Open rcsmit opened 2 years ago
Bonjour, Hello, Thanks for your question.
There is currently no calculation of confidence intervals for the vaccine effectiveness, in this repo. It could be done but one needs to be prudent because Farrington's method as said in slide 33 of https://eva.ecdc.europa.eu/mod/resource/view.php?id=5149:
"Key issue: Coverage data must be accurate – uncertainty not allowed for in 95% CIs"
The confidence intervals you found in the _00functions.R are used for the RR https://vaccine-effectiveness.datacovidvaud.ch/02a_VE_screening_method_overall.html#Incidence_rates_amongst_fully_vaccinated_and_others .
Hope this helps
Kind regards Olivier Duperrex
Bonjour, Hello,
How does your script calculate the Confidence intervals?
Farrington says:"This model is easily fitted by logistic regression using GLIM,with the number of vaccinated cases as the dependent variable, a binomial error structure with index equal to the number of cases, and the vector logit pi_i as offset. The fitting procedure produces standard errors for the parameters which may be used to calculate CI."
But I can only find Fisher and Wald in the script line 113 en 122 of 00_functions.R
Kind regards Rene Smit