Using a structured interview to obtain a lifetime history of criminal victimization, a community sample of 266 adult women who had experienced at least one incident of victimization was identified. These women were administered the Symptom Checklist-90-Revised, the Impact of Event Scale, and a structured clinical interview was used to identify Crime-Related Post-Traumatic Stress Disorder (CR-PTSD). A recently derived scale based on responses to items on the SCL-90-R was compared to the IES for predicting current diagnosis of Crime-Related PostTraumatic Disorder. Both the scale and the IES were found to improve prediction of CR-PTSD above base rates and to perform in a similar manner. The utility of each of these scales as a screening measure is discussed.
Post-Traumatic Stress Disorder (PTSD) is described in DSM-III-R (American Psychiatric Association, 1987) as a disorder where, following an event outside the range of usual human experiences, a person re-experiences the traumatic event (e.g., intrusive thoughts, nightmares, flashbacks) and has symptoms of emotional numbing/avoidance (e.g., avoidance of reminders of the trauma, diminished interest in activities) and increased physiological arousal (e.g., sleep difficulty, hypervigilance). While PTSD has traditionally been studied among Vietnam veterans (Fairbank, Keane, & Malloy, 1983; Foy, Sipprelle, Rueger, & Carroll, 1984; Keane, Malloy, & Fairbank, 1984), recent studies have found that many victims of violent crime also develop symptoms of PTSD following their victimization (Burgess & Holmstrom, 1985; Holmes & St. Lawrence, 1983; Steketee & Foa, 1987; Kilpatrick, Saunders, Veronen, Best, & Von, 1987; Kilpatrick et al., 1989). For example, victims of rape often experience fear and anxiety, intrusive imagery and nightmares, sleep disturbance, guilt, and impairment in social and sexual functioning following the assault (Steketee & Foa, 1987).
While the literature on sexual assault suggests that this symptom pattern is characteristic of rape victims, there has been little empirical work that has employed DSM-III-R diagnostic criteria in systematically assessing the incidence and prevalence of PTSD among victims of sexual assault and other violent crimes. The few studies which have assessed PTSD among these victims however, have identified high rates of PTSD with a substantial portion of rape victims describing symptoms of PTSD both immediately following the crime and during the weeks and months which follow the incident (Burge, 1988; Kilpatrick et al., 1987; Kramer & Green, 1989; Rothbaum & Foa, 1988). In studies of women who were sexually abused as children, it has been found that nearly all the women currently in treatment met DSM-III criteria for PTSD (Donaldson & Gardner, 1985; Lindberg & Distad, 1985).
Clearly, PTSD is a significant mental health problem among crime victims, especially victims of sexual assault. One of the major limitations in the study of CR-PTSD among victims of sexual assault and other crimes has been the paucity of adequate assessment measures. The majority of the studies cited above were able to obtain estimates of the prevalence of PTSD through the careful use of structured and semi-structured interviews which specifically addressed the symptoms required to meet diagnostic criteria for PTSD. While structured clinical interviews are recommended for diagnosing PTSD (Keane, Wolfe, & Taylor, 1987), interviews can be unreliable or invalid depending on the screening questions and interview techniques used, as well as the training and sensitivity of the interviewer (Best, Kilpatrick, Kramer, & McNeil-Harkins, in press). Best et al. (in press) suggested that low prevalence rates can be due, in part, to inadequate screening and interviewing techniques which may have failed to identify many crime victims. Saunders, Kirkpatrick, Resnick, & Tidwell (1989) found that most victims of serious violent crimes were not detected at intake interviews using normal methods. Consequently, there is a need for psychometrically sound measures of PTSD to establish concurrent validity of diagnoses based on interviews and to enable researchers and clinicians to efficiently screen for a history of victimization and the presence of PTSD.
Currently, the Impact of Event Scale (IBS; Horowitz, Wilner, & Alvarez, 1979) is one of the most frequently used self-report measures of PTSD symptoms (Burge, 1988; Kilpatrick et al., 1985; Rothbaum & Foa, 1988). Intrusion and avoidance subscales of the IES provide indices of some symptoms consistent with DSM-III-R criteria for PTSD (i.e., re-experiencing symptoms and symptoms of numbing/avoidance). However, a number of components of PTSD are not included in the IES, and the validity of the IES in detecting PTSD is only moderate (Kramer & Green, 1989). Also, the IES requires a prior knowledge that a respondent has experienced a traumatic event before it can be used.
In a previous report of different aspects of this study, Saunders, Arata, and Kilpatrick (1990) developed a scale or factor within the SCL-90-R for use as a screening and assessment measure for symptoms of CR-PTSD among women. This SCL-PTSD scale was found to improve significantly upon base rates in classifying respondents as CR-PTSD positive and CR-PTSD negative. Importantly, respondents do not need to identify a victimization experience to complete this measure. Thus, it can be used with general populations. The purpose of the analyses contained in this report was to compare the effectiveness of the SCL-PTSD scale to that of the IES in detecting symptoms of CR-PTSD. Performance of the new scale similar to that of the IES will provide evidence of concurrent validity for the SCL-PTSD scale.
Subjects in this study were 266 adult (age 18 years or older) female residents of Charleston County, South Carolina. These respondents were recruited from a representative probability sample of 2,004 adult women living in Charleston County who had been selected for a different study using random-digit dialing methods. A thorough description of the sampling techniques used in the parent study is presented in Kilpatrick et al. (1985) and Saunders et al. (1990). The final sample included women with a history of criminal victimization and complete data on the SCL-90-R and the IES.
Of the final sample, 24.1% of the women were between the ages of 18 and 29 years; 48.5% were between 30 and 49 years; and 27.4% were age 50 years or older. Their average age was 40.6 years (SD = 13.5). Most of the women (73.7%) were non-Hispanic whites, and the remaining women (26.3%) were black. Married women composed 63.2% of the sample, 20.3% were divorced, separated, or widowed at the time of the assessment; and 16.5% had never been married. The sample was relatively well educated, with only 14.3 % not being high school graduates, and 29.7% being college graduates. Most of the women (60.8%) were employed or students at the time of the assessment; 23.7 % were housewives; and 9.1% were retired or disabled. Only 6.4% of the sample was unemployed. Approximately one-quarter of the sample (26.3 %) had household incomes below $15,000; 43.7 % were between $15,000 and $35,000; and 26.3% were above $35,000.
Self-Report Instruments. The Symptom Checklist-90-Revised (SCL-90-R; Derogatis, 1977) is a 90-item self-report symptom inventory designed to assess current psychological symptom patterns. Each item is a description of a psychological symptom and is rated by respondents on a 5-point scale as having caused them no discomfort (0) to extreme discomfort (4) during the past week. Scores can be derived for nine symptom subscales: (1) somatization, (2) obsessive-compulsive, (3) interpersonal sensitivity, (4) depression, (5) anxiety, (6) hostility, (7) phobic anxiety, (8) paranoid ideation, and (9) psychoticism. Additionally, three indices of general distress are produced: (1) Global Severity Index, (2) Positive Symptom Distress Index, and (3) Positive Symptom Total. Derogatis (1977) reported adequate test-retest and internal consistency reliability, and concurrent and discriminant validity for the SCL-90-R.
Based on responses to the SCL-90-R, a PTSD scale was derived (Saunders et al., 1990) to use as an experimental screening measure for CR-PTSD. This scale consisted of 28 items which discriminated CR-PTSD negative and CR-PTSD positive respondents among the 355 subjects (noncrime victims were included). The SCL-PTSD scale had a high degree of internal consistency with a coefficient alpha of .93. In order to obtain a scale score, the item responses were summed and divided by 28 to obtain a mean item score for the 28 items. This is the scoring procedure used with standard SCL-90-R scale scores (Derogatis, 1977). The SCL-PTSD scale score could range from 0 to 4. The SCL-PTSD scale score was able to successfully discriminate between CR-PTSD negative and CR-PTSD positive subjects. Furthermore, a discriminant analysis indicated that the CR-PTSD scale score correctly classified 89.3% of the respondents.
The Impact of Event Scale (IES; Horowitz et al., 1979) is a 15-item self-report inventory designed to measure the extent to which a given stressful life event produces subjective distress. Items are designed to yield subscores for intrusive and avoidance symptoms. Respondents are requested to indicate on a 4-point scale the frequency with which each of 15 statements was true for them during the past seven days. Each statement is linked to the specified stressful life event. Scores are obtained for the avoidance, intrusion, and total distress scales. The scale has good split-half reliability (r = .86) and high test-retest reliability (r = .87) (Horowitz et al., 1979). Respondents completed the scale for their worst sexual assault experience, or for their worst victimization experience if it was not a sexual assault.
Structured Interviews. The Incident Report Interview (1RI) was used to identify incidents of criminal victimization. The 1RI is a highly structured, behaviorally specific interview schedule developed specifically for the larger project of which this study was a part (Kilpatrick et al., 1987). It was designed to screen for incidents of violent criminal victimization (sexual assault, aggravated assault, robbery, and burglary) that had occurred any time in a respondent's life (including childhood), whether or not such incidents had been reported to authorities, and even if they had not been labeled as crimes by the victims. The 1RI was also used to collect comprehensive information about up to three victimization experiences, as well as respondent demographic data.
The Diagnostic Interview Schedule (DIS; Robins, Helzer, Croughan, & Ratcliff, 1981) is a structured interview schedule designed to assess particular signs and symptoms of different types of mental disorders. A modified version of the DIS was used in this study to determine whether respondents currently met or had ever met over their lifetime, DSMIII (American Psychiatric Association, 1980) diagnostic criteria for several mental disorders. At the time of the development of the project, the available form of the DIS did not include adequate procedures for assessing Post-Traumatic Stress Disorder. A set of assessment questions was developed and added to the interview schedule to determine if respondents met the DSM-III criteria for PTSD, utilizing their criminal victimization as the stressor event. The questions developed are similar in content and form to those used to assess PTSD in the Structured Clinical Interview for DSM-III-R (Spitzer, Williams, & Gibbon, 1986).
Potential respondents were contacted by phone by female research assistants and asked to participate in the original study. If they agreed, an appointment was made to administer the assessment procedures at the project office. First, the respondents were administered the SCL90-R, in addition to other self-report instruments used in the parent study. Next, the 1RI was administered to respondents on an individual basis by female research assistants trained in issues of victimization and the use of the instrument. The IES was completed based on victimization experiences identified in the 1RI. Finally, the DIS was administered by doctoral-level clinical psychologists, clinical social workers, or clinical psychology interns trained in its use. During administration of the DIS, these interviewers were unaware of the victimization history of respondents until the last portion of the interview dealing with CR-PTSD. Since the CR-PTSD diagnostic criteria were specifically linked to previously identified assault incidents, it was necessary for interviewers to know the respondent's victimization status for this portion of the DIS interview. Non victims of crime were not administered the IES nor were they assessed for CR-PTSD since by the definition used in the study, they could not have it.
The 266 women had all been victims of at least one violent crime in their lifetimes. Of these women, 71.0% had been victims of at least one sexual assault, 12.6% of aggravated assault, 7.2% of robbery, and 61.7% of burglary. Furthermore, of these crime victims, 7.5% (n = 20) met DSM-I11 diagnostic criteria for CR-PTSD at the time of the interview and were labeled as CR-PTSD positive, while 82.5% (n = 246) were CR-PTSD negative.
Zero-order correlation coefficients indicated that the SCL-PTSD scale score was moderately associated with the IES intrusion score (r = .46, p < .001), avoidance score (r = 3S,p < .001), and total score (r = .44,p < .001). These results suggest that the SCL-PTSD share a substantial proportion of variation but may have important unique qualities.
Results of univariate analyses of variance reported in Table 1 revealed that the positive and negative CR-PTSD diagnostic groups differed significantly on their mean IES and SCL-PTSD scale scores. The degree of separation of the group mean scores was similar and substantial for both scales, approximately two of the CR-PTSD negative group's standard deviations for each factor. The eta coefficients for the SCL-PTSD scale and the IES total score were compared using t-tests with appropriate adjustments for dependent samples (Hinkle, Wiersma, & Jurs, 1979) and were not significantly different. These results suggest that the size of the respective relationships between the IES and SCL-PTSD scales and CRPTSD diagnosis were similar, and support the concurrent validity of the SCL-PTSD scale.
The moderate association between the IES and the SCL-PTSD scale and the lack of difference in the size of their respective relationships with CR-PTSD diagnosis suggested that they may measure somewhat different aspects of the same construct. In order to test if the SCL-PTSD scale possessed a unique ability to discriminate between the CR-PTSD groups over and above its shared or concurrent variation with the IES, a hierarchical discriminant function analysis was performed. These results are reported in Table 2.
The IES total score was entered in Step 1 and, as expected, resulted in a significant canonical variate being constructed with a squared canonical correlation coefficient of .227. The SCL-PTSD score was entered in Step 2 and significantly increased the squared canonical correlation coefficient by. 102, indicating that the SCL-PTSD made a significant unique contribution to discriminating between the CR-PTSD positive and negative groups. Inspection of the respective standardized canonical function coefficients and the canonical structure coefficients for the final model indicates that for this sample the SCL-PTSD made a slightly larger contribution to the discriminatory power of the canonical variate than did the IES. These findings imply thatthe unique contribution of the SCL-PTSD scale is larger than that of the IES, and that the SCL-PTSD scale alone may better discriminate between the CR-PTSD groups than does the IES alone.
Two follow-up discriminant function analyses using each scale as individual predictors were performed to examine this question, and the classification results for both analyses are presented in Table 3. Of the 20 CR-PTSD positive subjects, the IES score was able to correctly identify 17 for a Sensitivity rate of 85.0% compared to 15 (75.0%) for the SCLPTSD scale. The ffiS correctly classified 207 of the 246 CR-PTSD negative respondents for a Specificity rate of 84.1 % while the SCL-PTSD scale correctly placed 223 (90.7%). Overall, the IES correctly classified 84.2% of the subjects versus 89.5% by the SCL-PTSD. There were no significant differences between the SCL-PTSD scale and the IES on their Sensitivity or overall classification rates. However, the Specificity of the SCL-PTSD scale was significantly larger than that of the IES (McNemar Test, £2(1) = 5.11 ,p < .05, n = 246). This difference is accounted for by the tendency of the IES to classify a larger proportion of the respondents as CR-PTSD positive (21.1% vs. 14.3% for the SCL-PTSD), resulting in increased Sensitivity but decreased Specificity. The two scales agreed on the classification of 216 (81.2%) of the subjects, with most of the disagreements (68.0%) occurring when the IES classified subjects as CR-PTSD positive while the SCL-PTSD scale labeled them as negative.
Despite the fact that in this sample the SCL-PTSD scale classified a substantially smaller percentage of subjects as CR-PTSD positive, it was somewhat more efficient in its positive classification as indicated by its Positive Predictive Power. Positive Predictive Power is the percentage of respondents predicted by the SCL-PTSD scale to be CR-PTSD positive (n = 38) and who actually were (39.5%). The Positive Predictive Power of the IES was 30.4%. The Negative Predictive Power, the percentage of respondents predicted to be CR-PTSD negative and who actually were, was 93.4% for the SCL-PTSD scale and93.4% for the IES. Considering the low base rate of CR-PTSD (7.5% of the sample), the apparently low Positive and high Negative Predictive Power percentages are actually substantial improvements upon this base rate. Overall, the SCL-PTSD scale appears to be more conservative than the IES in classifying respondents as CR-PTSD positive, but more efficient in its positive predictions resulting in a larger Specificity rate.
The CR-PTSD scale on the SCL-90-R has recently been proposed as a screening device for detecting CR-PTSD, especially among large groups of female clients. As a new scale, there is a need for evidence of discriminant and concurrent validity, as well as a need for replication. The current study has demonstrated evidence of concurrent validity for the CRPTSD scale with the IES, a commonly used measure of symptoms of PTSD. The results suggested that the IES and the CR-PTSD did not differ significantly in their ability to correctly identify persons with symptoms of CR-PTSD. Furthermore, the IES misclassified more respondents as CR-PTSD positive than the CR-PTSD scale. Overall, though, the CRPTSD scale and the IES did not differ in the percentage of persons correctly classified.
These findings have several implications for clinicians and researchers wishing to identify clients with CR-PTSD. The CR-PTSD scale derived from the SCL-90-R was found to be as effective as the IES in correctly classifying individuals as CR-PTSD positive and somewhat better at identifying persons as CR-PTSD negative. The CR-PTSD scale can be administered without knowledge of victimization history and is as effective in predicting CR-PTSD when victimization status is not known (Saunders et al., 1990). The IES, on the other hand, requires the respondent to answer questions in reference to a particular victimization experience. While victimization status was known in this sample, in many clinical and research settings, victimization history is often not known, and the individual may not recognize the connection between previous victimization and current symptoms. Thus, the use of the IES requires a thorough individual assessment of prior victimization experiences. Furthermore, the IES requires the individual to focus more on the traumatic experience which may be upsetting and could be considered iatrogenically biasing. In addition, for respondents who have experienced more than one incident of victimization it may be necessary to complete the IES based on each of these experiences. Because the CRPTSD scale can be administered without knowledge of victimization status, the results from this scale can be used to alert the clinician to the need of a more thorough assessment of victimization history to determine the origin of the CR-PTSD symptoms. Thus, while there was little difference in the classification rates between the IES and the CR-PTSD scale, the CR-PTSD scale can be a more efficient screening device in general populations, is less intrusive to the respondent, and is as effective as the IES in samples of known victims.
There are several limitations of this study. First, the results were obtained using a subsample of the sample used to derive the CR-PTSD scale. Further replication is needed to determine the scale's generalizability to other samples. Also, the results are based on a community sample and may not be generalizable to clinical samples. Research comparing the scale in clinical and community samples is necessary. Finally, when data for this study were collected, DSM-III-R had not yet been published, so the results are based on DSMIII criteria for PTSD. Future research using current diagnostic criteria for PTSD is necessary to determine the CR-PTSD scale's usefulness in detecting CR-PTSD based on DSM-III-R.
American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3rd Edition). Washington, DC: American Psychiatric Association.
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd Edition revised). Washington, DC: American Psychiatric Association.
Best, C. L., Kilpatrick, D. G., Kramer, T. L., & McNeill-Harkins, K. (in press). Methodological weaknesses in sexual assault research. Journal of Interpersonal Violence.
Burge, S. K. (1988). Post-traumatic stress disorder in victims of rape. Journal of Traumatic Stress, 1, 193-210.
Burgess, A. W., & Holmstrom, L. L. (1985). Rape trauma syndrome and post traumatic stress response. In A. W. Burgess (Ed.), Rape and sexual assault: A research handbook (pp. 46-60). New York: Garland Publishing, Inc.
Derogatis, L. R. (1977). SCL-90: Administration, scoring, & procedure manual-l'for theR(revised) version. Baltimore: Johns Hopkins University School of Medicine.
Donaldson, M. A., & Gardner, R. (1985). Diagnosis and treatment of traumatic stress among women after childhood incest. In C. R. Figley (Ed.), Trauma and its wake (pp. 356-377). New York: B runner/M azel.
Fairbank, J. A., Keane, T. M., & Malloy, P. F. (1983). Some preliminary data on the psychological characteristics of Vietnam veterans with posttraumatic stress disorders. Journal of Consulting and Clinical Psychology, 51, 912-919.
Foy, D. W., Sipprelle, R. C., Rueger, D. B., & Carroll, E. M. (1984). Etiology of PTSD in Vietnam veterans: Analysis of premilitary, military, and combat exposure influences. Journal of Consulting and Clinical Psychology, 52, 79-87.
Hinkle, D. E., Wiersma, W., & Jurs, S. G. (1979). Applied statistics for the behavioral sciences. Chicago: Rand McNally.
Holmes, M. R., & St. Lawrence, J. S. (1983). Treatment of rape-induced trauma: Proposed behavioral conceptualization and review of the literature. Clinical Psychology Review, 2,417-433.
Horowitz, M. J., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: Measure of subjective stress. Psychosomatic Medicine, 41, 209-218.
Keane, T. M., Malloy, P. F., & Fairbank, J. A. (1984). Empirical development of an MM PI subscale for the assessment of combat-related posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 52, 888-891.
Keane, T. M., Wolfe, J., & Taylor, K. L. (1987). Post-traumatic stress disorder: Evidence for diagnostic validity and methods of psychological assessment. Journal of Clinical Psychology, 43, 32-43.
Kilpatrick, D. G., Best, C. L., Veronen, L. J., Amick, A. E., Villeponteaux, L. A., & Ruff, G. A. (1985). Mental health correlates of criminal victimization: A random community survey. Journal of Consulting and Clinical Psychology, 53, 866-873.
Kilpatrick, D. G., Saunders, B. E., Veronen, L. J., Best, C. L., Von, J. M. (1987). Criminal victimization: Lifetime prevalence, reporting to police, and psychological impact. Crime & Delinquency, 33, 479-489.
Kilpatrick, D. G., Saunders, B. E., Amick-McMullan, A., Best, C. L., Veronen, L. J., & Resnick, H. S. (1989). Victim and crime factors associated with the development of Crime-Related Post-Traumatic Stress Disorder. Behavior Therapy, 20, 199-214.
Koss, M. P. (1985). The hidden rape victim: Personality, attitudinal, and situational characteristics. Psychology of Women Quarterly, 9, 193-212.
Kramer, T. L., & Green, B. L. (1989). Post-traumatic stress disorder as an early response to rape. Manuscript submitted for publication.
Lindberg, F. H., & Distad, L. J. (1985). Post-traumatic stress disorders in women who experienced childhood incest. Child Abuse & Neglect, 2, 521-526.
Robins, L. N., Helzer, J. D., Croughan, J., & Ratcliff, K. S. (1981). The National Institute of Mental Health Diagnostic Interview Schedule: Its history, characteristics, and validity. Archives of General Psychiatry, 38, 381-389.
Rothbaum, B. O., & Foa, E. B. (1988). Responses following sexual and non-sexual assault. Paper presented at the 22nd Annual Association for the Advancement of Behavior Therapy Convention, New York.
Saunders, B. E., Arata, C. M., & Kilpatrick, D. G. (1990). Development of a crime-related Post-Traumatic Stress Disorder scale for women within the Symptom Checklist-90-Revised. Journal of Traumatic Stress, 3(3), 439-448.
Saunders, B. E., Kilpatrick, D. G., Resnick, H. S., & Tidwell, R. P. (1989). Brief screening for lifetime history of criminal victimization at mental health intake: A preliminary study. Journal of Interpersonal Violence, 4, 267-277.
Spitzer, R. L., Williams, J. B. W., & Gibbon, M. (1986). Structured clinical interview for DSM-III-R. Non-patient version (SCID-NP10/1/86). New York: Biometrics Research Department, New York State Psychiatric Institute.
Steketee, G., & Foa, E. B. (1987). Rape victims: Post-traumatic stress responses and their treatment: A review of the literature. Journal of Anxiety Disorders, 1, 69-86.
Acknowledgments. This research was partially supported by National Institute of Justice Grant No. 84-IJ-CX-0039. Points of view or opinions expressed in this article do not necessarily represent the official position or policies of the U.S. Department of Justice.
Catalina Mandoki Arata
Searcy Hospital, Mobile, Alabama
Benjamin E. Saunders
Dean G. Kilpatrick
Crime Victims Research and Treatment Center
Department of Psychiatry and Behavioral Sciences
Medical University of South Carolina at Charleston
Offprints. Requests for offprints should be sent to Catalina M. Arata, Ph.D., Department of Psychology, University of South Alabama, Mobile, AL 36688.…