when given absolute risk reduction, how do you express this in terms of
how many people that need to be treated in order to get a benefit from the treatment?
1/absolute risk reduction
ARR = event happening in control - event happening in new treatment
how do you calculate likelihood ratio for a positive result?
sens / (1 - spec)
it tells you how convincing a test result is
i want to see if there is a statistically significant change in something following treatment, what test is required?
wilcoxon signed rank score
what is confounding bias?
another factor that relates to main outcome + other factor
e.g. looking at vaping + URTI
but most people also smoke too
= smoking is confounder
how to calculate power
1 - probability of type II error ( %age of it NOT happening)
what are the phases of a clinical trial ?
0 - exploratory studies
1 - safety assessment
2 - efficacy
3 - effectiveness
4 - post marketing surveillance
what is absolute risk reduction?
difference in risk of an outcome between two groups
control - event
difference between type 1 and 2 error
type I = “I found something” (but it’s wrong)
Type 2 = ‘I missed it’ (but it’s right)
what is chi-squared test for
compare percentages
e.g. %ages of people who improved HTN following 2 different drugs
what plot is used to show publication bias
funnel plot (pyramids)
how to calculate relative risk reduction
(event rate - control)/control
how do you calculate likelihood ratio for a negative result?
(1 - sensitivity) / specificity
