What trial is a good example of stratified blocked randomization?

Examples of Clinical Trials Using Stratified Blocked Randomization

The MATRIX (Minimizing Adverse Haemorrhagic Events by Transradial Access Site and Systemic Implementation of AngioX) trial is an excellent example of stratified blocked randomization in clinical trial design. 1

Understanding Stratified Blocked Randomization

Stratified randomization is a technique used to ensure balance of participant characteristics between treatment groups. It involves performing separate randomization procedures within subsets of participants defined by important characteristics such as:

• Study center
• Age groups
• Disease severity
• Sex
• Other prognostic factors

Importantly, stratified randomization requires some form of restriction (such as blocking) to be effective. Without blocking, stratification alone is ineffective. 1

How Blocking Works Within Strata

Blocking ensures that comparison groups are generated according to a predetermined ratio (usually 1:1) within each stratum. For example:

• In a block of 8 participants, 4 would be allocated to each treatment arm
• This maintains balance throughout the trial, not just at the end
• Randomization is performed separately within each stratum

The MATRIX Trial Example

The MATRIX trial demonstrates sophisticated stratified randomization with a 3-level approach:

1. 8,404 patients with acute coronary syndrome were first randomized to radial versus femoral vascular access
2. Of these, 7,213 patients undergoing PCI were randomized again to heparin versus bivalirudin
3. Finally, 3,610 bivalirudin-assigned patients were randomized a third time to either post-procedural prolonged bivalirudin infusion or no infusion 1

This nested approach with multiple stratification levels ensured balance across important prognostic factors at each decision point.

Important Considerations for Stratified Blocked Randomization

Benefits:

• Ensures balance between treatment groups for important baseline characteristics
• Particularly valuable in smaller trials where simple randomization might result in imbalances
• Improves trial credibility by avoiding chance imbalances in prognostic factors 1

Limitations:

• Using too many stratification factors can be problematic:
  • Power can drop below 80% when four or more stratification factors are used 2
  • Complexity increases substantially with each additional factor
  • May be impossible to implement with many stratifying factors 1
• Block size considerations:
  • Small blocks improve balance but may reduce unpredictability
  • If block size becomes known, allocation might be predictable
  • Using larger block sizes or randomly varying block sizes can reduce this risk 1

Implementation Example

An example of proper implementation is described in the CONSORT guidelines: "Randomization sequence was created using Stata 9.0 statistical software and was stratified by center with a 1:1 allocation using random block sizes of 2,4, and 6." 1

Potential Pitfalls to Avoid

1. Overstratification: Using too many stratification factors can reduce statistical power, especially in phase 2 trials with smaller sample sizes 2

2. Predictability: When allocations are balanced by clinician or center, prediction rates can exceed 60-80%, potentially introducing selection bias 3

3. Block size disclosure: Knowledge of block size can compromise allocation concealment; using variable block sizes helps mitigate this risk 1

4. Overall imbalance: While stratified blocked randomization creates balance within strata, imbalance for the total trial may still occur, especially if the total sample size isn't appreciably larger than the product of the number of strata and block size 4

Stratified blocked randomization is a powerful technique for ensuring balance in clinical trials when properly implemented, with the MATRIX trial serving as an exemplary model of this approach.