Biostats

Question 1

Type 1 error

Answer 1

• finding a difference when one doesn’t really exist

- alpha = 0.05

Question 2

Type 2 error

Answer 2

• not finding a difference when one exists

- beta = 0.2

Question 3

Power =

Answer 3

1 - beta

Question 4

Tests - continuous variables, 2 groups

Answer 4

• t-test (parametric)

- Mann Whitney U or Wilcoxon Rank (nonparametric)

Question 5

Tests - continuous variables, 3 or more groups

Answer 5

• ANOVA (parametric)

- Kruskal Wallis (nonparametric)

Question 6

Tests - categorical variables

Answer 6

• Chi square

- Fisher’s exact (< 5 subjects in a square)

Question 7

Tests - linear regression

Answer 7

• dependent variable = continuous

Question 8

Tests - logistic regression

Answer 8

• dependent variable = dichotomous

Question 9

Relative Risk (RR)

Answer 9

• compares probability in each group

- used in prospective cohort studies, RCTs

Question 10

Odds Ratio (OR)

Answer 10

• used in case control studies

- reflects RR when probability of disease is low (~10%)

Question 11

RR =

Answer 11

a/(a + b) / c/(c + d)

Question 12

OR =

Answer 12

(a / c) / (b / d) OR (a x d)/(b x c)

Question 13

NNT =

Answer 13

1 / ARR (absolute risk reduction)

Question 14

ARR =

Answer 14

risk of outcome in control group - risk of outcome in intervention group

Question 15

Sensitivity =

Answer 15

a / (a + c)

Question 16

Specificity =

Answer 16

b / (b + d)

Question 17

Sensitivity

Answer 17

• proportion of people with disease who have a positive test
• good to rule out disease (SNOUT)
• screening test

Question 18

Specificity

Answer 18

• proportion of people without disease who have a negative test
• good to rule in disease (SPIN)