AbstractGambiense human African trypanosomiasis (gHAT) is a virulent disease declining in burden but still endemic in West and Central Africa. Although it is targeted for elimination of transmission by 2030, there remain numerous questions about the drivers of infection and how these vary geographically.In this study we focus on the Democratic Republic of Congo (DRC), which accounted for 84% of the global case burden in 2016, to explore changes in transmission across the country and elucidate factors which may have contributed to the persistence of disease or success of interventions in different regions. We present a Bayesian fitting methodology, applied to 168 endemic health zones (∼ 100,000 population size), which allows for calibration of mechanistic gHAT model to case data (from the World Health Organization HAT Atlas) in an adaptive and automated framework.It was found that the model needed to capture improvements in passive detection to match observed trends in the data within former Bandundu and Bas Congo provinces indicating these regions have substantially reduced time to detection. Health zones in these provinces generally had longer burn-in periods during fitting due to additional model parameters.Posterior probability distributions were found for a range of fitted parameters in each health zone; these included the basic reproduction number estimates for pre-1998 (R0) which was inferred to be between 1 and 1.19, in line with previous gHAT estimates, with higher median values typically in health zones with more case reporting in the 2000s.Previously, it was not clear whether a fall in active case finding in the period contributed to the declining case numbers. The modelling here accounts for variable screening and suggests that underlying transmission has also reduced greatly – on average 96% in former Equateur, 93% in former Bas Congo and 89% in former Bandundu – Equateur and Bandundu having had the highest case burdens in 2000. This analysis also sets out a framework to enable future predictions for the country.Author summaryGambiense human African trypanosomiasis (gHAT; sleeping sickness) is a deadly disease targeted for elimination by 2030, however there are still several unknowns about what factors influence continued transmission and how this changes with geographic location.In this study we focus on the Democratic Republic of Congo (DRC), which reported 84% of the global cases in 2016 to try and explain why some regions of the country have had more success than others in bringing down case burden. To achieve this we used a state-of-the-art statistical framework to match a mathematical gHAT model to reported case data for 168 regions with some case reporting during 2000–2016.The analysis indicates that two former provinces, Bandundu and Bas Congo had substantial improvements to case detection in fixed health facilities in the time period. Overall, all provinces were estimated to have reductions in (unobservable) transmission including ∼ 96% in former Equateur. This is reassuring as case finding effort has decreased in that region.The model fitting presented here will allow predictions of gHAT under future strategies to be performed in the future.