UC Irvine

Interventions to interrupt tuberculosis (TB) transmission are urgently needed in high burdensettings where recent transmission contributes substantially to disease incidence. Integrated analysis of genomic, geospatial, and social network data offers a promising approach to characterize TB transmission and identify places and groups of people at increased risk that could be targeted for public health interventions. However, these methods often exclude data for important locations such as work, school, or other community sites, i.e., activity spaces where transmission may occur. The goal of our analysis was to integrate genomic, geospatial, and social network data in analysis of activity space of participants diagnosed with TB in Botswana to contribute to increased understanding of community TB transmission. We analyzed data collected during 2012–2016 for the Kopanyo Study, a population-based study of TB transmission in Botswana. We included data from whole genome sequencing conducted on archived samples from the original study. We analyzed activity space of individuals belonging to TB outbreaks (genotypic groups with ≤ 5 single-nucleotide polymorphisms). We also included genotypically ungrouped participants as a comparison group. We conducted a spatial point pattern analysis to identify geographic hotspots of transmission. We conducted a geostatistical analysis to investigate potential spatial correlates of TB transmission. We conducted a location-based social network analysis to identify specific locations in the community where transmission may have occurred, and investigate social contact patterns that may be associated with transmission.

In a spatial point pattern analysis of full activity space, we analyzed six outbreak groupsand ungrouped participants residing in greater Gaborone. We identified spatially distinct core areas and found evidence of localized transmission in four of six outbreak groups. This could suggest geographically distinct chains of transmission.

In a geostatistical analysis using activity space median center points, we analyzed five outbreakgroups and ungrouped participants residing in greater Gaborone. Significant hotspots were detected for three groups. Activity spaces in these areas were associated with an increased risk of belonging to an outbreak group, even after adjusting for income and HIV coinfection status.

In location-based social network analysis, we identified a total of 42 shared locations amongparticipants in the same genotypic group, representing potential sites of transmission. These included shopping, transport, and alcohol-related venues, as well as two shared households and one work-related site. Five of the eight genotypic groups displayed a significant homophily effect, indicating location-based links in the network were more likely to occur between participants that matched by genotypic group.

The results of these analyses suggest potentially distinct social or spatial network groupsinvolved in ongoing transmission. Locations and groups of people associated with these groups could be targeted for outreach to interrupt ongoing transmission. Integrated analysis could be a useful tool for planning public health outreach to interrupt ongoing transmission in high burden settings.