Test 2 Feddock Flashcards
Risk of bias
Similar prognosis
Accurate measurements of exposures/outcomes
Complete and adequate follow up
SImilar prognosis
Accureate measurements of exposrues and outcomes
Complete and adequate follow up
Random and allocaiton concealment
Defined objective outcomes and appropritate blinding
Loss to follow up and intention to tx
Directness
Diff between pop of interest and study participants
Diff in tx between pop of interest and study part
Use of surrgoate outcome instead of outcome of interest
Indirect comparison betwen presence/absence of an exposure or risk factor
Cosistency
Effect estimates consistent across subgroups
Effect estimates consistent across outcome measurements
Heterogeneity is minla
Precision
Use CI
Publication bias
Look at the funnel plot
Alpha versus Beta
Power vs. confience
Alpha - prob of finding diff when NOT one
Confidence
Beta- false negative
Power - % that you find a true difference when there is one
What effects heterogeneity
Different populations, different doses/administration
Cointervention vs. surveillance bias
Cointervention - failure of blinding the physicians
Surveillance - think COHORT study
External validity effects
Must look at the population of the patients in the study
Question to ask about negative study
Was power high enoug h
Meta advantages
Generalizability, power, allows for comparison
