PLoS medicine
-
The current World Health Organization (WHO) pediatric tuberculosis dosing guidelines lead to suboptimal drug exposures. Identifying factors altering the exposure of these drugs in children is essential for dose optimization. Pediatric pharmacokinetic studies are usually small, leading to high variability and uncertainty in pharmacokinetic results between studies. We pooled data from large pharmacokinetic studies to identify key covariates influencing drug exposure to optimize tuberculosis dosing in children. ⋯ Children older than 3 months have lower rifampicin exposures than adults and increasing their dose by 75 or 150 mg could improve therapy. Altered exposures in children with HIV is most likely caused by concomitant ART and not HIV per se. The importance of the drug-drug interactions with lopinavir/ritonavir and efavirenz should be evaluated further and considered in future dosing guidance.
-
[This corrects the article DOI: 10.1371/journal.pmed.1003793.].
-
Vaccines have reduced severe disease and death from Coronavirus Disease 2019 (COVID-19). However, with evidence of waning efficacy coupled with continued evolution of the virus, health programmes need to evaluate the requirement for regular booster doses, considering their impact and cost-effectiveness in the face of ongoing transmission and substantial infection-induced immunity. ⋯ Our modelling suggests that regular boosting of the high-risk population remains an important tool to reduce morbidity and mortality from current and future SARS-CoV-2 variants. Our results suggest that focusing vaccination in the highest-risk cohorts will be the most efficient (and hence cost-effective) strategy to reduce morbidity and mortality.
-
Risk-based screening for lung cancer is currently being considered in several countries; however, the optimal approach to determine eligibility remains unclear. Ensemble machine learning could support the development of highly parsimonious prediction models that maintain the performance of more complex models while maximising simplicity and generalisability, supporting the widespread adoption of personalised screening. In this work, we aimed to develop and validate ensemble machine learning models to determine eligibility for risk-based lung cancer screening. ⋯ We present parsimonious ensemble machine learning models to predict the risk of lung cancer in ever-smokers, demonstrating a novel approach that could simplify the implementation of risk-based lung cancer screening in multiple settings.