Forest plot showing I² = 100% with only one study (Stata/BE 19.5)

[Sandra Martinez]

Hello,

I’m using Stata/BE 19.5, and when generating a forest plot with only one study (for visualization purposes only), I notice that Stata displays I² = 100% in the output. I understand that heterogeneity should not be calculated with a single study, as it's not statistically applicable.

My intention is not to perform a full meta-analysis but rather to visually present the effect estimate of that study, as is sometimes done in Cochrane reviews.

I would also like to know if there is any way to display several independent meta-analyses together in one forest plot, without treating them as subgroups or calculating an overall combined effect.

Is there any way to control or avoid this behavior in Stata, or any recommended workaround?

Thanks in advance for any suggestions.

Best regards,
Sandra

[Andrew Musau]

... when generating a forest plot with only one study (for visualization purposes only), I notice that Stata displays I² = 100% in the output. I understand that heterogeneity should not be calculated with a single study, as it's not statistically applicable.

My intention is not to perform a full meta-analysis but rather to visually present the effect estimate of that study, as is sometimes done in Cochrane reviews.

webuse bcgset, clear
meta forestplot if _n==1, nooverall noohetstats noohomtest

Res.:

I would also like to know if there is any way to display several independent meta-analyses together in one forest plot, without treating them as subgroups or calculating an overall combined effect.

The -nooverall- option suppress the overall effect size, so you can include all studies and add this option. The coefficients and CIs do not change, whether you add each individually or all together. But otherwise, AFAIK, there is no way to combine graphs generated by meta forestplot. You may want to take a look at coefplot from SSC to create similar graphs.