Systematic Review/Meta-analysis

#1
Hello, for a class I am taking I need to superficially understand an applied statistical approach in a research paper.

I was reading this one: https://onlinelibrary.wiley.com/doi/full/10.1111/acem.13502
That used systematic review and meta-analysis to test if low-dose ketamine could be used as an alternative to opioids in the treatment of acute pain.

To my understanding, the researchers used meta-analysis by forming an electronic search program with their established criteria to get a pool of studies that could answer the research question they were trying to study. The use of this statistical approach allows the researchers to come to a conclusion with greater statistical significance as it has a larger sample size, diversity of subjects, and results. This allowed them to have a better idea of how safe and effective ketamine is for treating acute pain, as well as giving a wider range of adverse effects that may have occurred. Overall meta-analysis gives a conclusion with more statistical power/credibility and makes it more applicable to the general population.

Is my understanding of this statistical approach correct?
@hlsmith
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Yes this seems correct. It is a form of pooling results together. I could flip a coin 1000 times to see if it is fair. However, that may take a long time and perhaps I flip it a little different than others or I am in a place with high humidity or wind, etc. Though another option may be to collect data on the results of say a dozen people who flip a similar coin and pool their results and make a conclusion. Now we substitute the coin for say an intervention (yes/no). Knowing how the intervention works across people is more informative than a single center's results due to their potential idiosyncrasies and the heterogeneous scenarios across the planet.

However, if the centers or studies vary enough they maybe looking at slightly different things, which would need to be addressed via random effects - which account for within study variability and between study variability. The latter part being the addition.