equivelance, non inferiority and superiority

Question 1

how do we test the hypothesis in a equivelance trial

when can we confirm equivelance is shown?

Answer 1

when the point estimate e.g., odds ratio or mean difference AND 95% CI lie within the equivelance margins

Question 2

an equivelance margin set to 20% what does this mean intuitively

Answer 2

for the intervention to be deemed equivelant there must be a mean difference between the arms of + or - 20%.

Question 3

explain intuitively the reuslts of an equivelance trial what a mean difference of (95% confidence interval):
> Per protocol: -0.3% (-7.5% to 6.5%)
> Intention to treat: -1.0% (-8.1% to 6.1%)

Answer 3

depends what the equivlance margin is? if 20% then equielance has been shown

Question 4

hypothesis testing for a non inferior trial

Answer 4

if the point estimate and 95% CI is smaller than non inferiority margin (single number this time, aka delta)

then we can say treatment A non inferior to treatment B

Question 5

lets say non inferiority margin of 10% is chosen, interpret the restults if the point estimate

Answer 5

to show non inferiority the difference between treatments and 95% CI must be less than 10%

Question 6

whats the difference between the equivelance margin and non inferiority margin

Answer 6

equivelance margin of 20% means the point estiamte and CI mmust be MORE than -20% or LESS than 20% to be deemed equivelance.

Non inferiority margin instead single number. For treatment to be deemed non inferior the point estimate and CI must be LESS than 20%