Critical appraisal and study designs

Question 1

Why is randomisation important?

Answer 1

balances patient characteristics between groups

Question 2

Best type of randomisation

Answer 2

Centralised computer randomisation with contact by phone or computer

Question 3

If a trial loses lots of patients during intervention this…

Answer 3

reduces sample size AND weakens findings

Question 4

Why shld patients who withdraw/lost to follow-up during a trial be analysed according to the groups they were originally assigned?

Answer 4

to preserve the value of randomisation

Question 5

When might blinding be unimportant in an RCT?

Answer 5

If the outcome is objective e.g. death

Question 6

Why type of study may be used to estimate prevalence?

Answer 6

Cross-sectional study

Question 7

From which type of study can we estimate risk of disease associated with an exposure using odds ratio?

Answer 7

Case control

Question 8

What’s the main feature of a cohort study?

Answer 8

it can determine the incidence rate of disease amongst exposed and unexposed individuals

Question 9

Disadvantage of excluding English language studies:

Answer 9

• relevant studies may be published in non-English language journals
• English-language only searches tend to overstate the effect of an intervention

Question 10

Disadvantage of using only 1 database

Answer 10

Not all relevant articles may be indexed in medline

Question 11

Advantages of getting two authors to review quality of selected studies for a systematic review (breaks this q down - why is it good to assess study quality and why is it good to use 2 authors to do this?

Answer 11

• Assessing quality is important because studies w weaker designs will be less valid and can overestimate effects.
• Using two independent reviewers to assess quality makes it less likely that errors will be made.
• Using pre-agreed criteria helps make the process objective and transparent.

Question 12

Why are RCTs least subject to bias compared to observational studies?

Answer 12

• randomisation btwn grps increases chance any difference in outcomes is due to intervention rather than differences btwn groups

Question 13

Meta-analysis

Answer 13

A statistical technique for quantitatively combining the results of multiple studies that measure the same outcome into a single pooled or summary estimate

Question 14

What’s the effect on sample size if a cohort study is used to investigate a rare outcome?

Answer 14

Sample size is excessive

Question 15

What is the primary purpose of controls in a case-control study?

Answer 15

• offer a comparison group

comprising people from the same population but without the outcome of interest

Question 16

Define power

Answer 16

the ability of the study to identify a true effect