The utility of routine incident reporting data is limited by the voluntary nature of these data, lack of reporting by physicians, and the lack of a standardized framework for interpreting the reports. This analysis of traditional paper-based incident reports from two hospitals sought to classify the underlying system and human factors that contributed to errors. The authors were able to determine at least one underlying factor in most reports, but found that incidents often were not described in enough detail to allow more specific classification. A prior study described an intervention to improve the number and quality of voluntarily filed incident reports.