UC Berkeley

Chinese American immigrants have one of the lowest rates of mental health services utilization among minority populations in the United States. With regard to depression, this disparity is associated with stigma, insufficient knowledge of depression and its treatment, help seeking in medical and community settings, and misdiagnosis. From the viewpoint that the manifestation of depression varies culturally, this research, composed of three studies, used an emic approach combined with item response analysis to explore cultural interpretations and expressions of distress, and their influence on assessment.

The first study investigated manifestations of social self-construal in the illness narratives of eleven men and nine women diagnosed with major depression. These narratives revealed a range of emotional, cognitive, and behavioral experiences that contained a social interpretation of the self. Also, psychological concepts of anxiety, irritability, and emotional control referred directly to a social self. This study demonstrated the social content in lived experiences of illness.

The second study examined the structure of depression. Depressed outpatients (n = 103) and community members (n = 124) completed a 47-item pilot scale containing Western and culture specific symptoms and a 21-item checklist of somatic symptoms. A multidimensional item response model of 59 items in three dimensions--psychological, somatic, and social--provided optimal fit of the data. Various patterns of dimensional expression within and across levels of overall distress were observed. Also, analysis of differential item functioning (DIF) showed significant influences of gender, age, education and acculturation on different areas of content. This study implicated a dimensional construct in illness presentation and progression.

The third study developed a self-report measure for screening in primary care and community settings. Based on responses to the 47-item pilot scale, 42 items fit a unidimensional polytomous item response model. Analysis of DIF by gender, age, and acculturation variables revealed non-equivalent items. A 9-item instrument was composed of items unbiased by gender and with content validity for the purpose of screening. This measure showed excellent reliability coefficients, and similar predictive validity but higher convergent and content validity than a commonly used Western measure. This study highlighted a role for culturally sensitive content for effective outreach and assessment.