In this post, we have exclusively written about Experimental Designs used for Randomized Controlled Trial (RCT). What is it? To know, read on…
(Previous posts about Clinical trials:
Clinical Trials (P-2): Randomization.,
Clinical Trials (P-3): Placebo effect and Blinding.)
In Random Controlled Trials, patients are randomly assigned to the groups. Depending on the number and nature of the treatment, characteristics of disease, the endpoint and many other factors, the subjects (patients) can be allotted to different groups. As the clinical trial tends to get very expensive and there is an attempt to minimize the funds requirement as well. The way the patients are allotted to different groups and the treatment procedure given to the different goups have to planned during the planning of the clinical trial. The ‘way’ an experiment is conducted is called experimental design. It is the outline of the different steps in the experiment to test the given hypothesis. Mentioned below are the different types of study design used for RCT:
- Parallel design
In this type of design, a group of subjects (participants) receive the intervention and another group (control group) receives the standard treatment or placebo. But each group receives a single treatment throughout the study. Each individual in the two intervention arms hence receives only one type of treatment throughout.
This design is considered to be ‘gold standard’ for the phase 3 trials. Bias is minimized as the randomization and blinding is possible. However, the between-subject comparision is confounded by various prognostic factors. No carryover occurs here.
Most trials have a parallel design such as
- Dose escalation studies
- Some acute and progressive conditions s/a cancer.
- In case of drugs which bring about changes in the course of disease and has permanent effects.
Example: A double-blind parallel-group clinical trial, patients with fibromyalgia were randomized to receive either saffron or duloxetine.
Read the experiment in detail (also find words like blinded, exclusion and inclusion criteria, etc): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6235666/
- Factorial design
Randomised controlled trials can be very expensive. Factorial trial is a great study when two or more different interventions are to be tested such that it affects the same target measurement. It can also be used to find the effect of two different interventions (treatments) together. However, it is important that there is no interaction between the two interventions and their effects are independent of the presence or the absence of the other intervention. This design can be randomized and blinded.
The 2 x 2 factorial design is the simplest one (Diagram), with only two treatments (interventions) involved. In this the participants are allocated to receive none of the intervention, two interventions singly and both together (none, A, B and AB).
Example of factorial trial to assess the effects of daily zinc, probiotic or placebo on the immunogenicity oral rotavirus vaccine. Infants were randomized equally between study arms as infants receiving (1) probiotic and zinc, (2) probiotic and placebo, (3) placebo and zinc, (4) two placebos.
Read the entire paper here: https://www.ncbi.nlm.nih.gov/m/pubmed/28874323/
Similarly, there can be 2 x 3, 3 x 3 factorial designs as well, depending on the factors (interventions) involved. Read more on these:
These studies enable efficient simultaneous investigation of two interventions by including all participants in both analyses.
Second, it is possible to consider both the separate effects of each intervention and both interventions together.
It is possible to answer two or more questions in a single trial with only a marginal increase in cost.
Needs smaller sample size (half of parellel trials)
Patients must be willing and to take any of the treatment combinations
Requires extra time, compliance and management in applying two treatments at same time. This design may require burdensome administration scheme if blinded.
Optimal dose modification strategy for toxicity may be hard to determine.
Interaction complicates interpretation of treatment effects.
- Crossover trials
Crossover trials involves trial participants receiving all the interventions involved in the trial at different specified times and sequence. This design may involve two or more interventions allocated to participants at random order. The order in which each individual receives them (A then B or B then A) should be determined by random allocation.
Each subject act as his/her own control by receiving at least two different interventions [e.g., a new drug (treatment A) versus the standard drug (treatment B)] at different times during the trial. There should be enough time gap between the two interventions to make sure that the effect of the first intervention completely fades out. This ‘carry-over effect’ or the ‘spill-over effect’ of the first intervention could result in error in the reading of the second intervention. Hence enough time is allowed to pass, known as ‘wash-out period’ between the two interventions. Consequently, this design is suitable only when neither of the interventions has long-term effects. Read this example: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5846618/#!po=1.25000
- Here each subject is compared with himself/herself, eliminating confounding factor from the comparison.
- Increased statistical precision in the comparison and nearly half the sample size is required to yield the statistical power as a two arm parellel design.
- Apt for early phases, to specially analyse pharmacokinetic properties in healthy volunteers.
Not suitable for drugs with long term effect or to test drugs for disease which progresses very fast or is at severe stage, as keeping a wash out period would be fatal in such case.
- Cluster design
This is similar in its layout as parellel design only that the patients are randomly assigned as a complete group or cluster, and not individuals. All the individuals in a cluster or groups are given same treatment, and hence simplifies the process of documentation and reduces confusion. Hence it is used for convenience sake. Clusters could be group of individuals from different places, eg. Different schools, colleges, villages, etc.
E.g. of a study involving a cluster design: In a study the allergic rhinitis patients were allotted based on their geographical location. Read the paper: https://www.karger.com/Article/Abstract/381339
This is all for this post. Next post I will write about the different phases of clinical trials. Read all the post of this series before you read that, to understand better.
Hope u like this post, if yes please comment, like and share!!
Also follow us on Facebook, Twitter, Instagram or send an email to firstname.lastname@example.org.
Have a nice day!
Read other notes by The Biotech Notes:
The Tuskegee Study: Ethics in Clinical Research.