In R4EPI/tuni: Tables for Epidemiological Analysis

Examples

Here is an example of the tables produced by {epitabulate} from a randomly generated data set based on an MSF data dictionary for Measles data.

library(epidict)
library(matchmaker)
library(epikit)
library(dplyr)
library(epitabulate)

linelist <- epidict::gen_data("Measles", numcases = 1000, org = "MSF")
measles_dict <- epidict::msf_dict("Measles", compact = FALSE)

# Cleaning linelist data
linelist_clean <- matchmaker::match_df(
  x = linelist,
  dictionary = measles_dict,
  from = "option_code",
  to = "option_name",
  by = "data_element_shortname",
  order = "option_order_in_set"
)
linelist_clean

Quick proportions with conficence intervals

There are three functions that will provide quick statistics for different rates based on binomial estimates of proportions from binom::binom.wilson()

attack_rate()
case_fatality_rate()
mortality_rate()

attack_rate(10, 50)
case_fatality_rate(2, 50)
mortality_rate(40, 50000)

In addition, it's possible to rapidly calculate Case fatality rate from a linelist, stratified by different groups (e.g. gender):

case_fatality_rate_df(linelist_clean, 
                      exit_status %in% c("Dead in facility", "Dead on arrival"), 
                      group = sex, 
                      add_total = TRUE, 
                      mergeCI = TRUE)

Quick tabulations

the_symptoms <- tab_linelist(linelist_clean,
  cough, nasal_discharge, severe_oral_lesions,
  transpose = "value"
) 
the_symptoms

the_symptoms %>%
  epikit::rename_redundant("%" = "proportion") %>%
  epikit::augment_redundant("(n)" = "n") %>%
  knitr::kable()

2x2 tables

In R, creating 2x2 tables is as simple as using the function table(), but unfortunately, it can be difficult to interpret the values of these tables because the dimensions are often flipped around for different analyses. The function data_frame_from_2x2() will present these values labeled unambiguously.

symptoms_tf <- linelist_clean %>%
  dplyr::transmute(
    pneumonia = grepl("Yes", pneumonia),
    cough = grepl("Yes", cough),
    nasal_discharge = grepl("Yes", nasal_discharge),
    oral_lesions = grepl("Yes", severe_oral_lesions),
    contact = grepl("known case", contact_history)
  ) 

symptoms_tf

print(pxc <- with(symptoms_tf, table(pneumonia, cough)))
print(pxcxc <- with(symptoms_tf, table(pneumonia, cough, contact)))

data_frame_from_2x2(pxc)
data_frame_from_2x2(pxcxc)

Odds / Risk / Incidence Rate Ratios

tu <- tab_univariate(symptoms_tf, 
  outcome = pneumonia, 
  cough, nasal_discharge, oral_lesions, contact,
  mergeCI = TRUE,
  extend_output = FALSE,
  measure = "OR"
)
tu %>% 
  select(-est_type) %>% 
  epikit::augment_redundant("exposed " = "exp_") %>%
  rename(odds = est_ci) %>%
  knitr::kable(digits = 3)