Skip to content

Using mechanistic model-based inference to understand and project epidemic dynamics with time-varying contact and vaccination rates: New Zealand’s 2021 Covid-19 outbreak

License

Notifications You must be signed in to change notification settings

michaelplanknz/model-inference-covid19-nz2021

Repository files navigation

model-inference-covid19-nz2021

Using mechanistic model-based inference to understand and project epidemic dynamics with time-varying contact and vaccination rates: New Zealand’s 2021 Covid-19 outbreak

Matlab code to reproduce the analysis in "Using mechanistic model-based inference to understand and project epidemic dynamics with time-varying contact and vaccination rates: New Zealand’s 2021 Covid-19 outbreak"

  • Run go.m with a specified choice of fittedToDate to fit the model using case data up to that date and plot the main figures in the article. go.m is the top-level Matlab script that calls three functions:
    1. fitToData.m [if the relevant results file /results/ABCSMC_cases_to_DD-MMM-YYYY.mat does not exist] - runs the ABCSMC algorithm described in the paper to generate a sample from the posterior distribution of the values of the time-varying control function C(t) by fitting to data on new daily cases up to the specified time point
    2. runForwardSims.m [if the relevant results file /results/results_fit_to_DD-MMM-YYYY.mat does not exist] - takes the posterior samples saved in ABCSMC_cases_to_DD-MMM-YYYY does not exist and runs the forward model for a longer period of time to generate projections of the future epidemic dynamics under the current estmiated values of the control function
    3. plotGraphs.m - takes the results of the forward simulations saved in /results/results_fit_to_DD-MMM-YYYY and plots graphs with data superimposed
  • Run drawVaccineGraphs.m to reproduce Fig. 1 in the article.
  • Run plotPriorDraws.m to reproduce Supplementary Figure S2.

RESULTS

Results are saved in /results/

  • results/ABCSMC_cases_to_DD-MMM-YYYY.mat contains the results of the ABCSMC parameter inference part of the method fitted to - data up TO DD-MMM-YYYY.
  • results/results_fit_to_DD-MMM-YYYY.mat contains the output of the forward model using the parameters saved in ABCSMC_cases_to_DD-MMM-YYYY
  • figures/results_fit_to_DD-MMM-YYYY.png contains grpahs of the model output with data superimposed

DATA

Data files are read in from /data/

POPULATION DATA

  • data/nzpopdist.xlsx New Zealand population size in 5-year age bands
  • data/nzcontmatrix.xlsx Contact matrix for the New Zealand popultion from Prem et al.

EPIDEMIC DATA

  • data/epiData.xlsx Number of new reported cases of Covid-19 and number of hopsital beds occupied with Covid-19 cases on each day from 17 Aug 2021 to 23 Jan 2022

VACCINE DATA

  • data/dates_vaccine.csv Dates covered by the vaccine data
  • data/firstDoseProp_akl_YYYY-MM-DD.csv Proportion of the population who had received their 1st (but not 2nd) dose in 5-year age bands and on each date shown in dates_vaccine.csv, according to MOH data available on YYYY-MM-DD
  • data/secondDoseProp_akl_YYYY-MM-DD.csv Proportion of the population who had received their 2nd dose in 5-year age bands and on each date shown in dates_vaccine.csv, according to MOH data available on YYYY-MM-DD

REFF ESTIMATES FROM EPINOW2

  • data/EpiNow2_estimates_AllRegions_MOHdata_YYYYMMDD.csv Output from the Epinow2 model including estimates of the time-varying reproduciton number

About

Using mechanistic model-based inference to understand and project epidemic dynamics with time-varying contact and vaccination rates: New Zealand’s 2021 Covid-19 outbreak

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages