For instance if we are studying mortality associated with a drug, and there are 15 studies with the outcome but we need only a subset of the population studied by them. If 5 studies were included in an IPD analysis. Can we take the estimate of this IPD analysis to combine with other studies to conduct a newer systematic review. (Intention :- instead of requesting subset data to 15 authors if this IPD estimate is taken, we have to approach only 10 authors for data)
Thanks in advance.
Here is some discussion about pooling and meta-analysis:
Article Simple Pooling versus Combining in Meta-Analysis
Article COVID-19 and Cancer: Lessons From a Pooled Meta-Analysis
Article Pooling, meta-analysis, and the evaluation of drug safety
You can in theory however would have to report this in the manuscript as you have raw data for 10 studies and 5 were extracted by some one else for their study. In the end calculation would be valid, It is just the matter of reporting how data was extracted-some peers may accept it, some may not.
HA
This may be possible theoretically in certain circumstances. It would be necessary for you to disclose and state this in the report/manuscript that you are sending for publication if you are looking to publish the work. This is important because you have 10 studies and it is stated that five publications were extracted by another person for his/her study. The calculations may be valid ultimately, but it is extremely important that you state clearly in your report the way that the data were obtained for reviewers to understand it in the process and then give their informed opinions of your work.
Good luck!
Daniella Kingsley-Godwin
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis