Identifying Risk Factors for PTSD in Women Seeking

Medical Help after Rape

Anna Tiihonen Möller1,2*, Torbjörn Bäckström3, Hans Peter Söndergaard1, Lotti Helström1,2
1 Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden, 2 Department of Obstetrics and Gynecology, Stockholm South Hospital,
Stockholm, Sweden, 3 Department of Clinical Science, Obstetrics and Gynecology, Umeå University, Umeå, Sweden

Abstract
Objectives: Rape has been found to be the trauma most commonly associated with Posttraumatic Stress Disorder (PTSD)
among women. It is therefore important to be able to identify those women at greatest risk of developing PTSD. The aims of
the present study were to analyze the PTSD prevalence six months after sexual assaults and identify the major risk factors
for developing PTSD.

Methods: Participants were 317 female victims of rape who sought help at the Emergency Clinic for Raped Women at
Stockholm South Hospital, Sweden. Baseline assessments of mental health were carried out and followed up after six
months.

Results: Thirty-nine percent of the women had developed PTSD at the six month assessment, and 47% suffered from
moderate or severe depression. The major risk factors for PTSD were having been sexually assaulted by more than one
person, suffering from acute stress disorder (ASD) shortly after the assault, having been exposed to several acts during the
assault, having been injured, having co-morbid depression, and having a history of more than two earlier traumas. Further,
ASD on its own was found to be a poor predictor of PTSD because of the substantial ceiling effect after sexual assaults.

Conclusions: Development of PTSD is common in the aftermath of sexual assaults. Increased risk of developing PTSD is
caused by a combination of victim vulnerability and the extent of the dramatic nature of the current assault. By identifying
those women at greatest risk of developing PTSD appropriate therapeutic resources can be directed.

Citation: Tiihonen Möller A, Bäckström T, Söndergaard HP, Helström L (2014) Identifying Risk Factors for PTSD in Women Seeking Medical Help after Rape. PLoS
ONE 9(10): e111136. doi:10.1371/journal.pone.0111136
Editor: Koustuv Dalal, Örebro University, Sweden
Received July 14, 2014; Accepted September 22, 2014; Published October 23, 2014
Copyright: ß 2014 Tiihonen Möller et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that, for approved reasons, some access restrictions apply to the data underlying the findings. The data underlying the
study are human subject sensitive clinical data. The data, however, will be shared with interested investigators upon request and in accordance with the
institutional ethics committee approval. Data requests may be sent to Anna Tiihonen Möller (anna.moller@ki.se).
Funding: The study was funded by grants from the Swedish Research Council (project number 90418601), and supported by grants provided by the Stockholm
County Council, Sweden (‘‘ALF-medel’’, number 20110192). The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* Email: anna.moller@ki.se

Introduction assault variables (e.g. demographics, earlier victimization, and

psychiatric morbidity) and the assault related variables (e.g. victim-
Rape has been found to be the trauma most commonly assailant relationship, injuries, perceived life threat) are all fairly
associated with Posttraumatic Stress Disorder (PTSD) among easily measurable circumstances that can be used in a clinical
women [1], and approximately one third of the victims will be setting to identify those victims with highest risk of developing
diagnosed with PTSD at some time following the assault [2,3,4]. PTSD. The post-assault variables (e.g. social support, coping
PTSD is characterized by persistent re-experiencing of the strategies, and the impact of contact with the legal system) are
traumatic event, persistent avoidance of stimuli associated with difficult to assess at an initial acute visit post-rape and therefore are
the event, numbing of general responsiveness, and symptoms of less suitable as predictors in clinical settings. In the current study,
increased arousal [5]. These are all disabling symptoms causing performed in a clinical setting, we therefore focus on the first two
significant impairment in the victim’s daily lives, and the burden to sets of factors.
the individual as well as to society has been thoroughly reported Identifying risk factors for the development of PTSD following
[6]. sexual assault and subsequently permit early interventions with
Empirical research on risk factors for developing PTSD and the those victims at greatest risk may reduce the incidence [8,9]. Thus,
impact of sexual assault has used a broad variety of ways quite a few studies have been conducted trying to clarify the
approaching the subject. Many studies tend to divide the relationship between various predisposing variables and the
numerous potential risk factors into three categories: pre-assault occurrence of PTSD after rape, and the literature shows highly
variables, assault variables, and post-assault variables [7]. The pre- discrepant results. Pre-assault victimization has been found in

PLOS ONE | www.plosone.org 1 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

some studies to increase the risk of developing PTSD [10] whereas the sum scores were: 0–9 (no depression), 10–16 (mild depression),
other studies have not found this association [11]. A majority of 17–29 (moderate depression), and scores $30 (severe depression).
the studies have not found any association between different The Stanford Acute Stress Reaction Questionnaire, SASRQ
victim-assailant relationships and PTSD [12,13] however, some [21], was used at baseline and at the 6-month visit post-rape for
have reported that victims of stranger rapes have an increased risk measuring anxiety and dissociation, and at baseline also for
of PTSD compared with victims of other assailants [11,14], measuring the occurrence of Acute Stress Disorder (ASD). The
whereas others have found that victims of current partners have SASRQ is a 30-item self-report instrument with 3 additional
the highest risk of developing PTSD [15]. questions relevant to the diagnosis of ASD. The instrument can be
A majority of the conducted studies have a cross-sectional used as a Likert-type scale (0–5) or dichotomously (0–2:0, 3–5:1)
design, often in a large sample of college students, based on the for the presence of a symptom. A diagnosis of ASD according to
participant’s ability to recall circumstances from past traumatic DSM IV requires at least three out of the five types of dissociative
events. These designs have a risk of causing a non-representative symptoms, one re-experiencing symptom, one avoidance symp-
sample with a high proportion of recall bias. Only a few tom, one marked anxiety/increased arousal symptom, and
longitudinal studies with acute trauma samples have been impairment in at least one important area of functioning. We
conducted and they often suffer from selection bias and problems also used SASRQ total score for measuring PTSD symptom
with too small sample sizes, making generalizations hazardous severity.
[2,4,10,16,17,18]. The Posttraumatic Stress Diagnostic Scale, PDS [22] was used
The aims of the present study were to: (a) estimate the incidence at baseline to assess PTSD symptom score (0–51), pre-existing
of PTSD 6 months following a sexual assault using a structured PTSD diagnosis and lifetime histories of traumatic events. A
clinical interview, and (b) by using self-report questionnaires at complete PTSD was diagnosed at baseline when the respondent in
baseline and after 6 months further explore psychiatric morbidity PDS part 1 reported having been exposed to or witnessed a
in this population, and (c) prospectively identify factors associated traumatic event that according to PDS part 2 involved threat to
with the development of PTSD at 6 months following a sexual life or physical integrity and according to PDS part 3 having at
assault. We hypothesized that about one third of the victims would least 1 re-experiencing symptom, more than 2 avoidance
symptoms, more than 1 arousal symptom, having a duration of
develop PTSD, and that those not diagnosed as having a full
mentioned symptoms over 1 month and that the symptoms
PTSD would have major disabling symptoms. We also hypoth-
according to PDS part 4 also caused impairment in the
esized that current PTSD status (i.e. already having PTSD at the
respondents’ daily life in at least one area.
time of the index assault) would be higher than in the general
population and that victims of sexual assault represent a group
with more psychiatric morbidity and history of earlier traumati- Procedures
zation. Furthermore, we hypothesized that earlier victimization, At a medical check-up appointment approximately 10 days
psychiatric morbidity, perceived life threat during the assault, and after the acute visit, eligible women were asked to participate in
more severe assaults would be predictive of PTSD at 6 months the study. Consenting women completed the 3 self-rating
post-rape. questionnaires: BDI, PDS, and SASRQ. Information about
history of earlier sexual assaults, sexual assault in childhood, and
number of other earlier traumas were taken from the PDS
Method questionnaire. Information on demographics, employment status,
Participants psychiatric treatment history, assault characteristics, and possible
Out of 1,047 eligible women, 317 female victims of rape or injuries were gathered from the clinic’s structured data files. Six
attempted rape who had been in contact with the Emergency months after the rape, study participants were diagnosed regarding
Clinic for Raped Women at Stockholm South Hospital, Sweden, PTSD using the Structured Clinical Interview for DSM-IV (SCID
I) and they filled in 2 of the self-rating questionnaires (BDI,
between February 2009 and December 2011 agreed to participate
SASRQ).
in the study (Figure 1). All women were seen within one month
after the index assault. The women were considered eligible if
being over the age of 18 years old and literate in Swedish. Further Statistical analyses
they had to be capable of participating in an interview for Group comparisons between completers and non completers,
diagnosis of PTSD and being able of filling in self rating PTSD versus non-PTSD participants, were performed using
questionnaires for mental health. independent t-tests for continuous variables and chi-square tests
for categorical variables. The main outcome of the study, the
occurrence of PTSD 6 months post-rape, was analyzed using
Measures logistic regression. Our model strategy was as follows: First,
The PTSD Module of the Structured Clinical Interview for correlates of PTSD were examined including psychometric
DSM-IV (SCID I) was used to establish current PTSD 6 months variables, victim characteristic variables, and assault characteristic
post-rape. The SCID-I is a widely used structured clinical variables in 3 separate regression models. Variables were entered
interview [5]. A diagnosis of full PTSD was made using DSM in these models if they were found significant in the crude analyses.
IV-TR (i.e. when clusters A and F were fulfilled). Second, a simultaneous multivariate regression model was tested
The level of physical violence used during the assault was with significant factors from the three domains analyzed to predict
defined according to the NorVald Abuse Questionnaire [19] as the occurrence of PTSD. Variables were considered significant if
none, mild (hitting, smacking your face, holding you firmly), the Wald test resulted in a p,0.05. The associations are presented
moderate (hitting with fist(s) or hard object, kicking, pushing as odds ratios (OR) and 95% confidence interval (CI). Hosmer-
violently) or severe (threat to life, strangulation, showing a weapon Lemeshow goodness-of-fit test was used to examine if the model
or knife). adequately fitted the data (p-value .0.05 suggests good fit) [23].
The Beck Depression Inventory, BDI [20], consists of 21 The ability of ASD to predict PTSD was assessed by calculating
questions measuring depressive symptoms. Cut-off points used for the positive predictive value of ASD according to the formula a/

PLOS ONE | www.plosone.org 2 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

Figure 1. Flow-chart of study population.

doi:10.1371/journal.pone.0111136.g001

(a+b), where a stands for the women with ASD that also developed anal or vaginal penetration). Twenty percent had experienced
PTSD and b stands for the women with ASD that did not develop other sexual assaults (including attempts), and 10.1% could not
PTSD [24]. The correlation coefficient between the SCID-I and remember what type of assault they had been subjected to
SASRQ was calculated and is described as the linear correlation (mainly because of influence of alcohol).
(dependence) between two variables, giving a value between +1
and 21 inclusive, where 1 is total positive correlation, 0 is no Attrition
correlation, and 21 is total negative correlation. All statistical Of the 317 women entering the study, 201 women completed
analyses were conducted using the statistical software version SPSS the six month assessment. The socio-demographics (age, ethnicity,
20.0. marital status, and occupation) of non completers were not
The study was approved by the local medical ethics committee different from completers (p..05), apart from current alcohol
in Stockholm (2008/759-31). abuse being more common among non-completers (16.4% vs.
7.5%, p = .013). Non-completers were more depressed
Results (M = 28.63 vs. M = 24.75, p = .005) and had more avoidance
symptoms at baseline (M = 22.04 vs. M = 20.33, p = .018) than
Study Sample
completers, but there was no difference in pre-existing PTSD at
The age range was 18 to 59 years (M = 26.36 years, SD = 8.86).
baseline (19.1% vs. 19.8%, p..05).
The average time between the assault and the initial acute visit was
3.06 days (SD = 5.56 days) and the time from the assault to the
initial psychiatric assessment was 19.9 days (SD = 6.7 days). A
majority (70.6%) had experienced a completed rape (i.e., including

PLOS ONE | www.plosone.org 3 October 2014 | Volume 9 | Issue 10 | e111136

 Table 1. Psychometrics at baseline of PTSD versus Non-PTSD.

PTSD Non-PTSD

(n = 74) (n = 105)

M SD M SD t OR 95%CI

BDI

PLOS ONE | www.plosone.org

Total score 30.1 11.9 23.0 9.9 24.3 1.06*** [1.03–1.09]
Severe Depression (%) 54.1 26.0 3.36*** [1.78–6.32]
SASRQ
Total score 104.7 22.1 89.6 25.2 24.1 1.03*** [1.02–1.05]
Dissociation 34.6 8.6 29.5 10.4 23.5 1.06** [1.02–1.10]
Re-experience 18.2 7.1 14.2 6.5 23.8 1.09*** [1.04–1.14]
Avoidance 21.7 5.9 19.9 6.5 21.9 1.05 [1.00–1.10]
Arousal 22.7 4.6 20.2 6.3 22.8 1.09** [1.02–1.15]
ASD Diagnosis (%) 91.8 74.0 3.92** [1.53–10.06]
PDS
Pre-existing PTSD (%) 27.0 16.5 1.87 * [0.90–3.89]

Note. Independent t-test for continuous variables presented in means (M) and standard deviations (SD). Pearsons chi-test for categorical variables presented in percent. OR = odds ratio. CI = confidence interval. BDI = Beck

4
Depression Inventory. SASRQ = The Stanford Acute Stress Reaction Questionnaire. PDS = The Posttraumatic Stress Diagnostic Scale. *p,.05. **p,.01.***,.001.
doi:10.1371/journal.pone.0111136.t001

October 2014 | Volume 9 | Issue 10 | e111136

Risk Factors for PTSD
 Risk Factors for PTSD

Table 2. Victim characteristics of PTSD versus Non-PTSD.

PTSD Non-PTSD

(n = 74) (n = 105)
OR 95% CI

Mean age (SD) 27.5 (9.7) 25.6 (7.7) 1.03 [0.99–1.06]

Ethnicity (%)
Swedish 78.4 85.7 0.60 [0.28–1.32]
European 9.5 4.8 2.09 [0.64–6.86]
Non European 12.2 9.5 1.32 [0.51–3.42]
Marital status (%)
Unmarried/No partner 68.9 61.0 1.42 [0.76–2.67]
Married/Co-habitor 16.2 21.9 0.69 [0.32–1.49]
Occupation (%)
Working 47.3 54.3 0.76 [0.42–1.37]
Studying 23.0 27.6 0.78 [0.39–1.56]
Sick leave/Un-employed .3 months 28.4 16.2 2.05 * [1.10–4.23]
Current alcohol abuse (%) 6.8 8.6 0.77 [0.25–2.41]
Psychiatric treatment history (%) 58.1 36.2 2.45** [1.33–4.50]
Life time depression (%) 36.5 19.0 2.44** [1.24–4.82]
History of sexual assault
.18 years old (%) 60.8 49.0 1.61 [0.88–2.95]
,18 years old (%) 45.8 30.8 1.90* [1.02–3.55]
Number of earlier traumatic events (%)
0 20.5 32.7 0.53* [0.23–0.99]
1 16.4 27.9 0.94 [0.38–2.32]
2 26.0 20.2 2.05* [0.86–4.89]
$3 17.8 11.5 2.46* [1.13–6.60]
.5 19.2 7.7 3.97** [1.57–11.45]
History of $2traumatic events 63.0 39.5 2.62** [1.41–4.85]

Note. Independent t-test for continuous variables presented in means and standard deviations (SD). Pearsons chi-test for categorical variables presented in percent.
OR = odds ratio. CI = confidence interval. *p,.05. **p,.01. ***,.001.
doi:10.1371/journal.pone.0111136.t002

Psychopathology at 6 months following rape (n = 105) on various psychosocial measures. The women in the
According to the SCID-I at six months, 36.8% (74/201) of the PTSD group reported being more depressed both at the baseline
women met all 6 criteria for PTSD. When women who had not assessment and at 6 months post-rape (p,.001). They also
experienced a Criterion A trauma (22/201) were excluded, the reported higher PTSD symptom scores from all symptom clusters
PTSD prevalence was 41.3% (74/179). After excluding the 39 at baseline according to the SASRQ. The correlation coefficient
women who, according to the PDS questionnaire, had signs of pre- between SASRQ and the SCID at 6 months was high,
existing PTSD at baseline, 38.6% (54/140) of the women (Rho = .721, p,.001). A summary of the psychometrics in the
developed PTSD six months post-rape. two groups is shown in Table 1.
Fifty-four percent of the women (97/179) had a high symptom Differences between the two groups concerning characteristics
load of B- (re-experience), C- (avoidance), and D- (arousal) of the victim and assault characteristics are shown in Table 2 and
symptoms at 6 months and 74.9.% (n = 134) met the duration Table 3. The only demographic difference between the two
criteria (E). Only 51% (n = 101) reported significant impairment in groups was that the women in the PTSD group more often
their daily lives (F). reported being unemployed or on sick-leave. More women in the
According to Beck Depression Inventory (BDI), 47.5% suffered PTSD group reported having a psychiatric treatment history, a
from moderate or severe depression 6 months after the assault history of a lifetime depression, and were more sexually
regardless of PTSD status. traumatized in childhood than in the non-PTSD group. One
out of five of the women in the PTSD group had experienced
more than 5 earlier traumas (both sexual assaults and other
Comparisons between women with and without PTSD 6 traumas). Furthermore, the women suffering from PTSD more
months after sexual assault often reported the current sexual assault as involving more
Women who fulfilled criteria for a full PTSD status (n = 74) physical violence, full penetration, multiple sexual acts, perceived
using the SCID-I at 6 months (regardless of pre-existing PTSD life threat and more often having been victims of group assault. A
status at baseline) were compared with those not having PTSD smaller proportion of women in the PTSD group reported

PLOS ONE | www.plosone.org 5 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

Table 3. Assault characteristics of PTSD versus Non-PTSD group.

PTSD Non-PTSD

(n = 74) (n = 105)
OR 95%CI

Relationship to assailant (%)

Partner 13.7 10.5 1.36 [0.54–3.38]
Acquaintance 50.7 49.5 1.05 [0.58–1.90]
Stranger 15.1 25.7 0.51 [0.24–1.11]
Group 17.8 6.7 3.10** [1.29–8.03]
Amnesia 2.7 7.6 0.34* [0.07–1.00]
Victim under influence of alcohol (%) 68.1 78.1 0.60* [0.30–0.90]
Moderate or severe violence during assault (%) 32.9 18.1 2.32** [1.15–4.45]
Perceived life threat (%) 48.6 29.1 2.30** [1.23–4.32]
Penetrating assault (%) 79.5 67.6 1.85* [1.20–4.02]
Multiple acts (%) 54.8 34.3 2.32** [1.26–4.29]
Physical injury (%) 87.7 69.4 3.11** [1.43–7.24]

Note. Pearsons chi-test for categorical variables presented in percent. OR = odds ratio. CI = confidence interval. *p,.05. **p,.01.***,.001.
doi:10.1371/journal.pone.0111136.t003

amnesia in conjunction with the assault. Notable was also that ASD as a predictor of PTSD
there was no difference in mean number of visits to a therapist at Suffering from ASD shortly after the sexual assault was as
the clinic between the women who developed PTSD and the ones mentioned associated with the development of PTSD, both in
who did not (M = 7.6 vs. M = 6.6, p = .175). crude measures (OR = 3.9, 95%CI [1.5–10.1], p = .001) and in the
final model. However, even though a vast majority of the women
Predictors of PTSD status developed ASD, only 43% of these women continued to develop
Three regression analyses were conducted separately that PTSD. Unlike what has been found after other kind of traumas
analyzed psychometric variables, victim characteristics, and assault [25], ASD as a predictor of PTSD after rape has a high sensitivity
characteristics. Variables in these regressions were all found (92%) however a low specificity (30%) and low positive predictive
significant in crude analysis (see Table 1, 2, and 3). Of the value (43%).
psychometric variables entered (depression, dissociation, re-expe-
rience, avoidance, arousal, and ASD), severe depression Discussion
(AOR = 2.75, 95% CI [1.55, 4.52], p = .002) and acute stress
disorder (ASD) (AOR = 2.61, 95% CI [1.14, 6.00], p = .031) at The major findings of the present study were as follows: (a) with
baseline were associated with the development of PTSD. considerations taken to pre-existing PTSD status at baseline 39%
Of the victim characteristics (life-time depression, psychiatric of the women developed PTSD 6 months following rape.
treatment history, history of sexual assault in childhood, history of Furthermore (b) we found that 19% of the study population
sexual assault in adulthood, history of $2 traumatic events, and already suffered from PTSD at the time of the current sexual
employment status), a history of $2 traumatic events (AOR = 2.02, assault due to earlier victimization. The major risk factors for
95% CI [1.10, 4.15], p = .040) and psychiatric treatment history developing PTSD found (c) were having been sexually assaulted by
(AOR = 2.01, 95% CI [1.05, 3.83], p = .034) were associated with a group, suffering from ASD shortly after the assault, having been
the occurrence of PTSD. subjected to several acts during the assault, having been injured,
and having a history of $2 traumatic events.
Of the assault variables (physical injury, victim-offender
relationship, perceived life threat, type of sexual assault, whether To our knowledge the present study is one of the largest
longitudinal studies performed in a clinical setting. The use of
the victim had been under the influence of alcohol, and severity of
parallel psychometrics (i.e. both self rating questionnaires and
physical violence during the assault), perceived life threat
clinical interview) to assess PTSD symptoms give strength to the
(AOR = 2.15, 95% CI [1.01, 3.76], p = .044), having been sexually
study. Unlike many other studies we used six months as the time
assaulted by a group (AOR = 3.84, 95% CI [1.16, 10.69],
for diagnosis of PTSD. This was based on evidence from earlier
p = .027), having been subjected to several sexual acts
studies that the number of people with PTSD decreases
(AOR = 2.71, 95% CI [1.39, 4.29], p = .004), and having been
substantially within the first three months and that symptoms
injured (AOR = 2.07, 95% CI [1.00, 4.54], p = .050) were found to
persisting beyond three to six months have a high probability of
be significant risk factors. Suffering from amnesia was found to be
becoming chronic [1,4]. Because of its representative sample of
a protective factor for PTSD (OR = 0.31, 95% CI [0.07, 0.95],
sexually assaulted women from a big catchment area our study
p = .038) in the crude analysis; however, it was not so in the
substantially adds to the literature concerning mental health in the
adjusted analysis.
aftermaths of sexual assaults.
The significant predictors from these three initial regressions
Our finding of a PTSD prevalence of 39% is in line with two
were then entered simultaneously into a final model predicting
earlier longitudinal studies from clinical settings [2,4]. Our results
PTSD (Table 4).
are also consistent with the PTSD prevalence found in two larger

PLOS ONE | www.plosone.org 6 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

Table 4. Factors associated with PTSD six months after sexual assault (final model).

B S.E Wald AOR 95% CI

Sexually assaulted by group 1.39 0.67 4.33 4.01* [1.18, 152]

ASD at baseline 1.61 0.68 5.56 4.96* [1.30, 18.6]
Multiple acts during assault 1.06 0.38 7.65 2.93** [1.49, 6.14]
Physical injury 0.91 0.45 4.23 2.48* [1.22, 6.50]
Severe depression at baseline 0.82 0.41 3.92 2.33* [1.03, 5.10]
History of $2 traumatic events 0.64 0.34 3.61 1.91* [1.14, 3.72]
Psychiatric treatment history 0.66 0.34 3.75 1.93 [1.00, 3.81]
Perceived life threat 0.27 0.40 0.44 1.34 [0.60, 3.89]
Pre-existing PTSD 0.15 0.53 0.08 0.89 [0.42, 3.33]

Hosmer and Lemeshow Test: 0.86. AOR = adjusted odds ratio. CI = confidence interval.
*p,.05. **p,.01. ***p,.001.
doi:10.1371/journal.pone.0111136.t004

cross sectional studies of representative population based samples In the present study dissociation was found a significant risk
[3,26], which further supports the accuracy of our finding. The factor for PTSD in univariate analysis, however, the association
high proportion of PTSD symptoms found, (e.g. 85% having re- did not remain significant in the multivariable model. This finding
experiencing symptoms), after 6 months even give further proof of was consistent with another study [17] who stated that dissociation
the dramatic nature of a sexual assault. As a majority of the victims was a poor predictor of PTSD in rape victims however a useful
present with disabling symptoms after 6 months, but only 51% predictor after other criminal assaults.
consider themselves as having a significant impairment in their Out of the victim characteristic variables entered in the
daily lives, the PTSD diagnosis can be regarded as a blunt regression model earlier victimization remained significant. In
instrument. The high co-morbidity between PTSD and depression this study we saw an almost linear effect between the number of
has been explored in numerous earlier studies [27], and is earlier traumas and the risk of developing PTSD. As two recent
supported also in the current study where patients in the PTSD studies have suggested [10,31], the link between childhood sexual
group were significantly more depressed and had a history of abuse and PTSD seems to be mediated through the increased risk
lifetime depression. of adult sexual abuse. This further suggests a cumulative effect of
The second finding in our study that one out of five women had traumas. Contrary to some studies [28,32] current alcohol abuse
a pre-existing PTSD at the time of the current sexual assault is was not found related to PTSD in the present study. This could be
interesting, yet not surprising. We know that earlier trauma explained by the high number of women with current alcohol
substantially increase the risk of re-victimization [10,28]. Half of abuse among the non-completers.
the women in the current study population had a history of at least When exploring the assault variables, the only factors that
one earlier sexual assault and as much as one third had remained associated with PTSD in adjusted analyses were having
experienced a sexual assault before the age of 18. These findings been subjected to several acts, having been sexually assaulted by a
support our second hypothesis that victims of sexual assaults differ group and having been injured during the assault. The association
from the general population with regard to earlier victimization between having been subjected to several acts is concordant with
and psychopathology. Having a pre-existing PTSD at the entrance another study [33] that found that victims of several rape types
of the study was not strongly associated with having PTSD at the 6 were more prone to develop PTSD. In the same line sexual assault
months assessment. Almost half of the women with PTSD at by a group probably can be looked upon as a more severe
baseline had recovered at 6 months, suggesting an ongoing interpersonal violence. In the current study women sexually
recovery process. This finding is also in line with the finding from assaulted by a group had the highest risk of developing PTSD out
an earlier study [29] that PTSD in response to prior trauma was of the different perpetrator categories. Surprisingly, sexual assaults
not predictive of PTSD following subsequent trauma. by single strangers did not increase the risk. Those studies
The main focus of the present study was to identify the factors reporting victims of stranger assaults being more prone to develop
associated with the development of PTSD at 6 months following a PTSD [11,14] however have not discriminated between stranger
sexual assault. There was an alarmingly high occurrence of assaults by a single or multiple assailants, having done so the
psychopathology found shortly after the sexual assault with almost results might not have been so distinct. In the present study
80% of the women suffering from ASD at the 2 weeks assessment. women sexually assaulted by a partner developed PTSD more
The use of ASD to predict development of PTSD after various often than women objected to acquaintance and single stranger
traumatic events have been supported in earlier studies [30], assaults. This supports our findings from an earlier study [34] that
however the use of ASD as a predictor of PTSD after sexual sexual assault by a partner is even more violent than assaults by
assaults has been questioned [2]. We found the presence of ASD strangers.
being highly associated with development of PTSD even in the There are some clinical implications of our findings. By focusing
final regression model. Even though majority of the victims on potential risk factors for PTSD, easily assessed in a clinical
develop ASD, far from all will continue developing PTSD, setting (e.g. factors already known at the time of the first acute
suggesting a substantial ceiling effect. Thus, ASD is not an optimal visit) they can be used even in small clinics where psychological
predictor of PTSD in samples of rape victims. treatment is not a matter of course for everyone. With increased
knowledge about the largest risk factors for PTSD these clinics

PLOS ONE | www.plosone.org 7 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

more easily can identify those women at greatest risk and resources Conclusion
could be directed to these women. ‘‘High risk’’ women could be
referred to specialized crisis centers more quickly and maybe some The overall finding confirms that sexual assault represents one
cases of development of PTSD can be avoided. of the most traumatic experiences a person can be exposed to
The present study has some limitations that need to be regarding the high incidence of PTSD in its aftermath compared
mentioned. The fact that almost 37% of the consenting patients to other traumas [1]. Further, the mean age among the women in
were lost to follow-up is of course a potential bias. However, the this study suggests that sexual assault is something that strikes
women in the middle of their development into adulthood. To
proportion of women by victim-assailant relationship were the
identify those women at greatest risk of developing PTSD we
same for the completers in this study as the consecutive patients
suggest health workers to identify those women with a potential
that sought help during the 13 months evaluated in another study
weakness in form of earlier victimization and pre-existing mental
from the same clinic [34], suggesting that our sample was
health conditions. Because of the ceiling effect, ASD alone is not
representative for the Stockholm area. In our attrition analysis
enough since far from all would develop PTSD. Therefore a
completers did not differ from non-completers regarding victim-
combination of ASD and other risk factors should be looked for.
and assault characteristics apart from non-completers more often
Although certain specific variables may vary between studies they
had a current alcohol-abuse. If this is caused by simple all point towards the same direction; increased risk of developing
forgetfulness, shame or other factors is unknown. Apart from PTSD is caused by a combination of victim vulnerability and the
non-completers being more depressed at baseline, they did not extent of the dramatic nature of the current assault. Future studies
tend to score higher on the psychometrics at baseline compared to of secondary preventions focusing on these women would be of
the completers. The only PTSD symptom cluster found signifi- interest.
cantly increased among the non-completers was the avoidance
symptoms which could explain the attrition. However, our results
Author Contributions
might still be biased by not having any information on women who
did not seek any help. One could speculate that the rape victims Conceived and designed the experiments: ATM TB HPS LH. Performed
the experiments: ATM LH. Analyzed the data: ATM TB HPS LH.
that do not seek care are in some aspects similar to the ones that
Contributed reagents/materials/analysis tools: ATM TB HPS LH.
did not complete the 6 months assessment. Contributed to the writing of the manuscript: ATM TB HPS LH.

References
1. Kessler RC, Sonnega A, Bromet E, Hughes M, Nelson CB (1995) Posttraumatic 17. Dancu CV, Riggs DS, Hearst-Ikeda D, Shoyer BG, Foa EB (1996) Dissociative
stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry 52: experiences and posttraumatic stress disorder among female victims of criminal
1048–1060. assault and rape. J Trauma Stress 9: 253–267.
2. Elklit A, Christiansen DM (2010) ASD and PTSD in rape victims. J Interpers 18. Darves-Bornoz JM, Lepine JP, Choquet M, Berger C, Degiovanni A, et al.
Violence 25: 1470–1488. (1998) Predictive factors of chronic Post-Traumatic Stress Disorder in rape
3. Masho SW, Ahmed G (2007) Age at sexual assault and posttraumatic stress victims. Eur Psychiatry 13: 281–287.
disorder among women: prevalence, correlates, and implications for prevention. 19. Swahnberg IM, Wijma B (2003) The NorVold Abuse Questionnaire (NorAQ):
J Womens Health (Larchmt) 16: 262–271. validation of new measures of emotional, physical, and sexual abuse, and abuse
4. Rothbaum BO, Foa EB, Riggs DS, Murdock T, Walsh W (1992) A Prospective in the health care system among women. Eur J Public Health 13: 361–366.
Examination of Post-Traumatic Stress Disorder in Rape Victims. Journal of 20. Beck ATS RA (1996) Beck Depression Inventory (2nd ed.). San Antonio: TX:
Traumatic Stress 5(3). Psycholocical Corp.
5. American Psychiatric Association (1994) Diagnostic and statistical manual of 21. Cardena E, Koopman C, Classen C, Waelde LC, Spiegel D (2000) Psychometric
mental disorders (4th ed.). properties of the Stanford Acute Stress Reaction Questionnaire (SASRQ): a
6. Kessler RC (2000) Posttraumatic stress disorder: the burden to the individual valid and reliable measure of acute stress. J Trauma Stress 13: 719–734.
and to society. J Clin Psychiatry 61 Suppl 5: 4–12; discussion 13–14. 22. Foa, Jaycox L Perry K (1997) The validation of a self-report measure of PTSD:
The Posttraumatic Stress Diagnostic Scale. Psychol Assess 9: 445–451.
7. Brewin CR, Andrews B, Valentine JD (2000) Meta-analysis of risk factors for
23. Horsmer DW LS (1980) A goodness-of-fit test for multiple logistic regression
posttraumatic stress disorder in trauma-exposed adults. J Consult Clin Psychol
model. Communications in Statistics A10: 1043–1069.
68: 748–766.
24. Fletcher RH FS (2005) Clinical epidemiology : the essentials (4th ed.). Baltimore:
8. Resnick H, Acierno R, Kilpatrick DG, Holmes M (2005) Description of an early
Lippincott Williams & Wilkins.
intervention to prevent substance abuse and psychopathology in recent rape 25. Bryant RA, Creamer M, O’Donnell M, Silove D, McFarlane AC (2012) The
victims. Behav Modif 29: 156–188. capacity of acute stress disorder to predict posttraumatic psychiatric disorders.
9. Rothbaum BO, Kearns MC, Price M, Malcoun E, Davis M, et al. (2012) Early J Psychiatr Res 46: 168–173.
intervention may prevent the development of posttraumatic stress disorder: a 26. Zinzow HM, Resnick HS, McCauley JL, Amstadter AB, Ruggiero KJ, et al.
randomized pilot civilian study with modified prolonged exposure. Biol (2012) Prevalence and risk of psychiatric disorders as a function of variant rape
Psychiatry 72: 957–963. histories: results from a national survey of women. Soc Psychiatry Psychiatr
10. Nishith P, Mechanic MB, Resick PA (2000) Prior interpersonal trauma: the Epidemiol 47: 893–902.
contribution to current PTSD symptoms in female rape victims. J Abnorm 27. Kilpatrick DG, Ruggiero KJ, Acierno R, Saunders BE, Resnick HS, et al. (2003)
Psychol 109: 20–25. Violence and risk of PTSD, major depression, substance abuse/dependence, and
11. Bownes IT, O’Gorman EC, Sayers A (1991) Assault characteristics and comorbidity: results from the National Survey of Adolescents. J Consult Clin
posttraumatic stress disorder in rape victims. Acta Psychiatr Scand 83: 27–30. Psychol 71: 692–700.
12. Ullman SE, Filipas HH (2001) Predictors of PTSD symptom severity and social 28. Acierno R, Resnick H, Kilpatrick DG, Saunders B, Best CL (1999) Risk factors
reactions in sexual assault victims. J Trauma Stress 14: 369–389. for rape, physical assault, and posttraumatic stress disorder in women:
13. Zinzow HM, Resnick HS, McCauley JL, Amstadter AB, Ruggiero KJ, et al. examination of differential multivariate relationships. J Anxiety Disord 13:
(2010) The role of rape tactics in risk for posttraumatic stress disorder and major 541–563.
depression: results from a national sample of college women. Depress Anxiety 29. Cougle JR, Resnick H, Kilpatrick DG (2009) Does prior exposure to
27: 708–715. interpersonal violence increase risk of PTSD following subsequent exposure?
14. Ullman SE, Filipas HH, Townsend SM, Starzynski LL (2006) The role of victim- Behav Res Ther 47: 1012–1017.
offender relationship in women’s sexual assault experiences. J Interpers Violence 30. Brewin CR, Andrews B, Rose S, Kirk M (1999) Acute stress disorder and
21: 798–819. posttraumatic stress disorder in victims of violent crime. Am J Psychiatry 156:
15. Temple JR, Weston R, Rodriguez BF, Marshall LL (2007) Differing effects of 360–366.
partner and nonpartner sexual assault on women’s mental health. Violence 31. Filipas HH, Ullman SE (2006) Child sexual abuse, coping responses, self-blame,
Against Women 13: 285–297. posttraumatic stress disorder, and adult sexual revictimization. J Interpers
16. Kaysen D, Simpson T, Dillworth T, Larimer ME, Gutner C, et al. (2006) Violence 21: 652–672.
Alcohol problems and posttraumatic stress disorder in female crime victims. 32. Kaysen DL, Lindgren KP, Lee CM, Lewis MA, Fossos N, et al. (2010) Alcohol-
J Trauma Stress 19: 399–403. involved assault and the course of PTSD in female crime victims. J Trauma
Stress 23: 523–527.

PLOS ONE | www.plosone.org 8 October 2014 | Volume 9 | Issue 10 | e111136

 Risk Factors for PTSD

33. Epstein JN, Saunders BE, Kilpatrick DG (1997) Predicting PTSD in women 34. Moller AS, Backstrom T, Sondergaard HP, Helstrom L (2012) Patterns of injury
with a history of childhood rape. J Trauma Stress 10: 573–588. and reported violence depending on relationship to assailant in female Swedish
sexual assault victims. J Interpers Violence 27: 3131–3148.

PLOS ONE | www.plosone.org 9 October 2014 | Volume 9 | Issue 10 | e111136