RCTS - validity and patient preferences (Statistics) Flashcards
Define External Validity
Generalisability of results
Ability to apply results from sample population to the broader population
To what extent are the sample population representative of the population of interest?
Consider ways in which sample population may differ from the population:
Eligibility criteria/Exclusion criteria
Volunteer bias
Loss to follow up
Clinician preference
Setting - healthcare system, recruitment from primary/secondary/tertiary centres, selection of centres
PATIENT PREFERENCE
Many trials exclude females particularly of reproductive age, comorbidities, by age
Define Internal validity
Accuracy of results.
The extent to which a study is free from systematic error and confounding. This arises from randomisation and its proper implementation (allocation concealment) and maintenance (minimal loss to follow up)
Threats to internal validity
PATIENT PREFERENCE
Attrition
Confounding variables
Researcher bias
Diffusion (if there is social interaction between intervention and control groups)
Statistical regression to mean
How does patient preference affect external and internal validity?
External - Selection bias/volunteer bias. Those who chose to take part may not be representative of the population
Internal - Less loss to follow up in preference group. If they knowingly receive preferred treatment –> better concordance, influence on therapeutic effect. If knowingly receives other treatment may perceive a negative effect, poorer compliance, loss to attrition
Allocation with patient preference in RCT can minimise attrition
Advantages of a partially randomised patient preference trial
Patients have preferred treatment if they want it
- may enhance recruitment
- better for mental and physical wellbeing
- reduced demoralisation
- reduced drop out
- may improve perceived benefit
- could be argued that it is more representative of real life
Is it more ethical?
Requires more patients, more time.
HOWEVER loss of randomisation
