Developing a Tuberculosis Transmission Model That Accounts for Changes in Population Health

Background. Simulation models are useful in policy planning for tuberculosis (TB) control. To accurately assess interventions, important modifiers of the epidemic should be accounted for in evaluative models. Improvements in population health were associated with the declining TB epidemic in the pre-antibiotic era and may be relevant today. The objective of this study was to develop and validate a TB transmission model that accounted for changes in population health. Methods. We developed a deterministic TB transmission model, using reported data from the pre-antibiotic era in England. Change in adjusted life expectancy, used as a proxy for general health, was used to determine the rate of change of key epidemiological parameters. Predicted outcomes included risk of TB infection and TB mortality. The model was validated in the setting of the Netherlands and then applied to modern Peru. Results. The model, developed in the setting of England, predicted TB trends in the Netherlands very accurately. The R-2 value for correlation between observed and predicted data was 0.97 and 0.95 for TB infection and mortality, respectively. In Peru, the predicted decline in incidence prior to the expansion of Directly Observed Treatment Short Course (The DOTS strategy) was 3.7% per year (observed = 3.9% per year). After DOTS expansion, the predicted decline was very similar to the observed decline of 5.8% per year. Conclusions. We successfully developed and validated a TB model, which uses a proxy for population health to estimate changes in key epidemiology parameters. Population health contributed significantly to improvement in TB outcomes observed in Peru. Changing population health should be incorporated into evaluative models for global TB control.