Patterns of COVID-19 Testing Preferences in Rural Underserved Populations (Preprint)

<sec> <title>BACKGROUND</title> A particular challenge during the Covid-19 pandemic was to provide testing and treatment for already disadvantaged and vulnerable populations. Media reports suggested socioeconomic and racial disparities in access to testing. Many states implemented testing in a sporadic and disorganized way and it is unclear to what extent this disproportionally affected populations experienced barriers to accessing care. It is also unclear that if potential barriers to testing were due to systemic challenges or whether there were underlying individuals motivations for not getting tested. </sec> <sec> <title>OBJECTIVE</title> The objective of this study was to understand the trade-offs individuals in rural and vulnerable populations make between attributes of individual diagnostic testing and how preferences and trade-offs vary across individuals. </sec> <sec> <title>METHODS</title> We used a mixed methods approach, first conducting focus groups to identify barriers to COVID-19 testing and then identifying optimal strategies to increase testing using hypothetical scenarios by developing a Discrete Choice Experiment. We analyzed the data using a conditional logit model (CL) and using latent class analysis (LC). </sec> <sec> <title>RESULTS</title> We found that respondents cared about select structural factors, but that these were not the primary drivers of choice for testing. We also found that the attributes of testing were all significant in the CL model, apart from home visit and walk in, and had the expected signs. However, when taking a closer look at preference heterogeneity and unobserved preferences, we concluded that some important covariates were driving preferences, including: age, gender, medical vulnerability, insurance status, trust in government organizations, and previous flu vaccination -which may be a proxy for compliance. In sum, these covariates helped explain the observed preference heterogeneity. Contrary to our hypotheses, rurality did not significantly impact preferences for testing. </sec> <sec> <title>CONCLUSIONS</title> The results suggest that important social, behavioral and even policy factors affect choice for testing. Contrary to our hypotheses, rurality did not significantly impact preferences for testing, but attitudes towards government and other beliefs did. Health care interventions intended to reduce rural health disparities that do not reflect the underlying values of individuals in those subpopulations are unlikely to be successful. </sec>