)
1.4 (1.2)
1.8 (1.1)
24.4 (13.1)

19.5‡
32.6‡
19.2‡
11.5‡
3.2‡
30.9‡
3.0‡
0.3‡
2.0‡
3.9‡
2.3‡
3.7‡

.001
.001
.001
.001
.04
.001
.053
.78
.15
.02
.10
.03

Abbreviations: MDE, major depressive episode; PTSD, posttraumatic stress disorder.
*Bonferroni-adjusted significance level, P = .002.
†␹22.
‡F2,227.
§Based on an ordinal variable: 1, primary school; 2, secondary school; 3, high school; and 4, university/postgraduate.

did the other groups, as indicated by more torture and
war stressors. However, greater overall severity of torture meant exposure to more events from all stressor categories and not only to physical torture events. The comparison of the number of stressors in each event category
across the 3 groups (0.5, 4.2, and 6.6, respectively) in
Table 2 shows that overall severity of torture is largely
accounted for by the stressors in the “deprivation of basic needs” category. Despite greater severity of trauma
exposure, group 3 did not report greater distress or loss
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
281

of control during the torture and did not have significantly higher rates of PTSD and depression than the other
groups. Significant differences were noted on only occupational disability and the Beck Depression Inventory. Caution needs to be exercised in interpreting these
latter results because multiple comparisons might have
led to type II error, and thus the between-group comparisons need to be assessed in relation to a Bonferroniadjusted P=.002. Furthermore, univariate analyses might
be misleading because they do not take into account posWWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

Table 3. Factors Related to PTSD and Depression
PTSD
␹

2

Step 1
Step 2
Step 3
Model
Factors
Step 1
Sex
Age
Education
Step 2
Sex
Age
Education
Severity of torture
Step 3
Sex
Age
Education
Severity of torture
Overall distress
Overall control

12.2
3.5
22.5
38.3

df
3
1
2
6
OR (95% CI)

Depression
2

P Value

R

.01
.06
.001
.001

0.07
0.09
0.21
0.21

␹

2

1.84
2.20
24.68
28.71

df

P Value

R2

3
1
2
6

.61
.14
.001
.001

0.01
0.03
0.20
0.20

OR (95% CI)

0.60
1.03
0.59

0.28-1.30
1.00-1.05
0.43-0.83

.19
.07
.01

NA
NA
NA

0.74
1.01
0.78

0.26-2.08
0.98-1.05
0.51-1.19

.56
.44
.25

NA
NA
NA

0.69
1.03
0.60
1.51

0.32-1.53
1.00-1.05
0.43-0.83
0.98-2.33

.36
.07
.01
.06

NA
NA
NA
NA

0.83
1.02
0.78
1.61

0.29-2.39
0.98-1.05
0.51-1.20
0.82-3.13

.73
.42
.26
.16

NA
NA
NA
NA

0.48
1.03
0.61
1.41
1.76
1.70

0.20-1.11
1.00-1.06
0.44-0.87
0.89-2.25
0.98-3.18
1.23-2.35

.09
.09
.01
.14
.06
.001

NA
NA
NA
NA
NA
NA

0.52
1.01
0.85
1.41
6.74
1.52

0.17-1.56
0.97-1.05
0.54-1.33
0.71-2.78
1.95-23.37
1.01-2.28

.24
.65
.47
.33
.01
.05

NA
NA
NA
NA
NA
NA

Abbreviations: CI, confidence interval; NA, not applicable; OR, odds ratio; PTSD, posttraumatic stress disorder.

sible intervening effects of other variables correlated with
the predictor variables of interest. Also note that any trend
toward significance does not necessarily reflect the effects of physical torture per se because group 3 included
the most severely tortured survivors, and more severe torture meant not only physical torture but also more nonphysical forms of torture, as noted previously herein.
To circumvent the problem posed by univariate analyses as much as possible, we conducted logistic regression analyses using the diagnosis of current PTSD and
major depression as the dependent variables in turn. Age,
sex, and education were entered into the equation at step
1. A variable representing the distinction between the 3
groups in Table 2 (1=nonphysical torture, 2=nonphysical tortureϩbeating, and 3=nonphysical tortureϩat least
1 form of physical torture other than or in addition to
beating) was entered at step 2 so that we could examine
its effects independent of demographic variables. For the
purposes of testing the study hypothesis, this variable constituted a better measure of severity of torture than total
number of torture stressors because it represented a continuum of torture severity not only in terms of the nature of stressors experienced (eg, nonphysical vs physical) but also quantitatively (eg, number of torture stressors
experienced). This variable showed a strong correlation
with total number of torture stressors (r=0.58; PϽ.001).
Thus, variability in this measure meant exposure to not
only physical torture but also to more stressors of all types.
Because this variable was based on observable or measurable aspects of the torture experience, it will be referred to as objective severity of torture.
The distress and control ratings, on the other hand,
reflected subjective severity of torture because they were
based on the survivors’ own appraisal of the torture events.
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
282

The global ratings of distress and control showed highly
significant correlations with the respective distress (range,
0.31-1.00) and control (range, 0.46-0.79) ratings for more
than 40 of the stressors. The correlation between global
ratings of distress and control was only moderately high
(r=0.42; PϽ.001), indicating that they did not measure
the same construct. These variables were entered at the
final step because we were interested in examining not
only the unique variance explained by the measure of objective severity (ie, controlling for all other variables) but
also the variance it explained before the effects of the subjective severity measures were taken into account. Although the measure of objective severity of torture did
not significantly correlate with the Global Sense of Control Rating (r=0.04; P=.59), it showed a marginal but significant correlation with the Global Distress Rating
(r =0.14; P =.04).
The results of logistic regression analyses are given in
Table 3. The independent variables explained 20.5% of
the variance in current PTSD and 19.6% of the variance
in depression. Objective severity of torture explained only
about 2% of the variance in PTSD and depression, which
was not statistically significant. On the other hand, the
proportion of variance explained by the subjective severity measures in PTSD (11.6%) and depression (16.7%)
at the final step was highly significant. In the full regression model only less education and greater perceived uncontrollability of the torture related to PTSD, whereas depression related to greater perceived distress and
uncontrollability. Objective severity of torture related to
neither PTSD nor depression. Because 38 (97.4%) of the
39 survivors with depression also had PTSD, the findings concerning depression also largely applied to depression comorbid with PTSD.
WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

Using social/occupational disability as the dependent measure in a multiple regression analysis and entering the independent variables in the same order as in
the previous analyses, the independent variables explained 11.9% of the total variance (F6,223 = 6.1;
PϽ.001). Objective severity of torture explained only
1.4% of the total variance when entered at step 2, which
was not significant (P = .07). On the other hand, the
subjective severity measures explained significant variance (8.7%) at the final step (PϽ.001). In the full regression model, significant predictors were greater loss
of control during the torture (␤ = .27; P = .001), male sex
(␤ = .13; P = .05), and less education (␤ = .16; P = .01).
Objective severity of torture did not show a significant
prediction (␤= .09; P = .16). These findings support the
third study hypothesis that perceived distress and uncontrollability of the torture stressors, rather than mere
exposure to them, would be associated with greater
likelihood of PTSD and depression.
COMMENT

The present study results suggest that psychological stressors cannot be easily distinguished from physical torture
in terms of their relative psychological impact. Although physical torture methods were rated as somewhat more distressing than some stressors that did not
involve severe physical pain, certain other stressors, such
as sham executions, threats of rape, sexual advances,
threats against self or family, witnessing the torture of
others, humiliating treatment, isolation, deprivation of
urination/defecation, blindfolding, sleep deprivation, and
certain forced stress positions, seemed to be as distressing as most physical torture stressors. These findings suggest that physical pain per se is not the most important
determinant of traumatic stress in survivors of torture.
The fact that physical torture did not contribute to longterm psychological outcome over and above the effects
of nonphysical stressors further supports this point.
These findings thus provide no support for the first 2
study hypotheses.
In this study, we could not examine the relative impact of individual stressors because in detention or interrogation settings various stressors occur concurrently or in clusters, leading to highly correlated trauma
exposure variables in assessment. We did, however, construct a measure that not only contrasted the 2 types of
stressors of interest (nonphysical vs physical) but also
represented a continuum of cumulative exposure severity in terms of the number of all forms of stressors experienced. This measure was based on the assumption
that exposure to each stressor led to the same degree of
impact (eg, 1) and that the effects of different stressors
were additive. The present results demonstrate that this
measure has no predictive value, which could be explained by the fact that it does not take into account 2
important factors: the subjective impact of the stressors
and the interactions between them. Consistent with the
present study findings, previous research19 has shown that
what determines traumatic stress in torture survivors is
perceived uncontrollability and stressfulness of the tor(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
283

ture stressors and not mere exposure to them. An example of this phenomenon comes from a previous study17
by our group that showed that 67% of political activists
(with high levels of psychological preparedness for torture) did not develop PTSD despite having endured a
mean of 23 different forms of torture (measured using
the same scale as in the present study) and a mean of 291
exposures to torture. That study also showed that higher
resilience levels, which also meant greater ability to
exercise control over torture stressors, were associated
with less perceived distress during torture and less PTSD
subsequently.19
The cumulative impact of torture stressors is also determined by the interactions among them. The distressing or helplessness-inducing effect of a particular stressor
might be compounded when combined with another
stressor.5 For example, the distressing effects of various
forms of physical torture, such as beating or electrical
shocks, might be augmented by blindfolding or hooding because the latter procedures remove visual control
over the stressors, thereby making them less predictable
and less controllable. Similarly, the traumatic impact of
physical torture might be maximized when coupled with
restriction of body movements, consistent with observations that restraint in animals potentiates the effects
of exposure to uncontrollable stressors.20 This might also
explain the distressing effects of various stress positions, such as forced standing or rope bondage. Thus, the
relative impact of each stressor needs to be considered
in the context of its interactions with other concurrent
stressors. A measure of mere exposure to torture stressors fails to capture such important information, and this
is the most likely reason it showed no prediction with
posttorture outcome in this study.
The present findings pointing to the important role of
uncontrollability of stressors in traumatic stress are consistent with learning theory formulations of traumatic stress
in torture survivors5 and similar evidence from studies of
other trauma survivors.21,22 These findings imply that various psychological manipulations, ill treatment, and torture during interrogation share the same psychological
mechanism in exerting their traumatic impact. All 3 types
of acts are geared toward creating anxiety or fear in the
detainee while at the same time removing any form of control from the person to create a state of total helplessness
(see the article by Bas¸og˘lu and Mineka5 for a review of some
of the control removal strategies used by torturers). Thus,
manipulations designed to remove control from the detainee might have a severe traumatic impact, even when
they do not involve physical torture. Evidence shows that
20% of the suspects detained for ordinary police interrogation experience abnormally high levels of anxiety because of uncertainty and lack of control over the environment23 and that some people develop PTSD after such an
experience.24
Humiliating treatment and attacks on personal integrity, cultural values, morals, or religious beliefs may induce feelings of helplessness in the individual through
not being able to act on anger and hostility generated by
such aversive treatment. Evidence25,26 shows that animals and humans respond with anger, hostility, and aggression to threats to physical and psychological wellWWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

being. Furthermore, the ability to aggress during
uncontrollable stress can dramatically reduce the impact of the stressor in animals.27 This idea is also supported by anecdotal reports of some torture survivors that
suggest that expression of anger and hostility toward the
torturers alleviates distress during torture.5
Because the study findings are highly relevant to the
current controversy surrounding the definition of torture, it is worth examining what they imply for the definition proposed in the US Justice Department memorandum,4 where it is argued that the definition of a
particular act as torture requires proof of “prolonged mental harm” associated with that act. This argument was
based on literature evidence showing that the most common psychiatric diagnosis among torture survivors is
PTSD and that torture survivors have elevated rates of
PTSD.4(p15) Because this document cited a review article28 by the main author of the present study (M.B.) and
a group of internationally recognized trauma experts (published in 2001 in a book sponsored by the US National
Institute of Mental Health), it is worth briefly examining herein whether the literature evidence supports the
argument concerning “prolonged mental harm.” In the
cited review article, the statement about PTSD being the
most common diagnosis among torture survivors28(p41) was
made in reference to the findings of a controlled study17
that was conducted by our research group. This was based
on the finding that torture survivors (political activists)
had significantly more lifetime and current PTSD than
did controls (33% vs 11% and 18% vs 4%, respectively).
Although that study pointed to an association between
torture and PTSD, the rates of PTSD in the sample were
surprisingly low despite the extremely severe torture experienced by the survivors (a mean of 23 different forms
of torture and a mean of 291 exposures to torture), as
noted earlier. Thus, although there is evidence that torture leads to PTSD in some cases, many people survive
extremely severe torture in relatively good psychological health and never develop PTSD. Conversely, some survivors develop PTSD after ostensibly milder forms of ill
treatment or psychological stressors that do not involve
physical torture. The fact that 60% of the present study
participants without any experience of physical torture
developed PTSD at some stage and 45% had current PTSD
suggests that such cases are not uncommon. These findings do not support a definition of torture based on evidence of “prolonged mental harm.” Such a definition does
not make logical sense given that it would disqualify many
severely tortured peoples’ experience as torture simply
because they did not develop PTSD.
This study was retrospective and thus subject to the
methodological limitations inherent in such research.
Problems in recall may have affected the survivors’ reports to a certain degree, but this is unlikely to invalidate the results. Global distress and control ratings showed
high and consistent correlations with the respective ratings for each stressor. The control ratings for each stressor
were based on detailed information about the survivors’
actual coping behaviors during the torture and thus were
unlikely to reflect substantial distortions in recall associated with high distress levels at the time of assessment. Furthermore, the distress and control ratings
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
284

seemed to have sufficient validity given that they varied
across different types of stressors in a way that made intuitive and theoretical sense, as noted previously herein.
In addition, evidence on the external validity of these measures from a previous study18 shows that they are associated with psychological resilience factors in torture survivors. The latter is also true for the present study, but
the data are not reported owing to space constraints.
In conclusion, aggressive interrogation techniques or
detention procedures involving deprivation of basic needs,
exposure to aversive environmental conditions, forced
stress positions, hooding or blindfolding, isolation, restriction of movement, forced nudity, threats, humiliating treatment, and other psychological manipulations conducive to anxiety, fear, and helplessness in the detainee
do not seem to be substantially different from physical
torture in terms of the extent of mental suffering they
cause, the underlying mechanisms of traumatic stress,
and their long-term traumatic effects. Such stressors satisfy the criterion of “severe mental suffering,” which is
central to the definition of torture in international conventions. Furthermore, these findings do not support the
distinction between torture and other cruel, inhuman,
and degrading treatment made by the Convention Against
Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment.1 Although both types of acts are
prohibited by this convention, such a distinction nevertheless reinforces the misconception that cruel, inhuman, and degrading treatment causes less harm and might
therefore be permissible under exceptional circumstances. These findings point to a need for a broader definition of torture based on scientific formulations of traumatic stress and empirical evidence rather than on vague
distinctions or labels that are open to endless and inconclusive debate and, most important, potential abuse.

Submitted for Publication: June 27, 2006; accepted September 18, 2006.
Correspondence: Metin Bas¸og˘lu, MD, PhD, Section of
Trauma Studies, Institute of Psychiatry, King’s College,
University of London, Box PO91, De Crespigny Park, Denmark Hill, London SE5 8AF, England (spjumeb@iop.kcl
.ac.uk).
Author Contributions: Dr Bas¸og˘lu takes responsibility
for the integrity of the data and the accuracy of the data
analysis. All the authors had full access to all the data in
the study.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grants
from the Bromley Trust.
Role of the Sponsor: The funding organization had no
role in the design and conduct of the study; collection,
management, analysis, and interpretation of the data; or
preparation, review, or approval of the manuscript.
Additional Information: This article is dedicated to the
memory of Keith Bromley.
Acknowledgment: We thank all of our colleagues who
contributed to the project and the participants who made
the study possible by sharing with us their experiences
during the war.
WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.

REFERENCES
1. United Nations Convention Against Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment. Geneva, Switzerland: Office of the High Commissioner for Human Rights; 1987. http://www.unhchr.ch/html/menu3/b/h_cat39
.htm. Accessed December 18, 2006.
2. Amnesty International Report 2005. Amnesty International Web site. http://web
.amnesty.org/report2005/usa-summary-eng. Accessed February 25, 2006.
3. US Defense Department. Working group report on detainee interrogations in the
global war on terrorism: assessment of legal, historical, policy and operational
considerations, April 4, 2003. In: Greenberg KJ, Dratel JL, eds. The Torture Papers. Cambridge, Mass: Cambridge University Press; 2005:286-359.
4. US Justice Department. Memorandum for James B. Comey, Deputy Attorney General, re: legal standards applicable under 18 USC §§ 2340-2340A. http://www
.usdoj.gov/olc/dagmemo.pdf. Accessed February 25, 2006.
5. Bas¸og˘lu M, Mineka S. The role of uncontrollable and unpredictable stress in posttraumatic stress responses in torture survivors. In: Bas¸og˘lu M, ed. Torture and
Its Consequences: Current Treatment Approaches. Cambridge, Mass: Cambridge University Press; 1992:182-225.
6. Bas¸og˘lu M, Livanou M, Crnobaric´ C, Francˇisˇkovic´ T, Suljic´ E, Ðuric´ D, Vranesˇic´
M. Psychiatric and cognitive effects of war in former Yugoslavia: the relationship between lack of redress for trauma and posttraumatic stress reactions.
JAMA. 2005;294:580-590.
7. Mineka S. Animal models of anxiety-based disorders: their usefulness and
limitations. In: Tuma AH, Maser J, eds. Anxiety and the Anxiety Disorders. Hillsdale,
NJ: Lawrence A Erlbaum Associates; 1985:199-244.
8. Mineka S, Kelly K. The relationship between anxiety, lack of control and loss of
control. In: Steptoe A, Appels A, eds. Stress, Personal Control and Health. Brussels, Belgium: Wiley; 1989:163-191.
9. Mineka S, Zinbarg R. Conditioning and ethological models of anxiety disorders.
In: Hope DA, ed. Nebraska Symposium on Motivation. Lincoln: University of Nebraska Press; 1996.
10. Mineka S. Animal models of clinical psychology. In: Smelser N, Baltes P, eds.
International Encyclopedia of the Social and Behavioral Sciences. Oxford, England: Elsevier Science; 2002:2020-2025.
11. Mineka S, Zinbarg R. A contemporary learning theory perspective on the etiology of anxiety disorders: it’s not what you thought it was. Am Psychol. 2006;
61:10-26.
12. Mineka S, Sutton J. Contemporary learning theory perspectives on the etiology
of fears and phobias. In: Craske MG, Hermans D, Vansteenwegen D, eds. Fear
and Learning: From Basic Processes to Clinical Implications. Washington, DC:
American Psychological Association; 2006:75-97.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64, MAR 2007
285

13. Watters JK, Biernacki P. Targeted sampling: options for the study of hidden
populations. Soc Probl. 1989;36:416-430.
14. Frank O, Snijders T. Estimating the size of hidden populations using snowball
sampling. J Off Stat. 1994;10:53-67.
15. First MB, Gibbon M, Spitzer RL, Williams JBW. User’s Guide for the Structured
Clinical Interview for DSM-IV Axis I Disorders, Research Version 2.0. New York:
Biometrics Research Dept, New York State Psychiatric Institute; 1996.
16. Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Charney DS, Keane T. ClinicianAdministered PTSD Scale (CAPS): Current and Lifetime Diagnostic Version. Boston, Mass: Behavioral Science Division, National Center for Posttraumatic Stress
Disorder; 1990.
17. Bas¸og˘lu M, Paker M, Paker O¨, Ozmen E, Marks I, Incesu C, Sahin D, Sarimurat
N. Psychological effects of torture: a comparison of tortured with non-tortured
political activists in Turkey. Am J Psychiatry. 1994;151:76-81.
18. Bas¸og˘lu M, Mineka S, Paker M, Aker T, Livanou M, Go¨k S. Psychological preparedness for trauma as a protective factor in survivors of torture. Psychol Med.
1997;27:1421-1433.
19. Bas¸og˘lu M, Paker M. Severity of trauma as predictor of long-term psychological
status in survivors of torture. J Anxiety Disord. 1995;9:339-350.
20. Mineka S, Kihlstrom J. Unpredictable and uncontrollable events: a new perspective on experimental neurosis. J Abnorm Psychol. 1978;87:256-271.
21. Regehr C, Cadell S, Jansen K. Perceptions of control and long-term recovery from
rape. Am J Orthopsychiatry. 1999;69:110-115.
22. Maes M, Delmeire L, Mylle J, Almatura C. Risk and preventive factors of posttraumatic stress disorder (PTSD): alcohol consumption and intoxication prior
to a traumatic event diminishes the relative risk to develop PTSD in response to
that trauma. J Affect Disord. 2001;63:113-121.
23. Gudjonnson GH. The Psychology of Interrogations and Confessions: A Handbook.
Chichester, England: John Wiley & Sons; 2003:26.
24. Gudjonnson GH, MacKeith JAC. False confessions: psychological effects of interrogation: a discussion paper. In: Trankell A, ed. Reconstructing the Past: The
Role of Psychologists in Criminal Trials. Deventer, the Netherlands: Kluwer; 1982:
253-269.
25. Baron RA. Human Aggression. New York, NY: Plenum Press; 1977.
26. Averill JR. Anger and Aggression: An Essay on Emotion. New York, NY: SpringerVerlag; 1982.
27. Weiss JM, Pohorecky L, Salman S, Gruenthal M. Attenuation of gastric lesions
by psychological aspects of aggression in rats. J Comp Physiol Psychol. 1976;
90:252-259.
28. Bas¸og˘lu M, Jaranson JM, Mollica R, Kastrup M. Torture and mental health: a
research overview. In: Gerrity E, Keane TM, Tuma F, eds. The Mental Health Consequences of Torture. New York, NY: Kluwer Academic/Plenum Publishers; 2001:
35-62.

WWW.ARCHGENPSYCHIATRY.COM

Downloaded from www.archgenpsychiatry.com on March 6, 2007
©2007 American Medical Association. All rights reserved.