Biostats Flashcards
Describe the table for RR and AR? RR and AR equations?
Relative Risk: incidence in experimental group DIVIDED BY incidence of control group
Absolute Risk Reduction: incidence of experimental group MINUS incidence in control group
What is the difference between RR and AR?
AR includes baseline risk while relative risk does not
Equation for number needed to treat/harm/screen?
1 DIVIDED BY Absolute Risk
1
_________
AR
NNT: control - experiemental group
NNH: experimental group - control group
Define confidence interval
A confidence interval states the certainty that a given value from statistical analysis falls in between two values.
Ex: RR = 1.33 (CI: 1.21 - 1.63) > this means that we are 95% sure that 1.33 is the true value because it falls between the CI
Ex 2: RR = 1.33 (CI: 0.62 to 1.4) > since CI crosses ONE, no difference in risk between control and experiemental group
Define p < 0.05
States the liklehood of receiving results of a study by chance alone. For p<0.05, the chance is 5%
Difference in axis for tables for RR/ARR vs Tables for SPIN/SNOUT
Risk
x axis: exposure
y axis: disease
SPIN/SNOUT
x axis: disease
y axis: test/screening
Table for SPIN/SNOUT
sensitivity - first column
specificity - 2nd column
PPV - first row
NPV - 2nd row
define sensitivity of a test
ability to get positive test in someone with disease
good tests for screening
more false positives
SNOUT - for highly sensitive tests, negative test rules out
Ex: think of hep c and CSR patient
define specificity
negative test for people without disease
SPIN - highly specific test, positive, rules in disease
confirmatory test
Ex: think of hep c and CSR patient
define PPV
likelihood that person with disease has positive test
define NPV
likelihood that person without disease gets negative test
which parameter is affected by prevalence? how does it change?
PPV and NPV are affected by prevalence inversely
prev high = high PPV and low NPV
prev low = low PPV and high PPV
