Randomized Controlled Trials Flashcards
What is a trial?
An experiment where the investigators manipulate the environment
What is a clinical trial?
Subjects are people and we are studying what happens to them
What is a controlled clinical trial?
A trial with two groups (control and experimental). Conclusions will be based on a comparison between them.
RCTs are considered to be the gold standard study design from which….
the evidence on the efficacy of drugs is assessed
What are the 3 possible types of trials?
- To test an intervention in patients
- Prevention studies in healthy patients (vaccines)
- Community trials where groups of people are studied
Define equipoise?
A state of genuine balance and uncertainty about the benefits or harms that may result from two or more regimens.
Why is a state of equipoise an indication for a RCT?
Because there are no ethical concerns about one regimen being better for a particular patient.
What are the 3 “layers” of populations when designing an RCT?
Reference population (entire population to whom the results of the study will be applicable.
Experimental population (sample chosen from the reference population, should be representative of the reference population).
Study population (participants who are eligible and willing to participate).
What is randomization?
Random sorting of participants into a treatment or controlled group (equal or balanced chance of being exposed to an intervention or placebo).
What are 2 reasons why we need a placebo group?
- To understand the placebo effect
- To correctly ascribe the effects of the intervention
A well-designed RCT should eliminate all sources of…
bias
Randomization means the allocation of treatments to patients is carried out using a chance mechanism, such that…
neither the physician nor the patient know in advance which therapy will be assigned
What is the best defence against selection bias?
Randomization
How does randomization minimize/prevent confounding bias?
By balancing the distribution of risk factors
Normally, randomization ensures that the groups are similar “on average” at baseline. However, due to chance, some variables may not be balanced. This is highly dependent on…
the size of the study
How do we maintain effectiveness of randomization?
By intent-to treat analysis
What is stratified randomization?
Participants are divided into subgroups (strata) based on key variables. Randomization is then performed within each subgroup.
Stratified randomization really ensures…
balance of important (prognostic) factors
Stratified randomization can be limited by… (2)
small sample size
too many strata
What are the 2 advantages of randomization?
- Eliminates physician selection and patient self-selection bias
- Balances known and unknown factors among treatment and control groups
What are 2 disadvantages of randomization?
- Patient or physician may not want to participate in an experiment involving a chance mechanism to decide treatment
- May influence patient-physician relationship
Say we select all patients born on Monday to be in group A and all patients born on a Frida to be in group B. Is this true randomization?
No, it is pseudo-randomization
What is the concealment of allocation
Procedure which protects the randomization process (recruiter does not know what treatment will be allocated to the patient before enrolling them into the study).
Why do we need concealment of allocation?
Because knowledge of the treatment group assignment can introduce bias
What is a Table 1 in an RCT
Summary of the characteristics (and known risk factors) of the study population - meant to show that the population has been randomized effectively.
What are 3 ways in which to maintain balance between groups during follow-up
- Blinding to treatment
- Intention-to-treat analysis
- Using strategies for high retention of study participants
What is blinding?
Process of hiding treatment allocation that ensures participants, researchers, or both don’t know which treatment a participant is receiving (treatment or placebo).
Should mean that it is not possible to distinguish between interventions.
Concealment of allocation prevents…
Blinding prevents…
Concealment of allocation prevents selection bias.
Blinding prevents performance and detection bias.
Blinding can be impractical because… (2)
- side effects are obvious (e.g. swelling or irritation at the site of injection)
- can increase pill burden
What are the 3 types of blinding?
Single blind
Double blind
Triple blind
Define single blind
Participants are unaware of the group they are in
Define double blind
Participants and caregivers are unware of the group they are in
Define triple blind
Participants, caregivers and data analysts are unaware of the participants’ group
What is an open-label trial?
A trial with no blinding
What type of biases are eliminated by randomization?
What type of biases are eliminated by blinded treatment?
Randomization: selection bias and confounding
Blinding: Misclassification of exposure status, misclassification of outcome status
Name the 6 validity criteria for RCTs
- Randomization
- Concealed allocation
- Blinding
- Comparability at baseline
- ITT analysis
- Low losses to follow-up
Intent-to-treat analysis
ALL randomized subjects are analyzed, regardless of whether they completed the study, followed the protocol, or switched or discontinued treatment.
What are the advantages of ITT?
It preserves the power of randomization by maintaining balance in the analysis; it ensures validity of the results and provides a pragmatic estimate
What are the disatvantages of ITT?
Introduces misclassification of exposure
Biologic efficacy is not measured
What is per-protocol analysis?
Only includes participants who fully adhered to the treatment protocol and completed the study as planned.
Which type of analysis gives an accurate measure of how active the drug is when taken according to exact protocol?
Per protocol
In order to have adequate power or ability to detect reliability, a trial must have …
sufficient sample size
What is a type I error?
false positive results
What is a type II error?
false negative results
What is the formula for power?
1-beta
The power of an RCT is..
the ability to find a difference between two treatments when a difference exists (usually 80%)
The larger the sample size, the … the power
The larger the sample size, the greater the power
Power needs to take into account…
loss to follow-up
What are interim analyses?
Analyses conducted by a Data and Safety Monitoring Board (independent of investigators conducting the trial) to assess safety and efficacy of the RCT. May recommend early termination.
When might the Data and Safety Monitoring Board recommend early termination?
If it finds strong statistical difference between groups or no difference (when there is no equipoise).
What is the number needed to treat (NNT)?
The average number of patients that need to be treated by a specific therapy for one person to benefit.
The number needed to treat (NNT) is the reverse of…
absolute risk reduction (ARR) or risk difference
Describe phase 1 trials
- First trial using human subjects
- Generally done in healthy volunteers
- To determine pharmacokinetics and pharmacological effects
- Often use preparation that is not the final marketed form
Describe phase 2 trials
- In limited number of individuals having the disease of interest
- To determine drug effectiveness, side effects
- Looks at short term effects
Describe phase 3 trials
- Larger number/spectrum of individuals with the disease is studied
- Done after the effectiveness and short-term safety have been demonstrated
- Used as it would be marketed
- Larger RCTs
- Highly regulated
Generally done in ideal populations - May compare new drug to placebo or to existing standard of care
Describe phase 4 trials
- Conducted post-marketing
- To obtain additional info to what was found in phase 3 trials (e.g. different dosages, durations, diseases, sub-populations)
Name 2 strengths of RCTs
- Minimize the potential for bias (uf well-executed)
- Best study design to address questions of therapy efficacy
Name 4 limitations of RCTs
- Expensive and time consuming
- Not reflective of routine practice
(highly selected populations, highly controlled practice) - Not enough power to study uncommon but serious adverse effects
- Infeasible in many instances because of ethical concerns
True or false: Clinical trials must be registered in a public registry to be considered for a publication.
True. Trials must register at or before the onset of patient enrolment
CONSORT guidelines stands for…
Consolidated Standards of Reporting Trials
