Background: Extensively drug-resistant (XDR) tuberculosis has spread among hospitalised patients in South Africa, but the epidemic-level effect of hospital-based infection control strategies remains unknown. We modelled the plausible effect of rapidly available infection control strategies on the overall course of the XDR tuberculosis epidemic in a rural area of South Africa.
Methods: We investigated the effect of administrative, environmental, and personal infection control measures on the epidemic trajectory of XDR tuberculosis in the rural community of Tugela Ferry. Assessments were done with a mathematical model incorporating over 2 years of longitudinal inpatient and community-based data. The model simulated inpatient airborne tuberculosis transmission, community tuberculosis transmission, and the effect of HIV and antiretroviral therapy.
Findings: If no new interventions are introduced, about 1300 cases of XDR tuberculosis are predicted to occur in the area of Tugela Ferry by the end of 2012, more than half of which are likely to be nosocomially transmitted. Mask use alone would avert fewer than 10% of cases in the overall epidemic, but could prevent a large proportion of cases of XDR tuberculosis in hospital staff. The combination of mask use with reduced hospitalisation time and a shift to outpatient therapy could prevent nearly a third of XDR tuberculosis cases. Supplementing this approach with improved ventilation, rapid drug resistance testing, HIV treatment, and tuberculosis isolation facilities could avert 48% of XDR tuberculosis cases (range 34-50%) by the end of 2012. However, involuntary detention could result in an unexpected rise in incidence due to restricted isolation capacity.
Interpretation: A synergistic combination of available nosocomial infection control strategies could prevent nearly half of XDR tuberculosis cases, even in a resource-limited setting. XDR tuberculosis transmission will probably continue in the community, indicating the need to develop and implement parallel community-based programmes.