<html><head></head><body><div style="font-family: Verdana;font-size: 12.0px;"><div>Hello all,</div>
<div> </div>
<div>I have a problem computing the treatment effects with 95%CI from a meta-regression output. I hope somebody can help me with this.</div>
<div> </div>
<div>###Code###</div>
<div> </div>
<div>require(robumeta)</div>
<div>DN <-robu(formula = es ~ 1, data = SMD,studynum = study, var.eff.size = var,rho = .8, small = TRUE, modelweights = "CORR")<br/>
DN<br/>
print(DN)<br/>
sensitivity(DN)<br/>
forest.robu(DN, es.lab = "Time", study.lab = "study","Effect Size" = es)</div>
<div> </div>
<div>###Model with covariates (follow-up time points)###<br/>
DN1 <-robu(formula = es ~ 1 + Time, data = SMD,studynum = study, var.eff.size = var,rho = .8, small = TRUE, modelweights = "CORR")<br/>
DN1</div>
<div> </div>
<div>###Output DN1###</div>
<div> </div>
<div>
<div>RVE: Correlated Effects Model with Small-Sample Corrections </div>
<div>Model: es ~ 1 + Time </div>
<div>Number of studies = 6 <br/>
Number of outcomes = 15 (min = 2 , mean = 2.5 , median = 2 , max = 4 )<br/>
Rho = 0.8 <br/>
I.sq = 47.59573 <br/>
Tau.sq = 0.08054052 </div>
<div> Estimate StdErr t-value dfs P(|t|>) 95% CI.L 95% CI.U Sig<br/>
1 X.Intercept. -0.5922 0.282 -2.1005 1.94 0.1747 -1.845 0.6604 <br/>
2 Time1.week -0.9762 0.282 -3.4623 1.94 0.0778 -2.229 0.2764 *<br/>
3 Time12.months 0.3944 0.282 1.3989 1.94 0.3006 -0.858 1.6471 <br/>
4 Time4.weeks -1.3214 0.282 -4.6867 1.94 0.0454 -2.574 -0.0688 **<br/>
5 Time5.8.weeks 0.0316 0.373 0.0846 3.04 0.9378 -1.147 1.2105 <br/>
6 Time6.months 0.2858 0.307 0.9325 1.95 0.4517 -1.066 1.6379 <br/>
7 Time9.12.weeks 0.1701 0.327 0.5208 3.21 0.6363 -0.832 1.1720 <br/>
8 Timeimmediate -0.2162 0.282 -0.7669 1.94 0.5256 -1.469 1.0364 <br/>
---<br/>
Signif. codes: < .01 *** < .05 ** < .10 *<br/>
---<br/>
Note: If df < 4, do not trust the results</div>
<div> </div>
<div>#######</div>
<div>I know that the output is not reliable as the dfs are too low. Still, I would like to know how to calculate the treatment effects und CIs.</div>
<div> </div>
<div>The estimate is easy. Intercept value + follow-up e.g. -0.5922 + 0.3944 = -0.1978</div>
<div> </div>
<div>I read that one could calculate the standard errors via the variance-covariance matrix and then calculate the treatment effects and CIs. But I do not know how to do it here. Can somebody help me with this?</div>
<div> </div>
<div> </div>
<div>Regards,</div>
<div> </div>
<div>Tobias</div>
<div> </div>
</div>
<div> </div></div></body></html>