T uberculosis (TB) is a global health emergency (1). The World Health Organization (WHO) End TB Strategy proposes a 90% reduction in TB incidence and 95% reduction in TB deaths by 2035 compared with 2015 (2). To reach this target, effective interventions are needed to interrupt transmission of Mycobacterium tuberculosis. Contact investigations help prevent M. tuberculosis transmission by identifying and treating persons in close contact with persons with TB disease (3). WHO recommends tuberculosis preventive treatment (TPT) for household members of bacteriologically confirmed pulmonary TB patients to prevent progression to active TB disease (4). Contact investigations are a major tenet of the End TB Strategy but remain ineffective for various reasons (2,5,6). Many TB programs in high-burden areas limit contact investigations to household members (6). Recent studies suggest that such restrictions might miss key exposures in the community (7,8). Targeted, population-based, geographic TB screening is a potential approach to augment contact investigations (9-11) but is resource and time intensive and rarely includes TPT (11,12). We used population-based, molecular epidemiologic data from Botswana to investigate potential use of a neighbor-based approach for contact investigations. The Study During August 2012-April 2016, we enrolled participants treated for TB disease at 30 healthcare facilities in Botswana for a prospective molecular epidemiologic study, Kopanyo. In brief, Kopanyo was designed to explore potential clinical, demographic, geographic, social relationships, and M. tuberculosis genotypic characteristics among persons with TB (13,14). We interviewed enrolled patients by using a standardized questionnaire and abstracted clinical data from medical records (13). We collected and processed sputum samples for culture and genotyped isolates with 24-locus mycobacterial interspersed repetitive unitsvariable-number tandem-repeats by using standard methods (15). We geocoded and validated the primary residence of each enrolled patient (Appendix, https:// wwwnc.cdc.gov/EID/article/26/5/19-1568-App1. pdf). We excluded patients without a validated primary residential geocode and those who resided in locations outside of the study area. The study area included all 11 neighborhoods in Gaborone and 3 villages in the Ghanzi District: Ghanzi, D'Kar, and Kuke. We defined index patients as the first culture-positive pulmonary TB patient identified and started on treatment in a household. We used residence plots to identify nearest neighbors, which we defined as those who lived immediately next door, and next-nearest neighbors, which we defined as those who lived 2 doors away (Figure). We enumerated all subsequent TB cases identified by bacteriologic confirmation and clinical diagnosis within the index home, nearest-neighbor homes, and next-nearest neighbor homes. We defined