Test-Retest Reliability of an Experienced Global Trigger Tool Review Team

During a comprehensive patient safety program at a 550-bed regional hospital in the Capital Region of Denmark, we observed an unexpected and unexplained doubling of the median patient harm rate from 56 to 109 harms per 1000 patient days measured by the Institute for Healthcare Improvement Global Trigger Tool (GTT). Meanwhile, other measures of patient safety, including hospital standardized mortality ratio, were stable or improving. Moreover, the review team was very experienced and stable during this period. Thus, we hypothesized that the increase in harm rate was not a true reflection of increased risk of patient harm but the result of the team getting better at identifying harms during GTT reviews.

We examined the ability of the GTT review team to reproduce the rate of harm of two separate periods in the same hospital: period 1 (January-June 2010) and period 2 (October 2011-March 2012). For each period, we examined two samples: the original sample that was drawn and used for the ongoing monitoring of harm at the hospital during the safety campaign and a second that we drew and analyzed for this study.

We found increased harm rates both between review 1 and review 2 and between period 1 and period 2. The increase was solely in category E, minor temporary harm.

The very experienced GTT team could not reproduce harm rates found in earlier reviews. We conclude that GTT in its present form is not a reliable measure of harm rate over time.