EBM

Question 1

What is the best trial design for adverse events?

Answer 1

Phase IV/ post marketing surveillance

Question 2

What is the best way to compare multiple interventions?

Answer 2

Network Meta analysis

Question 3

What is internal vs external validity?

Answer 3

Internal validity = looks at the quality of the RCT (eg. Blinding, compliance etc)

External validity = highly comparable to general population

Question 4

What is a non inferiority margin?

Answer 4

The acceptable number of inferior outcomes when demonstrating non inferiority in an RCT

Will vary depending on outcome eg. Death might be 1%, pain might be 10%

Question 5

What are the disadvantages of non inferiority trials?

Answer 5

• new drug often sponsored by pharma
• poor adherence in both trial arms will show non inferiority
• non inferiority margin may be set too high by investigators. Ensure its sensible

Question 6

What is a hazard ratio?

Answer 6

The relative risk between two groups averaged over the duration of the RCT

Question 7

What is a propensity trial?

Answer 7

Instead of randomisation - patients are matched by propensity score

Eg pt 1 and 10 have similar propensity score for cardiovascular risk based on age and comorbidities

Question 8

What are the three ways to measure effect size and the trials they come from?

Answer 8

Relative risk
> from RCT and cohort study

Hazard ratio
> relative risk averaged over time

Odds ratio
> from case controls, cohort studies etc

Question 9

What is the best statistical test for categorical data vs numerical data?

Best test for crossover trial?

Answer 9

Categorical = Chi2 (percentages are categorical) (use Fischers exact for small sample size)

Numerical = t-test (numbers on a spectrum, eg. BP, DEXA results, FEV1

Cross over trial = paired t test