search

MJomaba / flu-evidence-synthesis

Code for reconstructing influenza dynamics from surveillance data using evidence synthesis techniques
19 stars 10 forks source link

Calculating the number of new cases per year using the vaccination_scenario function #59

Closed jdawa closed 7 years ago

jdawa commented 7 years ago

Hi, in the vaccination vignette, I’ve tried to change the number of age groups and risk groups from 7 age groups, 3 risk groups to 6 age groups and 2 risk groups. However, when I run the vaccination_scenario to get cases.per.year it seems to provide the number of new infected cases for 14 age groups instead of 12 age groups. The code I used is found below. If you look at cases.per.year1 at the end, it has cases in 14 age groups and not 12. Thanks for the help.

Load package and provide inputs for vaccine calendar

library("fluEvidenceSynthesis")

efficacy <- rep(c(0.7, 0.7, 0.7, 0.7, 0.7, 0.3), 2) calendar <- matrix(rep(0,12*4),ncol = 12) dates <- c(as.Date("2010-10-01"), as.Date("2010-11-01"), as.Date("2010-12-01"), as.Date("2011-01-01"))

Set rate of vaccine uptake for different dates/age groups

In this case the elderly (age group 6) start with a coverage of 0.62 in the first month followed by 0.7 and 0.925 in the second and third month. After three months the total uptake will be.

calendar[1,c(6,12)] <- 0.0 calendar[2,c(6,12)] <- 0.62 calendar[3,c(6,12)] <- 0.77 calendar[4,c(6,12)] <- 0.925

Set coverage for high risk young children

calendar[1,c(7,8)] <- 0.0 calendar[2,c(7,8)] <- 0.62 calendar[3,c(7,8)] <- 0.62 calendar[4,c(7,8)] <- 0.62

vaccine_calendar <- as_vaccination_calendar(efficacy = efficacy, dates = dates, coverage = calendar, no_age_groups = 6, no_risk_groups = 2)

Calculate the outbreak size

data("demography") data("polymod_uk") data("inference.results")

Reduce the number of age groups in polymod_uk to 6 by combining the last two age groups

poly <- polymod_uk[,c(1,2,3,4,5, 6,7,8)] poly[,8] <- rowSums(polymod_uk[,8:9]) polymod_uk <- poly

This returns the total size of the outbreak given the vaccination scenario and the 1000th posterior (mcmc) sample. The outbreak sizes is separated by age group and risk groups

cases.per.year1 <- vaccination_scenario(demography = demography, vaccine_calendar = vaccine_calendar, polymod_data = as.matrix(polymod_uk), contact_ids = inference.results$contact.ids[1000,], parameters = inference.results$batch[1000,], verbose = F )

BlackEdder commented 7 years ago

I am afraid what you are trying to do won't work like that for two reasons:

  1. You are using the fitting results from the uk fitting (data("inference.results")). These results are meant to be used with seven age groups, so it will still return those.
  2. If no incidence_function is passed it currently uses the infectionODEs function, which again assumes you are using seven groups.

I recommend you first try to get the fitting working following the more in-depth inference example. That should get you new fitting results that you can then use for this step.

jdawa commented 7 years ago

Thanks. I tried it again using your advice and it worked once I:

  1. Used the inference results that were calculated from my own data using the "more in-depth" inference example
  2. Added information on age_group_limits and risk_ratios when using the vaccination_scenario function
  3. And adjusted the infectionODEs parameters within the vaccination_scenario function to cater for my age groups