EPI 590R - Functions, continued

Writing functions to reduce copy and paste

If you find yourself doing something over and over, you can probably write a function for it.

One is example is the regression tables we created yesterday.

Let’s imagine we are fitting multiple similar models

logistic_model <- glm(glasses ~ eyesight_cat + sex_cat + income,
  data = nlsy, family = binomial()
)
poisson_model <- glm(nsibs ~ eyesight_cat + sex_cat + income,
  data = nlsy, family = poisson()
)
logbinomial_model <- glm(glasses ~ eyesight_cat + sex_cat + income,
  data = nlsy, family = binomial(link = "log")
)

We have a model for a table we want to create

tbl_regression(
  poisson_model,
  exponentiate = TRUE,
  label = list(
    sex_cat ~ "Sex",
    eyesight_cat ~ "Eyesight",
    income ~ "Income"
  )
)

Even though this is already a function, we can wrap it in a new function!

New table function

We can refer generically to model and then put the model we want a table for as an argument

new_table_function <- function(model) {
  tbl_regression(
    model,
    exponentiate = TRUE,
    label = list(
      sex_cat ~ "Sex",
      eyesight_cat ~ "Eyesight",
      income ~ "Income"
    )
  )
}

Exercise

Copy and paste this function into your script
Create each of the models and run
Try to figure out how you could allow someone to pass the tidy_fun argument (from the regression exercises yesterday)