How to estimate OR and 95%CI using Predict funtion for svyglm model?

[jiangaimin1995]

design<-svydesign(data=data1115, id=~id, strata=~strata, weights=~weightage, nest=TRUE)
svy.fit<-svyglm(outcome ~ age + gender + drugs + educations + income, family=quasibinomial(link="logit"), design = design)
y<-predict(svy.fit)
y<-as.numeric(y)
#transform into probability
y<-exp(y)/(1+exp(y))

And then, how should I get the OR and 95%CI?