Randomised Control Trial, Analysis And Interpretation Flashcards
Definition of number needed to treat
No of patients who need to receive treatment in order to prevent AE in 1 patient
Definition of patient years
Total no of years that each participant contributed
Definition of internal validity
Lack of bias
Definition of external validity/generalisability
Whether the subjects are typical of patients normally seen
Whether treatment is typical of treatment delivered in most practises
Definition of explanatory trial
Tests efficacy
Can the intervention work under ideal conditions
Definition of pragmatic trial
Tests effectiveness
Does the intervention work in normal clinical practise
How do you calculate absolute risk reduction
How do you calculate relative risk reduction
ARR=Outcome in control - outcome in intervention
RRR=Outcome in control-outcome in intervention/outcome in control
How would you calculate number needed to treat (NNT)
NNT=1/ARR (when using percentages)
No of patients that need to receive treatment in order to prevent AE in 1 patient
If using patient years
NNT=patient years/ARR
What is the intention to treat
All randomised individuals included in analysis according to allocated treatment
This is regardless of drop outs/non compliance/loss in follow ups
Reduces ecological fallacy
Maintains balance between groups
Before a trial starts, who must approve of the trial
What info must be shared with the participants
What info must the participants give
What 2 independent oversight groups assess the conduct of the trial
What can they do
Ethics committee must approve
Informed consent must be obtained from participants
- Explain purpose and uncertainty
- Randomisation
- Alternative treatments
- May withdraw at any time
Trial steering committee
-ensures proper trial conduct
Data monitoring committee
- monitors safety analysis
- monitors data to see if intervention is safe
- If intervention found to be harmful prematurely => trial ends
- If intervention found to be beneficial prematurely => trial ends and control group also gets access to intervention
How is internal validity achieved
- Allocation of participants to trial arms
- trial performance
- outcome assessment
- method for analysis
- reporting bias
Allocation of participants to trial arms
- Randomisation
- Allocation concealment
Trial performance
-personnel blinding
Outcome assessment
-blinding
Method for analysis
-ITT
Reporting bias
- Publication of protocol
- report prespecified outcomes so all data is reported
What is external validity and generalisability
Whether the subjects are typical of patients usually seen
If treatment is typical of treatment delivered in most practices
What is the difference between explanatory and pragmatic trials
Explanatory
-tests efficacy, whether intervention works under ideal conditions
Pragmatic
-tests effectiveness, if intervention works in normal clinical practice
