I am currently planning to construct a meta-analysis on a supplement for knee osteoarthritis. The primary outcomes include parameters from the WOMAC scale.
However, from the trials that I gathered, some use Likert scale 0-4 for each question, some used visual analog scale (VAS).
How do I transform it to the same format/presentation so that meta analysis is possible?
Is there any formula?
Please help me.
I would recommend you to compute standardised mean difference (SMD). Cochrane handbook provides helpful guidelines on how to compute SMD.
Hi Dr Saurab, I do understand regarding role of computing SMD in trials that shared the same outcome using different tools.
However, one of the trials used mean change instead of mean SD. This will cause the data to be not analysable eventually.
In the first place do such thing exist? Interchange WOMAC scale score?
I will definitely refer to the Cochrane Handbook for further info on SMD.
Thanks again.
Can you explain what do you mean by mean change and mean SD? May be provide the link to the paper? If you are saying that authors did not report SD, you can still compute it by referring to Cochrane Handbook.
Make sure not to use SE and SD interchangeably.
Good luck
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis