Meta analysis- clinical interpretation

Question 1

How does blinding the studies affect meta-analysis (2)

Answer 1

• More consistent scores

- More than one person conducting the assesssment

Question 2

What biases are involved in meta analysis (4)

Answer 2

Assessing the risk of bias

• Random sequence generation (selection)
• Allocation concealment (selection)
• Selective reporting (reporting bias)
• Blinding of outcome assessment (detection bias)

Question 3

What are some different types of trials that may be used in a meta-analysis and what are the disadvantages of them? (7)

Answer 3

Randomised trials

• may not be ethical to randomise
• different levels of expertise
• strict eligibility - differs from ‘average’ patient

Observational trials

• confounding factors
• differences in healthcare and standard of care

Question 4

What are some standardised outcome measures used in meta-analysis? (6)

Answer 4

Binary- odds ratio, relative risk
Continuous- same units, can have much larger effect

Outcome measures may be:
Mortality
Hospital Admission
Biomarkers 
Sometimes side effects

Question 5

What are funnel plots and what can they show? (7)

Answer 5

A funnel plot can show bias
Increases in sample sizes can show a more clustered funnel plot

Symmetrical- normal
Asymmetrical- missing results 
- publication and reporting bias 
- heterogeneity
- intensity of intervention

Question 6

What are subgroup analyses and what are some disadvantages? (4)

Answer 6

Comparing overall effect to a particular subgroup

Can lead to

• False positives
• Multiple characteristics per person
• Different effect sizes

Question 7

What are sensitivity analyses and meta regression (2)

Answer 7

Sensitivity analyses
- Calculating outcomes under alternative assumptions

Meta regression
- Determines strengths and character of a relationship