Randomized trials stopped early for benefit: a systematic review.

Area of Science:

• Clinical Trials
• Medical Research Methodology
• Evidence-Based Medicine

Background:

• Randomized clinical trials (RCTs) stopped early for apparent benefit gain significant attention and influence clinical practice.
• Key aspects like the prevalence, effect magnitude, and reporting of early stopping decisions in RCTs remain unclear.

Purpose of the Study:

• To investigate the occurrence and evaluate the reporting quality of RCTs that were terminated prematurely due to observed treatment benefits.

Main Methods:

• A systematic review of literature databases (MEDLINE, EMBASE, Current Contents) up to November 2004 identified RCTs stopped early for benefit.
• Data abstraction by 12 independent reviewers covered trial characteristics, funding, stopping reasons, treatment effects, and reporting of interim analyses and stopping rules.

Main Results:

• 143 RCTs stopped early for benefit were identified, with most published in high-impact journals, often industry-funded drug trials in cardiology, oncology, and HIV/AIDS.
• The proportion of such RCTs in high-impact journals rose from 0.5% (1990-1994) to 1.2% (2000-2004).
• Reporting was poor, with 94% of trials failing to adequately report sample size, interim analysis details, stopping rules, or adjusted analyses. Trials with fewer events showed larger, potentially implausible, treatment effects.

Conclusions:

• RCTs stopped early for benefit are becoming more frequent.
• Inadequate reporting on early stopping decisions is common, particularly in trials with few events, which may show exaggerated treatment effects.
• Clinicians should approach results from early-stopped RCTs with caution and critical evaluation.