statistics Flashcards
highly sensitive test is useful for what?
SNOUT: rule out
helps for screening purposes, as it negative results on these tests will help rule out diagnosis
what is a likelihood ratio?
probability of a given test result occurring in a patient with a disorder compared to a probability of the same result occurring in a patient without the disorder
- does NOT change with prevalence of the disease
- useful for >2 different studies
- positive LR: sensitivity/ (1- specificity)
- negative LR: (1-sensitivity)/ specificity
sensitivity vs specificity
- sensitivity: TP/ (TP+FN)
- specificity: TN/ (TN+FP)
PPV vs NPV
- DOES depend on disease prevalence
- PPV: (TP/TP+FP)
- NPV: (TN/TN+FN)
which studies use odds ratio?
- case control studies–> exposure of participants with the disease vs exposure of participants without the disease
which studies use relative risk?
cohort studies
attributable risk percent
risk of exposed- risk of unexposed/ risk of exposed
What is the basic premise for intention to treat analysis
Randomization
Blinding
What is a type II error and what is its relation with sample size and power?
Occurs when the alternative hypothesis is rejected but is actually true
Power and sample size increase will decrease type II error
What is a measurement bias?
Poor data collection leads to inaccurate results, usually not seen in randomized clinical trials
Calculating number needed to treat
1/ARR
1/risk of control-risk of case
define standardized mortality ratio
observed n of death/ expected n of death
adjusted measure of overall mortality
used in occupational epidemiology
What is a factorial study design?
experimental study design with 2 or more interventions and all combinations of intervention
what is length time bias?
when survival benefits of a screening tests are overstated due to including lots of benign slow progressing cases
What does external validity measure?
generalizability of the study to other populations