Gambiense human African trypanosomiasis (gHAT) is a vector-borne infection transmitted by tsetse flies in West and Central Africa. The World Health Organization has targeted gHAT for elimination of transmission (EoT), defined as reducing incidence to zero with minimal risk of reintroduction. To achieve EoT, many high-risk regions have implemented combined approaches that add vector control (VC) to established medical interventions.
There is ongoing uncertainty about how effective VC interventions are in particular settings. Entomological monitoring—typically deploying traps to estimate local tsetse abundance—provides non-routine data that can reduce uncertainty about VC impact. The central question addressed in this work is whether collecting such entomological data can improve decision-making for gHAT control by informing whether to continue or stop VC and by increasing the cost-effectiveness of intervention strategies.
The study employs an active adaptive management (AAM) approach. AAM is the process of periodically updating the control strategy based on incoming monitoring data; in this context, operational decisions (for example, continuing or halting VC) are conditioned on observed entomological signals of tsetse abundance and inferred VC effectiveness. To link observations to parameter estimates, posterior distributions for VC effectiveness were obtained using an adaptive MCMC method. These posterior distributions quantify uncertainty about how well VC reduces tsetse populations in the health zones under study and feed into subsequent decision analysis.
Posterior estimates of VC effectiveness were integrated into an established gHAT transmission model. The economic evaluation used a policy objective of maximising the intervention's net monetary benefit. This allowed comparison of alternative monitoring strategies by estimating the monetary value gained from reduced disease burden and intervention costs under different data-collection regimes and decision rules.
The analysis focused on two illustrative health zones in the Democratic Republic of Congo, chosen because they differ in the level of uncertainty around VC effectiveness. The two zones modelled were Mulumba and Vanga. The study compared strategies that included VC with varying frequencies of entomological monitoring—specifically noting a strategy with twice-yearly monitoring—and strategies without such monitoring to assess whether the additional data altered optimal decisions and cost-effectiveness.
Under the model and economic assumptions used, the strategy that combined VC with twice-yearly entomological monitoring was found to be cost-effective in the health zone of Mulumba at a willingness-to-pay threshold of $153. In this setting, the monitoring-informed AAM process enabled selection of an optimal monitoring strategy. By contrast, the same monitoring strategy was not cost-effective in the health zone of Vanga at that willingness-to-pay threshold. These contrasting results illustrate that the value of collecting non-routine entomological data depends on local epidemiology, uncertainty about VC effectiveness, and the economic context captured by the willingness-to-pay threshold.
This study presents the first application of AAM to monitoring data for a vector-borne infection, demonstrating a quantitative framework for assessing how entomological monitoring might improve the cost-effectiveness of VC interventions. The findings suggest that in some settings—represented here by Mulumba—regular entomological monitoring can provide sufficient information to change management decisions in a way that is economically beneficial. In other settings—represented by Vanga—the same monitoring investment may not yield cost-effective benefits at the evaluated threshold.
The data used were aggregated from the World Health Organization's HAT Atlas and are not publicly shareable without WHO approval; researchers may request access from WHO. Full model code supporting the analyses is available at osf.io/8q5ug. Funders declared in the article include the Engineering and Physical Sciences Research Council, the Royal Society, and the Gates Foundation. The authors declared no competing interests and reported adherence to relevant ethical guidelines.
This work is a modelling and decision-analytic study based on available data and specified economic assumptions. It is presented as a preprint and has not been peer reviewed; the authors state it should not be used to guide clinical practice without further validation. Specific methodological details, parameter values beyond those summarised, and full analytical results are reported in the study's full text and supplementary materials; readers should consult the original preprint for those details.