Biostats Flashcards
sensitivity
TP/(TP+FN)
specificity
TN/ (TN+FP)
PPV
TP / (TP+FP)
NPV
TN / (TN+FN)
incidence vs. prevalence vs. attack rate
incidence = # new cases/ # vulnerable
prevalence = #existing cases/population at risk
attack rate = #sick/ #who were exposed
OR
(a/c)/(b/d) = ad/bc
RR
(a/(a+b)) / (c/(c+d))
relative risk reduction
1 - RR
attributable risk
AR = a/(a+b) - c/(c+d)
Absolute risk reduction
difference in risk compared to control
NNT
1 / ARR
number needed to harm
1 / AR
random error
reduces the precision in a test
systematic error
reduces the accuracy of a test
examples of selection bias
- berkson bias (study only looking at inpatients), loss to follow up, health care worker study
measurement bias
Hawthorne effect - pts who know they are being studied behave differently than they normally would
positive skew data
mean > median > mode
- curve leans to the left (tail goes positive)
negative skew data
mean < median < mode
- curve leans to the right (tail goes negative)
type I error (a)
- stating that there is an effect or difference when there isn’t one
- a is the probability of making a type 1 error, if p <0.05 then there is a less than 5% chance that the data will show something that is not really there
type II error (b)
- stating that there is not an effect or difference when one exists
- b is the probability of making a type II error
- 1-b is the statistical power, which is the probability of rejecting the null hypothesis when it is false (aka probability of finding a true relationship)
- increased power and decreased b by increasing sample size, increasing expected effect size and increasing precision of measurement
calculating 95% confidence interval
- 95% CI = Mean +/- 1.96*SD/sqrt n
t test
checks differences between the means of 2 groups
ANOVA
checks differences between means of 3 or more groups
chi-square test
checks differences between 2 or more percentages or proportions of categorical outcomes (not mean values)
