epiverse-trace / serofoi

Estimates the Force-of-Infection of a given pathogen from population based sero-prevalence studies
https://epiverse-trace.github.io/serofoi/
Other
17 stars 4 forks source link

Adding VPH use case to Use Cases article #212

Open ntorresd opened 2 days ago

ntorresd commented 2 days ago

A use case that has been discussed extensively during the development of the package is related to Human Papillomavirus (HPV). We've explored the use of age-varying forward random walk models with seroreversion to estimate the FOI in real-life scenarios. To exemplify this, consider the following simulated serosurvey:

foi <- data.frame(
  age = 1:70,
  foi = c(
    rep(0.0, 15),
    rep(0.1, 15),
    rep(0.01, 70-30)
  )
)

survey_features <- data.frame(
  age_min = seq(1, 70, 10),
  age_max = seq(10, 70, 10),
  n_sample = 100
)

serosurvey <- simulate_serosurvey(
  mode = "age",
  foi = foi,
  survey_features = survey_features,
  seroreversion_rate = 1.1
)

image

The options we've been using to model sexual debut from 15 years old on are similar to (custom FOI indexation can also be used):

seromodel <- fit_seromodel(
  model_type = "age",
  serosurvey = serosurvey,
  foi_prior = sf_normal(0, 1e-5), # set first FOI to zero
  foi_index = get_foi_index(serosurvey, 15),
  is_seroreversion = TRUE
)

Yielding the following results:

image

By specifying the seroreversion rate prior by means of the seroreversion_prior parameter is possible as long as the selection is justified.

We want to add a use case to the Real-Life Use Cases For Serofoi article (file). For this, we need to select and add a real-case dataset to the package and redact the use case in a similar manner to the others. @zmcucunuba, please branch from dev to work on this.