eparker12.github.io-oral_vaccine_interventions_metaanalysis_2018

Church et al (2018): meta-analysis of interventions to improve oral vaccine immunogenicity

This github page contains the statistical analysis code and input data for the meta-analysis of interventions to improve oral vaccine immunogenicity described by Church et al (2018).

The full output of these analyses is available in the study's e-appendix.

The final analysis includes 12 interventions (e.g., Breastfeeding withheld, Buffer, etc.). For each intervention, we include the following:

• An Rmd file containing the R code for the intervention-specific analyses.
• A corresponding html file with the output of these analyses.
• A folder containing the figures included in this html (used when rendering the e-appendix).

We also include the following:

• site.yml - used to generate the layout for the e-appendix;
• index.Rmd and index.html - used generate the Home tab of the e-appendix;
• footnotes.Rmd and Footnotes.html - used generate the Footnotes tab of the e-appendix;
• summary.Rmd and summary.html - used to generate the Summary tab of the e-appendix;
• Church_et_al_2018_metaregression_acronyms.csv - containing acronyms used in the plots;
• Church_et_al_2018_metaregression_footnotes.csv - containing plot footnotes.
• Church_et_al_2018_summary.csv - containing the data for the summary plot (comparing the estimated effects of different interventions).

Analysis code

In spite of the numerous files and folders, all of the analyses are derived from a basic script titled Church_et_al_metaregression.Rmd. This contains the phrase 'intervention_placeholder' on lines 9 and 26. Each of the intervention-specific Rmd files are generated simply by updating these two lines with the intervention of interest (e.g. 'Breastfeeding withheld').

This basic Rmd file requires the following:

• Church_et_al_2018_metaregression_input.csv – containing the input data;
• Church_et_al_2018_plot_settings.csv - containing the plotting parameters for each intervention (axis limits, plot height, plot width, and footnote coordinates).

Compiling the analysis from the command line

Although the final analyses can be generated solely in R using the files provided, we used the script contained in Church_et_al_2018_command_line_compiler.sh to generate the intervention-specific analysis scripts from the basic Church_et_al_metaregression.Rmd script in unix. This script also renders the analyses into the e-appendix.

Other resources

The systematic review protocol has been published on PROSPERO.

The extracted data has been published on the Open Science Framework.