ADAPTIVE RANDOMIZATION

ADAPTIVE RANDOMIZATION

Randomization in clinical trials is the process of ensuring that there is no bias while assigning treatments/therapies to group of patients. At Kreara, the randomization and blinding processes are carried out as per the standard operating procedures.

Over the years of experience, it has been observed that traditional randomization approaches may not be suitable in practical situations. According to the traditional methods, the randomization list is prepared at the initial stage of the study, even before the recruitment starts and the entire sample of patients is often not available at the time of randomization. So, it is very likely that at the end of randomization, bias might crop in.

This statistical imbalance can be addressed by adopting, adaptive randomization techniques wherein the assignment of treatment to a new subject is dependent on the current balance across treatment groups.

The different methods of Adaptive Randomization are

1. Adaptive biased coin design

In this approach the assignment of treatment group aims at attaining a balance among the treatment groups by assigning the next patient to the group with the smaller sample size with higher probability. In recent years, we have applied adaptive randomization in oncology study trials aiming at comparison of two treatments. In that case there were only 10 patients available at the time of study start but the required sample size was 50,so we developed an algorithm to implement the adaptive biased coin design by balancing sample sizes in each group.

2. Covariate adaptive randomization

This method aims to balance the covariates across the treatment groups by assigning the next patient to the group that causes the smallest maximum imbalance across the covariate groups.

This is one of the most important methods to minimize the imbalance when compared to other adaptive methods. Covariate-adaptive randomization has an advantage over stratified randomization, as it is able to achieve balance over a large number of covariates when the sample size is small to medium. We have applied this method for small to moderate size clinical trials with several prognostic factors or covariates.

We have resorted to the covariate-adaptive randomization for a phase 1 study that was conducted with the objective of assessing the treatment effect on patients with a rare genetic disease. After assignment of treatments to an initial count of patients, this method was adopted to assign remaining patients to the three treatment groups. The covariates considered during randomization were age, sex and hormone level.

3. Response adaptive

A response-adaptive randomization procedure changes the allocation probabilities for each subject according to previous treatment assignments and responses in order to meet some objective.

Response adaptive randomization is used in studies dealing with the severe disease conditions (treatments/therapies) and here the allocation to group is based on the previous patient response. A hypothetical scenario would be a study conducted on say respiratory failure (for phase 1).

The approach adopted will include considering the response to treatment of a subject while assigning the treatment to the next subject. If the patient response is success (S), then second subject receives the same treatment. The same treatment assignment on other patients is continued until a failure (F) occurs. Upon failure, the next subject will be assigned to the other treatment group. This method is controversial and not commonly used.

Post A Comment

Protected by WP Anti Spam