· Web viewTitle: The relationship between childhood adversities and dissociation in severe mental...
Transcript of · Web viewTitle: The relationship between childhood adversities and dissociation in severe mental...
Title: The relationship between childhood adversities and dissociation in severe mental
illness: A meta-analytic review.
Authors: Sonya Rafiq a, Carolina Campodonico a, and Filippo Varese a,b
a School of Health Sciences, Division of Psychology and Mental Health, Faculty of Biology,
Medicine and Health, Manchester Academic Health Science Centre, The University of
Manchester. 2nd Floor, Zochonis Building, Brunswick Street, Manchester. M13 9PL, UK.
b Complex Trauma and Resilience Research Unit, Greater Manchester Mental Health NHS
Foundation Trust, Manchester (UK).
Corresponding author:
a Filippo Varese. School of Health Sciences, Division of Psychology and Mental Health,
Faculty of Biology, Medicine and Health, Manchester Academic Health Science Centre, The
University of Manchester. 2nd Floor, Zochonis Building, Brunswick Street, Manchester. M13
9PL, UK.
Email: [email protected]
Telephone: 0161 306 0434
Running title: Childhood trauma and dissociation: Meta-analysis
This paper was published in Acta Psychiatrica Scandinavica
1
Abstract
Objective: Several studies have observed that dissociative experiences are frequently
reported by individuals with severe mental illness (SMI), especially amongst patients that
report a history of adverse/traumatic life experiences. This review examined the magnitude
and consistency of the relationship between childhood adversity (sexual abuse, physical
abuse, emotional abuse, neglect, bullying, natural disasters and mass violence) and
dissociation across three SMI diagnostic groups: schizophrenia, bipolar disorder and
personality disorders.
Method: A database search (EMBASE, PubMed, and PsycINFO) identified 30 eligible
empirical studies, comprising of 2199 clinical participants. Effect sizes representing the
relationship between exposure to childhood adversity and dissociation were examined and
integrated using a random-effects meta-analysis.
Results: The results indicated that exposure to childhood trauma was associated with
heightened dissociation across SMIs. Positive significant associations were also found
between specific childhood adversities and dissociation, with aggregated effect sizes in the
small-to-moderate range.
Conclusion: These findings support calls for the routine assessment of traumatic experiences
in clients with SMIs presenting with dissociative symptoms, and the provision of adequate
therapeutic support (e.g. trauma-focused therapies) to manage and resolve these difficulties.
Keywords: Trauma, dissociation, schizophrenia, bipolar disorder, personality disorder.
2
Summations
Small-to-moderate relationship between childhood trauma and dissociation across
SMIs.
Medium relationship between childhood trauma and dissociation in schizophrenia-
spectrum disorders.
All forms of childhood traumatic experiences (e.g. sexual abuse, physical abuse,
emotional abuse, emotional neglect, physical neglect) were found to be associated
with dissociation in people diagnosed with SMI.
Considerations
Only three studies were found eligible for the bipolar sample (i.e. low power),
therefore no strong conclusions can be drawn regarding the association between
childhood trauma and dissociative experiences in bipolar disorder.
The relationship between childhood trauma and dissociation in SMI requires further
corroboration using longitudinal designs to determine causality and directionality.
No firm conclusions can be drawn regarding which specific traumatic experience is
most strongly associated with dissociation, as the studies included in the analysis were
not independent (i.e. effects for sexual abuse and physical abuse estimated from the
same sample of participants).
Introduction
3
Dissociation has been defined by DSM-5 (1) as “a disruption of and/or discontinuity
in the normal integration of consciousness, memory, identity, emotion, perception, body
representation, motor control, and behaviour,” (1; p. 291). Dissociation can manifest as a
range of anomalous experiences, including dissociative amnesia, (e.g. inability to access
information about oneself/salient episodes of one’s life, which is qualitatively distinct from
mere forgetting) absorption, (e.g. losing awareness of one’s surroundings/passage of time
whilst being highly absorbed in certain activities) identity alteration, (e.g. experiencing one’s
self as comprising multiple distinct personas) derealisation, (e.g. feeling as if the world is
unreal) and depersonalisation (e.g. feeling disconnected/detached from one’s body/emotions;
(2, 3).
Dissociative experiences are common in the immediate aftermath of potentially
traumatic life experiences, and many researchers and clinicians regard dissociation as a
defence mechanism, that protects individuals against the overwhelming distress caused by
certain adverse life experiences (4). For example, at the time of a traumatic event,
dissociation allows for the individuals to “detach” from the source of distress, enabling them
to adapt and protect their internal and/or physical selves (4-6). However, the initial ‘adaptive’
response may become maladaptive over time, for example when used automatically or
indiscriminately in a variety of circumstances as a response to stress or reminders of trauma
(5), therefore, potentially impacting on the persons’ functioning and preventing the
integration of factors that may promote long-term recovery and well-being (e.g. the
correction of maladaptive trauma-related beliefs; 5, 7).
Growing empirical evidence has indicated that dissociative experiences are pervasive
within clients with SMIs; heightened dissociation is commonly reported by individuals who
received diagnoses of psychosis, (8, 9) bipolar disorder, (10) and personality disorders (11,
12). In line with evidence suggesting that dissociation is a common sequela of adverse and/or
4
traumatic life experiences, many empirical studies have demonstrated that individuals with
SMIs (13-16) exposed to childhood adversity (e.g. sexual abuse, physical abuse, emotional
abuse, neglect, bullying, etc.), experience greater dissociation compared to individuals
exposed to little or no trauma (17-21). However, the relationship between childhood trauma
and dissociation in individuals with SMIs remains contradictory, as some studies have found
no association between the two constructs (22, 23).
Several meta-analyses have attempted to investigate and clarify the relationship
between childhood trauma and dissociation across clinical and non-clinical individuals (4, 7).
However, previous evidence synthesis has largely neglected individuals with SMIs, either
because of the lack of sufficient empirical evidence at the time these syntheses were
conducted (7), or the use of study selection strategies that precluded the investigation of
dissociation in the context of specific mental health presentations (4). Also, previous reviews
have only considered a very limited range of traumatic exposures (i.e. physical abuse and
sexual abuse; 4, 7), consequently excluding the examination of other potentially traumatic
experiences that can plausibly impact on dissociation. Therefore, most of the empirical
studies that have investigated the relationship between childhood maltreatment and
dissociation have not been systematically reviewed.
Aims of the study
This meta-analysis examined and summarised evidence for an association between
childhood trauma and dissociation and considered the consistency of this relationship across
different SMIs (i.e. schizophrenia, bipolar and personality disorder). A secondary aim was to
investigate the relationship between specific types of childhood adversities and dissociation
and to determine the magnitude of these relationships across SMIs.
Method
Search procedure
5
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) guideline (24), alongside our PROSPERO review protocol (25) were followed for
the current meta-analysis (there was a deviation from the protocol in regards to quality
assessment, as it was not conducted because most quality assessment tools such as the
Effective Public Health Practise Project, are poorly suited to evaluate the population of
studies within this review). In the present review, we opted to employ the definition of SMI
(e.g. psychosis, personality disorder and bipolar disorder) employed in the UK as part of the
Improving Access to Psychological Therapies for SMI project (26). Additionally, the search
strategies and study selection criteria of previous relevant meta-analyses (27-31) were
examined to inform the selection of childhood adversities considered in the current evidence
synthesis, (i.e. physical abuse, sexual abuse, emotional abuse, physical neglect, emotional
neglect, natural disasters, mass violence and bullying) the eligible operational criteria used to
define dissociation, the diagnostic groups investigated (i.e. schizophrenia-spectrum disorders,
personality disorders and bipolar disorder), as well as the search terms used to identify
relevant papers considering the constructs of interest.
PubMed, EMBASE, and PsycINFO, were systematically searched between 1986 up
to and including October 2016 using the following search terms: Dissociation combined with
trauma related search terms and severe mental health related search terms (see supplementary
material for the complete list of search terms). Medical subject headings (MeSH) in PubMed
and subject headings in EMBASE and PsycINFO were used to further expand the literature
search.
To reduce file drawer effects the following steps were taken to identify all relevant
studies: 1) reference lists of eligible studies (i.e. backward search) and articles that cited
eligible studies (i.e. forward search) were scanned to locate studies not identified in the
6
database search; and 2) authors of all eligible studies were contacted regarding any relevant
unpublished studies (see supporting information; Table S1).
Inclusion and exclusion criteria
Studies published after 1986 were included as the first validated and widely used
measures of dissociation (e.g. DES) was published in 1986 (29). Also, studies utilising the
subsequent quantitative methodologies were eligible: 1) between-group comparisons
contrasting traumatised individuals (e.g. exposed to an eligible traumatic event, i.e. physical
abuse, sexual abuse, emotional abuse, physical neglect, emotional neglect, natural disasters,
mass violence and/or bullying) with non-traumatised comparable controls (e.g. patients with
identical diagnosis) on measures of dissociation; 2) between-group comparisons contrasting
individuals with high dissociation (e.g. scored in the pathological range of a dissociation
measure) with low dissociation comparable controls on measures of trauma; 3) correlational
studies examining the relationship between dissociation and trauma in an eligible clinical
sample or; 4) longitudinal studies investigating the temporal relationship between trauma
exposure and subsequent presence and/or severity of dissociative experiences. Studies written
in English, Spanish, Italian, Portuguese and Urdu were also eligible for inclusion (researchers
were bilingual). Moreover, studies had to employ: 1) validated self-report measures of
dissociation and; 2) validated self-report measures of exposure to potentially traumatic
experiences in childhood (i.e. prior to age 18). Furthermore, studies employing the following
diagnostic populations were included: 1) schizophrenia and related psychotic disorders; 2)
personality disorders or; 3) bipolar disorder based on ICD-9, ICD-10, DSM-III, DSM-III-R,
DSM-IV, DSM-IV-TR or DSM-5 diagnostic criteria.
Studies were excluded based on the subsequent criteria: 1) non-quantitative
methodology (e.g. qualitative studies, case studies, etc.); 2) measures of dissociation were
staff administered instead of self-report; 3) dissociative symptoms were either primarily drug
7
or alcohol induced or a medical (e.g. epilepsy) or organic cause was attributable; 4) non-
clinical sample; or 5) insufficient information was provided to calculate an effect size.
Eligibility was assessed independently by two researchers following a two-stage
procedure: title and abstract screening, and whole paper screening. In the first stage, S.R.
screened all the titles and abstracts, and C.C. screened 20% of all title and abstracts
(agreement ratings = 93.3%, adjusted kappa = .87). In the second stage, both S.R. and C.C.
reviewed all papers independently (agreement ratings = 96.0%, adjusted kappa = .92). Any
inter-coder discrepancy was resolved during consensus meetings and consultations with a
third researcher (F.V.).
Quality assessment
Eligible studies were quality assessed using the Effective Public Health Practise
Project tool (EPHPP; Thomas, 2003). The EPHPP has been found to be valid (Thomas et al.
2004) and reliable (Armijo-Olivio et al. 2012). The components included in the tool are: 1)
selection bias; 2) study design; 3) confounders; 4) blinding; 5) data collection methods and;
6) withdrawals and drop-outs (given that all studies were cross-sectional, this component was
found to be not applicable for the current analysis). Each component was rated as either
strong, moderate or weak. A global rating of strong (no weak rating), moderate (one weak
rating) or weak (two or more weak ratings) was given based on component ratings. All
studies were quality assessed by S.R and monitored by F.V; disagreements were discussed
and resolved during meetings
Effect size computation and statistical analyses
A series of random-effects meta-analyses were carried out using Comprehensive
Meta-Analysis, version 2. The random effects model was chosen as it allows for
heterogeneity and generalizability of results across studies (32). Pearson’s r was selected as
the main effect size metric for all analyses as most of the eligible studies reported
8
correlational effects. In the case of studies reporting correlational effects, correlation
coefficients and sample sizes on which correlations were estimated were extracted from the
primary studies. In the case of studies reporting between-group effects, the mean, standard
deviation and sample size were used to generate Cohen’s d, which were then converted to
effects of the r-family using appropriate computational methods (32). Similarly, when binary
data effects were reported, odds ratios (ORs) were estimated from appropriate descriptive
statistics (e.g. 2 x 2 tables), and then converted to r-family effects (32).
A number of studies included in the meta-analysis used different measures of trauma
and dissociation and reported both correlation and between-group analyses. To ensure
consistency and comparability across studies, and to avoid issues stemming from
dependency, a coding hierarchy was developed to guide the extraction of statistical
information. The following hierarchy was utilised: 1) when multiple separate effects for
childhood adversities (e.g. sexual and physical abuse) were reported within the same study
and no total trauma scores were provided, authors were contacted to provide information on
the summary score when appropriate, and when no information was provided, a composite
summary effect size was generated by merging the effects prior to the analysis (32); 2) when
measures of adversity exposure included multiple categories (i.e. none, low, moderate, or
severe trauma) we only extracted effects contrasting the no trauma exposure sample to the
most severe trauma sample; 3) when the dissociation scores were divided into different
severities (i.e. low, moderate, or high), the low dissociation sample was compared to the high
dissociation sample; 4) when the relationship between trauma and dissociation was
investigated at two different time points (i.e. baseline and post-assessments), only the
baseline results were selected; and 5) when between group and within group statistical
analyses were reported in the same study, within group information (i.e. correlation or
regression) was extracted; and 6) when multiple studies appeared to be drawn from the same
9
participant sample, the study with the greatest sample size was included to increase precision
(or if studies with overlapping samples had missing data, the study with the relevant data was
included in the meta-analysis).
A series of analyses were carried out to examine the main research questions of this
review. Firstly, analysis of the association between total childhood trauma (or composite
childhood trauma) and dissociation was performed for the whole sample. After the main
analysis, subgroup-analyses were conducted to investigate the differences between
schizophrenia, personality disorder and bipolar disorder (the statistical differences between
studies was also investigated). Secondly, subgroup analyses were conducted for different
types of childhood trauma (e.g. sexual abuse, physical abuse, emotional abuse, physical
neglect or emotional neglect, etc.) and dissociation when examined in primary studies. Again,
subgroup-analyses were conducted to investigate the differences between schizophrenia,
personality disorder and bipolar disorder. However, it was not possible to directly compare
the effects of different types of trauma (i.e. the effect of sexual abuse is stronger than that of
physical abuse) as in most cases these effects were estimated from the same sample of
participants (32).
Heterogeneity was examined in all analyses using the Q and I2 statistic, to determine
if heterogeneity was present, and to what degree the amount of statistical inconsistency
between studies existed. Following heterogeneity analysis, publication bias was assessed by
visual inspection of the funnel plots, and Egger’s test for funnel plot asymmetry (33) was
carried out to verify the influence of unpublished studies (i.e. no publication bias if studies
are located symmetrically around the mean effect size; 32). Duval and Tweedie’s, trim-and-
fill method (34) was applied to both the overall analysis and subgroup analyses when
appropriate to correct for the presence of publication bias. Furthermore, sensitivity analyses
10
(e.g. one study removed analysis, sub-group analyses) were conducted to determine the effect
of potentially influential studies on the meta-analytic results.
Results
Summary of studies
Figure 1 includes a flow chart summarising the search strategy for our review. The
overall number of studies retrieved from database searches totalled 6245. After title and
abstract screening, 332 full papers were screened, and 30 studies were included in the meta-
analysis. However, 31 effect sizes were extracted, as Braehler et al. (8) considered the
relationship between childhood adversity and dissociation in two separate diagnostic groups
of interest.
Demographic characteristics of eligible studies
The total number of participants included in the review totalled 2199. Most of the
studies had participants with a diagnosis of schizophrenia-spectrum disorders (n = 1192),
followed by personality disorder (n = 630), bipolar disorder (n = 303) and a mixed sample (n
= 74). The mean age was 35.28 (with information regarding age not reported for one study).
The ratio of male to females was 1156: 1108. Participants were recruited from a range of
countries including: USA (n = 630), Turkey (n = 373), Canada (n = 264), Germany (n = 263),
South Korea (n = 260), UK (n = 207), Spain (n = 116), Austria (n = 52), and Australia and
the UK (n = 34).
Measures of dissociation and childhood trauma
11
The details regarding measures of dissociation and childhood adversity are displayed
in Table 1. Seven different measures of dissociation were used to investigate the concept of
dissociation. The most commonly administered measures were the DES (k = 23) and
variations of the DES (DES-II, k = 2 and DES-T, k = 2). Furthermore, ten measures of
trauma were used to investigate childhood trauma. The most commonly administered
measures of childhood trauma were the CTQ (n = 16) and variations of the CTQ scale, i.e.
CTQ-SF (n = 6). Additionally, the specific types of adversity measured by the scales included
childhood sexual, physical and emotional abuse, and childhood emotional and physical
neglect. We found no eligible studies investigating the association between dissociative
experiences and exposure to mass violence, natural disasters or bullying in childhood.
Quality assessment
The global quality assessment ratings are presented in Table 1. The majority of
studies fell into the weak range (k = ), the remaining studies were rated as moderate (k = ). No
study achieved a strong global rating (component ratings displayed in Table S2). These
global quality ratings were affected by methodological limitations common in cross-sectional
literature (and quasi-experimental studies), specifically selection bias, study design and a
limited control of confounding variables
Design characteristics of eligible studies
The meta-analysis included between-group (n = 7) and correlational (n = 24) designs
to investigate the relationship between trauma and dissociation. No longitudinal design
studies were considered eligible for analysis, as the studies located did not meet the eligibility
criteria (i.e. did not investigate the relationship between childhood trauma and dissociation).
[Insert Figure 1 approximately here]
[Insert Table 1 approximately here]
Statistical analysis of the overall sample
12
The first set of analyses examined the relationship between total childhood trauma
and dissociation in the overall sample (i.e. irrespective of whether specific studies considered
samples of individuals with diagnoses of schizophrenia-spectrum disorders, personality
disorders or bipolar disorder). The results of the analysis are presented as a forest plot in
Figure S2 (see Figure 2). The analysis demonstrated a significant association between
exposure to potentially traumatic experiences in childhood and dissociation; k = 31, r = .33,
95% CI (.27, .39), p < .001. Based on Cohen’s (79) criteria (i.e. r = .10 indicates a small
effect; r = .30 indicates a medium effect and; r = .50 indicates a large effect), the summary
effect corresponded to a “medium-sized” association (when interpreting effect sizes, we urge
the reader to be cautious in associating larger effect sizes with more value, given the arbitrary
nature of criteria to evaluate the magnitude of summary effect sizes; 80).
Additionally, the Q and I2 analyses (Q = 97.50, p < .001, I2 = 69.23%) suggest a
significant amount of heterogeneity exists. Statistical heterogeneity was in the moderate
range according to widely used criteria to evaluate the amount of statistical inconsistency in
meta-analytic findings (i.e. I2: 25% = low, 50% moderate, 75% = high heterogeneity; 81).
However, the interpretation of I2 may be misleading given the arbitrary nature of the criteria.
[Insert Figure 2 approximately here]
Statistical analyses of clinical groups
The second set of analyses considered the association between childhood adversity
and dissociation within different diagnostic groups. Whereas the association between
childhood adversity and dissociation was found to be robust and significant in studies with
patients diagnosed with schizophrenia-spectrum (k = 20, r = .39, 95% CI [.31, .46], p < .001)
and personality disorders (k = 7, r = .24, 95% CI [.19, .29], p < .001), no significant
relationship was found in studies considering patients with bipolar disorder (k = 3, r = .15,
95% CI [-.04, .32], p = .114). Based on Cohen’s criteria (79), the relationship between
13
childhood adversity and dissociation appeared more robust within schizophrenia-spectrum
disorders studies (i.e. a medium summary effect size) than studies considering the other
diagnostic groups (i.e. small summary effect sizes). This was corroborated by a Q test
examining the difference between summary effects, which indicated that the magnitude of the
relationship significantly differed across clinical groups; Q = 10.03, df = 2, p = .007 (when
bipolar disorder is removed from analysis the significant difference between groups remains;
Q = 4.73, df =1, p = .030)
The Q test for schizophrenia-spectrum disorders was significant and the I2 statistic
indicated that heterogeneity was in the moderate range (Q = 50.06, p < .001, I2 = 64.04%).
Regarding personality disorders, no significant amount of heterogeneity was found (Q = 6.59,
p = .360, I2 = 9.01%). However, this result may be unreliable as the Q test may be biased in
meta-analyses that include a small number of studies (32).
Association between specific childhood adversities and dissociation across all SMIs
The second set of analyses investigated the relationship between specific childhood
adversities and dissociation within the combined SMI sample (see Table 2). The results
indicated that all types of childhood adversities for which it was possible to extract specific
effects, were positively associated with dissociation, with summary effect sizes within the
small-to-moderate range. Moderate statistical heterogeneity was observed in the analyses
considering sexual abuse, physical abuse, emotional abuse and aggregated neglect.
Associations between specific childhood adversities and dissociation within specific
diagnostic groups
The third set of analyses investigated the relationship between specific childhood
adversities and dissociation within different clinical groups. In the analyses considering the
14
schizophrenia-spectrum studies (see Table 2), all types of childhood adversity were
significantly associated with higher dissociation, with the notable exception of neglect
(results may be due to a few studies examined). Additionally, sexual abuse and aggregated
neglect demonstrated a medium amount of heterogeneity, while emotional abuse displayed
high levels of heterogeneity.
In the analyses focusing on personality disorders, small but significant associations
were identified between all types of childhood trauma investigated and dissociation, with the
exception of emotional neglect (see Table 2). Furthermore, aggregated neglect demonstrated
a significant moderate amount of heterogeneity, but heterogeneity statistics for this set of
analyses should be interpreted with caution given the small number of studies that examined
associations between specific types of adversities and dissociation in individuals with
personality disorder. No statistical integration was carried out for bipolar disorder as there
were too few studies for the analysis to be meaningful (32).
[Insert Table 2 approximately here]
Publication bias analyses
The results of our publication bias analyses, indicated the possible presence of
publication bias in the analysis examining the relationship between total childhood trauma
and dissociation within the overall sample (Egger’s test p = .012). As a result, Duval’s and
Tweedie’s trim-and-fill method was applied to correct for this potential bias. When seven
hypothetical studies were included in the analysis, the summary effect size was reduced, but
remained substantial; (r = .28, 95% CI [.22, .34]). Further analyses found no evidence of
publication or other selection bias for the analyses reported (see Table S2).
Sensitivity analyses
One study removed analyses found no evidence of potential influential cases across
the analyses reported (except for the bipolar subgroup; the relationship between childhood
15
trauma and dissociation became significant after the removal of Yilmaz et al., 2016). As a
few studies reported only significant relationships (i.e. selective reporting) between certain
specific childhood adversities and dissociation (12, 82, 83), sensitivity analyses were
conducted by excluding these studies from all analyses. This did not alter the pattern of
findings reported earlier. Furthermore, two studies contained childhood adversity measures
that examined both eligible and non-eligible (e.g. witnessing domestic violence, loss, etc.)
childhood traumatic experiences (43, 84). When the studies were excluded, the significant
relationships between childhood adversity and dissociation in the overall sample and
schizophrenia-spectrum studies were consistent with those reported in our original analyses
(see supporting information; Table S3 for all further sensitivity analyses and statistics).
Finally, to ensure that the design of the eligible studies did not influence our meta-
analytic findings, an additional subgroup analysis contrasting correlational and between-
group effect sizes was carried out for meta-analyses conducted on the overall sample. No
significant difference (Q = .05, df = 1, p = .830) was found between the summary effects of
studies which used between-group (k = 7, r = .32, 95% CI (.23, .41), p < .001) and
correlational designs (k = 24, r = .33, 95% CI (.25, .41), p < .001).
Discussion
Summary and discussion of the findings
The primary aim of the present meta-analysis was to examine the relationship
between childhood adversity and dissociation across three SMI diagnostic groups (i.e.
schizophrenia-spectrum disorders, personality disorders and bipolar disorder). Our findings
indicated a robust small-to-medium relationship between childhood adversity and
dissociation, when the effects of all studies which considered SMI patients were aggregated
in a single analysis. The magnitude of this relationship is broadly consistent with that
16
observed by the previous meta-analyses of the link between traumatic life experiences and
dissociation in other clinical and non-clinical populations (4, 7). The magnitude of this
relationship was found to differ across diagnostic groups according to widely used criteria to
evaluate effect sizes (79). Whilst a moderate and significant relationship was found within
schizophrenia-spectrum disorders studies, the association between childhood adversity and
dissociative experiences was somewhat smaller in studies with individuals who received
personality disorder diagnoses. Furthermore, the results found no evidence of an association
between life adversities and dissociative experiences in individuals with bipolar disorder.
Whilst, this negative finding might be in part due to the very small number of eligible bipolar
studies, this finding parallel those of recent meta-analytic synthesis, indicating that
dissociative experiences are not a prominent complaint of clients with diagnoses of bipolar
and related disorders (85), despite the high levels of childhood adversity observed in the
clinical group (86).
In regards to the secondary aim of the present review (i.e. examine the relationship
between specific types of childhood adversity and dissociation across different SMIs), it is
notable that our search strategy did not identify any eligible studies considering the
association between dissociation and bullying, mass violence and conflict exposure, despite
the fact that in previous research these potentially traumatic experiences have been linked to
both SMIs and stress-and trauma-related symptoms, including dissociative experiences. The
findings of our meta-analytic integrations indicated that all other types of childhood
adversities considered in this evidence synthesis showed a significant relationship with
dissociation. Across all diagnostic samples, the association between emotional abuse and
dissociation was particularly robust. While conclusions cannot be drawn as to which
childhood adversity might be most associated with dissociation based on our findings (e.g.
due to non-independent nature of the sample of participants), primary studies that employed
17
multivariate analyses to test the relative contribution of different life adversities have often
found that emotional abuse is one of the most robust predictors of dissociative experiences (8,
36). This may be explained by the interpersonal nature of emotional abuse. Repetitive
childhood emotional abuse in the form of verbal insults, criticism, humiliation and isolation,
may be seen as more threatening compared to other types of abuse, as emotional abuse is
often perpetrated by somebody whom the victim is close to and is reliant upon for protection
and support (87). It has been proposed that the contradictory nature of an emotionally abusive
“caregiver” (being at the same time the source of abuse and comfort/protection/survival) may
be particularly conductive to dissociative responses that protect the survivors’ internal selves
whilst maintaining attachment to significant others in order to ensure survival (87, 88).
In many cases, our findings demonstrated considerable statistical heterogeneity. Thus,
the reported summary effect sizes should be interpreted with some caution, as guidance (e.g.
GRADE approach) suggests that considerable statistical heterogeneity decreases the quality
of evidence (insert REF) and indicates t the “true” effect may be somewhat smaller or larger
than indicated in the current meta-analysis (32). Given the noticeable methodological and
clinical heterogeneity of the included studies, the results of the heterogeneity analyses are not
surprising. The quality of included studies may partly explain the heterogeneity reported. The
majority of studies quality assessed using the EPHPP achieved weak global ratings (k =); this
may affect the meta-analytic evidence supporting an association between childhood trauma
and dissociation in SMI. Studies were consistently rated as weak because most of the eligible
studies were correlational, did not take confounding variables into account and were
vulnerable to selection bias. Additionally, components of the EPHPP, in particular selection
bias were rated as weak because the necessary information required to assess bias was not
present. Per EPHPP guidance, inferences about what the authors intended to do could not be
made, and instead judgements were required to be made using the information contained in
18
the study. However, while studies may not be entirely transparent regarding methodology
(e.g. selection bias, blinding, etc.) due to word count restrictions in published papers, it does
not necessarily imply that the method was not used in the study (Mhaskar, Djulbegovic,
Magazin, Soares, & Kumar, 2012). Therefore, the results of quality assessment may be
somewhat conservative in reflecting the true methodological quality of included studies.
A series of subgroup and sensitivity analyses were conducted in an attempt to explain
the observed statistical heterogeneity. The analyses showed that the inconsistency between
studies is unlikely to be due to study design, as the effect sizes did not substantially differ
across correlational and between-group designs. However, the relationship between
childhood adversity and dissociation substantially varied across diagnostic groups, which
may explain some of the statistical inconsistency observed in the overall analysis. As a wide
range of empirical studies was included in the meta-analysis, other unmeasured variables not
measured in the current study may account for the observed heterogeneity. Factors such as
the detailed features of the potentially traumatic events under scrutiny (e.g. timing, frequency,
severity, relationship to the abuser, etc.) affect the trauma and dissociation relationship. For
example, individuals abused by parents (vs others e.g. relatives, strangers or friends), or those
exposed to more than one type of abuse demonstrated higher dissociative scores (48, 89), and
as the number of perpetrators increase; the greater the dissociative symptoms experienced by
the victim (90). Additionally, socio-demographic characteristics of the respondents (e.g. age,
gender, ethnicity, etc.) may plausibly moderate the association between trauma exposure and
dissociative experiences. For instance, age was found to be a significant predictor of
dissociation (13), and previous research has found that the relationship between trauma and
dissociation varied across ethnicities (91) and gender (92). Given that only a small amount of
studies explored such factors, it was not possible to examine the contribution of those
moderators systematically using meta-analytic methods. While many possible sources of
19
heterogeneity were not controlled for, some of the primary studies accounted for a range of
factors shown to be related to self-reported dissociative experiences, such as substance abuse,
cognitive functioning and medical conditions such as epilepsy (92-94). As the corpus of
evidence pertaining to the relationship between trauma and dissociation in SMIs is
increasing, future systematic reviews may attempt to further explore the impact of these
effect moderators.
An important consideration relevant to the findings of the present review concerns the
conceptualisation and measurement of dissociation in the available literature. Whilst the
majority of studies included in this meta-analysis (and previous meta-analytic reviews of
dissociation studies; 4, 7, 85) used the DES to assess dissociative experiences, several
researchers have criticised the use of this measure on the grounds that dissociation should be
regarded as a multifaceted rather than a unitary construct (2, 95). Critics have argued that the
unitary model of dissociation is over-inclusive and implies that dissociative experiences such
as absorption and derealisation are qualitatively similar and only differ in “amount” of
dissociation (2). Therefore, even though the DES has good psychometric properties and is
easy to administer, its over-inclusive and unitary nature may limit the validity of research
findings gathered using this measure. On the other hand, the proposal that dissociation should
be conceptualised as a multifaceted construct, although theoretically plausible, is yet to
translate into the development of psychometrically sound measures assessing qualitative and
aetiologically distinct forms of dissociation (e.g. dissociative detachment and
compartmentalisation; 2, 95) that could replace the DES. Furthermore, the inclusion in the
DES of items assessing absorption has received some criticism, as this experience is not
necessarily pathological and is common in the general population. Therefore, it has been
argued that the use of the DES might inflate and bias any assessment of “true” dissociative
experiences (4). Despite this, numerous studies have indicated that absorption is robustly
20
associated with other dissociative experiences and other dimensions of psychopathology, (96,
97) and it has been therefore argued that its exclusion from dissociation tools used in both
clinical and research settings may be problematic (4).
A number of additional limitations should be taken into account when interpreting our
findings. All the studies included in the meta-analysis used retrospective measures of
childhood trauma. Some researchers argue that the use of retrospective measures of
childhood trauma inflate the relationship between trauma and dissociation. Also, it has been
claimed that patients may, in fact, exaggerate traumatic memories (98) or confuse fantasy
with factual memories (99). However, Dalenberg et al. (4) found that when using more
objective measures of childhood trauma (e.g. confirmation by therapist, child protective
service reports, etc.), the relationship between childhood sexual abuse and dissociation
remained moderate (r = .30) and similar to that observed in studies where self-report
measures (r = .32) were administered. Furthermore, research suggests that traumatic
experiences measured using self-report measures tend to be under-reported rather than over-
reported (100).
In the present meta-analysis, we did not employ an “off the shelf” quality assessment
tool to evaluate primary studies. Instead, we extracted specific methodological variables and
conducted a range of sensitivity analyses to quantitatively assess for potential biases (e.g.
selective reporting, study design, measures, etc.). Multiple sensitivity analyses were
conducted and the analyses demonstrated no difference in effect sizes after methodological
variables were accounted for .
Alternatively, the omission of certain studies may have biased results. Despite
emailing authors of all eligible studies for any unpublished work they may have; no
unpublished study was located. However, in the current meta-analysis, the trim-and-fill
method was used to adjust for the potential influence of publication or other selection biases
21
in relevant analyses. When bias was accounted for the relationship between childhood
adversity and dissociation did not change substantially. Also, the literature search was
extended to languages other than English in order to minimise selection biases. Future meta-
analytic reviewers should strive to increase the comprehensiveness of the search and include
unpublished studies/reports to reduce the file drawer effect.
In the present meta-analysis, the initial screening phase (title and abstract) was
completed by one researcher. S.R screened all titles, abstracts and full papers, while C.C.
screened 20% of all titles and abstracts and all full papers. Research suggests the assessment
of eligibility at all stages should be completed by at least two researchers to minimise
bias/error (101, INSERT REF). To minimise error during the screening stage, S.R used
backward and forward screening on all 30 eligible studies to ensure the retrieval of all
possible studies.
Another potential limitation is the varied conceptualisation of which psychiatric
diagnoses should be regarded as “SMIs”. It is broadly recognised that the conceptualisation
of SMI lacks specificity and remains inconsistent (102). In the present review, we employed
the definition of SMI used in the UK as part of the Improving Access to Psychological
Therapies for SMI project (26), but we acknowledge that other diagnostic groups may have
qualified for inclusion if other SMI definitions were employed. For example, some
researchers have employed narrower definitions of SMIs (e.g. schizophrenia-spectrum
disorders and bipolar disorder only; 103, 104), whilst others used more inclusive ones (102,
105).
Future research
Regardless of the limitations mentioned above, the findings of this evidence synthesis
suggest that the dissociative experiences reported by individuals with SMIs, can be at least in
part represent psychological sequelae of exposure to adverse life experiences in childhood.
22
However, causality and directionality cannot be ascertained as only between-group and
correlational studies were included, as no eligible longitudinal studies were identified.
Further research is required to clarify the direction of the associations considered in the
present evidence synthesis. Other non-clinical longitudinal studies have demonstrated that
exposure to childhood trauma was associated with heightened levels of dissociation
compared to non-abused individuals (106); therefore, it is plausible that the same direction of
effects may be observed in the context of individuals who also received diagnoses of SMIs.
Also, many of the included studies in the review did not control for potential confounding
factors that may affect the relationship between childhood trauma and dissociation.
Therefore, to separate the effects of trauma from those possibly caused by other variables
(e.g. the sustained disturbances from the parent-child relationship) additional primary
research is required.
In relation to the secondary aim of the analysis, no firm conclusions can be drawn
regarding which traumatic experience is most strongly associated with dissociative
experiences. As the studies included in the analysis were not independent (32). Future studies
should explore further the relationship between specific types of childhood trauma and
dissociation. Using appropriate statistical methods and sufficiently large samples, to
disentangle the relative contribution of qualitatively different childhood adversities as well as,
the interplay between exposure to different adversities (e.g. the effect of
re-victimisation/polyvictimisation) and the impact of important trauma-related factors that
may further aggravate dissociative symptoms (e.g. the length and “severity” of potentially
traumatic experiences). Additionally, given the multifaceted nature of dissociation, future
research should consider the association between trauma/specific traumatic experiences and
specific features of dissociation. We initially planned to investigate those experiences (25),
however, too few studies were located (e.g. 43,44). Therefore, sufficient information could
23
not be extracted to undertake these analyses. Lastly, no studies were considered eligible for
natural disasters, mass violence, and bullying. Future studies should focus on these
adversities as well as, other common childhood adversities in the risk of developing
dissociation, as these traumatic experiences have been found to increase the risk of
developing dissociation in the general population (100, 107-109).
Clinical implications
These findings along with other meta-analytic results (4, 7), highlight the damaging
role of childhood trauma in the development of dissociative experiences. Therefore, it is
advisable for practitioners to routinely and sensitively enquire about abuse history in patients
who display dissociative symptoms (8, 11). Current research suggests that clinicians do not
routinely question individuals regarding trauma (110, 111). This may be due to clinicians’
beliefs about the biological aetiology of mental illness (112), the lack of awareness of the
adverse effects of social factors and life experiences on adult functioning (113), or concerns
that questioning service-users about potentially traumatic events may lead to further distress
and aggravation of symptoms (102). However, research indicates that most trauma survivors,
including those who received diagnoses of SMIs, do not experience any aggravation in their
mental health when asked about past traumatic experiences, and positively evaluate the
opportunity of talking about these experiences (114-116).
Additionally, clinicians should work collaboratively with service users to include and
understand the role played by trauma and dissociation in the development and maintenance of
their presenting difficulties, and incorporate appropriate interventions to ameliorate
dissociation and related difficulties in their treatment plans. Although a number of
psychological interventions for dissociative experiences are available (e.g. cognitive
behaviour therapy for depersonalisation; 117), their evaluation in clients with SMIs is in its
24
infancy (118). On the other end, a growing number of studies in recent years have evaluated
the safety and efficacy of trauma-focused interventions in people with SMIs, with promising
findings (119, 120). Research suggests that trauma-focused interventions can effectively
reduce the severity of dissociative symptoms in trauma survivors (4), and there is evidence
that trauma-focused therapies such as Prolonged Exposure and Eye Movement
Desensitisation and Reprocessing can lead to similar improvements in dissociative
experiences in those with SMIs (120). Thus, psychological treatments that address traumatic
experiences and incorporate dissociation into treatment plans, could potentially benefit
individuals with SMIs experiencing dissociation and other trauma-related symptoms.
Conclusion
In conclusion, our evidence synthesis of 30 clinical studies found evidence in support
of an association between childhood adversity and dissociation in individuals with SMI. To
ensure robust conclusions, publication and other selection biases have been taken into
account.
Acknowledgements
We thank Helga Masramon, Minyoung Sim, and Vedat Sar, for providing information
regarding their relevant published studies. We also thank the researchers who kindly provided
copies of their papers for the screening stage of the research synthesis: Alexander McGirr,
Camilla Callegari, Colin A. Ross, Harvey E. Dondershine, Park E. Dietz, and Joel Parris.
Declaration of interest
The authors declare that they have no conflict of interest.
25
References
1 American Psychiatric Association. Diagnostic and statistical manual of mental
disorders DSM-V. 5th ed. American Psychiatric Association, Arlington, 2000.
2 BROWN RJ. Different types of “dissociation” have different psychological
mechanisms. J Trauma Dissociation 2006;7:7-28. doi:10.1300/J229v07n04_02.
3 SPITZER C, BARNOW S, FREYBERGER HJ, GRABE HJ. Recent developments in
the theory of dissociation. WPA 2006;5:82-86.
4 DALENBERG CJ, BRAND BL, GLEAVES DH et al. Evaluation of the evidence
for the trauma and fantasy models of dissociation. Psychol Bull 2012;138:550-588.
doi:10.1037/a0027447.
5 CARLSON EA, YATES TM, SROUFE LA. Dissociation and the development of the
self. In DELL PF, O’NEIL JA, eds. Dissociation and the dissociative disorders,
Routledge, New York, 2009.
6 GENTILE JP, DILLON KS, GILLIG PM. Psychotherapy and pharmacotherapy for
patients with dissociative identity disorder. Innov Clin Neurosci 2013;10: 22-29.
7 VAN IJZENDOORN MH, SCHUENGEL C. The measurement of dissociation in
normal and clinical populations: meta-analytic validation of the dissociative
experiences scale. Clin Psychol Rev 1996;16:365-382. doi:10.1016/0272-
7358(96)00006-2.
8 BRAEHLER C, VALIQUETTE L, HOLOWKA D et al. Childhood trauma and
dissociation in first-episode psychosis, chronic schizophrenia and community
controls. Psychiatry Res 2013;210:36-42. doi:10.1016/j.psychres.2013.05.033.
9 GHOREISHI A, SHAJARI Z. Reviewing the dissociative symptoms in patients with
schizophrenia and their association with positive and negative symptoms. Iran
J Psychiatry Behav Sci 2014;8:13-18.
10 ERYILMAZ G, KESEBIR S, GUL IG, OZTEN E, KARAMUSTAFALIOGLU KO.
Dissociative experiences in bipolar disorder II: are they related to childhood trauma
and obsessive-compulsive symptoms? Arch Clin Psychiatry 2015;42:38-40.
doi:10.1590/0101-60830000000045.
26
11 SEMIZ UB, BASOGLU C, EBRINC S, CETIN M. Childhood trauma history and
dissociative experiences among Turkish men diagnosed with antisocial personality
disorder. Soc Psychiatry and Psychiatr Epidemiol 2007;42:865-873. doi:
10.1007/s00127-007-0248-2.
12 ZANARINI MC, RUSER TF, FRANKENBURG FR, HENNEN J, GUNDERSON
JG. Risk factors associated with the dissociative experiences of borderline patients. J
Nerv Ment Dis 2000;188:26-30. doi:10.1097/00005053-200001000-00005.
13 SAR V, TAYCAN O, BOLAT N et al. Childhood trauma and dissociation in
schizophrenia. Psychopathology 2012;43:33-40. doi:10.1159/000255961.
14 SCHALINSKI I, TEICHER MH. Type and timing of childhood maltreatment and
severity of shutdown dissociation in patients with schizophrenia-spectrum disorder.
PLoS One 2015;10. doi:10.1371/journal.pone.0127151.
15 HARIRI AG, GULEC MY, ORENGUL FFC, SUMBUL EA, ELBAY RY, GULEC
H. Dissociation in bipolar disorder: relationships between clinical variables and
childhood trauma. J Affect Disord 2015;184:104-110. doi:10.1016/j.jad.2015.05.023.
16 WATSON S, CHILTON R, FAIRCHILD H, WHEWELL P. Association between
childhood trauma and dissociation among patients with borderline personality
disorder. Aust N Z J Psychiatry 2006;40:478-481.
doi:10.1111/j.14401614.2006.01825.x.
17 CHAE S, SIM M, LIM M, NA J, KIM D. Multivariate analysis of the relationship
between childhood trauma and psychotic symptoms in patients with schizophrenia.
Psychiatry Investig 2015;12:397-401. doi:10.4306/pi.2015.12.3.397.
18 GOODMAN M, WEISS DS, MITROPOULOU V et al. The relationship between
pathological dissociation, self-injury and childhood trauma in patients with
personality disorders using taxometric analyses. J Trauma Dissociation 2003;4:65-88.
doi:10.1300/J229v04n02_05.
19 KONG SS, BAE JH. Impact of childhood trauma on post-traumatic stress
symptoms and dissociation: mediating effect of bullying. J Korean Acad Psychiatr
Ment Health Nurs 2013;22:193-204. doi:10.12934/jkpmhn.2013.22.3.193.
27
20 OH HY, KIM D, KIM Y. Reliability and validity of the Dissociative
Experiences scale among South Korean patients with schizophrenia. J Trauma
Dissociation 2015;16:577-591. doi:10.1080/15299732.2015.1037040.
21 RUIZ MA, POYTHRESS NG, LILIENFELD SO, DOUGLAS KS. Factor structure
and correlates of the Dissociative Experiences Scale in a large offender sample.
Assessment 2008;15:511-521. doi:10.1177/1073191108315548.
22 JOHNSTON C, DORAHY MJ, COURTNEY D, BAYLES T, O'KANE M.
Dysfunctional schema modes, childhood trauma and dissociation in borderline
personality disorder. J Behav Ther Exp Psychiatry 2009;40:248-255.
doi:10.1016/j.jbtep.2008.12.002.
23 SIMEON D, KNUTELSKA M, SMITH L, BAKER BR, HOLLANDER E. A
preliminary study of cortisol and norepinephrine reactivity to psychosocial stress in
borderline personality disorder with high and low dissociation. Psychiatry Res
2007;149:177-184. doi:10.1016/j.psychres.2005.11.014.
24 LIBERATI A, ALTMAN DG, TETZLAFF J et al. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that evaluate health care
interventions: explanation and elaboration. Ann Intern Med 2009;151:65-95.
doi:10.7326/0003-4819-151-4-200908180-00136.
25 RAFIQ, S, CAMPODONICO C, VARESE, F. The relationship between childhood
trauma and dissociation in people with severe mental health difficulties: a meta-
analytic review. PROSPERO: International prospective register of systematic reviews.
2016. CRD42016048705. Available from:
http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42016048705.
26 JOLLEY, S, GARETY P, PETERS E et al. Opportunities and challenges in
Improving Access to Psychological Therapies for people with Severe Mental Illness
(IAPT-SMI): evaluating the first operational year of the South London and Maudsley
(SLaM) demonstration site for psychosis. Behav Res Ther 2015;64:24-30.
doi:10.1016/j.brat.2014.11.006.
27 KRAMER DN, LANDOLT MA. Characteristics and efficacy of early psychological
interventions in children and adolescents after single trauma: a meta-analysis. Eur J
Psychotraumatol 2011;2:7858. doi:10.3402/ejpt.v2i0.7858.
28
28 LUSZCZYNSKA A, BENIGHT CC, CIESLAK R. Self-efficacy and health-related
outcomes of collective trauma a systematic review. European Psychol 2009;14:51-62.
doi:10.1027/1016-9040.14.1.51.
29 PILTON M, VARESE F, BERRY K, BUCCI S. The relationship between
dissociation and voices: a systematic literature review and meta-analysis. Clin
Psychol Rev 2015;40:138-155. doi:10.1016/j.cpr.2015.06.004.
30 STEEL Z, CHEY T, SILOVE D, MARNANE C, BRYANT RA, VAN OMMEREN
M. Association of torture and other potentially traumatic events with mental health
outcomes among populations exposed to mass conflict and displacement: a
systematic review and meta-analysis. JAMA 2009;302:537-549.
doi:10.1001/jama.2009.1132.
31 VARESE F, SMEETS F, DRUKKER M et al. Childhood adversities increase the risk
of psychosis: a meta-analysis of patient-control, prospective-and cross-sectional
cohort studies. Schizophr Bull 2012;38:661-671. doi:10.1093/schbul/sbs050. 30
32 BORENSTEIN M, HEDGES LV, HIGGINS J, ROTHSTEIN HR. Introduction to
meta- analysis. John Wiley & Sons, Ltd, Chichester, UK, 2009.
33 EGGER M, DAVEY-SMITH D, SCHNEIDER M, MINDER C. Bias in meta-analysis
detected by a simple, graphical test. BMJ 1997;315:629-634.
doi:10.1136/bmj.315.7109.62.
34 DUVAL S, TWEEDIE R. A nonparametric “trim and fill” method of accounting for
publication bias in meta-analysis. J Amer Statist Assoc 2000;95:89-98.
doi:10.2307/2669529.
35 ROSS CA, ANDERSON G, CLARK P. Childhood abuse and the positive symptoms
of schizophrenia. Psychiatr Serv 1994;45:489-491.
36 HOLOWKA DW, KING S, SAHEB D, PUKALL M, BRUNET A. Childhood abuse
and dissociative symptoms in adult schizophrenia. Schizophr Res 2003;60:87-90.
doi:10.1016/S0920-9964(02)00296-7.
37 ROSS CA, KEYES B. Dissociation and schizophrenia. J Trauma Dissociation
2004;5:69-83. doi:10.1300/J229v05n03_05.
29
38 LYSAKER PH, DAVIS LW, GATTON MJ, HERMAN SM. Associations of anxiety
related symptoms with reported history of childhood sexual abuse in schizophrenia
spectrum disorders. J Clin Psychiatry 2005;66:1279–84. doi:10.4088/JCP.v66n1010.
39 SCHAFER I, HARFST T, ADERHOLD V et al. Childhood trauma and dissociation
in female patients with schizophrenia spectrum disorders: an exploratory study. J
Nerv Ment Dis 2006;194:135-138. doi:10.1097/01.nmd.0000198199.57512.84.
40 DORAHY MJ, SHANNON C, SEAGAR L et al. Auditory hallucinations in
dissociative identity disorder and schizophrenia with and without a childhood trauma
history: similarities and differences. J Nerv Ment Dis 2009;3:892-898.
doi:10.1097/NMD.0b013e3181c299ea.
41 VOGEL M, SPITZER C, KUWERT P, MOLLER B, FREYBERGER HJ, GRABE
HJ. Association of childhood neglect with adult dissociation in schizophrenic
inpatients. Psychopathology 2009;42:124-130. doi:10.1159/000204763.
42 LADDIS A, DELL PF. Dissociation and psychosis in dissociative identity disorder
and schizophrenia. J Trauma Dissociation 2012;13:397-413.
doi:10.1080/15299732.2012.664967.
43 PERONA‐GARCELAN S, CARRASCOSO‐LOPEZ F, GARCÍA‐MONTES JM et al.
Dissociative experiences as mediators between childhood trauma and auditory
hallucinations. J Trauma Stress 2012;25:323-329. doi:10.1002/jts.21693.
44 SCHAFER I, FISHER HL, ADERHOLD V et al. Dissociative symptoms in patients
with schizophrenia: relationships with childhood trauma and psychotic symptoms.
Compr Psychiatry 2012;53:364-371. doi:10.1016/j.comppsych.2011.05.010.
45 VARESE F, BARKUS E, BENTALL RP. Dissociation mediates the relationship
between childhood trauma and hallucination-proneness. Psychol Med 2012;42:1025-
1036. doi:10.1017/S0033291711001826.
46 KIM D, BAE H, HAN C, OH HY, MACDONALD K. Psychometric properties of
the Childhood Trauma Questionnaire-short form (CTQ-SF) in Korean patients with
schizophrenia. Schizophr Res 2013;144:93-98. doi:10.1016/j.schres.2012.12.020.
30
47 LAFERRIERE-SIMARD MC, LECOMTE T, AHOUNDOVA L. Empirical testing of
criteria for dissociative schizophrenia. J Trauma Dissociation 2014;15:91-107.
doi:10.1080/15299732.2013.834860.43.
48 ALVAREZ MJ, MASRAMON H, PENA C et al. Cumulative effects of childhood
traumas: polytraumatization, dissociation, and schizophrenia. Community Ment
Health J 2015;51:54-62. doi:10.1007/s10597-014-9755-2.
49 SIMEON D, NELSON D, ELIAS R, GREENBERG J, HOLLANDER E.
Relationship of personality to dissociation and childhood trauma in borderline
personality disorder. CNS Spectr 2003;8:755-762.
50 LOFFLER-STASTKA H, SZERENCSICS M, BLUML V. Dissociation, trauma,
affect regulation and personality in patients with a borderline personality
organization. Bull Menninger Clin 2009;73:81-98. doi:10.1521/bumc.2009.73.2.81.
51 YILMAZ O, ATES MA, SEMIZ UB et al. Childhood traumas in patients with bipolar
disorder: association with alexithymia and dissociative experiences. Anatolian J
Psychiatry 2016;17:188-195. doi:10.5455/apd.188358.
52 VAN DER KOLK BA, PERRY JC, HERMAN JL. Childhood origins of self-
destructive behaviour. Am J Psychiatry 1991;148:1665-1671.
doi:10.1176/ajp.148.12.1665.
53 ZANARINI MC, GUNDERSON JG, MARINO MF, SCHWARTZ EO,
FRANKENBURG FR. Childhood experiences of borderline patients. Compr
Psychiatry 1989;30:18-25. doi:10.1016/0010-440X(89)90114-4.
54 BERNSTEIN DP, FINK L. Manual for the Childhood Trauma Questionnaire: a
retrospective self-report. Harcourt Brace & Company, New York, 1998.
55 BERNSTEIN DP, FINK L, HANDELSMAN L et al. Initial reliability and validity of
a new retrospective measure of child abuse and neglect. Am J
Psychiatry 1994;151:1132-1136.
56 SAR V, OZTURK E, IKIKARDES E. Validity and reliability of the Turkish version
of Childhood Trauma Questionnaire. Turkiye.Klinikleri J of Med Sci 2012;32:1054–
1063. doi:10.5336/medsci.2011-26947.
31
57 WINGENFELD K, SPITZER C, MENSEBACH C et al. The German version of the
Childhood Trauma Questionnaire (CTQ): preliminary psychometric properties.
Psychotherapie, Psychosomatik, Medizinische Psychologie 2010;60:442-450.
doi:10.1055/s-0030-1247564.
58 BERNSTEIN DP, STEIN JA, NEWCOMB MD et al. Development and validation of
a brief screening version of the Childhood Trauma Questionnaire. Child Abuse Negl
2003;27:169-190. doi:10.1016/S0145-2134(02)00541-0.
59 HERNANDEZ A, GALLARDO-PUJOL D, PEREDA N et al. Initial validation of the
Spanish Childhood Trauma Questionnaire-short form: factor structure, reliability and
association with parenting. J of Interpers Violence 2013;28:1498–1518.
doi:10.1177/0886260512468240.
60 KIM D, PARK SC, YANG H, OH DH. Reliability and validity of the Korean version
of the Childhood Trauma Questionnaire-short form for psychiatric outpatients.
Psychiatry Investig 2011;8:305-311. doi:10.4306/pi.2011.8.4.305.
61 ROSS CA. Dissociative identity disorder. Diagnosis, clinical features, and treatment
of multiple personality. 2nd ed. John Wiley & Sons, New York, 1997.
62 GRAY MJ, LITZ BT, HSU JL, LOMBARDO TW. Psychometric properties of the
Life Events Checklist. Assessment 2004;11:330-341.
doi:10.1177/1073191104269954.
63 SANDERS B, BECKER-LAUNSEN E. The measurement of psychological
maltreatment: early data on the Child Abuse and Trauma Scale. Child Abuse Negl
1995;19:315–323. doi:10.1016/S0145-2134(94)00131-6.
64 CUSACK KJ, FRUEH BC, BRADY KT. Trauma history screening in a community
mental health center. Psychiatric Serv 2004;55:157-162.
doi:10.1176/appi.ps.55.2.157.
65 DAVIDSON J, SMITH R. Traumatic experiences in psychiatric outpatients. J
Trauma Stress 1990;3:459-475. doi:10.1002/jts.2490030314. 66 HERMAN JL, VAN DER KOLK B. Traumatic antecedent of borderline personality
disorder. In VAN DER KOLK B, ed. Psychological trauma, American Psychiatric
Press, Washington, 1987.
32
67 NIJENHUIS ERS, VAN DER HART O, VANDERLINDEN J. The traumatic
experiences checklist. In NIJENHUIS ERS, ed. Somatoform dissociation:
Phenomena, measurement, and theoretical issues, Van Gorcum, Assen, 1999.
68 BERNSTEIN EM, PUTNAM FW. Development, reliability, and validity of a
dissociation scale. J Nerv Ment Dis 1986;174:727–735. doi:10.1097/00005053-
198612000-00004.
69 FREYBERGER HJ, SPITZER C, STIEGLITZ RD, KUHN G, MAGDEBURG N,
BERNSTEIN- CARLSON E. Fragebogen zu dissoziativen Symptomen (FDS).
Deutsche Adaption,Reliabilität und Validität der deutschen Dissociative Experiences
Scale (DES). Psychotherapie, Psychosomatik, Medizinische Psychologie
1998;48:223–229.
70 SPITZER C, FREYBERGER HJ, STIEGLITZ RD et al. Adaptation and psychometric
properties of the German version of the dissociative experience scale. J Trauma Stress
1998;11:799–809. doi:10.1023/A:1024457819547.
71 PARK J, CHOE BM, KIM M et al. Standardization of Dissociative Experiences
Scale-Korean version (I). Korean J Psychopathol 1995;4:105–125.
doi:10.1080/15299732.2015.1037040.
72 YARGIC LI, TUTKUN H, SAR V. Reliability and validity of the Turkish version of
the Dissociative Experiences Scale. Dissociation: Progress Dissociative Disord
1995;16:577-591. doi:10.1080/15299732.2015.1037040.
73 CARLSON EB, PUTNAM FW. An update on the Dissociative Experiences Scale.
Dissociation: Progress Dissociative Disord 1993;6:16–27.
74 ICARAN E, COLOM R, ORENGO-GARCIA F. Validation study of the dissociative
experiences scale in Spanish population sample. Actas Luso Esp Neurol Psiquiatr
Cienc Afines 1996;24:7–10.
75 WALLER N, PUTNAM FW, CARLSON EB. Types of dissociation and dissociative
types: a taxometric analysis of dissociative experiences. Psychol Methods
1996;1:300-321. doi:10.1037/1082-989X.1.3.300.
76 DELL PF. The Multidimensional Inventory of Dissociation (MID): a
comprehensive measure of pathological dissociation. J Trauma Dissociation
2006;7:77–106. doi:10.1300/J229v07n02_06.
33
77 SCHALINSKI I, SCHAUER M, ELBERT T. The Shutdown Dissociation scale (Shut-
D). Eur J Psychotraumatol 2015;6:25652. doi:10.3402/ejpt.v6.25652.
78 BRIERE J. Trauma Symptom Inventory professional manual. Psychological
Assessment Resources, Odessa, 1995.
79 COHEN J. A power primer. Psychol Bull 1992;112:155-159. doi:10.1037/0033-
2909.112.1.155. 33.
80 GLASS GV, MCGAW B, SMITH ML. Meta-Analysis in Social Research. Sage,
Beverly Hills, 1981.
81 HIGGINS JP, THOMPSON SG, DEEKS JJ, ALTMAN DG. Measuring inconsistency
in meta-analyses. BMJ 2003;327:557-560. doi:10.1136/bmj.327.7414.557.
82 LOFFLER-STASTKA H, SZERENCSICS M, BLUML V. Dissociation, trauma,
affect regulation and personality in patients with a borderline personality
organization. Bull Menninger Clin 2009;73:81-98. doi:10.1521/bumc.2009.73.2.81.
83 SCHAFER I, HARFST T, ADERHOLD V et al. Childhood trauma and dissociation
in female patients with schizophrenia spectrum disorders: an exploratory study. J
Nerv Ment Dis 2006;194:135-138. doi:10.1097/01.nmd.0000198199.57512.84.
84 VAN DER KOLK BA, PERRY JC, HERMAN JL. Childhood origins of self-
destructive behaviour. Am J Psychiatry 1991;148:1665-1671.
doi:10.1176/ajp.148.12.1665.
85 LYSSENKO L, SCHMAHL C, BOCKHACKER L, VONDERLIN R, BOHUS M,
KLEINDIENST N. Dissociation in psychiatric disorders: a meta-analysis of studies
using she Dissociative Experiences Scale. Am J Psychiatry 2017;175:37-46.
doi:10.1176/appi.ajp.2017.17010025.
86 PALMIER-CLAUS J, BUCCI S, BERRY K, VARESE F. The relationship between
childhood adversity and bipolar disorder: a systematic review and meta-analysis. Br J
Psychiatry 2016;209:454-459. doi:10.1192/bjp.bp.115.179655.
87 FREYD JJ. Betrayal trauma: traumatic amnesia as an adaptive response to childhood
abuse. Ethics Behav 1994;4:307-329. doi:10.1207/s15327019eb0404_1.
34
88 CARLSON EA. A prospective longitudinal study of attachment
disorganization/disorientation. Child Dev 1998;69:1107-1128. doi: 10.1111/j.1467-
8624.1998.tb06163.x.
89 Greenfield SF, Strakowski SM, Tohen M, Batson SC, Kolbrener ML. Childhood
abuse in first-episode psychosis. Br J Psychiatry 1994;164:831-834.
doi:10.1192/bjp.164.6.831.
90 VAN DEN BOSCH LM, VERHEUL R, LANGELAND W, VAN DEN BRINK W.
Trauma, dissociation, and posttraumatic stress disorder in female borderline patients
with and without substance abuse problems. Aust N Z J Psychiatry 2003;37:549-555.
doi:10.1046/j.1440-1614.2003.01199.x.
91 ANGLIN DM, POLANCO-ROMAN L, LUI F. Ethnic variation in whether
dissociation mediates the relation between traumatic life events and attenuated
positive psychotic symptoms. J Trauma Dissociation 2015;16:68-85.
doi:10.1080/15299732.2014.953283.
92 SEEDAT S, STEIN MB, FORDE DR. Prevalence of dissociative experiences in a
community sample: relationship to gender, ethnicity, and substance use. J Nerv Ment
Dis 2003;191:115-120. doi:10.1097/01.NMD.0000050940.16782.6B.
93 ALPER K, DEVINSKY O, PERRINE K et al. Dissociation in epilepsy and
conversion non epileptic seizures. Epilepsia 1997;38:991-997.
doi:10.1111/j.15281157.tb01481.x.
94 PERZOW SE, PETRENKO CL, GARRIDO EF, COMBS MD, CULHANE SE,
TAUSSIG HN. Dissociative symptoms and academic functioning in maltreated
children: a preliminary study. J Trauma Dissociation 2013;14:302-311.
doi:10.1080/15299732.2012.736928.
95 HOLMES EA, BROWN RJ, MANSELL W et al. Are there two qualitatively distinct
forms of dissociation? A review and some clinical implications. Clin Psychol Rev
2005;25:1-23. doi:10.1016/j.cpr.2004.08.006.
96 ALLEN JG, COYNE L, CONSOLE DA. Dissociation contributes to anxiety and
psychoticism on the Brief Symptom Inventory. J Nerv Ment Dis 1996;184:639-641.
doi:10.1097/00005053-199610000-00010.
35
97 LEAVITT F. MMPI profile characteristics of women with varying levels of normal
dissociation. J Clin Psychol 2001;57:1469-1477. doi:10.1002/jclp.1110.
98 MERCKELBACH H, MURIS P, RASSIN E. Fantasy proneness and cognitive
failures as correlates of dissociative experiences. Pers Individ Dif 1999;26:961-967.
doi:10.1016/s0191-8869(98)00193-7.
99 GIESBRECHT T, MERCKELBACH H, GERAERTS E. The dissociative experiences
taxon is related to fantasy proneness. J Nerv Ment Dis 2007;195:769-772.
doi:10.1097/NMD.0b013e318142ce55.
100 WILLIAMS LM. Recall of childhood trauma: prospective study of women's
memories of child sexual abuse. J Consult Clin Psychol 1994;62:1167-1176.
doi:10.1037/0022-006X.62.6.1167.
101 HIGGINS JP, GREEN S, eds. Cochrane handbook for systematic reviews of
interventions (Vol. 4). John Wiley & Sons, Chichester, 2011.
102 MAURITZ MW, GOOSSENS PJ, DRAIJER N, VAN ACHTERBERG T. Prevalence
of interpersonal trauma exposure and trauma-related disorders in severe mental
illness. Eur J Psychotraumatol 2013;4. doi:10.3402/ejpt.v4i0.19985.
103 HARDOON S, HAYES JF, BLACKBURN R et al. Recording of severe mental
illness in United Kingdom primary care, 2000–2010. PLoS 2013;8.
doi:10.1371/journal.pone.0082365.
104 OSBORN D, LEVY G, NAZARETH I, KING M. Suicide and severe mental
illnesses. Cohort study within the UK general practice research database. Schizophr
Res 2008;99:134-138. doi:10.1016/j.schres.2007.11.025.
105 RUGGERI M, LEESE M, THORNICROFT G, BISOFFI G, TANSELLA M.
Definition and prevalence of severe and persistent mental illness. Br J Psychiatry
2000;177:149-155. doi:1192/bjp.177.2.149.
106 TRICKETT PK, NOLL JG, REIFFMAN A, PUTNAM FW. Variants of intra familial
sexual abuse experience: implications for short-and long-term development. Dev
Psychopathol 2001;13:1001–1019.
107 LAOR N, WOLMER L, KORA M, YUCEL D, SPIRMAN S, YAZGAN Y.
Posttraumatic, dissociative and grief symptoms in Turkish children exposed to
36
the 1999 earthquakes. J Nerv Ment Dis 2002;190:824-832.
doi:10.1097/01.NMD.0000041959.54021.A7.
108 NEUNER F, PFEIFFER A, SCHAUER-KAISER E, ODENWALD M, ELBERT T,
ERTL V. Haunted by ghosts: prevalence, predictors and outcomes of spirit possession
experiences among former child soldiers and war-affected civilians in Northern
Uganda. Soc Sci Med 2012;75:548-554. doi:10.1016/j.socscimed.2012.03.028.
109 PENNING SL, BHAGWANJEE A, GOVENDER K. Bullying boys: the traumatic
effects of bullying in male adolescent learners. J Child Adolesc Ment
Health 2010;22:131-143. doi:10.2989/17280583.2010.528580.
110 CUNNINGHAM T, SHANNON C, CROTHERS I et al. Enquiring about traumatic
experiences in psychosis: a comparison of case notes and self-report questionnaires.
Psychosis 2016;8:301-310. doi:10.1080/17522439.2015.1115542.
111 ROSSITER A, BYRNE F, WOTA AP et al. Childhood trauma levels in individuals
attending adult mental health services: an evaluation of clinical records and structured
measurement of childhood trauma. Child Abuse Negl 2015;44:36-45.
doi:10.1016/j.chiabu.2015.01.001.
112 READ J, FRASER A. Abuse histories of psychiatric inpatients: to ask or not to ask?
Psychiatr Serv 1998;49:355-359. doi:10.1176/ps.49.3.355.
113 READ J, DILLON J, LAMPSHIRE D. How much evidence is required for a
paradigm shift in mental health? Acta Psychiatr Scand 2014;134:287-294.
doi:10.1111/acps.12216.
114 CUNNINGHAM T, SHANNON C, CROTHERS I et al. The perspectives of people
with psychosis about participating in trauma-related research. Psychosis 2016;9:82-
85. doi:10.1080/17522439.2016.1145729.
115 NEWMAN E, KALOUPEK DG. The risks and benefits of participating in trauma-
focused research studies. J Trauma Stress 2004;17:383-394.
doi:10.1023/B:JOTS.0000048951.02568.3a.
116 GRIFFIN MG, RESICK PA, WALDROP AE, MECHANIC MB. Participation in
trauma research: is there evidence of harm? J Trauma Stress 2003;16:221-227.
doi:10.1023/A:1023735821900.
37
117 HUNTER ECM, BAKER D, PHILLIPS ML, SIERRA M, DAVID AS. Cognitive-
behaviour therapy for depersonalisation disorder: an open study. Behav Res Ther
2005;43:1121-1130. doi:10.1016/j.brat.2004.08.003.
118 FARRELLY S, PETERS E, AZIS M, DAVID A, HUNTER EC. A brief CBT
intervention for depersonalisation/derealisation in psychosis: study protocol for a
feasibility randomised controlled trial. Pilot Feasibility Stud 2016;2.
doi:10.1186/s40814-016-0086-7.
119 DE BONT PA, VAN DEN BERG DP, VAN DER VLEUGEL BM et al. A multi-site
single blind clinical study to compare the effects of prolonged exposure, eye
movement desensitization and reprocessing and waiting list on patients with a current
diagnosis of psychosis and co-morbid post-traumatic stress disorder: study protocol
for the randomized controlled trial treating trauma in psychosis. Trials 2013;14:151.
doi:10.1186/1745-6215-14-151.
120 VAN DEN BERG DP, DE BONT PA, VAN DER VLEUGEL BM et al. Prolonged
exposure vs eye movement desensitization and reprocessing vs waiting list for
posttraumatic stress disorder in patients with a psychotic disorder: a randomized
clinical trial. Jama Psychiatry 2015;72:259-267.
doi:10.1001/jamapsychiatry.2014.2637.
38
Excluded (n = 302)
Non-eligible clinical sample: 90
Non-clinical sample: 33
Non-valid/or no dissociation measure: 20
Non-valid/or no childhood trauma measure: 38
Combined adulthood and childhood trauma: 13
No analysis between trauma and dissociation: 82 a
Non-quantitative methodology: 8 b
Studies in non-eligible language: 3
Case studies: 3
Conference/dissertation abstracts: 5
Un-retrieved studies: 1
Overlapping samples: 2 c
Duplications: 4
PubMed (n = 1553)
PsycINFO (n = 2119)
EMBASE (n = 2573)
Total number of retrieved studies (n = 6245)
Stage 1: Studies following duplicate removal (n = 4859)
Stage 2: Studies following title and abstract screening
(n = 332)
Stage 3: Studies following full paper screening (n = 30)
Forward and backward screening: No studies meeting eligibility criteria d
Figure 1. Flowchart of search stratergy and studies to be included in the systematic reviewa Researchers only explored the difference in childhood trauma or dissociation scores between groups (i.e. psychosis group reported higher levels of dissociation compared to the bipolar group), they did not analyse the relationship between childhood trauma and dissociation.
b Non-quantitative methodology included reviews, qualitative and non-empirical studies.
c The same participants were recruited in more than one study.
d The forward and backward screening of all 30 eligible studies led to the retrieval of 12 studies. None of the 12 studies located met the eligibility criteria for the following reasons: non-eligible clinical sample (9 studies); no analysis between childhood trauma and dissociation (2 studies) and; non-clinical sample (1 study).
39
Table 1. Summary of eligible study characteristics
Author, date Country Design Diagnosis Sample size
n (F:M) Mean age
Measures
Dissociation TraumaQuality rating
PsychosisRoss et al. 1994 (35) Canada Between Schizophrenia 83 25 F: 56 M a 41.9 DES DDIS Moderate
Holowka et al. 2003 (36) Canada Correlation Schizophrenia 26 7 F: 19 M 30.8 DES CTQWeak
Ross and Keyes, 2004 (37) USA Between Schizophrenia 60 23 F: 37 M 40.1 DES DDISModerate
Lysaker et al. 2005 (38) USA Between Schizophrenia, schizoaffective disorder 40 40 M 47.6 d TSI
TAA (brief revised version)
Moderate
Schafer et al. 2006 (39) Germany Correlation Schizophrenia-spectrum disorders 15 15 F 34.6 b
DES (German version)
CTQWeak
Dorahy et al. 2009 (40) UK and Australia Between Schizophrenia 34 7 F: 27 M 41.6 c DES-T CTQ
Moderate
Vogel et al. 2009 (41) Germany Between Schizophrenia-spectrum disorders 80 27 F: 53 M 22.2 d
DES (German version)
CTQModerate
Sar et al. 2010 (13) Turkey Correlation Schizophrenia 70 38 F: 32 M 38.3DES (Turkish version)
CTQWeak
Laddis and Dell, 2012 (42) USA Correlation Schizophrenia 40 14 F: 26 M 42.1 MID TEQ Weak
Perona-Garcelan et al. 2012 (43) Spain Correlation Schizophrenia-spectrum disorders 71 17 F: 54 M 39.1
DES-II (Spanish version)
TQWeak
40
Schafer et al. 2012 (44) Germany Correlation Schizophrenia-spectrum disorders 93 48 F: 97 M a 34.0 b
DES (German version)
CTQWeak
Varese et al. 2012 (45) UK CorrelationSchizophrenia, schizoaffective disorder, delusional disorder
45 10 F: 35 M 44.6 d DES CATWeak
Braehler et al. 2013 (8) Canada Correlation First episode psychosis 62 16 F: 46 M 23.2 DES CTQWeak
Braehler et al. 2013 (8) Canada Correlation Schizophrenia, schizoaffective disorder 43 11 F: 32 M 31.5 DES CTQ
Weak
Kim et al. 2013 (46) South Korea Correlation Schizophrenia 100 59 F: 41 M 37.8 DES-T
CTQ-SF (Korean version)
Weak
Laferriere-Simard et al. 2014 (47) Canada Correlation
Schizophrenia, Schizophreniform disorder, schizoaffective disorder, brief psychotic disorder
50 16 F: 34 M 32.0 DES CTQ
Weak
Alvarez et al. 2015 (48) Spain Correlation Schizophrenia, schizoaffective disorder 45 20 F: 25 M 41.1
DES-II (Spanish version)
CTQ-SF (Spanish version)
Weak
Chae et al. 2015 (17) South Korea Correlation Schizophrenia 92 47 F: 51 M a 43.0 b
DES (Korean version)
CTQ-SF (Korean version)
Weak
Oh et al. 2015 (20) South Korea Correlation Schizophrenia 68 42 F: 26 M 34.5
DES (Korean version)
CTQ-SF (Korean version)
Weak
Schalinkski and Teicher, 2015 (14) Germany Correlation Schizophrenia-spectrum
disorders 75 26 F: 49 M 31.0 Shut-D LECWeak
Personality disorders
Zanarini et al. 2000 (12) USA Between Borderline personality disorder 290 233 F: 57 M 26.9 DES CEQ-R Moderate
41
Goodman et al. 2003 (18) USA Correlation Personality disorders 93 33 F: 62 M a 38.5 b DES CTQ Weak
Simeon et al. 2003 (49) USA Correlation Borderline personality disorder 20 8 F: 12 M 37.9 DES CTQ-SF Weak
Watson et al. 2006 (16) UK Correlation Borderline personality disorder 132 105 F: 34 M
a 32.6 b DES CTQ Weak
Simeon et al. 2007 (23) USA Correlation Borderline personality disorder 13 6 F: 7 M 39.2 DES CTQ-SF Weak
Loffler-Stastka et al. 2009 (50) Austria Correlation Personality disorders 52 25 F: 27 M 38.0DES (German version)
CTQ (German version)
Weak
Johnston et al. 2009 (22) UK Correlation Borderline personality disorder 30 27 F: 3 M 40.0 WDS CTQ Weak
Bipolar disorder
Eryilmaz et al. 2015 (10) Turkey Correlation Bipolar disorder-II 33 19 F: 14 M 32.4DES (Turkish version)
CTQ (Turkish version)
Weak
Hariri et al. 2015 (15) Turkey Between Bipolar disorder 200 122 F: 78 M 38.5 dDES (Turkish version)
CTQ (Turkish version)
Moderate
Yilmaz et al. 2016 (51) Turkey Correlation Bipolar disorder 70 23 F: 47 M 34.7DES (Turkish version)
CTQ (Turkish version)
Weak
Mixed sample
van der Kolk et al. 1991 (52) USA Correlation
Bipolar disorder and borderline, antisocial, schizotypal personality disorder
74 39 F: 35 M Not reported DES TAQ
Weak
a Participants did not complete all measures of either childhood trauma or dissociation. Therefore, the gender distributions displayed in the table were based on the whole
sample (i.e. including the participants whose data was later excluded as they did not complete the measures).
42
b Participants did not complete all measures of either childhood trauma or dissociation. Therefore, the mean age displayed in the table was based on the whole sample (i.e.
including the participants whose data was later excluded as they did not complete the measures).
c The mean age was not reported for the schizophrenia group, instead, the mean age for the whole sample (i.e. schizophrenia and dissociative identity disorder) was reported.
However, as they were no significant differences in age between the groups, the whole sample mean age was reported in the table.
d In the original studies the mean age was reported for separate groups (e.g. male and female, abused group vs non-abused group). Therefore, to calculate the mean age of the
sample of interest, the mean age for separate groups were combined using a formula.
Abbreviations: DDIS = Dissociative Disorders Interview Schedule.
Childhood trauma measures: Childhood Experiences Questionnaire-Revised (CEQ-R; 53), Childhood Trauma Questionnaire (CTQ; 54, 55 [English version]; 56 [Turkish
version]; 57 [German version]), Childhood Trauma Questionnaire-short form (CTQ-SF; 58; [English version]; 59 [Spanish version]; 60 [Korean version]), Dissociative
Disorders Interview Scale (DDIS; 61), Life Events Checklist (LEC; 62), The Child Abuse and Trauma Scale (CATS; 63), Trauma Assessment for Adults (TAA- brief revised
version; 64), Trauma Questionnaire (65), Traumatic Antecedent Questionnaire (TAQ; 66), Traumatic Experiences Questionnaire (TEQ; 67).
Dissociation measures: Dissociative Experience Scale (DES; 68 [English version]; 69, 70 [German version]; 71 [Korean version]; 72 [Turkish version]), Dissociative
Experience Scale-II (DES-II; 73 [English version]; 74 [Spanish version]), Dissociative Experience Scale-Taxon (DES-T; 75), Multidimensional Inventory Scale (MID; 76),
Shutdown Dissociation Scale (Shut-D; 77), Trauma Symptom Inventory Scale (TSI; 78)
43
Figure 2: Forest plot of the relationship between total childhood trauma and dissociation across SMIs.
45
Table 2: The relationship between specific childhood trauma and dissociation in the overall sample and subgroups
Childhood trauma k r (95% CI), p value Q Statistic I2 (%)
Overall sample
Sexual abuse 20 r = .26 (.17, .34), p < .001 Q = 58.41, p < .001 67.47
Physical abuse 19 r = .27 (.20, .34), p < .001 Q = 32.34, p = .021 44.34
Emotional abuse 18 r = .34 (.23, .44), p < .001 Q = 65.93, p < .001 74.21
Physical neglect 17 r = .23 (.17, .29), p < .001 Q = 21.33, p = .166 24.98
Emotional neglect 16 r = .10 (.03, .18), p = .009 Q = 22.86, p = .087 34.38
Neglect 3 r = .30 (.22, .38), p < .001 Q = 1.93, p = .382 0.00
Aggregated neglect a 36 r = .19 (.14, .24), p < .001 Q = 63.73, p = .002 44.08
Schizophrenia spectrum disorder
Sexual abuse 13 r = .30 (.18, .41), p < .001 Q = 37.60, p < .001 68.09
Physical abuse 13 r = .32 (.24, .40), p < .001 Q = 19.18, p = .084 37.45
Emotional abuse 12 r = .41 (.27, .54), p < .001 Q = 47.11, p < .001 76.65
Physical neglect 11 r = .30 (.23, .37), p < .001 Q = 9.74, p = .467 0.00
Emotional neglect 11 r = .11 (.02, .20), p = .014 Q = 12.57, p = .249 20.46
Neglect 1 r = .23 (-.07, .49), p = .129 Q = 0.00, p = 1.000 0.00
Aggregated neglect a 23 r = .22 (.15, .28), p < .001 Q = 34.66, p = .042 36.53
Personality disorder
Sexual abuse 5 r = .24 (.13, .35), p < .001 Q = 6.69, p = .153 40.20
Physical abuse 4 r = .23 (.13, .32), p < .001 Q = 0.42, p = .936 0.00
Emotional abuse 4 r = .25 (.14, .36), p < .001 Q = 2.65, p = .449 0.00
Physical neglect 4 r = .14 (.05, .24), p = .004 Q = 1.86, p = .603 0.00
Emotional neglect 3 r = .06 (-.21, .31), p = .679 Q = 7.63, p = .022 73.78
Neglect 1 r = .29 (.20, .37), p < .001 Q = 0.00, p = 1.000 0.00
Aggregated neglect a 8 r = .14 (.03, .25), p < .001 Q = 19.59, p = .007 64.26
Note. Cl = confidence intervals; a Physical neglect, emotional neglect and neglect were combined to generate the aggregated neglect effect size.
46
SUPPORTING INFORMATION
Sonya Rafiq a, Carolina Campodonico a, and Filippo Varese a,b
The relationship between childhood adversities and dissociation in severe mental illness: A meta-analytic review.
Full search stratergy: All search terms used in the database search
Table S1: Author contacts during the meta-analytic review
Table S2: Quality assessment
Table S2: Publication analyses
Table S3: Sensitivity analyses for the overall sample and subgroups
47
Search terms
Dissociation (multiple personalit*, OR dissociat*, OR absorption, OR depersonalisation, OR derealisation, OR depersonalization, OR derealization) AND
trauma (child abuse, OR maltreat*, OR victim*, OR trauma*, OR advers*, OR emotional abuse, OR physical abuse, OR psychological abuse, OR sexual
abuse, OR bullied, OR bully*, OR neglect*, OR persecution, OR mass violence, OR torture, OR combat*, OR terror*, OR war*, OR conflict*, OR refugee*)
AND severe mental health (SMI OR severe mental illness*, OR serious mental illness*, OR chronic mental illness*, OR enduring mental illness*, OR severe
and enduring mental illness*, OR persistent mental illness*, OR severe mental disorder*, OR serious mental disorder*, OR chronic mental disorder*, OR
persistent mental disorder*, OR schizophrenia, OR schizophrenic*, OR schizoaffective, OR psychoti*, OR psychosis, OR paranoi*, OR delusion*, OR
voice*, OR, hallucinat*, OR personality disorder*, OR affective illness*, OR affective disorder*, OR affective dysfunction*, OR bipolar*, OR bi polar*, OR
bi-polar*, OR mood disorder*, OR mood dysfunction*, OR manic, OR mania, mood swing*, OR hypomani*, OR mixed episode*, OR rapid cycling, OR
rapid cycle, OR mood disturbance*, OR Rcbd).
The following words were MeSH terms: schizophrenia, psychotic disorders, bipolar disorder, child abuse, physical abuse, sexual abuse, bullying, violence,
torture, absorption, depersonalization, and depersonalization. Subject headings in EMBASE and PsycINFO included: multiple personality, dissociation,
absorption, depersonalization, child abuse, victim, emotional abuse, physical abuse, sexual abuse, bullying, neglect, persecution, torture, terrorism, war,
conflict, refugee, schizophrenia, psychotic disorder, psychosis, paranoia, delusion, voice, hallucination, personality disorder, bipolar disorder, mood disorder,
mania, mood, hypomania, and rapid cycling bipolar disorder.
48
Table S1. Author contacts
Contact regarding Author responses Data retrieved
Missing data from eligible articles:
8 authors contacted:
5 responses
3 non-responders
2 data sources retrieved and included
1 author clarified data
Papers that were un-retrieved due to unavailable journal subscription
9 authors contacted:
8 responses
1 un-retrievable contact
8 papers retrieved
Unpublished data from authors of papers included in the analysis
30 authors contacted:
5 responses
25 non-responders
No data source retrieved
49
Table S2: Quality assessment – INSERT TABLE AND CHANGE OTHER TABLE NUMBERS
Table S2: Publication analyses for the overall sample and subgroups
Diagnostic group Egger’s test Trim-and-fill method Missing studies
SMI p = .012 r = .28, 95% CI (.22, .34) 7
Schizophrenia spectrum disorders p = .059 r = .35, 95% CI (.27, .42) 4
Personality disorders p = .964 r = .24, 95% CI (.20, .29) 2Note: CI = confidence intervals
50
Table S3: Sensitivity analyses conducted on the overall sample and subgroups
Sensitivity analyses r (95% CI), p value a Q Statistic, I2 (%) Change in
effect size b
1. Selective reporting: Overall analysis k = 28, r = .33, (.26, .39), p < .001 Q = 92.78, p < .001, I2 = 70.90 No
Selective reporting: Schizophrenia k = 19, r = .38, (.30, .45), p < .001 Q = 46.93, p < .001, I2 = 61.65 No
Selective reporting: Personality disorder k = 5, r = .19, (.12, .26), p < .001 Q = 2.46, p = .652, I2 = 00.00 No
2. Childhood trauma measure: Overall analysis k = 29, r = .32, (.26, .38), p < .001 Q = 94.71, p < .001, I2 = 70.44 No
Childhood trauma measure: Schizophrenia k = 19, r = .39, (.30, .46), p < .001 Q = 49.83, p < .001, I2 = 63.88 No
3. Reported total trauma: Overall analysis k = 26, r = .36, (.30, .42), p < .001 c Q = 48.86, p = .003, I2 = 48.83 No
Reported total trauma: Schizophrenia k = 18, r = .40, (.32, .47), p < .001 c Q = 35.00, p = .006, I2 = 51.42 No
Reported total trauma: Personality disorder k = 5, r = .21, (.07, .34), p = .003 c Q = 2.85, p = .584, I2 = 00.00 No
Reported total trauma: Bipolar disorder k = 2, r = .24, (.11, .35), p < .001 c Q = 0.06, p = .800, I2 = 00.00 Yes
4. CTQ measures: Overall analysis k = 22, r = .33, (.25, .41), p < .001 Q = 79.56, p < .001, I2 = 73.61 No
Non-CTQ measures: Overall analysis k = 9, r = .33, (.25, .42), p < .001 Q = 14.35, p = .073, I2 = 44.27 No
5. DES measures: Overall analysis k = 27, r = .34, (.28, .40), p < .001 Q = 91.48, p < .001, I2 = 71.58 No
Non-DES measures: Overall analysis k = 4, r = .22, (.02, .41), p < .033 Q = 5.65, p = .130, I2 = 46.85 YesNote. CI = confidence intervals. a Findings after the sensitivity analyses. b Whether the effects sizes from the analyses investigating the relationship between childhood
trauma and dissociation, altered as a result of the sensitivity analyses (e.g. medium effect size to small effect size). c When total trauma scores generated from composite
summary effect sizes generated by CMA2 (five studies) were removed from the overall and subgroup analysis, only total trauma scores reported in the studies were
51