Back

Survival of people with untreated tuberculosis: effects of time, geography and setting

Rodriguez, C. A.; Leavitt, S.; Bouton, T.; Horsburgh, C. R.; zur Wiesch, P. A.; Nichols, B. E.; Jenkins, H. E.; White, L.

2022-12-16 infectious diseases
10.1101/2022.12.15.22283231 medRxiv
Show abstract

BackgroundAn estimated 40% of people who developed tuberculosis in 2021 were not diagnosed or treated. Pre-chemotherapy era data are a rich resource on survival for people with untreated TB. We aimed to identify heterogeneities in these data to inform more precise use of them. MethodsWe extracted survival data from pre-chemotherapy era papers reporting TB specific mortality and/or natural recovery data. We used Bayesian parametric survival analysis to model the survival distribution, stratifying by geography (North America versus Europe), time (pre-1930 versus post-1930), and setting (sanitoria versus non-sanitoria). ResultsWe found 12 studies with TB-specific mortality data. Ten-year survival was 69% in North America (95 CI: 54%-81%) and 36% in Europe (95% CI: 10%-71%). Only 38% (95% CI: 18%-63%) of non-sanitorium individuals survived to 10 years compared to 69% (95% CI: 41%-87%) of sanitoria/hospitalized patients. There were no significant differences between people diagnosed pre-1930 and post-1930 (five-year survival pre-1930: 65%; 95% CI: 44%-88% versus post-1930: 72%; 95% CI: 41%-94%). ConclusionsMortality and natural recovery risks vary substantially by location and setting. These heterogeneities need to be considered when using pre-chemotherapy data to make inferences about expected survival of people with undiagnosed TB.

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.