Briere, John PhD
Although dissociation is often described as a posttraumatic response, the actual statistical association between trauma exposure and dissociative symptoms is surprisingly small. This suggests that that some dissociative phenomena may be unrelated to trauma, or may be related in more complex ways. Analysis of the normative data for the Multiscale Dissociation Inventory revealed significant dissociative symptoms in only 8% of trauma-exposed individuals from the general population. However, 90% of those with at least one clinically significant dissociation scale on the Multiscale Dissociation Inventory reported a trauma history, and significant dissociation was found in only 2% of nontraumatized individuals. A history of interpersonal violence, number of different types of trauma exposure, posttraumatic stress, and affect dysregulation were univariate predictors of dissociative symptomatology in trauma-exposed participants, but only posttraumatic stress and affect dysregulation were multivariate predictors. Trauma is probably an important, but insufficient, condition for the development of dissociative symptomatology. Additional risk factors, such as high posttraumatic stress and/or reduced affect regulation capacities, may determine whether trauma exposure results in clinically significant dissociation.
A variety of studies indicate that dissociative symptoms are associated with traumatic life experiences, including child abuse (e.g., Chu et al, 1999), combat (e.g., Branscomb, 1991), adult interpersonal violence (e.g., Feeny et al., 2000), fires (Koopman et al., 1994), and natural disaster (e.g., Cardeña and Spiegel, 1993). Dissociation appears especially more likely when such traumas are severe and/or chronic in nature, and occur in combination (Chu and Dill, 1990; Lipschitz et al., 1996), and is often associated with posttraumatic stress (Briere et al., 2005a; Feeny et al., 2000). Given these multiply replicated findings, a number of writers consider dissociation to be a posttraumatic response (e.g., Putnam, 1997; Spiegel, 1993).
Interestingly, however, the actual statistical association between dissociation and trauma exposure history is far less than might be expected. For example, van Ijzendoorn and Schuengel (1996) report that, across 26 studies, the average variance in Dissociative Experience Scale (Bernstein and Putnam, 1986) scores accounted for by participants sexual or physical abuse history was 6%. Similarly, in a large sample of clinical and nonclinical individuals, Briere et al. (2005b) found that trauma exposure accounted for an average of only 4% of the unique variance in dissociative symptoms.
Such data suggest that the relationship between trauma and dissociation may be more complex than is sometimes assumed. For example, there may be a range of variables that determines the extent to which an adverse event results in a dissociative response (Tillman et al., 1994). There also are other possible etiologies of dissociation: a smaller literature indicates that dissociative symptoms may arise from early insecure attachment relationships (e.g., Ogawa et al., 1997), emotional neglect (Simeon et al., 2003), neurobiological disturbance (e.g., Bechara et al., 1995), panic attacks (Krystal et al., 1991), and substance abuse (Krystal et al., 1996). These variables also may interact: early trauma and emotional neglect, for example, are often associated with attachment difficulties (Shapiro and Levendosky, 1999), all three of which are thought to lead to later neurobiological hypersensitivity to stress (van der Kolk, 2003) and affect regulation deficits (Pynoos et al., 1999). Some combination of these variables, in turn, may motivate the use of dissociation (and other avoidance strategies, including substance abuse) in the face of unmodulated posttraumatic distress (Allen, 2001; Briere, 2002a).
This paper reports on a further examination of the trauma-dissociation relationship, with special attention to the cross-evaluation of trauma exposure and clinically significant dissociation. It was hypothesized that trauma would be a necessary but insufficient condition for the development of dissociation, such that clinically significant dissociative symptoms would have (1) high specificity (elevated dissociation scores would be much more common among traumatized than among nontraumatized individuals) and, yet, (2) low sensitivity (a substantial number of trauma-exposed individuals would not report significant dissociative symptomatology). It was further hypothesized that trauma-exposed individuals would be more likely to experience significant dissociation if they had several risk factors previously noted in the literature: a history of interpersonal violence, multiple forms of trauma exposure, a reduced capacity to regulate negative affective states, and especially high levels of posttraumatic stress.
Data for the present paper were taken from the Multiscale Dissociation Inventory (Briere, 2002b) standardization study. The mean age of the 618 participants in this sample was 47.0 years (SD = 17.0), ranging from 18 to 91 years. Of the total sample, 328 (53%) were male. Racial composition was 494 (80%) Caucasian, 36 (6%) African American, 20 (3%) Hispanic, 16 (3%) Asian, 17 (3%) Native American, and seven (1%) other; 28 (5%) did not indicate their race.
Based upon their responses to the Detailed Assessment of Posttraumatic Stress (DAPS; Briere, 2001), 446 participants (73%) reported some form of trauma exposure in their lives. Per the DAPS, an event was considered a trauma if it met criteria A1 and A2 of DSM-IV-TR (American Psychiatric Association, 2000), involving experienced or witnessed threats to bodily integrity (or any developmentally inappropriate childhood sexual experience) and peritraumatic distress involving significant helplessness, terror, or horror. Of the traumatic experiences described as bothering participants the most at the time of data collection, the most common were motor vehicle accidents (N = 69, 15.5%), witnessing someone be seriously hurt or killed (N = 67, 15.0%), childhood sexual abuse (N = 34, 7.6%), and some form of physical assault (N = 33, 7.4%). An additional 108 participants (24.2%) reported one or more traumas but did not specify which was currently most upsetting.
The Multiscale Dissociation Inventory (MDI) is a standardized self-report test of dissociative symptomatology, with scales measuring six types of dissociative responses: disengagement, depersonalization, derealization, emotional constriction, memory disturbance, and identity dissociation. MDI scales can be converted to T-scores, thereby allowing for the determination of clinically significant (versus normal) levels of each type of dissociation (Briere, 2002b).
The DAPS has two validity scales and 10 scales that evaluate lifetime exposure to traumatic events, immediate responses to a specified trauma, symptoms of posttraumatic stress (reliving, avoidance, and hyperarousal), and associated features of PTSD (e.g., suicidality, substance abuse). The DAPS yields a total posttraumatic stress score (posttraumatic stress-total), and diagnoses of PTSD and ASD.
The Inventory of Altered Self Capacities (IASC; Briere, 2000) is a standardized test of difficulties in the areas of relatedness, identity, and affect regulation. The scales of the IASC assess the following domains: interpersonal conflicts, idealization-disillusionment, abandonment concerns, identity impairment, susceptibility to influence, affect dysregulation, and tension reduction activities. Scores on the IASC predict adult attachment style, childhood trauma history, interpersonal problems, suicidality, and substance abuse in various samples (Briere, 2000).
After approval by the institutional review board of the University of Central Florida, a national sampling service was used to generate a random sample of registered owners of automobiles and/or individuals with listed telephones in the US general population. Participants were mailed a questionnaire containing the MDI and several other measures, including the DAPS and IASC. Participants received $5.00 upon mailing back the questionnaire. In addition, 70 university students were tested with the same protocol (but without financial compensation) to provide additional participants in the lower age ranges. All questionnaires were anonymous, although financial compensation in the general population sample was tied to names and addresses that were destroyed before data analysis. Overall, 623 of 5485 potential participants (11.4%) returned the MDI, of which 618 were substantially complete. Although the actual normative sample used for the MDI was limited to participants who reported exposure to at least one trauma, the present analyses used the entire sample.
Two analyses were conducted in this study. In the first, participants scores on each of the scales of the MDI were dichotomized into clinically significant (a T-score that met or exceeded the clinical cutoff for that scale in the MDI manual) versus normal (a T-score lower than the clinical cutoff), and then were cross-tabulated with trauma exposure history (coded as present  or absent ). The resultant set of 2 × 2 matrices allowed determination of four types of criterion-related predictive validity (Weathers et al., 1997): (1) sensitivity (the probability that a MDI scale is clinically significant, given that trauma is present), (2) specificity (the probability that a MDI scale is not clinically significant, given that trauma is not present), (3) positive predictive power (the probability that trauma is present, given that a MDI scale is clinically significant), and (4) negative predictive power (the probability that trauma is not present, given that a MDI scale is not clinically significant).
The second analysis examined the extent to which additional variables moderated the relationship between trauma exposure and clinically significant dissociation. Logistic regression analysis was performed to predict the presence of at least one clinically elevated MDI scale, using the following variables: age, sex, history of exposure to at least one interpersonal trauma (e.g., rape, child abuse), history of exposure to at least one noninterpersonal trauma (e.g., disaster, motor vehicle accident), total number of different types of trauma experienced (the relative trauma exposure scale of the DAPS, consisting of the number of trauma exposure items endorsed positively in that measure), the affect regulation scale of the IASC, and the posttraumatic stress-total scale of the DAPS. Also tested was the statistical interaction between posttraumatic stress and affect dysregulation on dissociation status.
As shown in Table 1, dissociative symptoms were highly specific for trauma exposure and had substantial positive predictive power: clinically significant levels of dissociation were found in 0% to 2% of nontraumatized individuals (depending on the type of dissociation), whereas the presence of one or more clinically significant MDI scales predicted a trauma history in 90% of all cases. As predicted, however, specificity and positive predictive power did not translate into sensitivity or negative predictive power: traumatized individuals were not especially likely to report dissociative symptoms, and clinically nonsignificant MDI scores did not indicate the absence of a trauma history. This relationship was statistically significant for three of six MDI scales (derealization, memory disturbance, and emotional constriction), and in the expected direction for the remaining three. Statistical significance was also found for the overall presence of at least one MDI scale in the clinical range.
|TABLE 1. Cross-Tabulations of MDI Scales and History of Trauma Exposure -- PLEASE SEE ORIGINAL ARTICLE FOR TABLES|
As indicated in Table 2, prior to entry into the logistic equation (i.e., at step 0), posttraumatic stress, affect dysregulation, exposure to interpersonal trauma, and number of trauma exposure types were significant univariate predictors of at least one MDI scale elevation among trauma-exposed individuals. When all predictors were entered into the logistic regression equation at step 1, posttraumatic stress and affect dysregulation remained the only significant variables, [chi]2 (7) = 71.67, p < 0.001.1 The interaction between posttraumatic stress and affect dysregulation was not significant when entered at step 2, [chi]2 (1) = 1.15, NS.2
|TABLE 2. Logistic Prediction of at Least One Elevated MDI Scale Among Those With a History of Trauma Exposure --PLEASE SEE ORIGINAL ARTICLE FOR TABLES|
As hypothesized, although most individuals with significant dissociative symptoms reported a trauma history in the current study, the majority of trauma-exposed individuals did not experience significant dissociative symptomatology. These data suggest that the small relationship between trauma and dissociation may not be due to the prevalence of nontraumatic etiologies in the development of dissociation, but rather to the complexity of the trauma-dissociation relationship. Specifically, dissociative symptoms appear to occur most commonly in a subset of trauma survivors who have additional risk factors for dissociation, much in the same way that PTSD occurs only in a minority of trauma-exposed individuals (Yehuda and McFarlane, 1995).
Although the current data suggest that dissociative symptomatology is related to trauma exposure, the cross-sectional nature of this research precludes definitive causal conclusions. It is possible that other phenomena associated with a trauma history mediate or moderate its relationship to dissociative symptoms. For example, trauma is known to produce effects in addition to posttraumatic stress, including panic and other anxiety symptoms, that in turn have been implicated in dissociative responses. Also, as noted, early parent-child attachment disturbance and/or parental emotional abuse or neglect appear to increase the likelihood that an individual will be traumatized later in life, and both are associated with dissociative symptoms. A reverse relationship is also possible; dissociative symptoms may produce trauma: by virtue of decreased alertness to danger, for example, individuals with dissociative symptoms might be more vulnerable to interpersonal violence. Each of these scenarios might cause trauma to appear more causally related to dissociation than actually might be the case. The literature to date suggests, however, that dissociation is not a significant predictor of subsequent victimization (e.g., Sandberg et al., 1999), and reveals only moderate relationships between attachment disturbance, early emotional abuse or neglect, and either later trauma exposure or dissociative symptomatologyseemingly less than what would be required to create a 90% likelihood of trauma in those with dissociative symptoms. As well, from a broader perspective, emotional abuse or neglect that is of sufficient magnitude to disrupt parent-child attachment may be considered traumatic itself (Allen, 2001), even if not formally deemed so by DSM-IV.
If trauma is, in fact, integrally related to clinically significant dissociative symptoms, but only under certain conditions, it would be important to identify those factors that moderate the trauma-dissociation relationship. In the current study, multivariate analysis revealed that two variablessymptoms of posttraumatic stress and impaired capacity to regulate negative emotional statessubstantially increased the likelihood that a trauma-exposed individual would experience dissociative symptoms. Interestingly, however, no statistical interaction between posttraumatic stress and affect dysregulation was found in the present study, even though dissociation would appear to be most necessary for those with a combination of low affect regulation and high traumatic stress. Instead, these variables appear to be independent, additive moderators of trauma-related dissociation.
The role of traumatic stress in dissociation is well-described by the literature. The independent contribution of affect dysregulation, however, is a newer finding. It supports the notion that dissociation is a compensatory response to extreme affective experience, not only in individuals undergoing high levels of posttraumatic stress, but also for trauma-exposed individuals with low affect regulation capacities, above and beyond their level of trauma-related symptomatology. In either case, dissociative symptoms may attenuate the subjective experience of overwhelming emotional distress, and thus may provide some level of emotional equilibrium. This remains speculation, however; further research is clearly needed to explicate the role of affect regulation capacity in the relationship between trauma and dissociation. In fact, affect regulation itself is likely to reflect a variety of underlying phenomena, including the availability of emotional regulation skills (e.g., Wagner and Linehan, 1998) and neurobiological variables such as hypothalamic-pituitary-adrenal axis stability and level of adrenergic activation and/or sensitivity (Krystal et al., 1996; Yehuda, 1992). In light of this complexity, the finding of affect regulation difficulties among those with significant dissociative symptoms may require more in-depth exploration.
The present findings should be viewed in light of methodological limitations associated with the use of the MDI normative data. First, this study involved nonclinical participants, and thus may not be completely generalizable to clinically presenting individuals. Although the trauma exposure rate was relatively high in this sample, as has been shown in other general population studies (e.g., Elliott, 1997), it may be loweror the average trauma exposure may be less severethan what is typically present in clinical groups (Briere, 2004). Further, dissociation is known to be more prevalent in clinical than nonclinical samples (Spiegel, 1993). As a result, the relationship between trauma exposure and dissociative symptoms found in the current study (as well as other nonclinical investigations) may be less than what would be demonstrated in a more clinically acute group of individuals. It may be argued, however, that data from general population participants are preferable for studies in this area, since status as a clinical participant may involve a range of extraneous and relatively uncontrolled variables, some of which may be related to dissociation (and trauma) in unmonitored ways. A second potential limitation is the relatively low return rate for this study. Because only a subset of potential participants volunteered for the MDI normative project, the current findings may not be entirely representative of the general population, although the final MDI normative sample was statistically similar to relevant United States census data (Briere, 2002b).
These results indicate that clinically significant dissociation is relatively uncommon in the general population. Yet, when it occurs, such symptomatology is a substantial marker for trauma exposure. The relationship between trauma and significant dissociative symptomatology is probably moderated by a number of other variables, including, in the current study, level of posttraumatic stress and existing affect regulation capacities. Such findings may have clinical implications. As noted by Wagner and Linehan (1998) and others, for example, the treatment of dissociative symptomatology may be most effective when it supports the cognitive/emotional processing of trauma-related material and encourages the development of greater affect regulation capacities.
The author would like to thank Psychological Assessment Resources for access to the Multiscale Dissociation Inventory standardization study dataset.
American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders (4th ed, Text Revision). Washington DC: American Psychiatric Association. [Context Link]
Allen JG (2001) Traumatic Relationships and Serious Mental Disorders. New York: Wiley. [Context Link]
Banscomb LB (1991) Dissociation in combat-related post-traumatic stress disorder. Dissociation. 4:1320. [Context Link]
Bechara A, Tranel D, Damasio H, Adolphs R, Rockland C, Damasio AR (1995) Double dissociation of conditioning and declarative knowledge relative to the amygdala and hippocampus in humans. Science. 269:11151118. [Context Link]
Bernstein EM, Putnam FW (1986) Development, reliability and validity of a dissociation scale. J Nerv Ment Dis. 174:727734. [Context Link]
Briere J, Scott C, Weathers FW (2005a) Peritraumatic and persistent dissociation in the presumed etiology of PTSD. Am J Psychiatry. 162:22952301. [Context Link]
Briere J, Weathers, FW, Runtz M (2005b) Is dissociation a multidimensional construct? Data from the Multiscale Dissociation Inventory. J Trauma Stress. 18:221231. [Context Link]
Briere J (2004) Psychological Assessment of Adult Posttraumatic States: Phenomenology, Diagnosis and Measurement (2nd ed). Washington DC: American Psychological Association. [Context Link]
Briere J (2002a) Treating adult survivors of severe childhood abuse and neglect: Further development of an integrative model. In JEB Myers, LBerliner, J Briere, CT Hendrix, T Reid, C Jenny (Eds), The APSAC Handbook on Child Maltreatment (2nd ed, pp 175202). Newbury Park (CA): Sage Publications. [Context Link]
Briere J (2002b) Multiscale Dissociation Inventory. Odessa (FL): Psychological Assessment Resources. [Context Link]
Briere J (2001) Detailed Assessment of Posttraumatic Stress (DAPS). Odessa (FL): Psychological Assessment Resources. [Context Link]
Briere J (2000) Inventory of Altered Self Capacities (IASC). Odessa (FL): Psychological Assessment Resources. [Context Link]
Cardeña E, Speigel D (1993) Dissociative reactions to the San Francisco Bay Area Earthquake of 1989. Am J Psychiatry. 150:474478. [Context Link]
Chu JA, Dill DL (1990) Dissociative symptoms in relation to childhood physical and sexual abuse. Am J Psychiatry. 147:887892. [Context Link]
Chu JA, Frey LM, Ganzel BL, Matthews JA (1999) Memories of childhood abuse: Dissociation, amnesia and corroboration. Am J Psychiatry. 156:749755. [Context Link]
Elliott DM (1997) Traumatic events: Prevalence and delayed recall in the general population. J Consult Clin Psychol. 65:811820. [Context Link]
Feeny NC, Zoellner LA, Fitzgibbons LA, Foa EB (2000) Exploring the roles of emotional numbing, depression and dissociation in PTSD. J Trauma Stress. 13:489498. [Context Link]
Koopman C, Classen C, Spiegel D (1994) Predictors of posttraumatic stress symptoms among survivors of the Oakland/Berkeley, California, firestorm. Am J Psychiatry. 151:888894. [Context Link]
Krystal JH, Bennett A, Bremner D, Southwick SM, Charney DS (1996). Recent developments in the neurobiology of dissociation. In LK Michelson, WJ Ray (Eds), Handbook of Dissociation: Theoretical, Empirical and Clinical Perspectives (pp 163190). New York: Plenum. [Context Link]
Lipschitz DS, Kaplan ML, Sorkenn J, Chorney P, Asnis GM (1996) Childhood abuse, adult assault and dissociation. Compr Psychiatry. 37:261266. [Context Link]
Ogawa JR, Sroufe LA, Weinfield NS, Carlson EA, Egeland B (1997) Development and the fragmented self: Longitudinal study of dissociative symptomatology in a nonclinical sample. Devel Psychopathol. 9:855879. [Context Link]
Putnam FW (1997) Dissociation in Children and Adolescents: A Developmental Perspective. New York: Guilford Press. [Context Link]
Pynoos RS, Steinberg AM, Piacentini JC (1999) A developmental psychopathology model of childhood traumatic stress and intersection with anxiety disorders. Biol Psychiatry. 46:15421554. [Context Link]
Sandberg DA, Matorin AI, Lynn SJ (1999) Dissociation, posttraumatic symptomatology and sexual revictimization: A prospective examination of mediator and moderator effects. J Trauma Stress. 12:127138. [Context Link]
Shapiro DL, Levendosky AA (1999) Adolescent survivors of childhood sexual abuse: The mediating role of attachment style and coping in psychological and interpersonal functioning. Child Abuse Neglect. 23:1175 1191. [Context Link]
Simeon D, Nelson D, Elias R, Greenberg J, Hollander E (2003) Relationship of personality to dissociation and childhood trauma in borderline personality disorder. CNS Spectr. 8:755762. [Context Link]
Spiegel D (Ed) (1993) Dissociative Disorders: A Clinical Review. Lutherville (MD): Sidran. [Context Link]
Tillman JG, Nash MR, Lerner PM (1994) Does trauma cause dissociative pathology? In SJ Lynn, JW Rhue (Eds), Dissociation: Clinical and Theoretical Perspectives (pp 395414). New York: Guilford Press. [Context Link]
van der Kolk BA (2003) The neurobiology of childhood trauma and abuse. Child Adolesc Psychiatr Clin North Am. 12:293317. [Context Link]
Van Ijzendoorn M, Schuengel C (1996) The measurement of dissociation in normal and clinical populations: Meta analytic validation of the Dissociative Experiences Scale (DES). Clin Psychol Rev. 16:382. [Context Link]
Wagner AW, Linehan MM (1998) Dissociative behavior. In VM Follette, JI Ruzek, FR Abueg (Eds), Cognitive-Behavioral Therapies for Trauma (pp 191225). New York: Guilford. [Context Link]
Weathers FW, Keane TM, King LA, King DW (1997) Psychometric theory in the development of posttraumatic stress disorder assessment tools. In JP Wilson, TM Keane (Eds), Assessing Psychological Trauma and PTSD. New York: Guilford. [Context Link]
Yehuda R (1992) Neuroendocrinology of post-traumatic stress disorder. Curr Opin Psychiatry. 5:109112. [Context Link]
Yehuda R, McFarlane A (1995) Conflict between current knowledge about posttraumatic stress disorder and its original conceptual basis. Am J Psychiatry. 152:17051713. [Context Link]
1Because the MDI emotional constriction scale shares some similarity with the numbing symptoms of PTSD, the inclusion of this scale in the at least one MDI scale elevation variable may have increased its overall correlation with posttraumatic stress. However, logistic regression using a variable that indexed elevations on any MDI scale other than emotional constriction resulted in equivalent findings. [Context Link]
2In order to evaluate any effects of dichotomizing MDI scores, multiple regression analyses were also performed on the full (continuously measured) MDI scales. The same results were found for all individual scales: total posttraumatic stress and affect dysregulation were the only significant multivariate predictors, and there was no interaction between posttraumatic stress and affect dysregulation in any instance. R2 values associated with these analyses ranged from .27 (for identity dissociation) to .41 (for emotional constriction). [Context Link]
Key Words: Dissociation; MDI; assessment; trauma; regulation