Stats/Pharm Cram Flashcards
Type 1 error?
Rejecting a null hypothesis that is actually true
Type 2 error?
Accepting a null hypothesis that is actually false
Best way to analyse data with multiple variables?
Linear regression - if outcome is continuous
Logistic regression - if outcome is binary
Sensitivity?
If disease is present how likely they will have a positive test
a / a + c
Specificity?
If a disease is not present how likely they are to have a negative test
b / b + d
Positive predictive value?
Proportion of people with a positive test who have a disease
a / a + b
Negative predictive value?
Proportion of people with a negative test who do not have a disease?
d / c + d
Number needed to treat?
1 / absolute risk reduction
Best way to reduce type 2 error (accepting a null hypothesis that is actually false)?
Sample size
Omission vs commission bias?
Omission bias - inaction is safer than action
Commission bias - action is safer than inaction
Measure of association used for survival analysis?
Hazard ratio
Cross sectional study design?
Observational
Assess frequency of disease and related RF at a certain time point
Cohort study design?
Group with given exposure/RF to group without
See if exposure changes risk of disease
Case control study design?
Disease vs no disease (case vs control)
Prior exposure or risk factor
Ecological study design?
Observational
Compare different groups or at different time periods