Feat/vaccines #209

giovannic · 2023-01-03T14:40:33Z

Updates the vaccine modelling to allow arbitrary vaccine modelling parameters through the VaccineProfile class.

As an example of how VaccineProfiles are created, see the exported rtss_profile and rtss_booster_profile objects in vaccine_parameters.R

A future PR could be made to export new vaccine profiles.

Changes:

rename vaccine strategy functions from set_mass_rtss, set_rtss_epi -> set_mass_vaccination, set_vaccine_epi
add vaccine profile parameters to vaccine strategy functions
remove vaccine antibody variables rho/ds/dl...
change vaccination processes to sample antibody values at simulation time (each timestep) from the corresponding vaccine profile
fix regression tests and vaccine vignette

Review requests/notes:

Code review: (@pwinskill ?) most of this is renaming rtss -> vaccine, sorry if that makes most of the changes tedious. The most critical changes are in human_infection.R, where the sampling code is a bit different from before.
End-user review: (@htopazian ?) does the vaccine vignette make sense? Is there anything else we would want to document in this PR?

pwinskill

Hi @giovannic
This looks good to me, just one small comment and the following genral one:

Renaming RTSS to vaccine throughout has the possibility to cause seme confusion/ambiguity as we also have other types of vaccine in the model, tbv. We could leave this for now (as it isn't urgent) and make sure that the documentation states clearly that vaccine options refer to pre-erythrocytic vaccines only. I guess our other options would be to change all the new vaccines to pev, or to somehow include the tbv implmentation under the vaccine implementation (both of which seem a bit onerous to do!)

pwinskill · 2023-01-05T15:18:55Z

R/human_infection.R

  if (length(vaccinated) > 0) {
-    vaccinated_index <- source_vector[vaccine_times > -1]
-    antibodies <- calculate_rtss_antibodies(
+    vaccinated_index <- source_vector[vaccinated]


I can't see where vaccinated_index gets used

pwinskill · 2023-01-05T15:39:17Z

@htopazian it might be helpful if you could re-run at least one of the previous RTSS scenarios with the new code as a check that we're getting the same result?

htopazian

Nice Giovanni! A few comments in the file changes

man/set_mass_vaccination.Rd

R/vaccines.R

R/vaccine_parameters.R

R/parameters.R

R/vaccine_parameters.R

R/variables.R

htopazian · 2023-01-06T11:49:26Z

@htopazian it might be helpful if you could re-run at least one of the previous RTSS scenarios with the new code as a check that we're getting the same result?

@pwinskill great idea - looking pretty similar for the mass vax strategy with 100,000!

* correlation test passing * VaccineProfile class executing Fix infection integration tests Fix EPI regressions Fix regression tests Fix Vaccines vignette: * update vignette to use new vaccine API * update process and event creation for new vaccine code Fix R CMD check issues: * Fix correlation example * Update documentation for RTSS vaccine profiles update vaccine vignette Rename vaccine to pev Update vignettes for new function names Bump version and update NEWS Remove unused line

htopazian

Thanks Giovanni, it is looking good and problems with epi boosters are fixed!

pwinskill

Looks good - thank ❤❤
A couple of sticking points highlighted in the comments

pwinskill · 2023-03-10T12:52:04Z

R/pev_parameters.R

+#' @param profile primary vaccine profile of type PEVProfile
+#' @param coverages a vector of coverages for the primary doses
+#' @param timesteps a vector of timesteps associated with coverages
+#' @param age for the target population, (in timesteps)


Not clear to me what this means - is it a single value (upper bound)?

Sorry...presumably age of first dose? Would be helpful to clarify in the help

pwinskill · 2023-03-10T12:57:12Z

R/pev_parameters.R

+#' @export
+set_pev_epi <- function(
+  parameters,
+  profile,


Perhaps for a seperate PR, but is there a reason why the primary profile is a single fixed profile, while the booster can vary? Would we not want flexibility for both?

Ah, is the idea that we could have seperate profiles for the primary doses? We would need a new model for that, since we only model protection after the final dose.

Not seperate for the 3 doses in the primary series, but seperate vaccines given for the primary series at different timepoints, eg.
Year 1: primary series = 3 doses of vaccine A
Year 2: primary series = 3 doses of vaccine B

Oh I see! Yes, we could implement this feature if we want to model transition between vaccine types. Made an issue here: #228

pwinskill · 2023-03-10T14:32:27Z

R/pev.R

-      not_recently_vaccinated <- variables$rtss_vaccinated$get_index_of(
-        a = max(timestep - parameters$rtss_epi_min_wait, 0),
+      not_recently_vaccinated <- variables$vaccinated_timestep$get_index_of(
+        a = max(timestep - parameters$pev_epi_min_wait, 0),


The following was failing for me, I think in relation to this section:

p <- get_parameters(overrides = list( human_population = 1000, individual_mosquitoes = FALSE )) p <- p |> set_pev_epi( profile = rtss_profile, coverages = c(0.99, 0.99, 0.99), timesteps = 365 * (2:4) + 1, age = 5 * 365, min_wait = 180, booster_profile = list(rtss_profile, rtss_profile, rtss_profile), booster_timestep = c(190, 190, 190), booster_coverage = c(0.8, 0.8, 0.8) ) s <- run_simulation(timesteps = 365 * 7, parameters = p)

Setting min_wait = 0 resolved the issue. Is the a reasonable check we can make on what the user inputs here?

pwinskill · 2023-03-10T14:58:43Z

R/pev_parameters.R

+#' @description Parameters for a primary dose of RTS,S for use with the
+#' set_mass_pev and set_pev_epi functions
+#' @export
+rtss_profile <- create_pev_profile(


Really like how these profiles are set up btw 👍

pwinskill

Thanks @giovannic!

giovannic requested review from pwinskill and htopazian January 4, 2023 08:42

pwinskill reviewed Jan 5, 2023

View reviewed changes

htopazian reviewed Jan 6, 2023

View reviewed changes

Update vaccine parameterisation

22ddaa6

giovannic force-pushed the feat/vaccines branch 4 times, most recently from 29d3203 to fdf4d8b Compare February 28, 2023 11:17

giovannic added 2 commits March 7, 2023 11:27

Remove make PEVProfile for list based alternative

00d5383

giovannic force-pushed the feat/vaccines branch from fdf4d8b to 00d5383 Compare March 7, 2023 11:27

Spelling errors in pev changes

2f72e7d

giovannic requested review from pwinskill and htopazian March 7, 2023 11:55

htopazian approved these changes Mar 8, 2023

View reviewed changes

pwinskill reviewed Mar 10, 2023

View reviewed changes

Fix min_wait > 0 bug in pev_epi

b4093eb

giovannic force-pushed the feat/vaccines branch from 9b80d68 to b4093eb Compare March 10, 2023 15:24

giovannic requested a review from pwinskill March 13, 2023 09:15

pwinskill approved these changes Mar 13, 2023

View reviewed changes

giovannic mentioned this pull request Mar 13, 2023

Time-varying vaccine profiles for set_pev_epi #228

Open

giovannic merged commit 03cd3d5 into dev Mar 13, 2023

giovannic deleted the feat/vaccines branch March 17, 2023 11:13

htopazian mentioned this pull request Mar 30, 2023

Vignettes/vaccine #226

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feat/vaccines #209

Feat/vaccines #209

giovannic commented Jan 3, 2023 •

edited

Loading

pwinskill left a comment

pwinskill Jan 5, 2023

pwinskill commented Jan 5, 2023

htopazian left a comment

htopazian commented Jan 6, 2023

htopazian left a comment

pwinskill left a comment

pwinskill Mar 10, 2023

pwinskill Mar 10, 2023

pwinskill Mar 10, 2023

giovannic Mar 10, 2023

pwinskill Mar 13, 2023

giovannic Mar 13, 2023

pwinskill Mar 10, 2023

pwinskill Mar 10, 2023

pwinskill left a comment

Feat/vaccines #209

Feat/vaccines #209

Conversation

giovannic commented Jan 3, 2023 • edited Loading

pwinskill left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pwinskill commented Jan 5, 2023

htopazian left a comment

Choose a reason for hiding this comment

htopazian commented Jan 6, 2023

htopazian left a comment

Choose a reason for hiding this comment

pwinskill left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pwinskill left a comment

Choose a reason for hiding this comment

giovannic commented Jan 3, 2023 •

edited

Loading