stats/research Flashcards
sensitivity
few false negatives, good screening test.
TP/(TP+FN)
specificity
few false positives.
TN/(TN+FP)
PPV
prob that someone who tests positive has the disease
TP/(TP+FP)
NPV
prob that someone who tests negative does not have the disease
TN/(TN+FN)
type I error
reject null hypothesis in error - think something is significant when it really isn’t
type II error
accept null hypothesis in error - think something isn’t significant when it really is
validity
whether test measures what it’s intended to.
internal validity = accuracy
external validity = generalizability
reliability
consistency of results from a test
intention to treat
all patients must remain in the group to which they were randomly assigned - maintains power
NNT
NNT = 1/absolute risk reduction ARR = rate in untreated - rate in treated
incidence
number of newly diagnosed cases
prevalence
total number of cases existing at a given time
likelihood ratio
likelihood that a person with condition will have a positive test.
relative risk
probability of outcome in exposed group vs unexposed group
odds ratio
odds of an outcome in exposed vs unexposed group
cohort study
comparison of group with exposure to control w/o exposure
needs large sample size and long time, can lose people to follow up
case control
start with patients with and without DISEASE and check for hx of risk factors
cross sectional study
looking at associations at one point in time
longitudinal studies
studying one gropu over time
systematic review
summarizes articles on one topic
- specific clinical question
- thorough search for studies
- clear explanation for which studies were considered
- descriptive results section summarizing
meta-analysis
- summary statistic using data from all studies
- report whether combining all statistics is possible
- assess for publication bias
