|
Clinician-Administered PTSD Scale (CAPS)
Please note that this section is an archive (last updated in June 2006). [disclaimer]
Sections:
Overview |
Instrument Reviews |
Construct Overviews |
Book Compendium Reviews |
Internet Site Reviews
|
|
|
|
Practical Information
|
Instrument Name:
|
Clinician-Administered PTSD Scale (CAPS)
|
Instrument Description:
|
The CAPS’ authors claim that before the development of the CAPS, PTSD diagnostic interviews contained major limitations as to their reliability and validity. The CAPS’ developers therefore specifically attempted to overcome these limitations with their new instrument. The focus of the first version of the CAPS was to assess the severity and frequency of 17 PTSD symptoms as described in the DSM-III-R. (Ref: 1) Thirteen additional items were included in the scale: Eight assessed associated symptoms, and 5 evaluated outcome variables. (Ref: 7)
Constructs measured by the 8 associated symptom items include hopelessness, memory impairment, and guilt, while the 5 outcome items assess response validity, global impairment, global severity, and social and occupational impairment. (Ref: 7) The CAPS evaluates both long-term and current PTSD symptoms and diagnostic status. These dual time references help to determine, for example, the impact that PTSD symptoms have had and are having on a subject’s social and occupational functioning. (Ref: 1)
The CAPS has undergone several revisions, due to both user feedback and changes in the diagnostic criteria of PTSD. (Ref: 1) Examples of such changes are: adding a 17-item self-report checklist of traumatic events (to assist the interviewer in associating a traumatic event with PTSD, required by DSM-IV definition of PTSD); rewording descriptors of the intensity scale; adding a 3-point rating scale to the last 9 PTSD symptoms; and replacing some of the associated symptoms questions. (Ref: 7)
|
Price:
|
Free. It must be obtained from the National Center for PTSD website. (See below)
|
Administration Time:
|
Varies from 40 to 60 minutes, depending on the extent of symptoms. (Ref: 6)
|
Publication Year:
|
1990
|
Item Readability:
|
The CAPS’ reading level is 4.7, indicating that a person who is close to completing the 4th grade should be able to comprehend the scale. CAPS language is simple and straightforward in order to accommodate subjects with different educational backgrounds.
|
Scale Format:
|
A dichotomous response format (yes/no) helps determine if specific life events match the DSM definition of PTSD, while a 5-point Likert scale is used to report on the frequency and intensity of each PTSD symptom. (Ref: 1)
|
Administration Technique:
|
The CAPS is a structured interview administered by experienced clinicians such as physicians, psychiatric nurses, clinical researchers, or by an appropriately trained paraprofessional who has a working knowledge of PTSD. (Ref: 1)
|
Scoring and Interpretation:
|
The PTSD symptom frequency response scale has 5 options that are scored from 0-4. The options are as follows: Never, Once or twice, Once or twice a week, Several times a week, and Nightly or almost every day. The intensity response scale also has 5 options, scored from 0-4: None, Mild, Moderate, Severe, and Extreme. An interviewer can marked an additional option, “questionable validity,” for any patient response if reason exists for doubting that patient’s report. (Ref: 1) Revisions of the CAPS have included adding a 3-point rating scale (unlikely, probable, definite) to the last 9 items of the scale, thus allowing interviewers to determine if the symptom reported is the result of a specific traumatic event.
CAPS users have numerous results available to them in the form of different component scores, to be interpreted dependent on interview purpose (e.g., diagnosis, treatment plan). (Ref: 7) After an interview is complete, data are coded on a “summary sheet” to ease scoring and interpretation. There are 9 different diagnostic scoring rules, to be employed per interviewer need. CAPS users can find psychometric information for these 9 scoring rules in an article by Weathers et al. (Psych Assessment, 1999;11(2):124-33), which is not reviewed here. Higher scores indicate more intense and more frequent symptoms of PTSD, while lower scores represent the absence of or less intense, less frequent symptoms of PTSD. (Ref: 2) One study has assessed cut-scores for the CAPS, based on 3 DSM-III-R criteria. The percentages of participants for each cut-score are presented below: (See Ref: 3 for a description of the subject population.)
| Number of Symptoms
|
Intrusion
|
Avoidance
|
Hyperarousal
|
| >=1
|
76.3
|
75.0
|
84.2
|
| >=2
|
59.2
|
60.5
|
69.7
|
| >=3
|
26.3
|
48.7
|
57.9
|
More recently, two scoring rules seem to dominate practice. With one, a PTSD symptom must receive a frequency score of at least 1 (i.e., occur once or twice a month) and an intensity score of at least 2 (i.e., be at least moderate) for the symptom to be considered present. The second scoring rule states that a patient must attain a CAPS total severity score of at least 65 to receive a PTSD diagnosis. (Ref: 6)
|
Forms:
|
A new CAPS version (CAPS-DX) is used to assess adolescents and children; it underwent psychometric evaluation and field-testing in 2001. (Ref: 1) Other CAPS versions are: Caps-1, for current and lifetime diagnostic purposes; and CAPS-2 (or CAPS-SX), for assessing one-week symptom status. (Ref: 1) These two versions, CAPS-1 and CAPS-2, have been combined in the latest version of the CAPS. (Ref: 2) CAPS-2 is good for repeated assessments and evaluating treatment effects. The CAPS has been translated into 10 different languages, including Dutch. (Ref: 1,3,6)
|
|
|
Research Contacts
|
Instrument Developers:
|
DD Blake, FW Weathers, LM Nagy, DG Kaloupek et al.
|
Instrument Development Location:
|
National Center for Posttraumatic Stress Disorder
|
Instrument Developer Email:
|
weathfw@auburn.edu
|
Instrument Developer Website:
|
www.ncptsd.va.gov/publications/assessment/adult_interviews.html
|
|
|
Annotated Bibliography
|
1. Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Gusman FD, Charney DS, & Keane TM. The development of a Clinician-Administered PTSD Scale. J Trauma Stress. 1995 Jan;8(1):75-90. Review.[PMID: 7712061]
|
Purpose:
This is the Clinician-Administered PTSD Scale (CAPS) developmental article, which discusses the CAPS’ psychometric properties. The CAPS and other similar PTSD scales are compared.
|
|
|
Sample:
Not applicable.
|
|
|
Methods:
Description of the CAPS instrument and other competing PTSD assessment instruments.
|
|
|
Implications:
Overall, the authors reported that the CAPS met or exceeded their expectations for a new PTSD scale incorporating high psychometric standards. However, the CAPS interview instrument was not found to be perfect. Some obvious areas for revision are (a) the length of the interview and (b) strengthening its ability to assess Criterion A of the DSM criteria for PTSD (i.e., the presence of a recognizable traumatic stressor). The CAPS will likely need periodic revision as diagnostic criteria of PTSD evolve.
|
2. Blake DD, Weathers FW, Nagy LM, Kaloupek D, Klauminzer G, Charney DS, Keane TM, & Buckley TC. Clinician-Administered PTSD Scale (CAPS) Instruction Manual. Boston, MA: National Center for PTSD. 2000 Nov:1-49.
|
|
Purpose:
Serves as the user manual. Discusses developmental, administration, and scoring information for the CAPS.
|
|
|
Sample:
Not applicable.
|
|
|
Methods:
Not applicable.
|
|
|
Implications:
Not applicable.
|
3. Hovens JE, van der Ploeg HM, Klaarenbeek MT, Bramsen I, Schreuder JN, & Vladar Rivero V. The assessment of posttraumatic stress disorder: With the Clinician Administered PTSD Scale: Dutch results. J Clin Psychol. 1994 May;50(3):325-40. [PMID: 8071438]
|
|
Purpose:
To validate the CAPS among (a) Dutch civilian participants, (b) military veterans from World War II and conflicts in Indonesia and Korea, and (c) recently traumatized civilian from the local community.
|
|
|
Sample:
Participants were recruited from 3 treatment centers and 1 rehabilitation center. The Center 45 (the Dutch National Institute of Psychotherapy and Psychiatric Treatment for Veterans and Survivors of WWII) sample was composed of 27 participants, including 12 females. The mean age of the total sample was 56.1 years. Eighteen participants were married, 8 were divorced, and 1 was widowed. Educational background was as follows: Two had primary school, 10 attended vocational school, 10 attended secondary school, and 4 had university training.
The BNMO center (a private foundation to improve the well-being of disabled veterans) recruited 28 participants; all participants were male and had a mean age of 68.9 years. Twenty-three participants were married, 3 were divorced, and 2 were widowed. Participant educational training was as follows: seven participants with primary school, 9 attended vocational school, 7 attended secondary school, and 4 had university training.
Twenty-one participants were recruited from the third and fourth Centers (i.e., community mental health centers); 11 were in treatment (from the RIAGG Center) and 10 were in long-term treatment (from the JAGGZ Center). The sample contained 13 females and 8 males; 12 participants were married, 4 divorced, 4 widowed, and 1 was single. Three participants had primary school, 10 attended vocational school, and 7 had high school.
|
|
|
Methods:
The 3-month study identified eligible patients at Center 45 as those who received less than 1 month of treatment or who were requesting treatment from the center. Participants were examined by a psychologist or psychiatrist and received a letter describing the purpose of the study. At the second visit to the center, eligible patients were asked if they were willing to participate in the study. Those who consented received questionnaires (including various PTSD self-report measures) to answer at home and an interview appointment with an investigator scheduled for 1 week later. This study utilized a Dutch translation of the CAPS (translation performed by a study author).
The second set of subjects was randomly selected from the BNMO Center. Those that were willing to participate in the study received questionnaires and an appointment for a structured interview. Lastly, subjects who were being treated for traumatic experiences during WWII at the JAGGZ and those who fulfilled the DSM-III-R stressor criterion at the RIAGG were recruited. Psychologists assessed subjects to determine if they were psychologically stable. Researchers the recruited stable individuals for study participation. Willing subjects completed PTSD questionnaires and were invited for a structured interview.
|
|
|
Implications:
A main interest of this study was the use of the CAPS to assess participants who had been traumatized in either early childhood or adolescence. Assessment and diagnosis of such participants is problematic by DSM-III-R criteria because participants traumatized in early childhood, for example, by definition cannot readily differentiate between arousal before and after trauma, a DSM-III-R criterion for PTSD diagnosis. The CAPS was found to perform adequately in diagnosing PTSD symptom frequency and intensity; however, information obtained from associated symptom items did not add to an understanding of the disorder, nor could a diagnosis of lifetime PTSD be considered anything other than a rough measure of symptomatology.
|
4. Hyer L, Summers MN, Boyd S, Litaker M, & Boudewyns P. Assessment of older combat veterans with the Clinician-Administered PTSD Scale. J Trauma Stress. 1996 Jul; 9(3): 587-93. [PMID:8827658]
|
|
Purpose:
The study focused on validating the CAPS-1 on an older veteran population whose war exposures were associated with experiences from WWII or the Korean Conflict. The presence of possible PTSD symptoms and PTSD diagnosis were determined, and the CAPS’ psychometric properties were assessed. The CAPS had previously been validated only with Vietnam veterans.
|
|
|
Sample:
A sample composed of 125 World War II and Korean War veterans took part in the study. The average age of the sample was nearly 70 years, and the majority was married (80%). Participants had an average of 11.5 years of education, 83% were White, and 69% had taken part in World War II. Most study participants had not been wounded in war (72%); however, most had experienced heavy combat (79%).
|
|
|
Methods:
Participants were recruited from several medical and psychiatric outpatient clinics. Investigators reviewed patient charts to determine criteria eligibility. Participants had to be WWII or Korean War veterans, aged 60 or older, and exposed to combat. Exclusion criteria included having a life-threatening illness, being unable to read English, having been a prisoner of war, and having a psychotic or organic brain syndrome diagnosis. Participants were given a battery of PTSD self-report instruments and underwent a CAPS interview.
|
|
|
Implications:
Using SCID-PTSD diagnosis as a gold standard, researchers found that the CAPS had high sensitivity (.90), specificity (.95), and positive (.92) and negative (.93) predictive power. Results indicated that the scale could adequately assess older veterans with a war exposure other than the Vietnam War. A weakness of the study pertained to researchers not knowing if the sample, a convenience sample, adequately or sufficiently represented all types of treatment-seeking veterans.
|
5. King, DW, Leskin, GA, King, LA, & Weathers, FW. Confirmatory factor analysis of the Clinician-Administered PTSD Scale: Evidence for the dimensionality of posttraumatic stress disorder. Psychological Assessment 1998; 10(2): 90-96. [No PMID]
|
|
Purpose:
This study assessed the factor structure of the CAPS in order to investigate the possible dimensionality of the PTSD construct itself.
|
|
|
Sample:
Participants were 524 male veterans. Average age was approximately 51 years. Ethnic distribution was as follows: 82% White, 13% African American, 2% Hispanic, 2% American Indian/Alaskan Native, and <1% Asian American. The majority of the sample was employed (42%), while the others were students (9%), unemployed (33%), or retired/disabled (16%).
|
|
|
Methods:
Participants were evaluated at the National Center for PTSD (Boston) between 1990 and 1996 using the CAPS. Confirmatory factor analyses were performed, with a series of 4 nested models fit and then compared.
|
|
|
Implications:
The model of best fit was the 4-factor, first-order solution. This finding suggests that emphasis should be placed on the symptom cluster of the CAPS when interpreting scores. Although total score may assist in determining a diagnosis of PTSD, subscale scores corresponding to the 4-factor solution may prove more useful in designing a treatment plan. The source of the 4th factor was the splitting up of the avoidance/emotional numbing subscale. Severity scores were analyzed, obtained by adding frequency and intensity responses per symptom.
|
6. Foa EB & Tolin DF. Comparison of the PTSD Symptom Scale-Interview Version and the Clinician-Administered PTSD Scale. J Trauma Stress. 2000 Apr;13(2):181-91.[PMID:10838669]
|
|
Purpose:
This study compared the psychometric properties of the CAPS and the PTSD Symptom Scale-Interview Version (PSS-I) in civilian trauma survivor patients with and without PTSD. The PSS-I is an alternative measure of PTSD severity that requires less assessment time than the CAPS (i.e., 20-30 minutes per interview versus 40-45 minutes).
|
|
|
Sample:
There were 64 participants in the study: 12 clinic patients and 52 adult non-clinical volunteers. The mean age of all participants was 37 years. Ethnicity of the sample was 52% Caucasian, 39% African American, 3% Hispanic, 5% Asian American, and 1% “other ethnicities.” Fifty-three percent were females; 47% were males. All participants reported experiencing a traumatic incident that met Criterion A of the DSM-IV PTSD diagnosis. There were no combat veterans in the sample; all participants were civilians.
|
|
|
Methods:
First, 39 participants were interviewed by 2 clinicians. The first interview focused on pinpointing a specific traumatic event for further analysis in later interview phases. Participants were instructed to refer to the single event throughout all interviews; discussions were videotaped to verify compliance. The second interview focused on the CAPS and PSS-I evaluations. Participants were administered both tests, separately, and in a randomized order. Videotapes of the structured interviews were then reviewed by 2 new raters who had no access to the original interviewer ratings; the new raters then completed the CAPS and PSS-I based on participants’ videotaped responses. Patient ratings were then compared. Next, an additional 25 participants was administered the SCID-PTSD module in addition to the other 2 PTSD instruments. The order of instrument presentation in study phases 1 and 2 was randomly assigned. |
|
|
Implications:
The 2 measures, PSS-I and CAPS, were similar in regards to psychometric properties. Differences noticed between the 2 included that the PSS-I appeared to have slightly higher sensitivity, giving it an advantage in detecting actual PTSD, while the CAPS had slightly higher specificity, giving it an advantage in ruling out false positives. Limitations to this study included the small sample size and the inability to generalize findings to a veteran population (i.e., the sample consisted of civilians alone).
|
7. Weathers FW, Keane TM, & Davidson JR. Clinician-Administered PTSD Scale: A review of the first ten years of research. Depress Anxiety. 2001;13(3):132-56. Review. [PMID:11387733]
|
|
Purpose:
This paper reviews the psychometric properties, strengths, and weaknesses of the CAPS and its ability for assessing PTSD after its first 10 years of reported use in published research.
|
|
|
Sample:
Not applicable.
|
|
|
Methods:
Review of the literature.
|
|
|
Implications:
In 10 years of use, the CAPS has shown to be a PTSD instrument whose measures tend to exhibit sound psychometric properties. The CAPS provides a flexible structured interview and has been used with many different traumatized populations. Some criticisms of the instrument are: its length; the complexity of administration and scoring; and the PTSD symptom frequency-intensity overlap. The authors noted that the interview may appear long on paper but does not take long to administer. The authors also refuted the difficulty of learning the administration process, as new interview administrators who have undergone a 2-hour training have been able to consistently produce diagnoses and symptom assessments having high inter-rater agreement. Lastly, at the item level, correlations between symptom frequency and intensity are only moderate, indicating that these symptom facets are indeed distinct, and that combining their measures would result in a loss of information.
|
8. Mueser KT, Salyers MP, Rosenberg SD, Ford JD, Fox L, & Carty P. Psychometric evaluation of trauma and posttraumatic stress disorder assessments in persons with severe mental illness. Psychol Assess. 2001 Mar; 13(1): 110-7. [PMID:11281032]
|
|
Purpose:
Three measurement scales - the CAPS, the Traumatic History Questionnaire (THQ) for lifetime exposure to traumatic events, and the PTSD Checklist (PCL) - were compared, with regards to their psychometric properties when used for assessing trauma and PTSD in patients with severe mental illness.
|
|
|
Sample:
The sample was comprised of 30 participants, all of whom were able to provide signed consent forms. Of the 30 participants, 53% were women, 90% were White, 27% were never married, 23% were married, and 50% were either widowed, separated, or divorced. The majority of participants (83%) had graduated from high school, and 90% lived independently. Participant age range was 25-66 years, with an average age of 45 years. The participants’ first hospitalization was at an average age of 27 years, and each had an average of 7.58 hospitalizations in a lifetime, the last occurrence averaging 53 months ago. The sample contained a diagnostic mix of PTSD and non-PTSD patients.
|
|
|
Methods:
Thirty participants were randomly selected from outpatient psychiatric clinics in the Mental Health Center of Greater Manchester. Eligible participants were those with severe mental illness. After recruitment, participants underwent 2 separate interviews. The first interview focused on obtaining demographic and background characteristics and then completing the THQ, the CAPS, and the PCL. After demographic and background information was obtained, the THQ was administered. Next, the CAPS and PCL were administered, in random order. A second interview was then scheduled for 2 weeks later. In the follow-up session, the same interviewer readministered the 3 questionnaires in the same order as they had originally been administered.
|
|
|
Implications:
Internal consistency for the CAPS subscales and total score were consistently good. Test-retest reliabilities using standard cutoffs were moderate, and there were also moderately low kappa scores (i.e., .63 and .60) for participants with more severe psychotic symptoms as opposed to less severe symptoms. However, when using a CAPS severity score of 65 as a cut-score for PTSD diagnosis, kappa values were consistently high (e.g., .90) regardless of sample subpopulation. In general, the psychometric properties of the CAPS scales developed for the general population were comparable in clients with severe mental illness. Three limitations were noted by the author: 1) the study’s sample size was small and lacked ethnic heterogeneity, 2) PTSD diagnosis was based on chart review rather than on clinical interview, and 3) psychiatric symptoms and cognitive impairment were evaluated on the basis of clinician consensus rating, rather than by administration of standardized measures.
|
|
|
top
|
Factors and Norms
|
Factor Analysis Work:
|
A series of confirmatory factor analyses were performed, using a 4-factor 1st-order solution, a 2-factor higher order solution, a single-factor higher order solution, and a single-factor 1st order solution; chi square difference tests compared the 4 models. Changes across models in chi-square values were all significant. The 4-factor first-order model was a superior model, compared to the other 3 models. It had a goodness of fit index of 0.95, parsimony normed fit index of 0.76, and an incremental fit index of 0.96. (Ref: 5) All factor loadings exceeded 0.50 except for the amnesia symptom (0.27). Furthermore, Amos critical ratios were evaluated, and all scores exceeded 9.0 except for amnesia (5.66). (A critical ratio over 2.00 is considered evidence of a substantial relationship.) In assessing the relationships among the preferred model’s 4 factors, factor correlations were observed to range from 0.58 (effortful avoidance and emotional numbing) to 0.92 (re-experiencing and hyper arousal). (Ref: 5)
|
Normative Information Availability:
|
No information found
|
|
|
Reliability Evidence
|
Test-retest:
|
Test-retest reliabilities for the 3-symptom cluster ranged from .77-.96; they ranged from .90-.98 for the 17-item core symptom scale. (Ref: 1) Using a standard F1/I2 symptom-scoring rule, test-retest reliability (here, kappa) for a CAPS PTSD diagnosis was .63, with 83% agreement; using a severity cut-score of 65, kappa was .90, with 97% agreement. (Ref: 7) One study cited the CAPS pilot study of Blake et al., which analyzed the test-retest reliability of the scale over a 2-3 day period. Frequency and intensity scores were evaluated, and reliability coefficients ranged from 0.92 to 0.99. (Ref: 8) In the severe mental illness study, ICCs ranged from 0.60 to 0.80 and averaged 0.77. (Ref: 8)
|
Inter-rater:
|
Two interviewers simultaneously interviewed 9 participants to seek to determine the accuracy of CAPS score reporting. The interviewers consisted of the senior author and 1 of 3 clinical psychologists. Frequency of symptoms' scores correlated from 0.59 to 1.00, with a mean of 0.92 for core symptoms, and intensity scores correlated from 0.52 to 1.00, with a mean score of 0.86 for core symptoms. Total scores for the frequency items had correlations that ranged from 0.92 to 1.00, and for intensity items, 0.93 to 0.98. Lastly, the global severity correlation was 0.89. (Ref: 3) These findings are similar to scoring agreements documented in the study by Mueser et al. According to Mueser, the kappa for their PTSD diagnosis was 1.0 (100% acceptance), and ICC values for frequency and intensity scores ranged from 0.91 to 0.99. The median inter-rater score was 0.96 and the ICC for the total CAPS score was 0.99. (Ref: 8) In another study, inter-rater reliability for CAPS total score was .89 (92.5 agreement); for the subscales it was as follows: experience .86 (88.5% agreement), avoidance .81 (93.4% agreement), and arousal .95 (86.6% agreement). (Ref: 6) In a further study, CAPS PTSD diagnosis had a kappa of 1.0 (perfect agreement). ICCs for frequency scores ranged from .91-.99; for intensity scores the median ICC was .96. CAPS total severity score ICC was .99. (Ref: 7)
|
Internal Consistency:
|
Coefficient alpha for the 3 symptom clusters ranged from .85-.87; total score alpha was .94. (Ref: 1) Total CAPS current intensity scores had a coefficient alpha of 0.89; alpha was 0.63 for the intrusive items, 0.78 for the avoidance and numbing items, and 0.79 for the hyperarousal items. Similar results were found for the CAPS PTSD lifetime diagnosis (i.e., alphas of .86, .70, .86, and .67, respectively). (Ref: 3) A study assessing older combat veterans also evaluated the internal consistency of CAPS and reported a total score alpha value of 0.95, with alphas for each subscale ranging from 0.87 (avoidance) to 0.88 (arousal and experiencing). (Ref: 4) These findings were similar to other studies, including a sample of nonveterans (alpha ranged from 0.70 to 0.88 for total and subscale scores) and a sample of 30 participants with severe mental illness (subscale alphas ranged from 0.63 to 0.85 at baseline and 0.71 to 0.81 at follow-up, total score alpha was 0.94 at baseline and 0.95 at follow-up). (Ref: 6,7)
|
Alternate Forms:
|
No information found
|
|
|
Validity Evidence
|
Construct/ Convergent/ Discriminant:
|
The CAPS, the PTSD Symptom Scale-Interview (PSS-I), and the Structured Clinical Interview for DSM-IV (SCID) were compared, and convergent validity was assessed. Spearman correlations showed that the CAPS was strongly correlated with the SCID (ñ (23)= 0.83, p < 0.001) and the PSS-I (ñ (23) = 0.73, p < 0.001). (Ref: 6) Similarly, a study assessing PTSD in patients with severe mental illness assessed convergent validity by comparing the CAPS to the PTSD Symptom Checklist (PCL). In the initial assessment phase, there was a moderately high agreement (83%) between instrument diagnosis (kappa=0.67), while the follow-up phase produced a stronger agreement (93%, kappa=0.85). Diagnosis of PTSD based on the CAPS, an interview-based measure, and a self-report measure (PCL) were all significantly associated. (Ref: 8) In an additional study, CAPS and PCL diagnoses at baseline had a kappa of .67 (83% agreement), while at follow-up kappa was .85 (93% agreement). (Ref: 7)
The CAPS-1 also correlated with the SCID-PTSD (Structured Clinical Interview for DSM-III-R), SCID-DTREE (computerized version of the SCID) PTSD module, SCID-DTREE depression diagnosis (0.56), SCID-DTREE anxiety disorder diagnosis (0.39), and SCID-DTREE substance abuse diagnoses (0.32). Compared to the SCID-DTREE, the CAPS-1 had a sensitivity of .90, specificity of .95, positive predictive power of .92, and negative predictive power of .93. (Ref: 4) Comparing CAPS to SCID diagnosis of PTSD and using a cut-score of 65, the CAPS was found to have a sensitivity of .84, a specificity of .95, with a kappa of .78. (Ref: ?)
The CAPS scores were able to discriminate between PTSD and non-PTSD diagnosis groups: symptoms were more frequent and more intense in participants with a PTSD diagnosis. (Ref: 3) Similarly, CAPS scores for patients with PTSD averaged 82.34 (SD=16.76), while patients without PTSD averaged 48.31 (SD=20.21). (Ref: 5)
|
Criterion-related/ Concurrent/ Predictive:
|
Using a clinician’s diagnosis of PTSD as a gold standard, the overall agreement of the clinician-based DSM-III-R diagnosis and CAPS was 79%; sensitivity was .74, while specificity was .84 (kappa=0.58). (Ref: 3) In a study sample with a PTSD prevalence of 40%, CAPS PTSD diagnosis (based on yes/no responses to the presence of symptoms) versus the “gold standard” SCID-PTSD diagnosis, sensitivity was .90, specificity was .95, positive predictive value was .92, and negative predictive value was .93. For this study kappa was .75. (Ref: 4)
The correlations of frequency and intensity scores for intrusive items (r=0.93), avoidance items (r=0.94), and increased arousal items (r=0.88) were significant (p< 0.0001). (Ref: 3) A study comparing the PSS-I and the CAPS found similar results supporting concurrent validity (ñ= 0.87, p < 0.001). (Ref: 6) In a last study, clinicians’ diagnosis of PTSD was used as a gold standard. The CAPS showed a sensitivity of .74 and a specificity of .84; the overall agreement between the two measures was 79% (kappa=0.58). (Ref: 8)
Statistically significant correlations (at the p<0.001 level) were found between the CAPS intensity score and other PTSD self-report instruments, including the Mississippi Scale for PTSD (r=0.73), MMPI-PTSD (0.74), and Impact of Event Scale (0.62). The CAPS subscales and IES subscales were also significantly correlated (p<0.001). (Ref: 3) The point-biserial correlation between the CAPS-1 and Combat Exposure Scale was 0.26, while correlations with the Impact of Event Scale, and Mississippi Scale were .81 and .61, respectively. (Ref: 4) In another study, CAPS-1 total score was significantly correlated with the Mississippi Scale for Combat-related PTSD score (r=.91) and with the PK scale of MMPI (r=.77). (Ref: 1)
|
Content:
|
Researchers comparing the CAPS to the PCL found that avoidance, hyperarousal and intrusive symptoms appeared more frequently and more intensely in participants with PTSD; the number of symptoms was also greater in the PTSD group (Ref: 3). Frequency and intensity scores of avoidance items, hyperarousal items, and intrusive items were significantly correlated (r= 0.94, 0.88, and 0.93, respectively; p < 0.001). (Ref: 3) Primary evidence for the CAPS’ content validity lies with its use of DSM diagnostic criteria for assessing PTSD symptoms and the contribution of trauma experts in the instrument’s development and refinement. (Ref: 6)
|
Responsiveness Evidence:
|
No information found
|
Scale Application in VA Populations:
|
Yes. (Ref: 3-5)
|
Scale Application in non-VA Populations:
|
Yes. (Ref: 3,6,8)
|
|
|
Comments
|
The CAPS has over a decade’s worth of published research studies involving its use to attest to the acceptability in psychometric terms of measures obtained from it and to its appropriateness for a wide range of populations, varying in terms of geography, language, and source of originating traumatic stressor. It appears to meet suggested instrument standards for measuring PTSD; that is, (a) its content corresponds with current diagnostic criteria; (b) it provides both dichotomous and continuous data per symptom and disorder; (c) it can be administered by trained paraprofessionals; (d) its measures possess adequate reliability and validity; (e) it employs explicit behavioral anchors; (f) it treats symptom severity as multidimensional, measuring both frequency and intensity; and (g) it has a clear timeframe. (See, for example, the Watson et al standards, Ref: 1)
The CAPS assesses the presence of the 17 core PTSD symptoms as described by the DSM-IV. It assesses these symptoms in terms of both their frequency and their intensity and, in addition, provides a measure of overall severity (i.e., frequency response plus intensity response) per symptom and across symptoms.
In general, there are two classes of issues to be dealt with when considering use of the CAPS. The first class includes issues specific to the CAPS instrument itself. The CAPS is an interview instrument – not a self-reporting one – and must be administered by either a professional health care provider or a trained paraprofessional. Furthermore, whoever administers the CAPS must have experience in working with PTSD-afflicted patients. The interview itself, expected to take some 40+ minutes to conduct, is lengthy, while scoring of the CAPS is perhaps equally lengthy, if not more complex. Results are available per PTSD symptom, per symptom cluster, and for the full set of 17 core PTSD symptoms overall. Because both dichotomous and continuous data are collected per symptom, results are able to indicate 1) the presence or absence of a symptom, and 2) the extent or degree to which a symptom is present. The complexity of the frequency by intensity scoring leads to a sometimes-complex determination of the presence (or absence) of symptoms and PTSD condition. (See, for example, the liberal, moderate, and conservative scoring rules presented in Ref: 6). Although a “questionable validity” rating is available for interviewer use when patient responses are deemed unlikely or unreasonable, it is uncertain how helpful this feature has been in producing “correct” diagnostic assessments. Finally, as suggested above by the existence of multiple scoring rules, there remains an ongoing issue with identifying a single cut-score to use for determination of PTSD diagnosis.
The second class of issues concerns the PTSD construct itself. The CAPS is based on – one could say built around - current PTSD symptomatology as presented by the DSM-IV. If and when PTSD diagnostic criteria change, the CAPS itself will require revision. Such was indeed the case when the DSM-III-R, the original “symptom gold standard” for the CAPS, was replaced by the DSM-IV. PTSD construct change seems more likely than not. Assuredly, questions continue to exist in the field as to a full identification of PTSD and its associated symptoms. For example, are PTSD source exposures (incidents) always from outside the range of usual human experience, or are they sometimes within normal experience, but rather experienced in an abnormally traumatic way? Have all PTSD symptoms been identified (e.g., those originating from traumatic childhood experiences, from sudden incidents of trauma, or from repeated trauma)? Does PTSD only manifest itself via a stable and recognizable pattern of symptoms, or are other patterns and instability possible? Are PTSD symptoms that have been recognized as being “appropriate” for combat veterans equally “appropriate” for other PTSD sufferers? Can a traumatic stressor, the etiologic factor required for a diagnosis of PTSD, always be identified (e.g., after a childhood exposure)? It is clear that, as the PTSD construct itself undergoes change, so must any instrument seeking to measure it.
In addition to the 17 core PTSD symptoms, the CAPS also assesses the presence of what are called associative features (or symptoms) of PTSD (i.e., symptoms outside of the core symptoms but which are nevertheless related to and manifested with the condition). While assessing these associative symptoms has not added directly to an improved understanding of PTSD, it does seem to have been useful for prescribing and then following a process of treatment. In fact, focusing on symptom clusters and their associates for the treatment of PTSD is directly implied by CFA results, which have indicated that a first order 4-factor “symptom cluster” model best represented the study data at hand. PTSD is thus a multidimensional condition; attention to its multiple symptoms, both in assessing and treating them, is not only warranted but may in fact be “best practice.”
In summary, using the CAPS requires a time allotment for a lengthy interview, a professional or paraprofessional experienced with PTSD to administer it, and a detailed if not complex preparation of results. CAPS results, though, are not disappointing, for they can provide useful diagnostic as well as treatment plan information, and on a level of detail ranging from single symptom, to symptom cluster, to overall symptom set. The construct of PTSD, however, appears to be still evolving. PTSD is not just a “combat veteran’s condition” but one that everyday civilians, too, may experience. It may arise from a single determinant event or from a source of trauma spread across time and/or across occurrence. The very face or faces of PTSD may be evolving, and, if so, PTSD measurement instruments will need to evolve, too.
|
|
|
|
