Abstract | The WHO recommends integrating interventions to address the devastating TB/HIV co-epidemics in South Africa, yet integration has been poorly implemented and TB/HIV control efforts need strengthening. Identifying infected individuals is particularly difficult in rural settings. We used mathematical modeling to predict the impact of community-based, integrated TB/HIV case finding and additional control strategies on South Africa's TB/HIV epidemics. We developed a model incorporating TB and HIV transmission to evaluate the effectiveness of integrating TB and HIV interventions in rural South Africa over 10 years. We modeled the impact of a novel screening program that integrates case finding for TB and HIV in the community, comparing it to status quo and recommended TB/HIV control strategies, including GeneXpert, MDR-TB treatment decentralization, improved first-line TB treatment cure rate, isoniazid preventive therapy, and expanded ART. Combining recommended interventions averted 27% of expected TB cases (95% CI 18-40%) 18% HIV (95% CI 13-24%), 60% MDR-TB (95% CI 34-83%), 69% XDR-TB (95% CI 34-90%), and 16% TB/HIV deaths (95% CI 12-29). Supplementing these interventions with annual community-based TB/HIV case finding averted a further 17% of TB cases (44% total; 95% CI 31-56%), 5% HIV (23% total; 95% CI 17-29%), 8% MDR-TB (68% total; 95% CI 40-88%), 4% XDR-TB (73% total; 95% CI 38-91%), and 8% TB/HIV deaths (24% total; 95% CI 16-39%). In addition to increasing screening frequency, we found that improving TB symptom questionnaire sensitivity, second-line TB treatment delays, default before initiating TB treatment or ART, and second-line TB drug efficacy were significantly associated with even greater reductions in TB and HIV cases. TB/HIV epidemics in South Africa were most effectively curtailed by simultaneously implementing interventions that integrated community-based TB/HIV control strategies and targeted drug-resistant TB. Strengthening existing TB and HIV treatment programs is needed to further reduce disease incidence. |