I have 10 studies who look at a surgery for facial paralysis. The main outcome is time to nerve recovery. I have the mean, SDs for the time in months to recovery but obviously there is no control or baseline data. Do I just use zero for the mean and SD for my pre-op values? This gives me a weighted average which is essentially what Im looking for, but want to make sure I present the date the correct way.

  • $\begingroup$ Please indicate n-size (number of cases for each study). How did you compute weighted average ? The formula and its assumptions.? $\endgroup$ – Subhash C. Davar Jul 12 '17 at 10:19

I see no reason why you could not use a meta-analytic approach when attempting to identify overall recovery time as you describe. See this article for a similar conceptual example. The authors examine mean weight loss in this instance - but the idea is similar in that it is a meta-analysis based on a single value that represents change over time. In your instance, it would be mean recovery time (however that is defined) instead of the mean change in weight. Additionally, based on your description there is no need for what you describe as pre-op values.

I hope that helps.

  • $\begingroup$ Great thank you. Another question. If I am using random effects to perform the pooled analysis, do I need to keep random effects model in place for sub group analysis within that study group or go back to a fixed effects $\endgroup$ – user168533 Jul 11 '17 at 22:08
  • $\begingroup$ Maybe this chapter has some answers about common approaches? $\endgroup$ – Matt Barstead Jul 12 '17 at 2:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.