Key Stats In Evidence Based Medicine

Question 1

Define incidence

Answer 1

Measure the number of new cases of a particular disease arising in a population at risk in a certain time period

Question 2

Define prevalence

Answer 2

All cases of disease existing in a given population at a given time
A proportion

Question 3

What is the risk ratio/relative risk?

Answer 3

Ratio of risk in group A vs risk in group B
RR = R_exposed / R_unexposed

Question 4

What is the odds ratio?

Answer 4

Ratio of odds of outcome in group A vs odds of outcome in group B
OR = odds of outcome_exposed / odds of outcome_unexposed

Question 5

How do you calculate the risk ratio?

Answer 5

Risk_exposed / Risk_unexposed
- group of 200 > split into 2 groups of 100
- group A: 100 > 25 die (75 survive)
- group B: 100 > 10 die (90 survive)
- group A risk: 25/100 > 0.25
- group B risk: 10/100 >0.1
- 0.25/0.1 > 2.5
- RR = 2.5

Question 6

How do you calculate odds ratio?

Answer 6

odds of outcome_exposed / odds of outcome_unexposed
- group of 200 > split into 2 groups of 100
- group A: 100 > 25 die (75 survive)
- group B: 100 > 10 die (90 survive)
- group A odds: 25 : 75 > 1 : 3
- group B odds: 10 : 90 > 1 : 90
- 0.333/0.111 > 3
- OR = 3

Question 7

What is the absolute risk?

Answer 7

The risk of acquiring a given disease over a given period of time

Question 8

What is the absolute risk difference?

Answer 8

Absolute risk in exposed group - absolute risk in unexposed group

Question 9

Relationship between the p value + null hypothesis

Answer 9

• the smaller the p value, the stronger the evidence against the null hypothesis
• the larger the p value, the weaker the evidence against the null hypothesis

Question 10

Issues with the p value

Answer 10

• gives no indication as to the size of the effect
• no idea to the range of uncertainty
• if there is bias within the study, you can still get a low p value

Question 11

What is a confidence interval?

Answer 11

Gives a range of values in which you are reasonably confident that the true values

Question 12

What has to occur for something to be statistically significant in confidence intervals?

Answer 12

Range doesn’t cross 1

Question 13

What has to occur for something to be statistically significant in absolute difference?

Answer 13

Range spans 0