EBM

Question 1

What is Relative risk? Formula

What is RRR? Formula

Answer 1

Estimates the magnitude of an association between exposure and disease.

Formula = D+ve in New drug/D+ve in placebo or standard therapy

Relative risk reduction is the reduction in disease by a % with the new drug compared with the placebo/old drug.
e.g. RRR 80% means there is an 80% reduction in the disease when compared with placebo/standard therapy

Formula = 1 - RR

Question 2

What type of bias is this. Systemic difference in the way that disease incidence data are collected between populations and compared.

Answer 2

Information bias

Question 3

What type of bias? Differential recall of previous exposures to RF

Answer 3

Information bias

Question 4

What is the Odds ratio?

How does this differ from Hazard ratio?

Answer 4

Number of times an event happens divided by the number of time it does not within a group.
The RR if the outcome of interest is rare.

HR:
Relative risk of events occurring over time (Kaplein meir curves)
= survival over time
e.g. HR 0.79 means apixiban was associated with a 79% risk reduction in he risk of the primary outcome.
Measured by kaplein meir curve.

Question 5

What is absolute risk (AR). Formula?

What is ARR?

Answer 5

The absolute risk of getting a disease if you take new drug.

Formula for risk of taking drug: D+ve events in drug X/ total number of people in Drug X category
OR
Formular for risk of not taking drug: D+events in Placebo(standard therapy)/total number of people in placebo category

ARR = AR of not taking drug - AR of taking drug-

Question 6

Stratifying in clinical trials is aimed at reducing?

Answer 6

Confounding due to known variables

Question 7

A randomized clinical trial investigating a new drug would be most susceptible to which bias?

Answer 7

Publication bias

Question 8

What is a type 1 error aka false positive error?

Answer 8

Finding something there that isn’t there. E.g. D-ve but T+ve, false +ve result.
Null hypothesis incorrectly rejected in favour of alternate.

denoted by alpha.

Due to chance, bias.

If p

Question 9

What is a type II error aka false negative error?

Answer 9

Not finding somethere there when there is something there. e.g. D+ve, T-ve, false -ve result.
Null hypothesis is not rejected when it is in fact false.

Denoted by beta.

You can increase power and decrease beta by:

• increasing sample size
• increase expected effect size
• increase precision of measurement i.e. increase alpha from 0.05 to 0.10

Question 10

What 3 factors affect the power of a study?

Answer 10

Sample size
magnitude of difference between the 2 groups ( relative risk)
The p value required for a significant result i.e. alpha

Question 11

How does lowering the alpha from 0.05 to 0.02 affect error?

Answer 11

Increase in Type 2 error (false negative) and decreased type 1 error

Question 12

When can a type 1 error (false -ve) arise when a trial is reported?

Answer 12

If the null hypothesis is falsely rejected

Question 13

What is a cross sectional study?

How is it measured?

Answer 13

Collects data from a GROUP of people to assess the FREQUENCY of the disease (and related risk factors) at a particular point in time.

Measures disease prevalence.
Can show a risk factor association with disease, but does not establish causality.

Question 14

What is a case control study?

How is it measured?

Answer 14

Compares EXPOSURE or RF
Aa group of people with disease to a group of people without disease.
A retrospective study.
Good for rare disease.
e.g. patients with COPD had a higher odds of a history of smoking.

Measured by OR.

Question 15

What is a cohort study?

How is it measured?

Answer 15

TESTS EXPOSURE
Prospective and retrospective study.
Compares a group of people with a given exposure or RF to a group without such exposure.
Looks to see if exposure increases likelihood of developing disease.
e.g. smokers had a higher risk of developing COPD

Measured by RR.

Question 16

Cross over study?

Answer 16

Longitudinal study.
Subjects recieve a sequence of diff Tx exposures.
Repeated measures design in which each patient is randomly assigned to a sequence of Tx including at least 2 Tx.

Question 17

False negative calculation?

Answer 17

1- sensitivity

Question 18

False positive calculation?

Answer 18

1- specificity

Question 19

How do you avoid selection bias?

Answer 19

Randomisation and intention to treat analysis

Question 20

How do you avoid recall bias?

Answer 20

Decrease teh time from exposure to follow up

Question 21

What is confounding bias?

How do you avoid confounding bias?

Answer 21

When a factor is related to both the exposure and the iutcome, but not on the causal pathway.
E.g. pulmonary diseae is more common in coal workers however coal mine workers are more likely to smoke.

Avoid by multiple repeated studies and cross over studies (subjects as as own control).
matching patient with similair characteristics in both groups.

Reduce confounding by multi variable analysis

Question 22

What is lead time bias?

How to avoid?

Answer 22

Early detection is confused with increased survival.

Avoid by measuring back end survival (adjust survival according to disease)

Question 23

What is the null hypothesis?

What is the alternate hypothesis?

Answer 23

Null hypothesis:
N difference or relationship between disease and risk factor int he population
(1-B).

Alternate hypothesis:
Some difference or relationship between disease and RF

Question 24

What is teh confidence interval?

Answer 24

Measure of the precision of the results of a study. 95% certain the true effect lines in.
Short lines I—-I = more precice

If CI includes 1 it is not significant.

Question 25

T test measures?

Answer 25

Checks diff between means of 2 groups.

Tea is meant for 2

Question 26

ANOVA measures?

Answer 26

Checks the diff between means of 3 or mroe groups.

3 words: Analysis of variance

Question 27

Chi-square measures?

Answer 27

Checks diff between 2 or more percentages or proportions if a categorical outcomes. e.g. comparing percentages of members if 3 diff ehtnic groups who have essential HT

Question 28

Multivariate analysis measures?

Answer 28

Used to determine whether or not confounding is occurring due to other factors.

Question 29

Phases of clincial trials?

Answer 29

1 = Dose
2 = Safety and efficacy
3 = Efficacy compared to Gold Standard
4 = post marketing

Question 30

Lack of blinding and masking. What type of bias?

Answer 30

Information bias