InstituteforDiseaseModeling / covasim

COVID-19 Agent-based Simulator (Covasim): a model for exploring coronavirus dynamics and interventions
https://covasim.org
MIT License
255 stars 224 forks source link

"Time for disease recovery" parameters #426

Open lequ02 opened 5 months ago

lequ02 commented 5 months ago

Describe the bug

1) "Time for disease recovery" parameters in "covasim/covasim/README.rst" is not up to date. 2) Not sure if this is a bug but I don't understand why crit2rec == sev2rec == 18.1 3) And for crit2die, I think 9.7 days would be a better number

Expected behavior

1) The README should match the code in covasim/covasim/parameters.py 2) I think crit2rec should be this: 16.6 days = 24.7 onset-to-recovery - 6.6 sym2sev - 1.5 sev2crit 3) In the calculation provided in the code, 18.8 days total was used instead of 17.8, which is a more precise number. In the study cited in the code (Verity et al.), the authors said that: "In the subset of 24 deaths...early in the epidemic...we estimated the mean time from onset to death to be 18·8 days [15·7–49·7]" "However...our posterior estimate...informed by the analysis of aggregated data from China, is MORE PRECISE (mean 17·8 days [16·9–19·2])"

Additional context 1) The code parameters were updated to use more accurate numbers from recent research but the README was not updated.