The use of culturally adapted and translated depression screening questionnaires with South Asian haemodialysis patients in England
Sharma S., Norton S., Bhui K., Mooney R., Bansal T., Day C., Davenport A., Duncan N., Kalra P., Da Silva- Gane M., Randhawa G., Warwick G., Wellsted D., Yaqoob M., Farrington K.
Background: Depression is common amongst patients receiving haemodialysis (HD). Assessment and intervention when faced with language and cultural barriers is challenging. To support clinician decisions, we assessed the use of culturally adapted and translated versions of commonly-used depression screening questionnaires with South Asian patients receiving HD in England. Methods: Patients completed adapted versions of the Patient Health Questionnaire (PHQ-9), the Centre for Epidemiological Studies Depression Scale Revised (CESD-R), and the Beck Depression Inventory II (BDI-II). All questionaries were available in Gujarati, Punjabi, Urdu, and Bengali. A comparative sample of white-Europeans completed the questionnaires in English. The research was based across 9 National Health Service (NHS) Trusts in England. Structural validity of translated questionnaires was assessed using confirmatory factor analysis. Diagnostic accuracy was explored in a subgroup of South Asians against ICD-10 categories using the Clinical Interview Schedule Revised (CIS-R) with receiver operating curve (ROC) analysis. Results: 229 South Asian and 120 white-European HD patients participated. A single latent depression factor largely accounted for the correlations between items of the PHQ-9, CESD-R and BDI-II. Issues with measurement equivalence implied that scores on the translations may not be comparable with the English language versions. Against CIS-R based ICD-10 diagnosis of depression, sensitivity was modest across scales (50-66.7%). Specificity was higher (81.3-93.8%). Alternative screening cut-offs did not improve positive predictive values. Conclusions: Culturally adapted translations of depression screening questionnaires are useful to explore symptom endorsement amongst South Asian patients. However, data indicate that standard cut-off scores may not be appropriate to classify symptom severity. Use of the CIS-R algorithms for optimal case identification requires further exploration in this setting. Strategies to encourage recruitment of under-represented groups in renal research are also warranted, especially for in-depth discussions related to psychological care needs.