Please enable JavaScript.
Coggle requires JavaScript to display documents.
Randomised Controlled Trial (RCT) - Coggle Diagram
Randomised Controlled Trial (RCT)
Ethical Justification and Feasibility:
Not ethical justifiable if:
Lack of equipoise (insufficient belief to justify intervention and insufficient doubt to justify placebo)
Intervention already standard of care without evidence
Too expensive or difficult to be implemented afterwards
Not economically / practically feasible if:
Very expensive and time consuming
Outcome rare or has long induction time
Possible, but difficult with non-pearmaceutical interventions
-surgery, zone/physio-therapy, change in habits etc.
Definition: A study to compare the effect of a new intervention to none or another
Intervention is assigned by the investigator
A random principle is applied to allocate the interventions
--> randomisation makes groups similar, known and unknown confounders will be equally distributed
--> any subsequent difference can only be attributed to the intervention
Aim:
Causal Inference To establish causality we need to know relative risk (risk "if" exposed/ risk "if" not exposed)
Terminology
Terminology: Sub-types
Parallel group (simple RCT)
Special/alternative designs
-factorial (tests effects of 2 or more interventions in one study)
-cluster (groups rather than individuals are randomised)
-cross-over (all participants given all interventions, one after the other)
Efficacy vs Effectiveness:
Efficacy's the extent to which a specific intervention produces a beneficial result under ideal conditions
Effectiveness: the extent to which a specific intervention does what it is intended to do when deployed in the field
SO
,
Efficacy depends on biology (e.g. vaccines, drugs, nutrition etc.)
Effectiveness depends on efficacy and storage, compliance etc.
Superiority and equivalence / non-inferiority
:
Superiority:
Assess if a new intervention is superior to none or a standard intervention
(NOTE: lack of difference in trial is NOT the same as equivalence)
Equivalence and non-inferiority:
To assess if a new intervention is equivalent or non-inferior to standard
Not possible to prove "equivalence" of new vs standard intervention
But possible to test if the effect excludes pre-specific margins
Terminology: Other Classifications
Place of recruitment
Phase in drug development
Purpose: demonstrate biological or public health effect, prevent or treat a disease or to assess if new intervention is better/equal or worse
Key Features of RCT
Intervention and Outcomes:
Intervention: any exposure likely to affect humans
Outcomes:
Anything that can be measured in an individual
Primary, Secondary or Explorative outcomes
Randomisation:
Is the main strength of the RCT as it creates baseline equivalence for known and unknown risk factors
A) Done incorrectly: then it is not randomisation and introduce selection bias
B) Done correctly, nut baseline equivalence not achieved --> confounding factors
Key Factors & Issues Relating to `Randomisation
Baseline Inequivalence:
Baseline differences may occur (by chance) and cause confounding
Don't do significance tests for baseline differences: potentially misleading with respect to confounder control
Assess for baseline differences
Allocation Concealment
: Randomisation aims to ensure all have the same probability of receiving exposure so it is critical that allocation is concealed for the staff
Post-Randomisation Exclusions:
Participants are occasionally excluded after the start of a trial after randomisation --> may cause selection bias
Loss to Follow-up
: If large and differ by intervention groups (I.e. the control group had a 5% loss to follow up vs 10% in the experimental group) then it may also cause selection bias
Non-ITT Comparison
:
Essential to compare the groups as randomised (i.e intention-to-treat). If not, the benefits of randomisation are lost and bias is introduced
Blinding:
A double-blinded RCT is preferred as this would ensure participants do not change behaviour or lifestyle related to the exposure and outcome. Assessors would also not be biased towards a certain outcome when crying out the analysis etc.
Biases
Effect Modification (subgroup):
The effect of an intervention may be different in subgroups, ie certain characteristics (sex, age, severity) may modify the effect
Power - Type 2 Error:
Sample size should ensure 80-90% power to detect an effect on the primary outcome of a relevant magnitude or greater, while allowing for loss to follow-up
Multiplicity Type 1 Error:
To reduce the risk of chance findings the number of tests should be reasonable
If no true difference, for 1 test, the risk of type 1 error is α=0.05
Nutrition Intervention RCT
Micronutrients: Supplementation/fortification
Advantages:
-Placebo/blinding possible
-Not replacing normal diet
-Isoenergetic
Disadvantage:
-Unphysiological --> single nutrient in high doses
-Bioavailability, metabolism and effects may be modified by lack of other nutrients
Macronutrients:
When designing a macronutrient nutrient intervention, consider the following:
The background diet?
Will the intervention change the diet?
Will the energy or specific nutrients be constant?
What is the contrast between the groups?
Specific Challenges of Nutrition Intervention RCT
:
-Population may already be exposed to specific foods or nutrients
-Intervention may affect dietary intake - supplement or replace (?)