I am performing a meta analysis in which there are some studies which are RCTs and cohorts. There are two studies which are non-randomized controlled trials. Can I include those two studies in my systematic review and meta analysis? Is it fine to pool RCTs and non-RCTs?
If the trials you have provide the required data, you CAN include them in a meta-analysis. Whether or not you SHOULD do so is another matter. In a traditional hierarchy of evidence, non-randomised studies are much lower. They are much more prone to bias - so including them risks including highly biased results - rendering the results of your meta-analysis biased. One approach might be to undertake sensitivity analyses to see whether or not the inclusion of such studies alters the point estimate (much) and overall conclusions - but if it does, where does that leave you and if it doesn't what have you gained by including them?
IF this is a topic where non-randomised studies are, pragmatically, the only approach in many circumstances you might proceed with caution - but you must use a suitable approach to assessing the risk of bias for such studies (ROBINS) - and I would suggest doing the sensitivity analysis in any case. Bit remember the old adage - garbage in, garbage out.
It always depend on your study question and protocol. The risk of bias is generally higher in observational study than in RCT. However, sometimes depending on the field of interest, the RCT can be of poor quality of reporting which makes them at risk of bias too.
I think there is no one size fits all answer.
Ideally you cannot pool them together because of the inherent difference in bias between the study designs. Additionally, whether or not you can include nRCTs in your systematic review and meta-analysis depends on the quality of the nRCTs, the similarity of the nRCTs to the RCTs (homogeneity, Population vs statistical), and the impact of the nRCTs on the overall results. If you do include nRCTs, it is important to be aware of the limitations of this approach and to conduct a sensitivity analysis.
Here are some additional tips for conducting a meta-analysis that includes nRCTs:
- Use a random-effects model to account for the heterogeneity between the RCTs and nRCTs.
- Conduct a subgroup analysis to compare the results of the RCTs and nRCTs separately.
- Assess the impact of the nRCTs on the overall results by conducting a sensitivity analysis.
- Badges
- Science topic