Epidemiology & Technology

Easy Estimates Table after Logistic Regression

We need to address several concerns when running logistic regression in Stata.

  • You need to have estimates only for certain decimal point .
  • The 95% CIs should follow the ORs in parenthesis
  • Should be easy to identify the Reference category
  • Ideally would want to see the labels of coded categorical values

For such situation, the fantastic esttab and estout packages are incredibly useful. An example follows

logistic outcome i.ageGrp   i.sex  i.rural , baselevels cformat(%5.2f) pformat(%5.3f) sformat(%8.3f)
est store e1

estout e1 ,   cells (" b(fmt(2))    ci(par( ( , ) )) p(fmt(3)) ") label   eform  Code language: Stata (stata)

One thing to note, the Variables are not displayed, only their categories. It may prove to be problematic if many variables use same value tables. So instead of using Yes/No value labels, it is helpful to label categories as Hypertensive/Nonhypertensive, diabetic/non-diabetic, etc

Related posts