Stats and epi

Question 1

How to calculate sensitivity?

Answer 1

a/a+c

true positive/true positive + false negative

Question 2

How to calculate specificity?

Answer 2

d/b+d

True negative/true negative + false positive

Question 3

How to calculate PPP? What does PPP of 95% mean?

Answer 3

a/a+b

A positive test means they are 95% sure they don’t have disease

Question 4

How to calculate NPP

Answer 4

d/c+d

Question 5

What is the difference between incidence and prevalence?

Answer 5

Incidence is new case each year. Prevalence is total number in time.

Question 6

If incidence of placebo is 20% and therapy is 15%, what is the absolute risk?

Answer 6

20-15 = 5%

Question 7

What is NNT?

Answer 7

1/AR (%)

Question 8

What is Relative risk?

Answer 8

Difference in rate/baseline rate
i.e. placebo 20%, therapy 15%
RRR = 5/20 = 25%`

Question 9

What is type 1 error?

Answer 9

false positive. rejecting Ho when true. (alpha. Usually 0.05)

Question 10

What is type 2 error?

Answer 10

False negative. failure to reject Ho when false. Happens when study underpowered.

Question 11

What is power?

Answer 11

Probagility test will reject false negative Ho
Power = 1- beta = sensitivity. Usually 0.8 Determined by population size, size of effect, variance within populations, test used.

Question 12

What is phase 0 clinical trial?

Answer 12

Very low dose on humans.

Preliminary data on pharmacokinetics and pharmocodynamics.

Question 13

What is phase 1 clinical trial?

Answer 13

SMALL group of HEALTHY volunteers

Looking for toxicity (safety)

Question 14

What is phase 2 clinical trial?

Answer 14

LARGER groups
IIa - look at dosing requirements
IIb - efficacy.

Question 15

What is phase 3 clinical trial?

Answer 15

RCT

Definitive assessment of how effecitve a drug is in comparison with gold standard.

Question 16

What is phase 4 clinical trial?

Answer 16

Post marketing surveillance trial (pharmacovigilance = adverse drug reactions with other drugs)

Question 17

What are receiver operator curves?

Answer 17

Graph of sensitivity vs 1-specificity (i.e. true positive vs false positive)
Binary classifier system.
Provides tools to select optimal models and to discard suboptimal ones.

Question 18

Odds ratio. If 25% of babies are girls and 75% boys, what are the odds of being a girl?

Answer 18

Events occurring: Events not occurring

1:3

Question 19

What is the greatest strength of hte case-control study design?

Answer 19

Multiple risk factors can be assessed. good for rare diseases