Univariate regression (Odds, Risk, or Rate Ratios)

Performs univariate regression for each exposure on a binary, continuous, or count outcome. Depending on `approach`, returns either Odds Ratios (OR), Risk Ratios (RR), or Incidence Rate Ratios (IRR).

Usage

uni_reg(data, outcome, exposures, approach = "logit")

Arguments

data: A data frame containing the variables.
outcome: outcome variable (binary, continuous, or count).
exposures: A vector of predictor variables.
approach: Modeling approach to use. One of: `"logit"` (OR), `"log-binomial"` (RR), `"poisson"` (IRR), `"robpoisson"` (RR), `"linear"` (Beta coefficients), `"negbin"` (IRR)

Value

A list of class `uni_reg` and `gtsummary::tbl_stack`, including:

A publication-ready regression table (`tbl_stack`)
Accessor elements:
- `$models`: Fitted regression models for each exposure
- `$model_summaries`: Tidy model summaries
- `$reg_check`: Diagnostics (only for linear regression)

Details

This function requires the following packages: `dplyr`, `purrr`, `gtsummary`, `risks`.

Examples

data(PimaIndiansDiabetes2, package = "mlbench")
library(dplyr)
pima <- PimaIndiansDiabetes2 |>
  dplyr::mutate(diabetes = ifelse(diabetes == "pos", 1, 0))
uni_reg(pima, outcome = "diabetes", exposures = "age", approach = "logit")


  Characteristic
      N
      OR
      95% CI
      p-value
    
age
768
1.04
1.03, 1.06
<0.001
Abbreviations: CI = Confidence Interval, OR = Odds Ratio

Characteristic	N	OR	95% CI	p-value
age	768	1.04	1.03, 1.06	<0.001
Abbreviations: CI = Confidence Interval, OR = Odds Ratio