What’s a #UnionConf without the TB modelling course? Packed room again for a long day of intro to modelling.
@lshtm-tbmod.bsky.social @tb-mac.bsky.social #WCLH2025 #TBSky

Wrapping up the #CausalIntroCourse.

Kicking off the #CausalIntroCourse in Leeds. Hoping this will finally give me the tools to convince my mum that walking around the house barefoot ^^doesn’t^^ cause a cold.

I’m thrilled to share that I’ve submitted my PhD thesis at LSHTM (@lshtm.bsky.social) with the TB Modelling Group (@lshtm-tbmod.bsky.social)! The past three years have been an incredible journey, and I’m so proud of all that’s been accomplished. Excited to continue working in TB research! #TBSky

12/ Estimated ICERs against the BAU for all algorithms were below the estimated cost-effectiveness threshold range for Viet Nam. The two-step CXR+NAAT algorithm averted 4.29m DALYs (95%UI:2.86-6.14) at US$225 (95%UI:85-520) per DALY averted compared with BAU.
NOT YET PEER-REVIEWED

11/ Total intervention costs were substantial. However, they yield persistent cost savings, likely continuing beyond the 25-year time horizon. Most costs were for screening in NAAT-based algorithms and treatment in the CXR-only algorithm.
NOT YET PEER-REVIEWED

10/ All screening algorithms resulted in lower cumulative TB incidence, TB deaths, and DALYs compared with BAU between 2025 and 2050. Reductions were similar for NAAT-based algorithms but greater with the CXR-only algorithm.
NOT YET PEER-REVIEWED

9/ NAAT-based algorithms required at least six rounds to reach the prevalence threshold, while CXR-only required three. The threshold was not reached within the time horizon under BAU. NAAT-only achieved a prevalence reduction consistent with the ACT3 trial after three rounds.
NOT YET PEER-REVIEWED

7/ We calibrated a deterministic transmission model to TB epidemiology in Viet Nam. We designed three population-wide screening algorithms from 2025: sputum nucleic acid amplification tests (NAAT, Xpert MTB/RIF Ultra) only; CXR followed by NAAT; and CXR-only without microbiological confirmation.

9/ Of all recent infections, over 87% were found in the three regions of Southeast Asia, Western Pacific, and Africa. At the country level, India, China, and Indonesia together account for about 50% of recent global Mtb infections.
NOT YET PEER-REVIEWED

8/ We found that in 2022, 156 million people (95%UI:127-199) were recently infected with viable Mtb, equating to 2.0% (95%UI:1.6-2.5) of the global population, with substantial regional variability.
NOT YET PEER-REVIEWED

7/ By integrating these insights, our model generated estimates for 171 countries (covering 99.6% of the world population), all six WHO regions, and globally.
NOT YET PEER-REVIEWED

5/ Secondly, we also account for the evidence that many individuals may self-clear Mtb infection, calibrating these rates using data from @kchorton.bsky.social et al. - tinyurl.com/ykap23kx - on infection and disease pathways in a simulated cohort.

4/ Firstly, we adjusted annual infection risk (ARI) trends for immunoreactivity reversion and age-based social mixing to address ARI underestimation, offering a more accurate infection force estimate.
NOT YET PEER-REVIEWED.

Starting off #Union2024 with the introduction to TB modelling course. This time commemorating 10 years of the course with cake!

Livin' la vida loca, indeed
#AcademicSky #EpiSky