Psychometric properties of the Korean version of the Short Post-Traumatic Stress Disorder Rating Interview (K-SPRINT)
Jeong-Ho Chae, MD, PhD, Department of Psychiatry, St Mary's Hospital, The Catholic University of Korea, College of Medicine, 62 Yoido-dong, Youngdeungpo-gu, Seoul 150-713, Korea. Email: email@example.com
Aims: The Short Post-traumatic Stress Disorder (PTSD) Rating Interview (SPRINT) is a validated, eight-item, brief global assessment scale for PTSD. This report investigated the psychometric properties of the Korean version of the SPRINT (K-SPRINT).
Methods: Eighty-seven PTSD patients, 47 other psychiatric patients, and 63 healthy control subjects were enrolled in the study. All subjects completed a psychometric assessment package that included the K-SPRINT and the Korean versions of the Clinician-Administered PTSD Scale (CAPS), the Beck Depression Inventory (BDI), and the State Trait Anxiety Inventory (STAI).
Results: The K-SPRINT showed good internal consistency (Cronbach's α = 0.86) and test–retest reliability (r = 0.82). K-SPRINT showed moderatecorrelations with CAPS (r = 0.71). An exploratory factor analysis produced one K-SPRINT factor. The optimal diagnostic efficiency (91.9%) of the K-SPRINT was found at a total score of 15, at which point the sensitivity and specificity were 90.8% and 92.7%, respectively.
Conclusions: The present findings demonstrate that the K-SPRINT had good psychometric properties and can be used as a reliable and valid instrument for the assessment of PTSD.
POST-TRAUMATIC STRESS disorder (PTSD) is a syndrome that develops after a person has expe rienced or witnessed a life-threatening traumatic stressor.1 It is characterized by the re-experiencing of the traumatic event accompanied by symptoms of increased arousal and by the avoidance of stimuli associated with the event.2
Since PTSD was first introduced in DSM-III3 a number of interviewer-administered scales have been developed for the assessment of the symptoms of PTSD. These structured interviews include the Structured Clinical Interview for DSM-III-R,4 the PTSD Interview,5 the Clinician-Administered PTSD Scale (CAPS),6 the Structured Interview for PTSD,7 and the Post-traumatic Stress Scale.8 All of these scales show good psychometric reliability and validity and are comprehensive in assessing the core symptoms of PTSD. However, a major drawback of these structured interviews is that they are time-consuming, which may be a clinical problem when the scales are frequently or urgently filled out.
The Short PTSD Rating Interview (SPRINT), which was developed by Conner and Davidson,9 is a validated eight-item interview-based scale used to evaluate the severity of the PTSD symptoms. The SPRINT items include questions assessing the core symptoms of PTSD, as well as related aspects of somatic malaise, stress vulnerability, and functional impairment. The original SPRINT has been proven to have good internal consistency (Cronbach's α = 0.77) and test–retest reliability (intraclass correlation coefficient [ICC]: 0.78). The major strengths of the SPRINT are that it takes only 5–10 min to complete on average; its short but comprehensive coverage of the major domains of concern in PTSD; and its consideration of the important features associated with PTSD.
The aim of the present study was to develop a Korean version of the SPRINT (K-SPRINT) through the translation of the original English version into Korean while maintaining its basic structure. Another aim was the elucidation of the psychometric properties of the K-SPRINT by assessing its reliability and validity and to determine its usefulness in evaluating patients with PTSD.
The subjects consisted of 87 PTSD patients, 47 other non-psychotic psychiatric patients (non-PTSD patients), and 63 healthy controls. The patients and controls were recruited from 18 institutes in all provinces and territories of Korea. All of the subjects were between 18 and 65 years of age. PTSD and other psychiatric disorders were diagnosed by the Korean version10 of MINI International Neuropsychiatric Interview11 using the criteria of the DSM-IV. The diagnoses of non-PTSD patients included major depressive disorder (n = 24), panic disorder (n = 10), generalized anxiety disorder (n = 5), phobic disorder (n = 3), adjustment disorder (n = 3), and undifferentiated somatoform disorder (n = 2). Each of the healthy controls demonstrated that they did not have a lifetime history of psychiatric and medical disorders in a semi-structured interview. Each of the subjects provided informed consent for their participation in the present study after the procedure had been fully explained and the institutional review board approved the study.
The SPRINT consists of four items that correspond to each of the four PTSD symptom clusters (intrusion, avoidance, numbing, and hyperarousal), as well as four additional questions to assess somatic distress, being upset by stressful events, interference with work or daily activities, and relationships among family and friends, respectively. Each item consists of a 5-point Likert scale from ‘not at all’ (0) to ‘very much’ (4). The maximum score is 32, which would indicate the worst PTSD symptom state. The SPRINT also contains two additional questions to measure global improvement by percentage change and severity rating. The K-SPRINT was developed with the permission of the original author, using conventional techniques of translation and back-translation by bilingual psychiatrists, maintaining equivalence. The K-SPRINT was modified until the back-translated version was comparable with the original English version.
The CAPS is a comprehensive, psychometrically sound, structured clinical interview designed to assess adults for the 17 symptoms of PTSD outlined in the DSM-IV along with five associated features (guilt, dissociation, derealization, depersonalization, and reduction in awareness of surroundings). It consisted of CAPS-1 and CAPS-2, which were designed to assess the current or lifetime PTSD status and PTSD symptoms experienced during the previous week, respectively. The Korean version of CAPS had an excellent internal consistency of 0.95.12 In the present study we used only CAPS-2, in which the assessment period corresponded to that of SPRINT.
The Beck Depression Inventory (BDI), a 21-item self-administered questionnaire, was developed to assess the severity of subjective depressive symptoms.13 The Korean version of the BDI demonstrated good psychometric properties, and its internal consistency coefficient was reported to be 0.85.14
The State–Trait Anxiety Inventory (STAI), which is a self-reporting questionnaire, was designed to evaluate the severity of the anxiety symptoms and is composed of 20 questions for anxiety and 20 questions for trait anxiety.15 The anxiety state refers to the type of anxiety felt at the time the subject is completing the questionnaire, and the term ‘trait anxiety’ refers to the type of anxiety felt in general. The Korean version of STAI was previously shown to exhibit excellent psychometric properties and its internal consistency was reported at Cronbach's α = 0.91.16 The BDI and STAI were selected to prove the divergent validity of the SPRINT, which indicated that the results obtained by this instrument do not correlate too strongly with measurements of a similar but distinct trait.
The raters in the present study were seven experienced board-certified psychiatrists who participated in formal consensus meetings concerning the use of the K-SPRINT and the Korean version of CAPS. The consensus meetings consisted of an observation of the administration of the evaluations by an experienced supervisory psychiatrist and an actual administration via videotapes featuring two standard PTSD patients each time. All three meetings were held in the study period and the average attendance rate was 90.5%. The interrater reliabilities of the K-SPRINT and the Korean version of CAPS were high with intra-class correlation coefficients of 0.88 and 0.76, respectively.
Group comparisons were performed using analysis of variance (anova) and χ2 test to compare the quantitative and categorical variables, respectively. Internal consistency was assessed using Cronbach's α. Test–retest reliability was calculated by means of the ICC. The interrater reliability was also analyzed on the basis of ICC. The convergent and divergent validity between the K-SPRINT and other measures were evaluated using Pearson correlation coefficients. An anova and Tukey's post-hoc test were applied to identify the between-group difference by the severity measured by the K-SPRINT. The factorial validity of the K-SPRINT was examined using an unrotated principal components factor analysis in the PTSD patients. The sensitivity, specificity, predictive values, and diagnostic efficiency were calculated according to the standard formulae. We analyzed a receiver operating characteristic (ROC) curve to obtain the optimal cut-off score to detect PTSD in all subjects. The significance level was set at P < 0.05. All statistical analyses were conducted using SAS (Statistical Analysis system) version 9.1 (SAS Institute Inc., Cary, NC, USA).
Demographics and clinical characteristics
The mean ages of the PTSD patients, non-PTSD patients and healthy controls were 42.1 ± 13.3 years), 43.0 ± 11.7 years, and 40.9 ± 15.5 years, respectively. The numbers of men in the three groups were 36 (41.2%), 21 (44.7%), and 38 (60.3%), respectively. No significant differences were found in the age (F = 0.325, P = 0.723) or gender ratio (χ2 = 3.372, P = 0.185) among the three groups.
The mean duration of symptoms in the PTSD patients was 4.9 ± 10.0 years (range, 0.5–41.0 years). The worst experienced traumas in the PTSD patients were accident (n = 54), physical abuse (n = 22), sexual abuse (n = 6), and military combat (n = 5).
Cronbach's α was used to evaluate the internal consistency of the K-SPRINT in 87 PTSD patients. The coefficient was 0.86 at baseline for the eight items in K-SPRINT. Based on the criterion of 0.30 as an acceptable corrected item-total correlation,17 all eight items performed adequately (range, 0.55–0.78).
The test–retest reliability was examined by comparing the baseline K-SPRINT score with a K-SPRINT assessment performed 2 weeks later. The PTSD patients included only those who claimed that there was no change in their PTSD symptoms and agreed to a second K-SPRINT assessment. Among the 87 PTSD patients, 51 subjects were recruited for the evaluation of the test–retest reliability. The test–retest reliability was determined to be 0.82 (P < 0.0001).
The total scores ± SE of K-SPRINT in the PTSD and the non-PTSD patients and controls were 22.0 ± 0.6, 7.6 ± 1.1, and 1.2 ± 0.3, respectively. These values significantly varied on one-way anova (overall F = 266.72, P < 0.0001). The Tukey's post-hoc test showed that there were significant differences among the three groups. These results showed the good construct validity of the K-SPRINT.
Pearson's correlations among K-SPRINT, CAPS, BDI, and STAI in the PTSD patients are presented in Table 1. Convergent validity was proven with significant correlations between the total K-SPRINT and the weekly CAPS (n = 47). Divergent validity was demonstrated with a lack of significance in the correlations between the total K-SPRINT and the state and trait anxiety of the STAI (n = 87). A significant but intermediate correlation between the total K-SPRINT and the BDI (n = 87) suggested that the patients with PTSD might have depressive symptoms as major comorbidity, although symptoms of depression and PTSD were distinct categories.
Table 1. Pearson's correlations for PTSD measures
|CAPS||0.71*|| || || |
|BDI||0.62*||0.30|| || |
Exploratory factor analysis of the K-SPRINT in PTSD patients (n = 87) produced only one factor with an eigenvalue of 4.09, accounting for 51.1% of the total variance. Item loading on this factor ranged from 0.64 to 0.75.
An ROC analysis was conducted to obtain the cut-off score on the K-SPRINT assessment for the diagnosis of PTSD. The sensitivity, specificity, predictive values, and diagnostic efficiency were calculated for every possible K-SPRINT cut-off score. Table 2 shows 11 different threshold scores and their corresponding sensitivity, specificity, positive and negative predictive values, and diagnostic efficiency. The highest diagnostic efficiency was found at a total score of 15, at which point the sensitivity and specificity were 90.8% and 92.7%, respectively.
Table 2. Efficiency of the K-SPRINT for 11 possible cut-off scores
The area under the ROC curve (AUC) is an overall index of the accuracy of the discrimination provided by the K-SPRINT. The AUC ranges from 0 to 1, with AUC >0.85 generally considered to be an indication of good diagnostic ability.18 The AUC of the K-SPRINT is 0.957 and its standard error is 0.015 (P < 0.0001). This AUC value confirmed the excellent diagnostic ability of the K-SPRINT.
The sprint, which was developed as an interview-based scale assessing the severity of the core symptoms of PTSD as well as the related aspects of somatic malaise, stress vulnerability, and functional impairment, demonstrated good reliability in the original English version. The Cronbach's α and test–retest reliability of the SPRINT were 0.77 and 0.79, respectively, and those of the K-SPRINT were 0.86 and 0.82, respectively. The Cronbach's α of the K-SPRINT was within the optimal range considering that an ideal value of α should be between 0.70 and 0.90.19 The test–retest interval in the present study was 2 weeks. In clinical settings, longer test–retest intervals mean greater possibilities of symptom changes. Most of the PTSD subjects assessed in the present study were chronic types whose mean duration of symptoms was 4.9 years, and all of the subjects claimed that there were no changes in their symptoms. However, some subjects refused to enter the retest due to painful experiences with individual stressors. Consequently, 51 PTSD subjects (58.6%) were enrolled in the test–retest study. The result of the test–retest reliability was good. Thus, the K-SPRINT showed good reliability as compared with the original SPRINT.
We compared the K-SPRINT total score with the weekly CAPS total score in order to assess the convergent validity of the K-SPRINT with another PTSD rating scale. The CAPS is widely used for the diagnosis of PTSD and is often considered the gold standard for the assessment of PTSD.20 The correlation coefficient between K-SPRINT and CAPS demonstrated a strong correlation (r = 0.71, P < 0.0001), which was comparable to the previous result (Spearman's correlation coefficient, 0.781; P = 0.0003) from the Davidson group who developed the original SPRINT.21 Thus, K-SPRINT as well as the original SPRINT might be more ideal for use in routine clinical or research assessments of PTSD considering that they performed similarly to the CAPS and took less time to administer than the CAPS.
In the present study the best diagnostic efficiency was 91.9% when the cut-off score of the K-SPRINT was 15. Despite the difference in the sample selection, this result is similar to the result of the original SPRINT, suggesting the best efficiency of 96% when the cut-off scores are in the range of 14–17. The determination of the optimal cut-off point for the diagnosis of a disorder of concern is an invaluable process in the validation study. However, the best cut-off point to achieve optimal sensitivity and specificity may have considerable variances from one clinical setting to another.22 The differences in the standard to define a case, the particular disorders included in the cases, the timing of the interview relative to the administration of the screening, the prevalence and severity of the disorder in the studied population and socioepidemiological variables such as age, gender, and education could be included as possible sources of variation in the threshold.23 There were also some variations between different cultures and even between different institutes within the same culture.24 Therefore, the optimal cut-off point for the scale may not be fixed, in accordance with various situations.
The present findings must be cautiously interpreted considering the following limitations. Even though the present PTSD subjects experienced various traumatic events including accidents, sexual abuse, physical abuse, and military combat, the sample sizes of the various categories were relatively small, thus making it difficult to draw any definite conclusions. In addition, most of the healthy controls and non-PTSD patients did not experience a traumatic event strictly according to DSM-IV diagnostic criterion A for PTSD. Instead, they recollected the most distressing traumatic event past 1 year and interviewed with a rater. Further studies will be needed to apply this assessment tool to the same traumatic event-related subjects. Finally, the possibility that the PTSD patients might also have had other comorbid psychiatric disorders was not considered. Comorbidity rates are high among PTSD patients, with approximately two-thirds having at least two other disorders.1
In conclusion, the present findings demonstrate that the K-SPRINT had good psychometric properties and can be used as a reliable, valid, and timesaving tool to diagnose and assess PTSD. Further studies are needed to fully evaluate the K-SPRINT, including research applying it to the general population and in a primary medical setting where the prevalence of PTSD is relatively low.
This study was supported by a grant of the Korean Academy of Anxiety Disorders, Korean Neuropsychiatric Association, and Korean Health 21 R & D Project, Ministry of Health and Welfare (A060273), and the Specific Research Program (M10644000013–06 N4400-01310), Ministry of Science and Technology, Korea.