Sensitivity analysis TBhotspotclustersplos: Analysis based on all notified TB cases, unlike the primary analysis(https://github.com/mcewenkhundi/TBhotspotclustersplos) which is based on only microbiologically-confirmed notified cases

The goal of TBhotspotclustersplos repository is to replicate the analysis of the manuscript:

“Neighbourhood prevalence-to-notification ratios for adult bacteriologically-confirmed tuberculosis reveals hotspots of underdiagnosis in Blantyre, Malawi”

Authors: McEwen Khundi^1,2, James R Carpenter¹, Elizabeth L Corbett^1,2, Helena R A Feasey^1,2, Rebecca Nzawa Soko^1,2, Marriott Nliwasa^1,3, Hussein Twabi³, Lingstone Chiume¹, Rachael M Burke^1,2, Katherine C Horton², Peter J Dodd⁴, Ted Cohen⁵, Peter MacPherson^1,2,⁶

1. Malawi-Liverpool-Wellcome Trust Clinical Research Programme, Blantyre, Malawi,
2. London School of Hygiene and Tropical Medicine, UK,
3. Helse Nord TB Initiative, College of Medicine, University of Malawi,
4. School of Health and Related Research, University of Sheffield, Sheffield, UK,
5. Yale School of Public Health, New Haven, USA,
6. Department of Clinical Sciences, Liverpool School of Tropical Medicine, UK

Replicable code (Rscript folder description)

Description of the analysis scripts and source script of each output in the analysis

1. 1_Part2_all_models_notif_rintc.R

   • Script for producing notification model based on the final model selected in the primary analysis applied to all notified TB cases.

2. 2_Part4_derive_posteriors_rintc_2021-09-29.Rmd

   • Script for deriving posterior samples from the selected final models and also calculating the prevalence to notification ratios
   • the prevalence model was copied from the primary analysis repository.

3. 3_Part4_exploratory_analysis_posterior_rintc_2021-09-29.Rmd

   • Exploratory analysis of the posterior samples generated in part 2 including graphs

   • Source of S4_Fig.tiff and S5_Fig.tiff

4. 4_Model_table_results.R

   • Source of table S8_Table_allnotified_models_rintc.rtf

Data

dat_scale.rds

Neighbourhood level variables

dat_scale_ln.rds

Neighbourhood level variables in longitudinal format