I’m no expert on meta-analyses, but if the total pooled samples is 785 and there are 12 studies. That seems like quite a small average sample size per study. Is this not underpowered? An average of junk is still junk.
Small individual study sample size does not necessarily mean junk. Junk is a product of poor methodology. The main concern with sample sizes may be overestimates of effects if the event is rare, or poor precision (ie large confidence intervals).
However you’re right to presume that small studies are likely to be lower quality, often because they’r observational as opposed to randomized studies.
However you’re right to presume that small studies are likely to be lower quality, often because they’r observational as opposed to randomized studies.