Logo Lanfrica

YisaAdeniyiAbolade/act-efficacy-latent-components

Domain:

healthcare
Creator:
Yis
Host:
A statistical framework for jointly modeling day-3 parasite positivity and PCR-corrected ACT treatment failure to separate shared early-clearance and regimen-specific late-efficacy signals in African WHO therapeutic-efficacy studies. # Component-resolved Joint latent-component modeling of ACT therapeutic-efficacy surveillance R code and analytic data for the study: **Component-resolved Joint latent-component modeling of early parasite response and final ACT efficacy in African therapeutic-efficacy surveillance** The analysis jointly models day-3 parasite positivity and PCR-corrected per-protocol treatment failure in African *Plasmodium falciparum* therapeutic efficacy studies of artemether-lumefantrine, artesunate-amodiaquine, and dihydroartemisinin-piperaquine. ## Requirements The reported analysis uses R 4.4.0-compatible code and the `Matrix` package. ```r install.packages("Matrix") ``` ## Repository structure - `R/joint_model.R`: joint latent-component likelihood, estimation, uncertainty, latent-component estimates, and prediction functions. - `data/who_tes_dual_endpoint.csv`: analytic WHO therapeutic-efficacy dataset used in the study. - `analysis/descriptive_summary.R`: cohort summaries, threshold states, era summaries, and cross-regimen concordance. - `analysis/fit_model.R`: primary model fit, likelihood-ratio test, latent-component estimates, and sample-size sensitivity analysis. - `analysis/robustness_checks.R`: five-fold final-endpoint validation, forward temporal validation, and leave-one-country-out stability analysis. - `simulation/run_simulation.R`: threshold-balanced Monte Carlo study. - `analysis/make_figures.R`: figure generation. - `tests/validate.R`: numerical and data-integrity checks. - `reference/`: numerical results from the reported analysis. - `run_all.R`: complete workflow. ## Quick reproducibility check From the repository root: ```bash Rscript run_all.R quick ``` The quick workflow fits the complete empirical model, runs a reduced threshold-balanced simulation, generates figures, and performs numerical validation. A successful run ends with: ```text All validation checks passed quick workflow completed successfully. ``` ## Complete reproduction Run the …

Licenses