Lars Mølgaard Saxhaug https://twitter.com/load_dependent
Last compiled on Friday 02 July, 2021
ve_true <- 0.8 # vaccine efficacy, 1-rr
sample_size <- 1e6 # size of each group
weeks <- 52 # maximum number of weeks of follow up
df <- tibble(
week = rep(1:weeks, 3),
base_risk = c(rep(0.01, weeks), rep(0.005, weeks), rep(0.001, weeks)), # base risk per week, three
vaccinated_risk = base_risk - ve_true * base_risk,
vaccinated_cases=sample_size-sample_size * (1 - vaccinated_risk)^week,
placebo_cases=sample_size-sample_size * (1 - base_risk)^week)
d <- df %>%
group_by(base_risk) %>%
nest()
df <- df %>%
mutate(
eer = (vaccinated_cases) / sample_size, # exposed event risk
cer = (placebo_cases) / sample_size, # placebo event risk
rr = eer / cer,
ve = 1 - rr,
ar = cer - eer,
nntv = 1 / ar
)
#### Variable
vaccine efficacy
rm(ve_true)
sample_size <- 1e6 # size of each group
weeks <- 52 # maximum number of weeks of follow up
base <- 0.005
df <- tibble(
week = rep(1:weeks, 3),
ve_true= c(rep(0.5, weeks), rep(0.7, weeks), rep(0.9, weeks)),
base_risk = rep(base, 3*weeks),
vaccinated_risk = base_risk - ve_true * base_risk,
vaccinated_cases=sample_size-sample_size * (1 - vaccinated_risk)^week,
placebo_cases=sample_size-sample_size * (1 - base_risk)^week)
df <- df %>%
mutate(
eer = (vaccinated_cases) / sample_size, # exposed event risk
cer = (placebo_cases) / sample_size, # placebo event risk
rr = eer / cer,
ve = 1 - rr,
ar = cer - eer,
nntv = 1 / ar
)
