Biostatistics Flashcards
What is outcome measurement of a case-control study
Odds ratio
What is outcome measurement of a cohort study
Relative risk
Phase III vs Phase IV study
Phase IV is postmarketing surveillance (after approval); detects rare or long-term adverse effects
Sensitivity formula
True positive rate: TP/TP+FN
Specificity formula
True negative rate: TN/TN+FP
What parameter must be high for a screening test
Sensitivity
What parameter must be high to rule in a disease
Specificity
Positive predictive value formula
TP/TP+FP
What happens to PPV as prevalence increases
higher PPV
Negative predictive value formula
TN/FN+TN
What happens to NPV as prevalence increases
lower NPV
Incidence vs prevalence
Incidence= new cases
Prevalence= all current cases
(prevalence higher for chronic diseases)
Formula for Odds ratio
Odds that disease expose/odds without disease exposed
OR=ad/bc
Formula for relative risk
Risk developing disease in exposed/risk developing disease in unexposed
RR=a(a+b)/c/(c+d)
When are OR and RR similar
when prevalence is low
Formula relative risk reduction
1-RR
Attributable risk formula
difference in risk between exposed and unexposed
AR=a/a+b-c/c+d
Absolute risk reduction
difference in risk attributable to intervention as compared to control
Number needed to treat formula
1/ARR
Number needed to harm formula
1/AR
Random error reduces___
precision
systematic error reduces___
accuracy
Multiple/repeated studies, crossover studies, matching help reduce___ bias
confounding bias
Berkson bias
selection bias of hospitalized patients (have incr risk of exposure-throws off OR)
Hawthorne effect
groupes that know they are being studied behave differently; measurement bias
how prevent lead-time bias
measure “back-end” survival
Standard deviation vs SEM
SEM=mean/Square root(n)
Gaussian distribution
bell-shaped, mean=median=mode
Incr sample size, effect size, precision of meausrement
all incr power
Define Type 1 error
or a: stating there is an effect when none exists
Define Type II error
or b: stating there is not an effect when one exists
For 95% CI, Z=?
1.96
For 99% CI, Z=?
2.58
% within one SD
68%
% within two SD
95%
% within three SD
99.7%
What checks the difference between the mean of 2 groups
T-test
What checks the difference between the mean of 3+ groups
ANOVA
What checks the difference between 2+ categorical outcomes
Chi squared
