Journal of Traumatic Stress Studies, 9, 891-899.

Running Head: PTSD DIAGNOSIS IN AFGHAN REFUGEES

Modification of CAPS-1 for Diagnosis of PTSD in Afghan refugees

A. Samad Bazger Malekzai¹, John M. Niazi¹, Stephen R. Paige^1,3,

Shelton E. Hendricks^1,3, Denis Fitzpatrick¹, M. Patricia Leuschen², and C. Raymond Millimet³

¹Department of Psychiatry, ²Department of Pediatrics, University of Nebraska Medical Center and ³Department of Psychology, University of Nebraska at Omaha, Omaha, Nebraska

Corresponding author:

Dr. Stephen R. Paige

Dept. of Psychiatry

University of Nebraska Medical Center 5575

Omaha, NE 68198

Abstract

A DSM-III-R based instrument for the assessment of PTSD, the Clinician-Administered PTSD Scale (CAPS-1), was modified to accommodate cultural differences and translated into the Afghan languages Pushto and Farsi (Dari) and administered to 30 Afghan refugees living in the United States. The modified CAPS-1 was found to be practical and reliable. Inter-item correlations were calculated on the frequency and intensity scores for the 17 cardinal symptoms and the eight associated features items of the modified CAPS-1. The four reexperiencing items demonstrated significant independence from the avoidance and arousal symptom clusters. However, the avoidance and arousal symptom clusters were not found to be independent cardinal components of PTSD in our participants. The CAPS-1 criteria for diagnosis of PTSD were met by 50% of the subjects evaluated.

Keywords: PTSD assessment; Afghan, CAPS-1, Refugees

Over the past 15 years thousands of Afghans have been forced into refugee status. While there are recent descriptions of the general health status among Afghan refugees (Gessner, 1994; Wardak, 1993), there are no data as to the prevalence of posttraumatic stress disorder (PTSD ) among Afghan refugees, or any formal attempt to describe the syndrome in this population. The results of studies of PTSD among other populations exposed to war suggest a high prevalence of PTSD (Boehnlein, Lung, Moore, Riley, & Smith, 1990; Kroll, Habenicht, & Mackenzie, 1989; Mollica, Wyshal, & Lavelle, 1987; Saigh, 1989). In the present study we have attempted to develop a diagnostic instrument for the assessment of PTSD in Afghan refuges and provide an initial assessment of the extent of this problem among Afghan refugees living in a refugee community in the United States.

The Clinician-Administered PTSD Scale [CAPS-1] (Blake et al., 1990) was selected for the purpose of developing an assessment tool suitable for the evaluation of PTSD in Afghans. This instrument was developed by the National Center for PTSD to measure both cardinal and hypothesized associated symptoms of PTSD. The CAPS-1 not only provides specific questions for determination of the presence or absence of 17 core PTSD symptoms, but also uses a rating scale to indicate frequency and intensity of symptoms. A rating is also made of the overall impact of the PTSD symptoms on the patient's social and occupational functioning, making an important connection between assessed symptoms and quality of life.

Method

Translation and Modification of the Instrument

CAPS-1 was translated by the first author into the two most widely understood Afghanlanguages, Pushto and Farsi (Dari). The translated versions of CAPS-1 were evaluated for linguistic and cultural suitability for the target population by five Afghan physicians living in the United States. Adjustments were made by consensus. A model question, to guide the clinician, was added to the first part of the interview in order to establish the presence or absence of a traumatic event outside the normal range of human experience, translated as follows: "There are ups and downs in everybody's life, as we expect. In the past years we have faced much trauma, what is the worst trauma in your mind?"

Participants

Participants in the study were 30 native Afghans (15 males and 15 females) ranging in age from 19 to 75 years with a mean age of 42. All of the participants had been living in a metropolitan area on the West Coast of the United States for 10 years or less and had immigrated to the United States as a result of war and civil unrest in Afghanistan. It is estimated that 25,000 Afghan refugees live in this area (Omidian, 1993). The local Afghan community was made aware of the research project through advertisements and by word of mouth. Participants' formal education ranged from none to post-graduate studies. None of the participants reported a prior or current diagnosis of psychiatric illness. Those who volunteered to participate received and signed a consent form prepared in either Pushto or Farsi as appropriate. When necessary, the form was read to the participant by an investigator.

Procedure

The interviews were administered by two Afghan physicians, conversant in Pushto and Farsi with psychiatric experience among Afghans. Interviews were conducted in a private room at the home of the participant or in a room at an immigrant education center. Participants stated theirpreferred language for the interview. One physician interviewed the participant and scored responses while the second physician listened and also provided a score. Roles were reversed for alternate participants. The time taken to complete each interview was between 40 and 80 minutes. The average time was approximately 50 minutes.

Data Analysis

Scores (frequency and intensity) obtained on CAPS-1 by both the interviewer and observer were analyzed using a statistical package (SAS Institute Inc., 1989). Intraclass correlation coefficients were obtained for each scale item to calculate reliability of ratings between the two raters. Cronbach’s coefficient alpha (Cronbach, 1951) was calculated on the intensity and frequency scores averaged for the two raters as an index of the internal consistency for each of the three symptom clusters and the total 17 cardinal symptom items of the CAPS-1 composing the reexperiencing, avoidance, and arousal symptom clusters (criteria B, C, and D; American Psychiatric Association, 1987). Inter-item Pearson product-moment correlations were calculated on the intensity and frequency scores averaged for the two raters for the 17 cardinal symptom and the eight associated features items of the CAPS-1. The low number of participants in the study did not permit a meaningful factor analysis of the intercorrelation matrix. Instead, each of the correlation coefficients was transformed into a z-value using the Fisher z transformation procedure and then averaged. Tests of the difference between two average correlations using the Fisher z-transformed correlations (Kirk, 1990) were used to evaluate the independence of the three cardinal CAPS-1 item clusters. A symptom cluster was considered to be independent of another symptom cluster if the difference between the average within-cluster item correlations was significantly greater than the average between-cluster correlations of the items composing the two symptomclusters under consideration.

Results

Evaluation of the Instrument

The final Pushto and Farsi versions of CAPS-1 are essentially the same as the original English CAPS-1 except for some changes which were made for linguistic and cultural adequacy. In these instances one or more effective semantic alternatives could be found. Some psychological terms such as "dissociation" were left in English, as they were for the information of the clinician only. It is common practice for Afghan clinicians to use English medical terminology. All of the interview items appeared to be culturally appropriate except for part of Item 11, "Sense of Foreshortened Future," which asks if one expects his or her life to be cut short. The question for Item 11 was answered by the participants in the manner of "How would I know?" or "Only God knows." Many Afghans believe that life span is predetermined by God and is not for their judgment. Several participants viewed this question as offensive and/or unacceptable. Thus, Item 11 was maintained with reference to the need to plan for the future, but without reference to time of death. For Item 17, "Physiologic reactivity upon exposure to events that symbolize or resemble an aspect of the traumatic event," additional model questions were provided for the interviewer. This was necessary because participants with little or no education found it difficult to respond to the suggested question.

Intraclass correlation coefficients for symptom intensity ratings between interviewer and observer ranged from .90 to 1.0 (M = .98). Intraclass correlation coefficients for symptom frequency ratings between interviewer and observer ranged from .86 to 1.0 (M = .98). Alphacoefficients for the four reexperiencing items (.81), the seven avoidance items (.76), the six arousal items (.82), and for the total 17 CAPS-1 cardinal symptom items (.88) demonstrated satisfactory internal consistency.

Figure 1 presents the matrix of inter-item correlations for the intensity scores averaged across the two raters for the 17 cardinal symptoms and the eight associated features of PTSD. The intercorrelations among the items composing the reexperiencing, avoidance, and arousal symptom clusters and the associated features cluster are separated by bolded lines.

Comparisons of the reexperiencing within-cluster item correlations (average r=.59) with the reexperiencing-avoidance between-cluster items correlations (average r=.20; z= 1.75, p<.04, one-tail) and the reexperiencing-arousal between-cluster item correlations (average r=.32; z= 1.23, p<.11, one-tail) showed that the reexperiencing symptoms tend to be independent of the avoidance and arousal symptoms.

Comparisons of the arousal within-cluster item correlations (average r=.46) with the avoidance-arousal between-cluster item correlations (average r=.33; z= .57, p<.29, one-tail) and the avoidance within-cluster item correlations (average r=.33) with the avoidance-arousal between-cluster item correlations (average r=.33; z= 0.00, ns) showed that the symptoms composing the avoidance and arousal clusters are not independent of each other.

Of the 30 participants, 50% met the CAPS-1 criteria for a current diagnosis of PTSD. This included 52% of the male participants and 44% of the female participants. The incidence of PTSD increased from 10% in the 19 to 30 years age group to 100% in the 61 to 75 years age group. The Spearman Correlation Coefficient between age and PTSD incidence was .65 ( p <. 005).

Discussion

We conclude that the translated version of CAPS-1 is suitable for the assessment of PTSD in Afghans. The participants did not have difficulty with the interview format or the linguistic content except as noted for Item 11 which was adjusted to preserve a reference to planning for the future, but without reference to time of death. As assessed in this study, inter-rater reliability was high. While we recognize that the use of independent interviewers would have been the preferable method for assessing inter-rater reliability, it was not feasible because of our reluctance to require participants to undergo repeated and possibly stressful interviews. Internal consistency as indexed by Cronbach’s coefficient alpha was highly satisfactory for the 17 cardinal symptoms of the CAPS-1 and within each of the symptom clusters of reexperiencing, avoidance and arousal items.

The statistical analyses comparing the inter-item correlations among the three cardinal symptom clusters revealed that the four reexperiencing items demonstrated independence from the other two symptom clusters. However, the avoidance and arousal symptom clusters failed to reveal independent cardinal components of PTSD in our participants. We did not examine the internal consistency or the independence of the associated features items because there is no a priori prediction that these eight CAPS-1 items measure a single common component of PTSD, independent of other putative symptom clusters.

Watson, Kucala, Juba, Manifold, and Anderson (1991) published inter-item correlations for the 17 scales of the PTSD-I (Watson, Juba, Manifold, Kucala, & Anderson, 1991) on 131 Vietnam veteran PTSD patients. The intercorrelations for items that are common among their PTSD-1 and our modified CAPS-1 are generally of similar magnitudes. However, differences in the diagnostic instruments, populations studied, and diagnostic criteria make further comparisondifficult. Similar difficulties arise in comparing the results of their factor analysis with our results and with other factor analyses of various symptom rating scales in various populations of traumatized individuals including rape and nonsexual assault victims (Foa, Riggs, & Gerushney, 1995), Dutch resistance veterans of WWII (Hovens, et al., 1993), and Vietnam combat veterans (Silver & Iacono, 1984).

Scoring of the CAPS-1 interviews indicated a 50% incidence of PTSD in our participants. The lowest age range (19 to 30) assessed in this study was composed of individuals who were children at the time they were exposed to the worst trauma. An explanation of the lower PTSD rates among these youngest participants is offered by Garbarino, Kostelny, and Dubrow (1991) who propose that some children and youth may be protected from the harmful effects of trauma by the strong support of parents or care givers and by fulfillment of basic needs. However, the generalizability of these findings should be tempered by limiting factors of sample size, and possible selection biases such as the self-selection of participants in the present sample.

None of the participants showing evidence of PTSD had been previously diagnosed and none was receiving mental health services. Reasons for the lack of prior identification include reluctance to tell health care providers of symptoms other than physical complaints, inadequate screening or diagnostic systems for PTSD in Afghanistan and Pakistan, and a perception among some refugees of more pressing priorities than their own health. Our participants did not consider their symptoms extraordinary within the Afghan community and some Afghan refugees reported linguistic and cultural barriers in the medical system.

References

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd edit., rev.), Washington, DC: Author.

Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Klauminzer, G., Charney, D. S., & Keane, T. M. (1990). A clinician rating scale for assessing current and lifetime PTSD: The CAPS-1. the Behavioral Therapist, 13, 187-188.

Boehnlein, J. D., Lung, J. K., Moore, L. J., Riley, C., & Smith, D. (1990). The prevalence of PTSD and its clinical significance among South East Asian refugees. American Journal of Psychiatry, 147, 913-917.

Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrica, 16, 297-334.

Foa, E. B., Riggs, D. S., & Gershuny, B. S. (1995). Arousal , numbing, and intrusion: Symptom structure of PTSD following assault. American Journal of Psychiatry, 152, 116-120.

Garbarino, J., Kostelny, K., & Dubrow, N. (1991). What children tell us about living in danger. American Psychologist, 46, 376-383.

Gessner, B. D., (1994). Mortality rates, causes of death, and health status among displaced and resident populations of Kabul, Afghanistan. Journal of the American Medical Association, 272, 382-385.

Hovens, J. E., Falger, P. R. J., Op De Velde, W., Meijer, P., De Groen, J. H. M., & Van Duijn, H. (1993). A self-rating scale for the assessment of posttraumatic stress disorder in Dutch resistance veterans of World War II. Journal of Clinical Psychology, 49, 196-203.

Kirk, R. E. (1990) Statistics: An introduction (3rd ed.) Fort Worth, TX: Holt, Rinehart, & Winston.

Kroll, J., Habenicht, M., & Mackenzie, T. (1989). Depression and post-traumatic stress disorder in Southeast Asian refugees. American Journal of Psychiatry, 146, 1525-1530.

Mollica, R., Wyshal, G., & Lavelle, J. (1987). The psychosocial impact of war trauma and torture on South East Asian refugees. American Journal of Psychiatry, 44, 1567-1572.

Omidian, P. (1993). Aging and intergenerational conflict: Afghan refugee families in transition. Unpublished doctoral dissertation, University of California, San Francisco.

Saigh, P. A. (1989). The validity of the DSM-III-R post-traumatic stress disorder classification as applied to children. Journal of Abnormal Psychology, 98, 189-192.

SAS Institute Inc. (1989). SAS/STAT Users Guide (version 6), 4, Cary, NC: SAS Institute Inc.

Silver, S. M., & Iacono, C. U. (1984). Factor-analytic support for DSM-III's post-traumatic stress disorder for Vietnam veterans. Journal of Clinical Psychology, 40, 5-14.

Wardak, A. W., (1992) The psychiatric effects of war stress on Afghanistan society. In J.P. Wilson & B. Raphael (Eds.), International handbook of traumatic stress syndromes (pp. 349-364). New York: Plenum Press.

Watson, C. G., Juba, M., Manifold, V., Kucala, T., & Anderson, P. E. D. (1991). The PTSD Interview: Rationale, description, reliability, and concurrent validity of DSM-III-based technique. Journal of Clinical Psychology, 47, 179-188.

Watson, C. G., Kucala, T., Juba, M., Manifold, V., & Anderson, P. E. D. (1991). A factor analysisof the DSM-III post-traumatic stress disorder criteria. Journal of Clinical Psychology, 47, 205-214.

Author Note

This research was supported in part by the United States Agency for International Development and in part by Psychopharmacology Research Center of the Creighton-Nebraska Department of Psychiatry. This work is based upon a thesis submitted in partial fulfillment of the requirements for a master of science degree by the first author. The authors are indebted to the Afghan physicians who helped in developing the assessment instrument and Bonnie Green and three anonymous reviewers for their helpful comments on earlier drafts of this manuscript.

Figure Caption

Figure 1. Inter-item correlation coefficients for CAPS-1 intensity scores.