Biostats

Question 1

Equation for Sensitivity and Specificity

Answer 1

Sensitivity = true positive / (true positive + false negative)

Specificity = true negative / (true negative + false positive)

Question 2

PPV and NPV

Answer 2

Both change based on prevalence of disease in population

PPV (inc if higher prevalence) = if positive test, how likely is it that you have the disease?

NPV (inc if lower prevalence) = if negative test, how likely is it that you don’t have the disease

Question 3

Absolute Risk v. Relative Risk

Answer 3

Absolute risk is the #1 who experience the risk / total # people

Relative risk is risk in those w/ X minus risk of those w/o X

Question 4

NNH and NNT

Answer 4

NNH = 1 / AR

NNT = 1/ ARR

Question 5

If someone performs at 1 ST above mean, what percentile are they in?

Answer 5

84th (68 + 13.5 + 2.5)

Question 6

How do you double the precision of a test?

Answer 6

Increase the sample size by 4

Because standard error (measure of precision) is std deviation divided by the square root of N

Question 7

Cohort v. Case Control Study

Answer 7

Cohort = prospective; look at those with and without exposure then see if they develop disease (MEASURE RELATIVE RISK)

Case Control = retrospective; start with those with and without disease and see who had exposure in the past (MEASURE ODDS RATIO)

Question 8

Selection Bias

Answer 8

Having sicker patients in placebo group

Question 9

Bergson Bias

Answer 9

Using hospitalized patients in a study instead of general population

Question 10

Lead-Time Bias

Answer 10

Interpret earlier detection of a disease (screening) as an increase in survival

Question 11

Type I and II Error

Answer 11

Type I = alpha = false positive (reject null hypothesis when you should not)

Type II = beta = false negative (should reject the null but do not)