Create 95% CI’s from broom/modelr objects

[mbcann01]

Here are some functions I created in the WHI sexual function analysis to help get the results of logistic regression in the format I wanted

tidy_glm <- function(x, ...) {
 out <- tidy(x)
 ci  <- confint(x) %>% exp() %>% as.data.frame()
 out <- bind_cols(out, ci)
 out <- out %>%
   mutate(
     estimate = exp(estimate),
     p_value = round(p.value, 10) %>% format(nsmall = 6L)
   ) %>% 
   select(-p.value)
 out
}

tidy_output_mids <- function(x, ...) {
  x %>% 
    summary() %>% 
    tidy() %>% 
    mutate(
    est = exp(est) %>% round(2) %>% format(nsmall = 2L),
    lo.95 = exp(lo.95) %>% round(2) %>% format(nsmall = 2L),
    hi.95 = exp(hi.95) %>% round(2) %>% format(nsmall = 2L),
    p_value = round(`Pr...t..`, 10) %>% format(nsmall = 6L),
    or_95 = paste0(est, " (", lo.95, " - ", hi.95, ")")
  ) %>% 
  select(.rownames, est, lo.95, hi.95, or_95, p_value)
}

[mbcann01]

To be honest, I don't love the default output from Broom. I might want to make my own broom-like functions.