Calculate risk ratios and risk differences using a Poisson distribution for person-time data. Function works on individual level data or aggregated data p244 2nd Edition
rate(data, outcome, denominator, exposure, per_unit, ci_level = 95)A dataframe
Variable with the outcomes as a numeric variable
Variable giving the amount of time at risk
Variable giving whether exposed or not
Multiplier for rate values, e.g. 1000 for n outcomes per 1000 denominator
A numeric value giving the confidence interval
# Using individual level data
data(ebola)
library(dplyr)
#>
#> Attaching package: ‘dplyr’
#> The following objects are masked from ‘package:stats’:
#>
#> filter, lag
#> The following objects are masked from ‘package:base’:
#>
#> intersect, setdiff, setequal, union
ebola %>%
mutate(male = ifelse(sex == "male", 1, 0)) %>%
rate(outcome = died, denominator = days_at_risk, exposure = male,
per_unit = 100)
#> # A tibble: 2 × 10
#> male outcome denominator rate rate_ratio rate_ratio_lci rate_ratio_uci
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 133 1056 12.6 1.04 0.799 1.36
#> 2 1 91 754 12.1 1 0.766 1.31
#> # ℹ 3 more variables: rate_diff <dbl>, rate_diff_lci <dbl>, rate_diff_uci <dbl>
# Using aggregated data
# Table 14-2
cancer_xray <- data.frame(cases = c(41, 15), pyar = c(28010, 19017),
radiation = c(1, 0))
cancer_xray
#> cases pyar radiation
#> 1 41 28010 1
#> 2 15 19017 0
cancer_xray %>%
rate(outcome = cases, denominator = pyar, exposure = radiation,
per = 100000)
#> # A tibble: 2 × 10
#> radiation outcome denominator rate rate_ratio rate_ratio_lci rate_ratio_uci
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 15 19017 78.9 1 0.554 1.81
#> 2 1 41 28010 146. 1.86 1.03 3.35
#> # ℹ 3 more variables: rate_diff <dbl>, rate_diff_lci <dbl>, rate_diff_uci <dbl>