INTRODUCTION
In 2018, the World Health Organization updated the International Classification of Diseases, 11th Revision (ICD-11) to include complex posttraumatic stress disorder (cPTSD) as a separate diagnosis from posttraumatic stress disorder (PTSD), addressing the criticism that PTSD does not fully capture the variety of symptoms and responses to different traumas [1]. This distinction arises from the recognition that certain traumatic experiences lead to additional symptoms beyond those of PTSD. To be diagnosed with either PTSD or cPTSD, one must have directly experienced, witnessed, or been confronted with a traumatic event. Research indicates that exposure to multiple traumatic events increases the risk of PTSD among the general population [2], survivors of organized violence [3], and refugees [4]. The ICD-11 criteria for cPTSD include the core symptoms of PTSD plus three additional symptoms: difficulties in controlling emotions, a negative self-concept, and difficulties in maintaining relationships [1]. Recent studies have suggested a differential association between the type of trauma exposure and the diagnosis of PTSD versus cPTSD [5], with post-migration stressors and social stressors being more predictive of cPTSD than pre-migration or earlier traumas [3,4]. This signifies the complex nature of trauma and its impact on mental health, indicating the need for diagnostic criteria that capture the diverse experiences and symptoms of individuals affected by trauma.
The traumatic experiences of North Korean defectors (NKDs), who have fled their country for South Korea, are notably distinct and more complex compared to those encountered by the general population [6]. These traumas are not isolated incidents but are comprised of numerous, compounded events encountered throughout their journey of escape [7]. Recent findings suggest that the symptoms exhibited by NKDs in relation to potentially traumatic events (PTEs) align more closely with cPTSD than with conventional PTSD [8,9]. Despite this, the Trauma Scale for North Korean Refugees [10] and most commonly used to assess PTSD in NKDs [11], fails to address the core features of cPTSD, focusing instead on pre-migration trauma. To date, the Post-Traumatic Stress Symptom Scale for Displaced North Koreans by Yoon et al. [10] is the only scale that attempts to encompass cPTSD aspects in NKDs, yet it too is based on the criteria set forth in Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), indicating a gap in accurately diagnosing and understanding the full scope of trauma experienced by NKDs.
To effectively address the trauma-related mental health needs of NKDs, it is essential to adopt a comprehensive framework that encompasses pre-migration trauma, migration-related adversities, and post-migration stressors. The International Trauma Questionnaire (ITQ), recently validated for NKD) [12], is an effective tool for distinguishing PTSD and cPTSD symptoms in accordance with ICD-11 criteria. However, its section on PTEs lacks the cultural and contextual specificity necessary to accurately reflect NKDs’ lived experiences.
Existing trauma assessment tools developed for NKDs predominantly focus on pre-migration events and do not fully capture the expanded symptomatology and trauma timeline relevant to cPTSD. NKDs typically endure a series of cumulative and interrelated traumas across all phases of migration—ranging from extreme deprivation, political persecution, and forced labor in North Korea to life-threatening escape journeys and social exclusion in South Korea. These unique experiences necessitate the development of a tailored assessment tool that reflects both the nature and intensity of trauma specific to this population.
By refining the PTEs section of the ITQ to incorporate culturally relevant and migration-specific events, this study aims to improve the accuracy and clinical utility of trauma assessments for NKDs. A revised checklist that systematically documents trauma across temporal phases will serve as a critical foundation for targeted interventions, advancing trauma-informed mental health care for this vulnerable group.
METHODS
Objectives
The objective of this study was to develop a reliable and valid checklist designed to capture the traumatic experiences of NKDs. This study was approved by the Institutional Review Board of Severance Hospital (IRB No. 4-2021-0897).
Study design
The cross-sectional, self-administered questionnaire survey was conducted to evaluate the psychometric properties of the Trauma Experience Checklist developed for NKDs.
Participants
This study represents a secondary analysis of data collected in 2021 through an online cross-sectional survey targeting NKDs [12]. Participants were recruited through Naver Band and KakaoTalk chatrooms for NKDs, as well as offline counseling centers. Participant eligibility was verified by confirming NKD status through self-reported information on year of defection, year of arrival in South Korea, and Hanawon (resettlement support center) completion cohort. Trauma exposure was assessed through the survey questionnaire. Quality control measures included monitoring response patterns and completion times; responses were excluded for completion times <10 minutes (n=8) and duplicate participation (n=9). The final sample consisted of 520 participants, who were randomly divided into two equal groups (n=260 each) for analysis. One group was used for exploratory factor analysis (EFA), and the other for confirmatory factor analysis (CFA). All statistical analyses were conducted using SPSS Statistics 28.0 and AMOS 28.0 (IBM Corp.). The sociodemographic characteristics of the participants are summarized in Table 1.
Preliminary PTEs checklist
The 22-item Traumatic Event Checklist for NKDs was developed through a systematic multi-stage process to ensure cultural relevance and comprehensiveness for the NKD population.
First, we conducted a comprehensive literature review of existing studies documenting traumatic experiences among NKDs [12]. This review identified both universal traumatic events experienced across displaced populations (e.g., natural disasters, accidents, interpersonal violence) and experiences specific to the NKD population.
Second, in-depth interviews with NKDs (n=21) identified additional traumatic experiences not fully captured in existing literature, including betrayal (e.g., being reported to authorities, fraud after resettlement) and discrimination (e.g., marginalization based on social class in North Korea, stigmatization in South Korea).
Third, the preliminary item pool underwent a two-stage review process. Three experts with extensive professional experience working with the NKD population (a psychiatrist and two NKD-specialized researchers) evaluated each item for cultural appropriateness, relevance, clarity, and comprehensiveness. Subsequently, five NKDS reviewed the items to assess their comprehensibility and ensure the wording was accessible to the target population.
The resulting 22-item checklist reflects both general life events and culturally specific experiences relevant to the NKD population. General items include natural disasters, transportation accidents, serious illness or injury, and interpersonal violence. NKD-specific items capture experiences such as witnessing public executions, experiencing political persecution or imprisonment, forced labor, human trafficking or exploitation during escape, and forced repatriation from third countries.
Each item was phrased as either “Have you ever experienced~?” or “Have you ever witnessed~?” and included concrete examples (e.g., natural disasters such as earthquakes, typhoons, floods, and droughts). A distinctive feature of this measure is that for each endorsed event, participants were asked to specify whether the event occurred in North Korea, a third country during their escape journey, or after resettlement in South Korea. To assess the perceived severity of each traumatic experience, the survey was programmed with skip logic: participants who indicated no exposure to a specific traumatic event were not presented with the distress rating question for that item. For events to which participants were exposed, they rated the emotional impact on an initial 4-point scale (0=“not at all,” 1=“slightly painful,” 2=“very painful,” 3=“extremely painful”). To preserve the conceptual distinction between non-exposure and exposure without distress, responses were subsequently recoded to a 5-point scale: 0=“not applicable” (no exposure), 1=“not at all” (exposure with no distress), 2=“slightly painful,” 3=“very painful,” 4=“extremely painful.” The “not applicable” coding was assigned to all items for which participants indicated no exposure, distinct from “not at all,” which indicated that the event had occurred but was not perceived as emotionally painful. Total trauma scores ranged from 0 to 88 (based on the initial 22 items), with higher scores indicating greater cumulative exposure and severity of traumatic events.
Measures
The survey instrument consisted of the preliminary PTEs checklist described above, developed using a self-report format. Additional validated tools were administered to assess criterion validity, including the DSSS, and the World Health Organization of Life Assessment Instrument Bref (WHOQOL-BREF) for quality of life.
General characteristics of study participants
General characteristics such as age, sex, employment status, and educational level of the study participants were collected.
DSSS
The DSSS is a tool for evaluating depression and somatic symptoms, developed by Hung et al. [13]. In this study, the Korean version of DSSS adapted by Kim et al. [14] was used. This measurement includes total 22 items. It consists of 12 items for depressive symptoms (depression subscale; DSSS-D) and 10 items for somatic symptoms (somatic subscale; DSSS-S). Items are rated on a 0 to 3 Likert scale, and the higher the score, the more severe the symptom experience. Cronbach’s alpha coefficient of DSSS was 0.92 when it was developed and 0.95 in this study.
Quality of life
The Korean version of the WHOQOL-BREF standardized by Min et al. [15] is used. This tool consists of 26 questions in 5 domains, including physical health domain (7 questions), psychological domain (6 questions), social relationship domain (3 questions), environment domain (8 questions), and subjectively perceived quality of life It consists of two areas (2 items). Each item is rated on a 5-point Likert scale ranging from 1 “not at all” to 5 “strongly agree.” The higher score means a better quality of life. Cronbach’s alpha coefficient of WHOQOL-BREF was 0.89 in this study.
Statistical analysis
All statistical analyses were performed using SPSS Statistics 28.0 and AMOS 28.0. The sample of 520 participants was randomly divided into two groups of 260 each for independent EFA and CFA.
Descriptive statistics were used to summarize sociodemographic and clinical characteristics. Construct validity was assessed through a two-step process. EFA was first conducted using Group 1 data with principal component analysis and varimax rotation. Sampling adequacy was confirmed using the Kaiser-Meyer-Olkin (KMO) measure and Bartlett’s test of sphericity. A KMO value greater than 0.50 and a significant Bartlett’s test (p<0.05) indicated suitability for factor analysis. Items with commonality values below 0.40 were considered for deletion [16]. Factor retention was determined based on eigenvalues, scree plot, and parallel analysis.
To confirm the suitability of the model derived by EFA, CFA was performed with the data of Group 2. In CFA, model fit was evaluated using the following indices: χ2 statistic, normed χ2 (CMIN/DF), root mean square error of approximation (RMSEA), goodness-of-fit index (GFI), Tucker-Lewis index (TLI), comparative fit index (CFI), and incremental fit index (IFI). RMSEA values below 0.08 indicate acceptable fit, and below 0.05 indicate good fit. GFI, TLI, CFI, and IFI values of 0.70 or above indicate acceptable fit, with 0.90 or above indicating good model fit [17].
Criterion validity was examined using Pearson correlation coefficients between PTEs scores and DSSS and WHOQOL-BREF scores. Internal consistency reliability for all instruments and their subscales was evaluated using Cronbach’s alpha (α).
RESULTS
General characteristics
A total of 520 NKDs participated in the study, divided evenly into Group 1 (n=260) and Group 2 (n=260). The mean age was 46.33 years (SD=13.40) for Group 1 and 45.75 years (SD=13.43) for Group 2, with no significant differences across sociodemographic variables such as sex, employment status, education level, or religion. Approximately 50% of participants in both groups were employed, and the majority had completed high school or higher education in North Korea. Most participants had not experienced forced repatriation to North Korea. The majority entered South Korea between 2000 and 2010. A higher proportion lived alone rather than with a spouse. There were no statistically significant group differences in psychiatric treatment or monthly income. However, significant group differences were observed in three domains of quality of life: overall QOL (p=0.016), psychological wellbeing (p=0.033), and social relationships (p=0.011) (Table 1).
Study 1: EFA
To categorize the types of traumatic experiences reported by NKDs, an EFA was conducted using data from Group 1 (n=260). Item responses on the PTEs Checklist were found to be relatively evenly distributed on a 4-point Likert scale, with item means ranging from 0.39 to 2.55 (Table 2). Following Cramer and Barry’s16 recommendation, items with a communality value below 0.40 were considered for removal. Based on this criterion, the item “caring for a family member with a serious illness” (T18) was excluded due to a commonality of 0.286.
The remaining 21 items were subjected to EFA using principal components analysis with varimax rotation. Sampling adequacy was confirmed (KMO=0.819), and Bartlett’s test of sphericity was significant (χ2=1,462.99, p<0.001), supporting the factorability of the correlation matrix. During the analysis, three additional items—“life-threatening illness” (T7), “divorce of parents (in childhood)” (T14), and “betrayal” (T19)—exhibited factor loadings below 0.40 and were excluded.
EFA was then rerun with the final 18 items. The adjusted KMO value was 0.800 and Bartlett’s test remained significant (χ2=1,289.41, p<0.001), indicating continued suitability for factor analysis. The analysis yielded a four-factor solution, explaining a total of 52.02% of the variance. The four factors had eigenvalues of 4.186, 1.661, 1.624, and 1.248, accounting for 24.63%, 9.77%, 9.55%, and 7.34% of the total variance, respectively. Visual inspection of the scree plot indicated a clear inflection point after the fourth factor, and parallel analysis further supported the retention of a four-factor solution, as the observed eigenvalues for the first four factors exceeded those obtained from randomly generated data. The percentage of variance explained by each factor was as follows: Factor 1 (15.64%), Factor 2 (12.81%), Factor 3 (12.66%), and Factor 4 (10.90%). The four factors were conceptually labeled based on thematic similarity as: “death and life-threatening events,” “interpersonal violence and discrimination,” “violence by public authorities,” and “family dissolution.” based on the common themes emerging from the grouped items. This naming convention aims to elucidate the shared characteristics detected within each factor (Table 3).
Study 2: CFA
CFA was conducted using Group 2 (n=260) to validate the four-factor model identified in EFA (Figure 1).
Initial analysis indicated acceptable fit: RMSEA=0.061, GFI=0.911, TLI=0.895, CFI=0.916, and IFI=0.918. However, one item: “physical/emotional abuse and neglect by parents (in childhood)” (T13)—had a factor loading of 0.28 and was subsequently removed. After re-estimating the model with 17 items, model fit indices improved: RMSEA=0.059, GFI= 0.930, TLI=0.916, and CFI=0.930, indicating strong structural validity. Notably, the item “sexual assault” (T22) demonstrated stable factor loadings across both CFA models, with a standardized loading of 0.41 in the initial 18-item model and 0.41 in the final 17-item model following the removal of T13. This marginal increase indicates that T22’s contribution to Factor 4 (family dissolution) was minimally affected by the removal of T13, confirming its psychometric stability across model specifications.
Because the natural disaster item (T1) showed cross-loading on Factor 1 and Factor 2 in the EFA, additional CFA models were tested: one assigning T1 to Factor 2 and another excluding T1 entirely. Although the model excluding T1 demonstrated comparable model fit, T1 showed an adequate standardized loading (0.57) on the death and life-threatening events factor in the original CFA model. Given its conceptual relevance to life-threatening experiences and the aim of preserving content validity, T1 was retained in the final model.
Internal consistency
The Cronbach’s alpha value was 0.830 for total of 17 items. The 0.6 is often cited as the criterion value [18]. However, when examining the subscales, the alpha values varied, with the subscales for “death and Life-threatening events,” “interpersonal violence and discrimination,” and “violence by public authorities” recording satisfactory alpha levels of 0.791, 0.759, and 0.665, respectively. These values reflect a commendable level of internal consistency within these thematic areas. Conversely, the “family dissolution” subscale presented a lower alpha of 0.589, falling short of the conventional acceptability threshold.
Convergent and discriminant validity
Convergent and discriminant validity were assessed using Pearson’s correlation coefficients (Table 4). The four trauma subscales were significantly correlated with DSSS somatic and depression scores, as well as with the physical, psychological, and social domains of the WHOQOL-BREF (p<0.001), supporting convergent validity. In contrast, the correlations between trauma subscales and the environmental domain of QOL were weak or non-significant, thus supporting discriminant validity and confirming that the trauma constructs are conceptually distinct from environmental quality-of-life indicators.
DISCUSSION
The PTEs checklist, which initially consisted of 22 items, was refined to 17 items based on the results of exploratory and confirmatory factor analyses. This study demonstrates the strong structural validity and acceptable internal consistency of the revised 17-item checklist, which captures four core domains: “death and life-threatening events,” “interpersonal violence and discrimination,” “violence by public authorities,” and “family dissolution.” This checklist includes items that extend beyond DSM-5 trauma criteria but were consistently reported by NKDs as traumatic, particularly post-migration experiences such as betrayal, discrimination, and verbal abuse. While these may be debated as traumatic events versus exacerbating factors, the ICD-11 framework supports their inclusion by recognizing a broader range of potentially traumatic experiences, including severe discrimination and interpersonal betrayal. By incorporating these culturally relevant experiences, the revised checklist provides a more comprehensive assessment of trauma exposure among NKDs that can inform culturally tailored interventions [19].
In our analysis of internal consistency, the subscales for “death and life-threatening events,” “interpersonal violence and discrimination,” and “violence by public authorities” demonstrated acceptable reliability, with alpha values of 0.791, 0.759, and 0.665, respectively. In contrast, the “family dissolution” subscale recorded an alpha value of 0.589, below conventional thresholds. Future work should evaluate whether specific items affect reliability and whether additional items relevant to family dissolution may strengthen this subscale [20].
Items such as divorce and discrimination reflect experiences that, in the NKD context, carry consequences far exceeding those in the general population—family dissolution often resulting from forced separation or disappearance under life-threatening circumstances, and discrimination compounded by profound social marginalization—such that their subjective experience may well reach the threshold of extreme threat described by ICD-11. Indeed, the ICD-11 framework for cPTSD itself has been critiqued regarding the breadth of qualifying stressors, with some scholars arguing that cumulative and chronic adversities—even those not traditionally classified as “catastrophic”—can produce cPTSD symptomatology [1]. Our checklist does not claim that all included items definitively meet the ICD-11 Criterion A threshold for every individual; rather, it captures a comprehensive range of PTEs relevant to this population, whose subjective appraisal and context must be taken into account.
This is particularly relevant for the discrimination items (T20, T21), which do not meet DSM-5 Criterion A but correspond with ICD-11’s broader trauma framework. South Korean society’s pervasive discrimination is especially salient for NKDs who anticipate ethnic solidarity upon arrival, and its cumulative psychological impact may constitute traumatic stress as defined by ICD-11 [21]. Beyond item selection, the measurement approach itself merits note. PTEs were evaluated as a continuous variable capturing frequency and intensity rather than simple presence or absence, providing a more nuanced trauma profile. This approach ties to the concept of psychological resilience [22], and enables prioritization of those requiring immediate or intensive intervention.
The traumatic experiences captured in this checklist reflect both universal and NKD-specific dimensions of refugee trauma. While items such as witnessing death, physical violence, and discrimination are widely documented across refugee populations, experiences such as political persecution, forced labor, public executions, and forced repatriation are deeply embedded in the North Korean context. Comparative studies across refugee populations would help clarify which components of this checklist may be adapted for broader use.
Several limitations should be acknowledged. First, the subjective nature of factor labeling may have introduced classification ambiguities, as the interpretation of empirically derived factors inevitably involves researcher judgment. Second, recruitment via social networking platforms and offline counseling centers may not fully represent the NKD community, potentially excluding individuals without digital access or those less connected to NKD networks. The sample was further limited to adults aged 19 and older, excluding children and adolescents whose trauma profiles may differ substantially. Third, this study provides preliminary validity evidence only; convergent and discriminant validity with established PTSD and cPTSD measures remain to be established. Fourth, this study did not include a validated cPTSD symptom measure, such as the ITQ, limiting our ability to examine the relationship between trauma exposure and cPTSD symptomatology. Future studies should administer this checklist alongside established PTSD and cPTSD instruments to explore how different types of traumatic experiences relate to distinct trauma-related outcomes in this population.
Given these limitations, it should be noted that the total 17-item checklist demonstrated strong internal consistency (Cronbach’s α=0.830), exceeding the conventional threshold of 0.70. However, two subscales—”violence by public authorities” (α=0.665) and “family dissolution” (α=0.589)—fell below this threshold. We therefore recommend using the checklist in its entirety rather than relying on individual subscales. In particular, the family dissolution subscale warrants further investigation—including evaluation of whether existing items should be revised and whether additional items relevant to family dissolution experiences among NKDs would strengthen its reliability and content coverage.
Future research should employ diverse recruitment strategies to reach underrepresented subgroups, including younger populations and those with limited digital access. Comprehensive validity studies are needed to establish the full psychometric profile of this instrument and to examine which specific traumatic events are most strongly associated with ICD-11 PTSD and cPTSD diagnoses in this population.
In conclusion, this study developed a culturally adapted checklist of PTEs specifically for NKDs and provided preliminary evidence of its validity. The final 17-item tool demonstrated strong structural validity and acceptable internal consistency across most domains. Notably, it includes culturally and contextually relevant events, including discrimination and verbal abuse, that align with ICD-11 definitions of trauma but are often overlooked in standard diagnostic frameworks.
By adopting a multidimensional and continuous measurement approach, the checklist provides a more nuanced understanding of trauma exposure and its severity. This enables the identification of individuals at greater psychological risk and supports the development of tailored mental health interventions. Although the tool demonstrates promising psychometric properties, further refinement, particularly of the “family dissolution” subscale, and broader validation across diverse NKD populations are warranted.
The revised PTEs checklist serves as a practical and theoretically grounded tool applicable in both clinical and research contexts. It contributes to improving the assessment of trauma and supports the delivery of culturally responsive care for a highly vulnerable and underrepresented population.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
