MISCELLANEOUS Flashcards
Probability that a sick patient will have a positive test
Sensitivity
Probability that a healthy patient will have a negative test
Specificity
of true positives + false negatives divided by total number of tests done
Prevalence
True negatives divided by true negatives + false positives
specificity
Having a high false psoitive rate will lowera tests ability to be
specific
1 standard deviation =
68% of samples are around the mean
2 standard deviation =
95% of samples are around the mean
3 standard deviation =
99.7% of samples are around the mean
Positively skewed distribution (mean median mode order)
mean > median > mode
Negatively skewed distribution (mean median mode order)
mode > median > mean
TYpe 1 error
null hypothesis is rejected when it is actually true
Probability of a type 1 error
alpha. Often the p-value set to 0.05 (5% chance of a type 1 error)
Type 2 error
null is accepted when it is actually false. False negative
Beta
Probability of a type 2 error. 1-beta is the POWER often set to 80%
Compare numerical means of two different groups - use what test?
unaired t etest
example: average BMI in sleeve vs bypass pts
Compare two different numerical measurements taken from a single group of patients
paired t test
example: BMI in pts before vs after getting a sleeve
compare numerical means of three or more groups
ANOVA
average BMI in sleeve vs bypass versus gastric band pts
Compare categorical outcomes between two or more groups
Chi square test or Fischers exact test
example does VTE occur more often in pts who are obese?
Identify and adjust for multiple potential factors contributing a to a categorical outcome
multivariate logistic regression
example: in a large database, determine what variables contribute to incidence of infection
Identify and adjust for multiple potential factors contributing to a numerical outcome
multivariate linear regression
example in a large database, determine what variables contribute to total lengthj of stay
identify difference in survivorship over time betwen two or mrore groups
kaplan-meie analysis
example: cancer survival in pts who received neoadjuvamnt versus adjuvant chemotherapy
analyze a population at a particular moment in time to determine prevalence of factors and disease
cross sectional study
population of subjects are analyzed to associate certain factors with an outcome
cohort study
cohort study can determine what kind of risk
relative