Specify overall column and heterogenity statistic on forest plot in Stata 17

[Janine Stubbs]

Hello!

I'm working in Stata 17 with the following meta set, declared as follows:

meta set logrr logrrl logrru, civarlevel(95) civartolerance(0.008) studylabel(study) eslabel(logRR) studysize(participants)

* Example generated by -dataex-. For more info, type help dataex
clear
input str12 study float year str20 country float(cohorts participants cases rr rrlow rrup logrr logrrl logrru) byte _meta_id str12 _meta_studylabel float _meta_es double(_meta_se _meta_cil _meta_ciu _meta_weight)
"Ericson"      2015 "Sweden"               1  26930  2860  .93 .89  .98  -.07257068 -.11653383 -.020202687 1 "Ericson"       -.07257068 .024574722931540643  -.12073625344783068  -.02440510969609022  894.3519672236014
"Buijsse"      2015 "Multi-country (EPIC)" 8  25307 11059  .94  .9  .98  -.06187541 -.10536055 -.020202687 2 "Buijsse"       -.06187541 .021724342610256155   -.1044543357565999 -.019296477548778057 1014.0362069741808
"Guasch-Ferre" 2015 "USA"                  1 145087  9652  .98 .94 1.02 -.020202687 -.06187541   .01980261 3 "Guasch-Ferre" -.020202687 .020836611477435422 -.061041695065796375  .020636321045458317 1054.4347766510025
"Monbonen"     2005 "Finland"              1   4304   383 1.04 .94 1.15   .03922068 -.06187541   .13976192 4 "Monbonen"       .03922068 .051439039948394695  -.06159798987209798   .14003934152424335  316.4370854931246
end

The type of forest plot I need is:

meta forestplot, eform(risk ratios) random(dlaird)

I would also like to see each study size (given by the variable "participants").

The problem is that in specifying this, the "Overall" column title and the heterogeneity statistics are lost. Please compare:

meta forestplot study year participants _plot _esci _weight, eform(risk ratios) random(dlaird)

I would like help to specify that the "Overall" column title and heterogeneity statistics are displayed on the forest plot.
The manual apparently only instructs on how to suppress this information from the forest plot.

Thankyou,

Janine

[Andrew Musau]

Thanks for the data example. Here is a workaround that includes the extra information that you want. It appears that the default output just presents the contents of

meta summarize

There should be a way to replicate this output using other options. I do not do meta-analysis studies and the command's options do not appear to provide the flexibility that you want. You may want to send an email to Technical Services and they should be able to get in touch with someone who has detailed knowledge of the command. Meanwhile, see if the following is satisfactory.

* Example generated by -dataex-. For more info, type help dataex
clear
input str12 study float year str20 country float(cohorts participants cases rr rrlow rrup logrr logrrl logrru) byte _meta_id str12 _meta_studylabel float _meta_es double(_meta_se _meta_cil _meta_ciu _meta_weight)
"Ericson"      2015 "Sweden"               1  26930  2860  .93 .89  .98  -.07257068 -.11653383 -.020202687 1 "Ericson"       -.07257068 .024574722931540643  -.12073625344783068  -.02440510969609022  894.3519672236014
"Buijsse"      2015 "Multi-country (EPIC)" 8  25307 11059  .94  .9  .98  -.06187541 -.10536055 -.020202687 2 "Buijsse"       -.06187541 .021724342610256155   -.1044543357565999 -.019296477548778057 1014.0362069741808
"Guasch-Ferre" 2015 "USA"                  1 145087  9652  .98 .94 1.02 -.020202687 -.06187541   .01980261 3 "Guasch-Ferre" -.020202687 .020836611477435422 -.061041695065796375  .020636321045458317 1054.4347766510025
"Monbonen"     2005 "Finland"              1   4304   383 1.04 .94 1.15   .03922068 -.06187541   .13976192 4 "Monbonen"       .03922068 .051439039948394695  -.06159798987209798   .14003934152424335  316.4370854931246
end

meta set logrr logrrl logrru, civarlevel(95) civartolerance(0.008) studylabel(study) eslabel(logRR) studysize(participants)
replace _meta_studylabel = study + " (N = " + string(participants)+ ")"
meta forestplot, eform(risk ratios) random(dlaird)

Res.:

[Janine Stubbs]

Those are good ideas Andrew,
Thankyou for your help!
Janine