Strategic Planning for Tuberculosis Control in the Republic of Fiji

Abstract

The tuberculosis (TB) health burden in Fiji has been declining in recent years, although challenges remain in improving control of the diabetes co-epidemic and achieving adequate case detection across the widely dispersed archipelago. We applied a mathematical model of TB transmission to the TB epidemic in Fiji that captured the historical reality over several decades, including age stratification, diabetes, varying disease manifestations, and incorrect diagnoses. Next, we simulated six intervention scenarios that are under consideration by the Fiji National Tuberculosis Program. Our findings show that the interventions were able to achieve only modest improvements in disease burden, with awareness raising being the most effective intervention to reduce TB incidence, and treatment support yielding the highest impact on mortality. These improvements would fall far short of the ambitious targets that have been set by the country, and could easily be derailed by moderate increases in the diabetes burden. Furthermore, the effectiveness of the interventions was limited by the extensive pool of latent TB infection, because the programs were directed at only active cases, and thus were unlikely to achieve the desired reductions in burden. Therefore, it is essential to address the co-epidemic of diabetes and treat people with latent TB infection.

Keywords: disease modelling; epidemiology; health policy; public health; simulation; tuberculosis.

Figure 1

Model calibration results. Incidence and observed mortality (per 100,000 per year), prevalence (per 100,000) and the number of notifications by calendar year. The blue-shaded areas represent the calibrated model predictions obtained from the Metropolis simulation. The grey lines represent point estimates, and the hatched areas represent the confidence limits for each indicator from the Global TB Report 2017. The light grey line indicates the epidemic trajectory that would be required to achieve the different targets, and is a piecewise exponential function. M, Milestone; S, Sustainable Development Goal; E, End TB Strategy Target.

Figure 2

Intervention effectiveness. Incidence and observed mortality (per 100,000 per year), prevalence (per 100,000) and number of notifications by calendar year.

Figure 3

Diabetes prevalence counterfactuals. Incidence and observed mortality (per 100,000 per year), prevalence (per 100,000), and number of notifications by calendar year. Red lines represent the different levels of T2DM prevalence varying from 5% to 50%. The baseline scenario (15.6% T2DM prevalence) is shown in black.