Attendance of Underserved Populations at Field-Based Health Services Events: Application of Quasi-Experimental Methods that Accommodate the COVID-19 Pandemic.

Abstract

The COVID-19 pandemic disrupted the ability to receive health care services. Field-based health services became a logistically feasible alternative to medical center-based care. We compared two different field-based approaches to the delivery of SARS-CoV-2 testing and health education services for Latine communities using a quasi-experimental design that included propensity score matching to accommodate the challenges posed to research by the pandemic. From September 2021 through October 2022, we held 434 testing events, of which 234 used a geolocation approach and 200 used a partner-located approach to determine the location of the health services (n = 68 field sites in Oregon). We hypothesized that partner-located sites would obtain higher numbers of tests collected relative to geolocated sites, and that longer drive times to testing sites would be associated with lower testing rates. There were no differences in the number of tests collected by geolocated versus partner-located sites, controlling for population size and time-varying pandemic vulnerabilities measured as COVID-19 cases and deaths. Prior to propensity score weighting, a longer drive time to the testing site (both site types) was associated with a lower likelihood of total tests (IRR = .87, p < .01, CI [0.54, 0.92]), of Latine tests (IRR = .69, p < .001, CI [0.56, 0.84]), and of male tests collected (IRR = .67, p < .05, CI [0.47, 0.94]). The site's number of prior tests was associated with a significant 2% increase in tests collected and the prior week's number of county deaths was associated with a roughly 30% decrease in the likelihood of tests collected. However, the reduced testing rate when the death rate was higher was less likely in geolocated sites (IRR = 1.55, p < .001, CI [1.20, 2.01]). Implications for the utility of propensity score matching and time-varying covariates to accommodate pandemic challenges posed to research are discussed. Clinicaltrials.gov registration number: NCT05082935. Date of registration: 10/15/2021.