Reviewing the Evidence Flashcards
What three things do researchers need to consider when conducting trials?
Patient, treatment and outcomes
What do fixed effect models assume?
Assumes studies are estimating exactly the same effect size (i.e. one true effect)
What do random effects models assume?
Assumes the studies are estimating similar, not same, size
What is heterogeneity?
Fixed model effect versus random effects model, statistical for heterogeneity, sub-group analysis
How would you avoid publication bias?
Check protocol for search strategy, do a funnel plot, use a statistical test
What is primary research?
Studies e.g. RCT, literature reviews e.g. expert reviews and systematic reviews
What are the problems with expert reviews?
Implicit assumptions - can never be recreated
What are the advantages of systematic reviews?
Explicit assumptions, transparent methodology, reproducible, formal protocol
Describe a forest plot
Size of square is proportional to the weight given to it. Diamond shape is the collation of all of the studies. Lines are the confidence intervals
What is a funnel plot?
Plot of some measure of study size against measure of effects (e.g. odds ratio)
