The codes for simulations were written in Fortran and compiled with the Intel Fortran Compiler. Data analysis and figures were done Python 3.10 and the following open source libraries: pandas, matplotlib, and seaborn.

In this repository we show only the codes for the simulations.

The preprint is available at https://arxiv.org/abs/2201.02869. The following BibTeX code can be used to cite it:

@misc{schulenburg2022effects,
      title={Effects of infection fatality ratio and social contact matrices on vaccine prioritization strategies}, 
      author={Arthur Schulenburg and Wesley Cota and Guilherme S. Costa and Silvio C. Ferreira},
      year={2022},
      eprint={2201.02869},
      archivePrefix={arXiv},
      primaryClass={q-bio.PE}
}

See also Compartmental epidemic model to assess undocumented infections: applications to SARS-CoV-2 epidemics in Brazil and Outbreak diversity in epidemic waves propagating through distinct geographical scales.

All codes and data are available in the folder codes. Each *.f90 inside codes/main and codes/supp folders correspond to the respective figure in the main paper and supplementary material, respectively. The data (contact matrices and vaccine efficacies) are available at a common folder codes/data, with files:

• vaccines-data.dat: file with four columns - reduction in deaths with Vaxzevria, CoronaVac, and of infections with Vaxzevria and CoronaVac, respectively.
• <country>/contact_<contact>.dat: contact matrix C_ij (16x16) for each <country> (Brazil, Germany or Uganda) and contact scenarios <cenario> - all (unmitigated) or all-sd (social distancing).

The codes can be compiled with the Intel Fortran Compiler or with GNU Fortran compiler. Examples:

• Intel: ifort program.f90 -o program
• GFortran: gfortran program.f90 -o program

To execute, use ./program > output.dat or simply ./program.

Some of the *.f90 codes generates the time evolution data, and others the heatmaps. In each *.f90 file, enter the corresponding parameters after the !#Initial conditions comment.

The parameters used in the codes are the following:

1. R0: infection rate parameter ϖ
2. csipop: value of vaccination rate ξ
3. dia: days to start the vaccination (t_v)
4. ig: One of the following prioritization strategies:
   • 2: DAP strategy
   • 3: HVP strategy
   • 7: NP strategy
   • 8: DCP strategy
5. cenario: contact patterns scenario
   • 0: Unmitigated, 100% of all contacts
   • 2: Social distancing, considers the reduction of contact in the social distancing scenario
6. vacina: which vaccine data to use (see file codes/data/vaccines-data.dat)
   • 0: CoronaVac
   • 1: Vaxzevria

For the temporal evolution program, the results will have the value of each parameter in their name, as mentioned in the description of the main parameters. For the vaccine efficacies, CV refers to CoronaVac, but can be changed in the code for the corresponding name to the data used.

For the programs that generate the heatmap the name can have R0-tx, corresponding to the heatmap of vaccination rate ξ versus ϖ, or dia-tx, corresponding to the heatmap of delay t_v versus ϖ.

The first line of each *.dat file indicates what the values of the columns mean.

Here we have the result for the temporal evolution with the parameters:

• R0-13: R0 = 1.3
• tx-15: csipop = 0.0015
• fx4: results are for age group 4
• ig2: ig = 2 (DAP strategy)
• dia60: dia = 60 (t_v = 60)
• cenario-2: cenario = 2 (social distancing scenario)
• CV: vacina = 0 (CoronaVac)

Here we have the results for the heatmap of vaccination rate versus ϖ (represented by "R0-tx") with the parameters:

• T: the results are for total population
• ig8: ig = 8 (DCP strategy)
• dia30: dia = 30 (t_v = 30)
• CV: vacina = 0 (CoronaVac)
• cenario-0: cenario = 0 (unmitigated scenario)