Measurement invariance of the Clinician Administered PTSD Scale for DSM-5 between trauma-exposed non-Western refugee patients and Dutch patients seeking treatment for PTSD

Background: The Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) is the gold standard to assess post-traumatic stress disorder (PTSD). Although the CAPS-5 was developed and validated primarily in Western samples, it is often used to assess PTSD in non-Western refugees. Objective: This study investigated whether PTSD as assessed with the CAPS-5 has the same factor structure, item-factor relationships and symptom severity scores across trauma-exposed, treatment-seeking individuals with either a Dutch (i.e. Western) background or a non-Western refugee background. Methods: We evaluated four DSM-5 PTSD models and tested different levels of measurement invariance in non-Western refugee (nWr) patients (N = 710) and Dutch patients (N = 1162) using confirmatory factor analyses (CFA). Results: CFAs showed that the Hybrid model of PTSD fit the data best in the nWr group, the Dutch group, and in the combined sample, indicating that configural invariance was met. The model representing scalar invariance showed a worse fit compared to the model representing configural invariance, indicating that item-factor relationships and/or thresholds (representing the value above which an individual endorses the next symptom severity category) differed between the subsamples. Further exploration showed that this result could likely be attributed to nWr patients generally receiving higher symptom severity ratings on the CAPS-5 items than Dutch patients with similar underlying symptom severity levels. Conclusions: NWr patients showed a similar factor structure of DSM-5 PTSD symptoms as Dutch patients. Furthermore, the relative contribution of the majority of symptoms to its belonging symptom cluster was similar between nWr patients and Dutch patients, indicating that item-factor relationships were similar across groups. However, nWr patients generally received higher symptom severity ratings than Dutch patients. This is the first study to investigate measurement invariance of the CAPS-5 in a non-Western refugee (nWr) sample.The Hybrid model was the best fitting model in both the nWr sample and the Dutch (i.e. Western) sample and most item-factor relationships were similar between groups.Most item thresholds were lower for nWr patients, indicating that they generally received higher symptom severity ratings on the CAPS-5 items than Dutch patients with similar underlying symptom severity levels, either reflecting a different way of expressing complaints or an actual difference between the groups, e.g. in trauma load. This is the first study to investigate measurement invariance of the CAPS-5 in a non-Western refugee (nWr) sample. The Hybrid model was the best fitting model in both the nWr sample and the Dutch (i.e. Western) sample and most item-factor relationships were similar between groups. Most item thresholds were lower for nWr patients, indicating that they generally received higher symptom severity ratings on the CAPS-5 items than Dutch patients with similar underlying symptom severity levels, either reflecting a different way of expressing complaints or an actual difference between the groups, e.g. in trauma load.