The researchers analyzed transfusion-related near-miss data and found that near misses occur more frequently than adverse events causing harm. They conclude that near-miss data are a useful tool for identifying human and system failures and targeting improvements.