A multisite validity study of self-reported anesthesia outcomes.

The purpose of this study was to assess the validity of a multisite anesthesia voluntary adverse event reporting process. A data validation methodology was used through medical chart review on live records at 3 facilities (N = 600). The per-item aggregated error rate among all 42 data items was 0.3%: 0.1% for quality indicators, 1.3% for demographic/status variables, and 1.7% for administrative items. The per-patient error rate among all 42 data items was 6.3%: 3.0% for quality indicators, 1.7% for demographic/status variables, and 3.0% for administrative items. Trends such as better accuracy for more serious events continue, but observed error rates were lower than those found in previous surveys-an indication that, while further study is needed, nonpunitive voluntary reporting may reduce errors in anesthesia care.