The BlantyreSepsis R package contains data and code to replicate the analysis of the manuscript:

A longitudinal observational study of aetiology and long-term outcomes of sepsis in Malawi revealing the key role of disseminated tuberculosis

Joseph M Lewis^1,2,3, , Madalitso Mphasa¹, Lucy Keyala¹, Rachel Banda¹, Emma Smith¹,², Jackie Duggan⁴, Tim Brooks⁴, Matthew Catton⁴, Jane Mallewa⁵, Grace Katha⁵, Stephen B Gordon^1,2, Brian Faragher², Melita A Gordon^1,3, Jamie Rylance^1,2, Nicholas A Feasey^1,2

1. Malawi Liverpool Wellcome Clinical Research Programme, Blantyre, Malawi
2. Department of Clinical Sciences, Liverpool School of Tropical Medicine, Liverpool, UK
3. Department of Clinical Infection, Microbiology and Immunology, University of Liverpool, Liverpool, UK
4. Rare and Imported Pathogens Laboratory, Public Health England, UK
5. College of Medicine, University of Malawi, Malawi

Now published in Clinical Infectious Diseases here

Install the package from GitHub:

install.packages("devtools")
devtools::install_github("https://github.com/joelewis101/blantyreSepsis")

Or check out the source code at GitHub

Six datasets are included in the package; for details and variable definitions access the associated help file (e.g via ?BTparticipants). They are lazy loaded so will be available on loading the package.

• BTparticipants - Baseline characteristics and investigation results for included participants
• BTtreatments - Antimicrobial therapy and intravenous fluid therapy received by included participants
• BTaetiology - Results of malaria testing, aerobic blood culture, cerebrospinalfluid culture, tuberculosis diagnostic tests
• BTsera - Results of serologic tests for arboviruses, Rickettsioses, Leptospirosis
• BTarraycard - Results of multiplex PCR test for high consequence pathogens
• BTbc - Results of aerobic blood culture

This analysis is available as a package vignette; this can be built when downloading the package by typing:

devtools::install_github("https://github.com/joelewis101/blantyreSepsis", build_vignettes = TRUE )

The vignette will be then be available by typing vignette("analysis") and the code is accessed by typing edit(vigentte("analysis")). Be warned: building the vignette takes some time as it will run through the full missing data imputation and model fitting!

Alternatively the source code for the vignette is analysis.Rmd in the vignettes/ folder of this repo or the pkgdown site for this package has a rendered version, as well as variable definitions for the datasets.