Endo: EBM 2

Question 1

Absolute risk reduction/increase will be statisically INsignificant if the confidence interval includes ___

Answer 1

Absolute risk reduction/increase will be statisically INsignificant if the confidence interval includes 0

Question 2

Relative risk will be statisically INsignificant if the confidence interval includes ___

Answer 2

Relative risk will be statisically INsignificant if the confidence interval includes 1

Question 3

Define Type 1 error

Answer 3

Concluding that there is a difference when no true difference exists. (False positive)

*Telling a man that he’s pregnant

Question 4

Define Type II error

Answer 4

False negative.

Concluding there is no difference when truly there is one.

*Telling an obviously pregnant woman she’s not pregnant

Question 5

The risk of Type I error is equal to _______

Answer 5

The risk of Type I error (false positive) is equal to alpha (your p-value cut off)

Question 6

The risk of Type II error is equal to _______

Answer 6

The risk of Type II error (false negative) is equal to beta (1-power)

Question 7

Define power and beta.

What is the relationship between power and beta?

Answer 7

• Power = probability of detecting a difference if one exists
• Beta = probability of committing a Type II error (not detecting a difference if one exists)
• Power = 1 - beta

Question 8

What is the formula for NNT?

Answer 8

NNT = 1 / Absolute Risk Reduction

*NNT is an expression of the amount of benefit for this intervention

Question 9

Compare the time relevance of absolute and relative risk reduction

Answer 9

• Absolute risk reduction is based on a particular time frame
• Relative risk reduction does not need to be attached to a specific time frame

Question 10

Define sensitivity

Answer 10

Sensitivity = % of people who truly have disease who get a positive test result

(True positives)

Question 11

Define specificity

Answer 11

Specificity = % of specifically healthy people who get a negative test result

(True negatives)

Question 12

Explain SPIN, SNOUT

Answer 12

• A highly specific test rules IN disease when test is positive
• A highly sensitive test rules OUT disease when test is negative

Question 13

Define PPV

Define NPV

Answer 13

Positive Predictive Value = probability that you actually have disease given a positive test result

Negative Predictive Value = probability that you truly do not have disease given a negative test result