EBM

Question 1

What p value would cause one to accept the null hypothesis of a chi-squared analysis?

Answer 1

p > 0.05

Question 2

When would linear regression be used?

Answer 2

To help predict a variable from another variable

So you draw out a line chart and use one value to predict the other

Question 3

When would an un-paired t-test be used?

Answer 3

To compare the means of two unrelated populations

Question 4

When would a paired t-test be used?

Answer 4

To compare the means of two related populations (e.g. the same population before and after)

Question 5

What is a meta-analysis?

Answer 5

Combining results from different studies

Question 6

What is sensitivity?

Answer 6

Measures the proportion of +ve’s identified (e.g % of sick people identified as sick).
TRUE POSITIVE RATE

Question 7

What is specificity?

Answer 7

Measure the proportion of -ve’s identified (e.g. % of healthy people correctly identified as not sick).
TRUE NEGATIVE RATE

Question 8

The most important 3 factors when assessing a clinical trial are…

Answer 8

Validity, results and clinical relevance

Question 9

When would you use regression and what are the two types?

Answer 9

To predict dependant variable from independant
Can correct for confunding variables
Linear- when continous outcomes
Logistic- When binary outcomes

Question 10

What is the t-test used for?

Answer 10

Tests the change in mean between two groups

CONTINUOUS VARIABLE

Question 11

What is the chi-squared test used for?

Answer 11

Version of t-test to see if expected freq = anticipated frequency (testing null hypothesis)
BINARY VARIABLES ONLY

Question 12

What equation is used to calculate odds ?

Answer 12

Number of individuals without outcome

Question 13

What equation is used to calculate odds ratio?

Answer 13

odds of the outcome in another group (e.g. control)

Question 14

What is implied by an odds ratio of 1?

Answer 14

No difference between the two groups

Question 15

What is used to test the significance of odds ratio’s?

Answer 15

Chi squared
Allows us to compare proportions across two different groups.
Tells us whether the comparison is statistically significant (it gives a p value).

Question 16

What is Multiple (or multivariate) logistic regression used for?

Answer 16

Used to assess the predictive value of one or more variables on a YES/NO outcome

Question 17

You are asked: “How does smoking affect the likelihood of having pancreatitis, after accounting for (or ‘unconfounded by’ or ‘independent of’) alcohol consumption, BMI, etc.?”. What do you use to find the answer?

Answer 17

Multiple logistic regression

Question 18

What is positive predictive value?

Answer 18

Probability that the disease is present when the test is positive.

Question 19

What is negative predictive value?

Answer 19

Probability that the disease is not present when the test is negative.

Question 20

What equation is used to work out positive predictive value?

Answer 20

(true positives + false positives)

Question 21

What equation is used to work out negative predictive value?

Answer 21

(true negatives + false negatives)

Question 22

The true positive rate is also known as…..?

Answer 22

Sensitivity

Question 23

False positive rate is also known as….

Answer 23

(1- Specificity)

Question 24

What is found on the x and y axis of ROC curves?

Answer 24

x axis = (1- specificity)

y axis= Sensitivity

Question 25

What does ROC stand for?

Answer 25

Receiver Operating Characteristic

Question 26

What are ROC curves used for in a clinical context?

Answer 26

Setting test cut off points

Question 27

What does the null hypothesis usually state?

Answer 27

That the difference between the two means is 0

so intervention makes no difference

Question 28

What p value would cause you to reject the null hypothesis?

Answer 28

p less than 0.05

Question 29

What is a categorical variable?

Answer 29

Can take a fixed number of values

Question 30

What is a discrete variable?

Answer 30

Can take any of a number of values