In slevu/serpico2: Seroprevalence of SARS-CoV-2

knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE,
                       collapse = TRUE) #, comment = "#>")
knitr::opts_knit$set(root.dir = '../')
options(knitr.table.format = function() {
  if (knitr::is_latex_output()) 
    "latex" else "pandoc"
})
# options(scipen=999)

library(serpico2)

knitr::read_chunk("R/run_model.R")
run <- TRUE

Simulating a sample dataset

We start by simulating a sample dataset using function simulate_mock_data. By default it represents seropositivity measures in distinct individuals stratified by sex, age and region, and over 3 time periods. A baseline prevalence level is set for the second period and risks of infection are varied by period and individual covariates. Test positivity is a function of sensitivity and specificity. In this simulated example and for simplicity, response rate is considered fixed for now.

Sample looks like so:

rbind(head(d0), tail(d0))

And we can assess observed prevalence (test positivity) by categories

vars = c("time", "sex", "age", "reg")
print( lapply(setNames(vars, vars), function(v){
      round( tapply(d0[,"res"], d0[,v], mean), 3)
    }) )

We will also need poststratification population counts and region names (poststrat and code_region datasets). Following actual study input and results, we exclude one region (Mayotte).