JMIR Publications

ABSTRACT

Background:

The COVID-19 pandemic uncovered the dearth of resources and experience to respond effectively in local health departments, particularly in smaller communities. Publicly available surveillance data, the key information for local health departments, was not sufficiently timely or granular for targeted interventions. The City of Brownsville (COB) is located in a low-income south Texas border county plagued with severe health disparities. The COB, public health department shared local COVID-19 surveillance data weekly with academic partners that produced near real-time weekly geospatial maps of these data. Census tract level case maps were used to strategically target an educational outreach intervention named “Boots on the Ground” (BOG), and application of novel statistical methods were used to evaluate its impact.

Objective:

To evaluate the slope (sustained) and intercept (immediate) change in COVID-19 daily test counts 2 weeks pre and post BOG delivery.

Methods:

Using an interrupted time series design we evaluated the COB census tracts that received targeted BOG between April 21-June 8, 2020. A piece-wise Poisson regression analysis was used to quantify the sustained and immediate change between pre and post BOG COVID-19 daily test count trends. A sensitivity analysis of tracts that did not receive targeted BOG was conducted for comparison purposes.

Results:

During the intervention period, 18 of 48 COB census tracts received targeted BOG. Among these census tracts, significant difference in the slope coefficients from pre- and post-BOG daily test counts was observed in 5 tracts, with 2 tracts having a significant difference in the intercept. Additionally, 80% (4/5) of the significant slope changes showed an increase in pre- and post-slopes. This means the testing trend two weeks post BOG had a sustained increase from the trend two weeks pre-BOG. In the sensitivity analysis of the 30 census tracts not receiving BOG, the opposite was observed. In these tracts, 80% (8/10) of those with significant slope changes had a decrease in the pre- and post-BOG COVID-19 testing slopes.

Conclusions:

Targeting and evaluation of public health interventions is necessary and possible, particularly in small communities. This report highlights how collaboration between a school of public health and a local health department established and evaluated the impact of a real-time, targeted intervention delivering precision public health to a small community.