Background: Racial discrimination is recognised as a key social determinant of health and driver of racial/ethnic health inequities. Studies have shown that people exposed to racism have poorer health outcomes (particularly for mental health), alongside both reduced access to health care and poorer patient experiences. Most of these studies have used cross-sectional designs: this prospective cohort study (drawing on critical approaches to health research) should provide substantially stronger causal evidence regarding the impact of racism on subsequent health and health care outcomes.
Methods: Participants are adults aged 15+ sampled from 2016/17 New Zealand Health Survey (NZHS) participants, sampled based on exposure to racism (ever exposed or never exposed, using five NZHS questions) and stratified by ethnic group (Māori, Pacific, Asian, European and Other). Target sample size is 1680 participants (half exposed, half unexposed) with follow-up survey timed for 12-24 months after baseline NZHS interview. All exposed participants are invited to participate, with unexposed participants selected using propensity score matching (propensity scores for exposure to racism, based on several major confounders). Respondents receive an initial invitation letter with choice of paper or web-based questionnaire. Those invitees not responding following reminders are contacted for computer-assisted telephone interview (CATI). A brief questionnaire was developed covering current health status (mental and physical health measures) and recent health-service utilisation (unmet need and experiences with healthcare measures). Analysis will compare outcomes between those exposed and unexposed to racism, using regression models and inverse probability of treatment weights (IPTW) to account for the propensity score sampling process.
Discussion: This study will add robust evidence on the causal links between experience of racism and subsequent health. The use of the NZHS as a baseline for a prospective study allows for the use of propensity score methods during the sampling phase as a novel approach to recruiting participants from the NZHS. This method allows for management of confounding at the sampling stage, while also reducing the need and cost of following up with all NZHS participants.
Keywords: Health inequities; Health service utilisation; New Zealand; Prospective cohort study; Racism.