A review found small variable blocking schemes may not protect against selection bias in randomized controlled trials

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Research Integrity

Background:

• Randomized controlled trials (RCTs) are crucial for evidence-based medicine.
• Allocation concealment is vital for preventing bias in RCTs.
• Blocking and minimization are common randomization techniques used to ensure balance in treatment groups.

Purpose of the Study:

• To investigate if blocking strategies in RCTs increase baseline age heterogeneity, an indicator of potential subversion.
• To examine trends in the use of blocking over time.
• To assess the impact of different block sizes on age heterogeneity.

Main Methods:

• A hand search of four major medical journals (BMJ, JAMA, The Lancet, NEJM) was conducted for open RCTs published between 2001 and 2020.
• Meta-analyses were performed on trials using blocking, minimization, and simple randomization to evaluate baseline age heterogeneity.
• 179 open RCTs were identified and categorized by randomization method.

Main Results:

• Blocking was used in 58.1% of identified RCTs.
• Trials using blocking showed a baseline age heterogeneity of 24% (P=0.02).
• Trials with fixed block sizes of four exhibited 62% heterogeneity (P=0.001), while variable block sizes (including two) showed 40% (P=0.07). Simple randomization and minimization had 0% heterogeneity.
• Modern trials increasingly implement smaller block sizes.

Conclusions:

• Variable block sizes, particularly of size two, are linked to subversion and should be avoided.
• When stratification is necessary (e.g., by center), it should be combined with larger blocking schemes.
• Researchers should explore alternative methods to restrict randomization to ensure trial integrity.