Language Barriers to Healthcare for Linguistic Minorities: The Case of Second Language-specific Health Communication Anxiety

ABSTRACT

In this study we examined health communication anxiety (HCA) associated with language-discordant situations – that is, where people have to use their second language (L2) to communicate with health providers who are using their first language (L1). We adapted existing HCA scales in order to (1) assess L2 HCA in such situations separately for physical and mental/emotional health contexts and (2) control for potential confounds, such as HCA not related to L2 use and L2 communication anxiety not related to health, allowing us to obtain L2-specific measures of HCA. We examined the relationship between L2-specific HCA and willingness to use health services in language-discordant situations. English-speaking linguistic minority participants (N = 314) living in Québec, a predominantly French-speaking area of Canada, were recruited for online testing. The results revealed that, separately for both physical and mental/emotional health contexts, there were significant and meaningful L2-specific relations between HCA and willingness to use L2 health services – i.e., over and above general anxiety and discomfort about using an L2, and over and above general health communication anxiety. The effect was stronger for mental/emotional health contexts. The results are discussed in terms of their implications for understanding barriers to health services for linguistic minorities.

Acknowledgments

The authors would like to thank McGill University’s Training and Retention of Health Professionals Project for a Health Canada sub-award to NS, AGR and EC to support this project. The authors also thank members of the H-CALM (Health Care Access for Linguistic Minorities) Research team, and the Multilingualism Lab and the Culture, Health and Personality Lab groups at Concordia University for their input along the way. The authors also acknowledge help from AMI-Québec and from the Quebec Community Health & Social Services Network (CHSSN) for their assistance in reaching out to potential participants across Quebec. The authors also gratefully acknowledge the very helpful suggestions from several anonymous reviewers. Co-authors can be contacted at [email protected], [email protected], and [email protected].

Notes

1. The χ² test of difference indicated a significant difference between the configural and the weak models. The χ² test of difference, however, relies on the unreasonable assumption that the base model (in this case, the configural model) is a perfect fit for the data, and for this reason is no longer considered to be the preferred indicator of significant differences between two nested models (Cheung & Rensvold, 2002).