Article Text
Statistics from Altmetric.com
- Chart review methodologies
- Health services research
- Healthcare quality improvement
- Patient safety
- Performance measures
Schildmeijer et al's article1 leaves the impression that the inter-rater reliability of the Global Trigger Tool (GTT) is substantially lower than it really is. It compares the results of five teams who have examined the same 50 patient admissions using GTT to identify harmful adverse events (AEs). The abstract does not mention that one of the teams, with the highest harm rate, also used a different definition for harm than the GTT. This team found 99.7 AEs/1000 patient days. This is more than three times the results of the other four teams who used the GTT definition. They found AEs per 1000 patient days ranging from 27.2 to 33.2. The harm rates of the four teams had a combined weighted κ value estimate of 0.62, ranging from 0.38 to 0.81, which is a reasonably good inter reliability rate.
Although the intention for the GTT is not to compare results between hospitals, research has concluded that the GTT's inter-rater reliability rate is adequate for estimating regional and national rates of patient harm.2 This could have been undermined if the results in this abstract had been true.
This article would have been improved by excluding the team with the different definition of harm. The results would still have been interesting.
Footnotes
Competing interests None.
Provenance and peer review Not commissioned; internally peer reviewed.
Linked Articles
- PostScript