Plotting the odds ratio of multiple multivariate logistic regressions in one plot

[Zakia Nouri]

Hi everyone! I am trying to plot the odds ratios of multiple multivariate logistic regressions (y1, y2, y3, and y4) in one plot, and I hope someone can help me with that. I’d like to have a plot that looks like the one below where only the odds ratio of the variable of interest (x1) is plotted not the control variables (x2, x3, and x4). I am using Stata 14.2 and very much appreciate your help in advance.

Here is how my code looks like:

local dichotomies y1 y2 y3 y4
tempfile results
capture postutil clear
postfile handle str32 outcome_var str6 model float (x1 p lb ub) using `results'
 
foreach d of local dichotomies {
         logit `d' i.x1 x2 i.x3 i.x4
         matrix M = r(table)
         post handle ("`d'") ("logit") (M[1, 2]) (M[4, 2]) (M[5, 2]) (M[6, 2])
}
 
postclose handle
preserve
use `results', clear
gen odds_ratio = exp(x1)
replace lb = exp(lb)
replace ub = exp(ub)
list, noobs clean
export excel using "[…]\odds_ratio.xls", firstrow(variables) replace
restore

[Maarten Buis]

Take a look at ssc desc coefplot and https://www.stata-journal.com/articl...article=gr0059

[Tiago Pereira]

If the 95% CIs (on the log scale) are symmetrical, you can use -meta- with the option of equal weights. Otherwise, you can use -metan-, which has many options.

[Zakia Nouri]

Thanks Maarten for your response. I did check the coefplot, but I couldn't find anything about how to drop the x2, x3, and x4 odds ratio. I am only interested to plot the odds ratio of x1 for y1, y2, y3, and y4. Do you know if there is anyway to do that?

[Maarten Buis]

You can use the keep() and/or drop() options to select the coefficients you need to show

To display odds ratios you need to use the eform option, as the odds ratio is the exponentiated coefficient in a logit regression.

All that is explained in both the helpfile and the article I linked to in #2