Intro to RCTs Flashcards
What are observational studies?
- no intervention
- observe and record behaviour, Sx, attitudes
Association between risk factors and disease/outcomes
Cohort, cross sectional studies
What are experimental studies?
research intervenes deliberately to influence events
records effect of intervention
- effectiveness of new drugs, Rx (comparative)
- RCTs, cross-over trials
What 2 types of outcomes should be considered in a RCT?
- positive: good
- negative: bad
ideally, following I want to increase the abundance of positive ones and reduce negative ones
What kind of study design is RCT?
parallel design: patients randomised to a group, which are studies in parallel
then compare outcomes between Intervention and Control groups
What controls are used in a RCT?
- Standard Rx: active or positive control
- Placebo (negative control)
- No “additional” Rx (unlikely)
these are collectively known as the CONCURRENT CONTROLS: they reflect current natural epidemiology
What is the placebo in the RCT?
referred to as “treatment”
no active therapeutic ingredients
similar appearance and taste as active Rx
- administered in same way
What music new Rx or therapeutic regime RCT include?
a control group: to know how the condition would proceed normally (without new intervention)
allows comparison of how much recovery is down to new drug vs. natural improvement of condition
What is the aim of sampling for any study?
to minimise SELECTION BIAS
e.g. systematic difference between population and sample [characteristics]
What is the purpose of using random allocation to group the participants?
removes ALLOCATION BIAS and minimised confounding
What is the difference between ALLOCATION vs SELECTION bias?
ALLOCATION: bias using characteristics that designate participants to either control or intervention group (certain characteristics for control, certain for intervention)
SELECTION: bias made when choosing the initial sampling population. Results in systematic differences between sample and population and can make the study less representative
How does use of a placebo aid RCT design?
ENSURES BLINDING
- minimises ASSESOR BIAS
- minimises RESPONSE BIAS
(all needed to be able to infer causality)
What is the most ideal situation with sampling a population? Why is this not generally done?
investigate the ENTIRE population
all adults meeting inclusion criteria, potentially infinite
-> v. expensive and time consuming
What are the methods of sampling a population?
- convenience sampling (people taken as they are referred or present)
What is a method of making the sample more representative of the population?
using patients from multiple centres or countries or regions
What the importance of making the sample representative of the general population?
affects ability to GENERALISE study results to rest of population
If this is not possible to do, then there is SELECTION BIAS present
What is a bias?
= systematic difference
(not down to chance, but also may not be deliberate)
e.g. study contains more women or >65yo individuals than general population
this may be influenced by other non-deliberate factors: people who are more likely to volunteer after selection
What is the benefit of random allocation?
probability of an individual being allotted to Rx group is equal
eliminates allocation bias
minimises confounding (there should be no differences in baseline characteristic between groups to explain differences in outcomes
How does random allocation facilitate blinding?
- requires placebo
- double blind
What is the benefit of performing a RCT with double-blinding?
= group allocation blinded to researches and participants
reduces ASCERTAINMENT BIAS: - Response bias (patients)
- Assessment bias (researchers)
What are the advantages of using a placebo?
- RESPONSE BIAS
participants may alter responses if they know which group they are in - may report responses to please researchers
- difference between patient response and “truth”
- ASSESSMENT BIAS
researchers may bias outcomes: subjective interpretation to favour their theory - LOSS TO FOLLOW UP
participants less likely to drop out since they are unaware if they’re in the placebo group or not
What is the ‘TREATMENT RESPONSE’?
the response in the outcome for the intervention group
components: natural epidemiology of condition, placebo effect PLUS treatment effect
What is the ‘PLACEBO RESPONSE’?
response in outcome for the placebo group
components: natural epidemiology of condition PLUS placebo effect
What is the empirical difference between the ‘Treatment’ and ‘Placebo’ responses?
the inclusion of the ‘treatment effect’ in the treatment response
this is the pharmacology (of the intervention)
What is the Hawthorne effect?
It is a non-specific treatment effect;
Change in behaviour as motivational response to interest, care, or attention received through observation and assessment.
What is the PLACEBO EFFECT?
NOT the placebo response FYI
psychological effect, seen especially in subjective outcomes
Components:
- response to investigation and therapeutic ritual
- subsequent response to observation and assessment
- response to doctor-patient interactoin
What effect is the Hawthorne effect a part of?
part of the placebo effect
How is the treatment effect calculated?
= pharmacological effect (of intervention)
= Treatment response - Placebo response
What component in present in all RCT groups response?
- natural epidemiology of condition
- placebo effect
difficult to quantify these effects and separate one from the other
For valid consent, what should be included?
[information should be understood]
communication of
- risk
- random allocation
- use of placebo
- pros/cons of Rx
Why is informed and valid consent important for RCTs?
participants need to make an “INFORMED DECISION” as to whether or not they want to be involved
this can only be done if they are provided with and fully understand the info on all its aspects
What is a cross-over trial?
derivative of the RCT design
patients receive both Rx (one could be a placebo) but ORDER of Rx allocation is random
washout period exists between 1st Rx and 2nd Rx
What is the role of a WASH OUT period in the cross-over trial design?
allow any residual effect of previous Rx to be removed before 2nd Rx
What is the benefit of receiving the Rx in a random order in a cross-over trial?
minimises an ORDER EFFECT
calculate: control Rx - Inv Rx
across participants and then average
AVG: removes the variation from within an individual
When can Cross over trials NOT be used?
for conditions that are cured by Rx
must be chronic or Rx only relieves Rx short term
e.g. pain relief efficacy in cancer
pain relieved NOT cured
Can blinding occur in a RCT where the placebo is no Rx?
no because blinding is dependent on randomisation of Rx (but for this to work there must be Rx for both groups (placebo must be a treatment, just not active)
