EBM II Flashcards
two groups, split by exposure to disease outcome, cannot comput risk because irrelevant, can manipulate the number of subjects in the two groups, used to calculate odds ration, measures probability and an even occurence bs the probability of no event occurence
case control study
outcome A - exposure A/B or outcome B exposure A/B
case control calculations
odds that a case had been exposed to the risk factor/odds that a control had been exposred to the risk factor, interpreted as among the people in case control study an injured child was 5 times likely to have been exposed to high surgar instead of low sugar
odds ratio
need to review the mathetmatics behind these questions
risk can be computed for randomized control trials and cohort studies but not for case control studies, odds can be computed for case control studies ***
summary
testable and predictve statement that there is a certain relationship or differnece between populations regarding some parameter
hypothesis testing
states that there is a difference or relationship between groups but doesn’t specifc the direcdtion, there is a difference between the remdesivir and placebo treatments in shortneing the time to recovery in hisoitaplizes COVID9 patients with lower respiratory infections
nondirectional research hypothesis
two sided, one sided, fail to reject the null or rejection of the null, null states that there is no difference between the two groups
null hypothesis and alternative hypothesis
area to the right of the test statistic, probability of getting a value of the test statistic that it as least as extreme as one representing the sample data asssuming that the null hypothesis is true
p value
threshold for the p value, if below findings are significant, commonly 0.05, 0.01, and 0.10, represented by alpha
statistical significance
reject the null hypothesis when there is no differnece between the two groups and the null hypothesis true
type I error
fail to reject the null hypothesis when it is actually false, therefore the difference is real and you missed it, probability of a type II error is beta
type II error
fail to reject the null hypothesis when it is actually false, therefore the difference is real and you missed it, probability of a type II error is beta
type II error
for any fixed type I error alopha, an increase in the sample size n will cause a decrease in type II error B
controlling type I and type II errors
probability 1-B of rejecting false null hypothesis
power of a hypothesis test