Re-analysis of clinical trial data can change conclusions, say Stanford researchers