Understanding clinical trial designs Flashcards
What is a clinical trial?
- a type of clinical research that compares one treatment or intervention with another
- May involve patients or healthy people or both
- any research that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcones
What is bias?
- aims to obtain groups where the only systematic difference are the interventions under investigation
- Flaws in the design and conduct of a trial can introduce bias - i.e. Other systematic differences between the treatment groups
- can severly hamper the interpretation of the results of the trial
What are the different types of bias?
- Selection
- Performance
- Attrition
- Detection
- Reporting
What is selection bias?
- systematic difference between groups in how participants receibe intervention/control
Methods of minimising bias:
- Randomisation allocation concealment
What is performance bias?
- sys difference between groups in how particpants are cared for, co-interventions
- Methods of minimising bias:
- Blinding of participants, team and assessors
to minimise bias
What is attrition bias?
- Numbers of participants dropping out , non-random withdrawl
- minimise bias: intention to treat analysis
What is detection bias?
- How outcomes assessed
- Objective outcome measures
- Blinded outcome assessors
What is reporting bias?
- Which outcomes are reported
- minimise bias: registration of trials
Why should we randomise?
- problems with:
- historical controls
- non-randomised concurrent controls
- Factors (known and unknown) that might affect outcome similar in each arm of the trial
- Basis of statisrical inference
How do we avoid performance and detection bias?
- Blinding : witholding knowledge of treatment AFTER assignment
- Trial participants
- Healthcare providers
- Data collectors
- Outcome adjudicators
- Data analysts
What is ITT analysis?
- Exclusion of participants from the analysis can undermine randomisation
- Any exclusion should be carefully justified
- Where possible include all participants and “analyse as randomised - the intention to trear principle
- This provides an unbiased estimate of the effect of being assigned to an intervention
What is the registration of trials?
- Registration = the first step towards research transparency and future dissemination of outcomes
- Ensures all healthcare decisions are informed by all available evidence
- Provides opportunity for collaboration and reduces duplication of research efforts
- Improves awareness of trials for clinicians , researchers, patients and public
- TRIALS MUST BE REGISTERED
What are different trial designs?
- parallel group
- crossover
- factorial
- cluster
- platform trials
What are parallel groups ?
- groups receive different interventions in the same time period
- all participants have the same outcome assessments
- the groups are compared
What are pros and cons of parallel group trials?
Pros:
- simple
- robust
- both groups treated in the same time period
Cons:
- only provides differences between groups
- large studies required to study rare events
What are crossover trials?
- Participant receives all treatments
- Randomised to an order of treatnebt
- can compare the responses for each subject and compare the responses between the two groups
- Not suitable for all trials- useful for chronic conditions
- Need to include a washout period
What are pros of crossover trials?
- smaller number of patients required
- Each participant receives both treatments
- Can look at responses witin each patient : each patient acts as their own control
What are cons of crossover trials?
- Not suitable in certain disease areas
- Takes longer
- Potential for effect of first treatment to carry over
- Potential for a period time effect
- Greater risk of drop out
What is factorial design?
- different type of multi-arm design where participants can receive one or a combo of treatments
- For examole, two separate treatments and a combination of treatnents being tested together
- Patients receive none, one or bot of the treatments
What are pros of factorial trials?
- two trials in one
- can potentially determine if treatments work bettwe together (i.e. if there is an interaction)
What are cons of factorial trials ?
- more complicated
- bigger sample sizes might be required depending on the strength of interaction
What are cluster randomised trials?
- Groups of participants are randomised to a treatment/intervention e.g.
- GP surgeries
- Communities
- Hospitals
- Intervention naturally applied at a cluster level
- Convenient administration
- Avoids potential contamination between treatments
- Can enance recruitment
What are pros of cluster trials?
- prevents contamination between individuals assigned to different treatments
- Useful when individual randomisation is impossible
- May be easier to recruit to if consent is only for follow-up
What are cons of cluster trials?
- May need many participants
- Can be less efficient than individually randomised designs
- May be more complex to set up
What are platform trials?
- A master protocol which covers multiple research questions or comparisons
- The flexibility to make adaptations
- Effcient way to answer multiple questions in a short period of time
What are pros of platform trials?
- Efficient way to answer multiple research questions quickly
- Useful when there are many untested interventions
- Potential to greatly shorten time to patient benefit
What are cons of platform trials?
- Complex and expensive to set up and manage
- Not necessarily always more efficient than an individual parallel group trial
What is superiority vs non-inferiority?
S: Designed to show that a new treatment is better than the control
NI: Designed to show that a new treatment is not worse than te control
What is pragmatic vs explanatory?
-Explanatory trials : Trials which will demonstrate whether an intervention could work in idela circumstances
- Pragmatic trials
- Trials where te design mimics routine clinical practice as closely as possible with the exception that patients are randomly allocated to treatment
