Hello everyone, I am currently working on my first meta-analysis using revman. I would really appreciate any advice on how to proceed.
For some of my articles, I have the raw data (e.g out of 12 exposed, 6 has outcome. out of 12 not exposed, 6 has the outcome). For other articles, I have unadjusted and adjusted odds ratio without any raw data.
I would like to ask 3 questions.
1. is it possible for me to combine both the raw data and unadjusted odds ratio using the generic inverse variance function to generate a forest plot?
2. can i combine unadjusted odds ratio, adjusted odds ratio and raw data in generic inverse variance?
3. is it okay if i only use the manual dichotomous function to put in raw data and leave out the studies without raw data?
Thank you very much
1. Yes, you can combine raw data and unadjusted odds ratios in the generic inverse variance function in RevMan. You would input the log odds ratio and its standard error for studies with unadjusted odds ratios, and compute it for studies with raw data.
2. You can combine unadjusted odds ratios, adjusted odds ratios, and raw data, but it’s not recommended. Adjusted odds ratios account for confounding variables, making them different from unadjusted data. Mixing them without adjustment may bias the results.
3. It is acceptable to use the dichotomous function for studies with raw data and exclude studies without raw data, but you will lose information from the excluded studies, potentially reducing the overall robustness of your meta-analysis.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey