chest

Change-in-Estimate Approach to Assess Confounding Effects

Applies the change-in-effect estimate method to assess confounding effects in medical and epidemiological research (Greenland & Pearce (2016) <doi:10.1146/annurev-publhealth-031914-122559> ). It starts with a crude model including only the outcome and exposure variables. At each of the subsequent steps, one variable which creates the largest change among the remaining variables is selected. This process is repeated until all variables have been entered into the model (Wang Z. Stata Journal 2007; 7, Number 2, pp. 183–196). Currently, the 'chest' package has functions for linear regression, logistic regression, negative binomial regression, Cox proportional hazards model and conditional logistic regression.

Versions across snapshots

Dependencies (latest)