The persistence of tuberculosis in the age of DOTS: reassessing the effect of case detection.

Abstract

To investigate whether short-term annual declines of 5-10% in the incidence of tuberculosis (TB) can be sustained over the long term by maintaining high case detection rates (CDRs). We constructed a compartmental difference-equation model of a TB epidemic in a hypothetical population of constant size with a treatment success rate of 85%. The impact of CDR on TB incidence was then investigated by generating an equilibrium population with no TB case detection and increasing the smear-positive CDR under two scenarios: (i) rapid expansion by 10% per year to a CDR of 80% after 8 years, and (ii) gradual expansion by 1% per year to a CDR of 90% after 90 years. The model was applied in two hypothetical populations: one without HIV and the other with a stable HIV incidence representative of the African Region. The CDR for smear-negative TB was assumed to be a constant fraction of the smear-positive CDR. In the absence of a TB control programme, the projected annual incidence of TB was 513 cases per 100 000 population, with a point prevalence of 1233 per 100 000 and an annual TB-specific mortality rate of 182 per 100 000. Immediately increasing the TB CDR from 0% to 70% caused a 74% reduction in TB incidence within 10 years. However, once case detection stabilized at any constant level < 80%, projected TB incidence also stabilized. Ten years after a CDR of 70% was reached, the annual decline in TB incidence was < 1.5%, regardless of how rapidly case detection was scaled up and despite wide variation of all model parameters. While improved CDRs have a dramatic short-term effect on TB incidence, maintaining those rates, even at current target levels, may not reduce long-term incidence by more than 1-2% per year. TB control programmes and researchers should vigorously pursue improvements in case detection, regardless of current CDRs.