1
$\begingroup$

I am conducting a meta-analysis with 3 treatments (manual therapy, dry needling and placebo) and 3 outcomes (VAS, QoL and ROM, all of them are continuous variables). With metafor package I have R code to perform multivariate ma with multiple outcomes, and on the other hand, to perform a network ma with multiple treatments, but I don't know the more appropriate R code to combine both.

My data set looks like this:

data <- data.frame(study = c("Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Nicolas, 2022", "Eleuterio, 2021", "Eleuterio, 2021", "Eleuterio, 2021", "Eleuterio, 2021", "Oliver, 2020", "Oliver, 2020", "Oliver, 2020", "Oliver, 2020", "Oliver, 2020", "Oliver, 2020", "Camilo, 2021", "Camilo, 2021", "Camilo, 2021", "Camilo, 2021", "Camilo, 2021", "Camilo, 2021"), treatment = c("Manual therapy", "Dry needling", "Placebo", "Manual therapy", "Dry needling", "Placebo", "Manual therapy", "Dry needling", "Placebo", "Manual therapy", "Dry needling", "Manual therapy", "Dry needling", "Dry needling", "Placebo", "Dry needling", "Placebo", "Dry needling", "Placebo", "Manual therapy", "Dry needling", "Placebo", "Manual therapy", "Dry needling", "Placebo"), outcome = c("VAS", "VAS", "VAS", "QoL", "QoL", "QoL", "ROM", "ROM", "ROM", "VAS", "VAS", "QoL", "QoL", "VAS", "VAS", "QoL", "QoL", "ROM", "ROM", "VAS", "VAS", "VAS", "QoL", "QoL", "QoL"), n = c(10, 15, 12, 10, 15, 12, 10, 15, 12, 20, 21, 20, 21, 25, 25, 25, 25, 25, 25, 32, 31, 32, 32, 31, 32), mean = c(4, 3, 6, 15, 20, 10, 120, 140, 90, 5, 3, 25, 30, 3, 6, 20, 12, 145, 100, 5, 4, 8, 23, 20, 18), sd = c(2.5, 1.3, 1.5, 3.5, 4.6, 2.8, 10.2, 8.5, 9.1, 1.3, 1.8, 6.3, 9.2, 2.1, 1.6, 5.8, 6.4, 20.3, 18.6, 2.6, 2.1, 2.3, 6.5, 7.1, 6.8))

Many thanks in advance.

Improve this question
$\endgroup$
2
  • $\begingroup$ Instead of a traditional frequentist model, I would suggest you to have a look at Bayesian modeling in BUGS (directly or via R)... Another approach could be applying a multivariate meta-analysis framework $\endgroup$
    – Giuseppe Biondi-Zoccai
    Commented Dec 30, 2022 at 12:31
  • 1
    $\begingroup$ Hello Giuseppe, perhaps bugsnet r package can do it? I have no experience with Winbugs or Openbugs and this r package appears to work with bugs also; many thanks!! $\endgroup$
    – Nicolas Cuenca Zaldivar
    Commented Dec 30, 2022 at 13:29

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.