Analyzing voluntary medical incident reports.

Voluntary medical incident reports lacking consistency and accuracy impede the ultimate use of the reports for patient safety research. To improve this, two coders examined harm score usage in a voluntary medical incident reporting system where the harm scores were selected from a predefined list by different reporters. The two coders inter-rater agreement percent was 82%. The major categories and reviewed harm score jointly demonstrate that this process is critical and necessary in preparing the voluntary reports for further content and semantics analysis.