Methods: Data came from the National Immigrant Survey (NIS) 2003 cohort, a multi-cohort prospective-retrospective panel study of 8,573 new legal immigrants in the US. The final sample consisted of 2,119 Asian immigrants belonging to one of three groups: Southeast Asians (Philippine and Vietnamese, 35%), East Asians (Chinese and Korean, 29%), and Indians (37%). Health service use was categorized into four arms: only western (20%), only traditional (11%), both (7%), and no health service use (62%). Western health service users were immigrants who answered the question as to doctor's visits in the US. Traditional health service users were immigrants who answered at least one of ten questions regarding other source of health care in the US. For all analyses, this study used logit regressions with state-fixed effects, with the assumption that there is heterogeneity across states in terms of immigrants' assimilation and decision of health service use. Multiple Imputation method was adopted to address missing information.
Results: Overall, the behavioral model fitted well as a predictor of western health service use among Asian immigrants but was less suited at predicting their traditional and mixed health service use. Both English proficiency and US experience were associated with higher probability of using western (OR=1.23, p<.001; OR=1.03, p<.01) and mixed (OR=1.22, p<.05; OR=1.04, p<.05) health services. However, different patterns were found in sub-ethnic group analyses. For Philippine and Vietnamese immigrants, both assimilation factors were related to higher likelihood of using western health services (OR=1.20, p<.05; OR=1.05, p<.05). For Chinese and Korean immigrants, US experience was positively associated with their traditional health service use (OR=1.07, p<.01), meaning they prefer using their traditional medicine, regardless of degree of assimilation. For Indians, English proficiency was positively related with the probability of using mixed health services (OR=1.79, p<.01). Lastly, the interaction results showed Indian immigrants were more likely to use mixed heath services when possessing better English proficiency than were Philippine and Vietnamese immigrants. Also, Indian and East Asian immigrants were more likely to use mixed heath services with longer US stay, compared to Philippine and Vietnamese immigrants.
Conclusions and Implications: This study shows Andersen's behavioral model to be a suitable explanation for western health service use among Asian immigrants. Also, degree of their assimilation is an important enabling factor in use of western and mixed health services. However, the pattern substantially varies by ethnicity, suggesting the inappropriateness of treating Asian immigrants from diverse countries of origin as one group. Based on the findings, we suggest that health service programs for immigrants be tailored to each immigrant group in ways that reflect cultural preference.