The original version of GEMS, along with the additional error categorisations proposed in the aforementioned study are defined and illustrated in Table 1.> Whilst the classification of errors occurring in acute care contexts according to the amplified version of GEMS is of academic interest, it is of limited value in developing educational strategies aimed at reducing error.
5 In clinical practice, this will be particularly important near the threshold for diagnosis, where most errors in categorisation will occur if ambulatory monitoring is not used.
However, MoST provides for the possibility of a human and/or SNOMED CT categorisation error.
We also found a slight negative correlation between individuals' categorisation error scores (given as the number of questions answered incorrectly on the categorisation test) and their false memory susceptibility during recognition (given as the number of critical non-presented words recognised); however this trend was not significant (r s=-0.202 s=-0.202p=0.219, Figure 3).
Finally, the numbers of recall, recognition and categorisation errors were compared between males and females using Wilcoxon rank sum tests to look for an effect of gender.
Categorisation errors ranged from 7%to78%8% in different individuals (Mean 23%, SD 14%), showing that even though the test we had chosen was originally designed for pre-teens, the task was sufficiently challenging for the tested population to capture a large range of inter-individual variation.
Our result of a significant negative correlation between individuals' errors on a categorisation test and their susceptibilities to semantic type false memories during free recall demonstrates that false memories, to some extent, might be a by-product of our ability to learn rules, categories and concepts.
Three main topics surfaced: 1. Categorisation of errors, 2. Management of errors, and, 3. Ethical approach towards error disclosure.
In order to limit interobserver bias, all grading and severity of errors identified in the main study were independently reassessed by two members of the study team (KS and SK), with discordant assessments collectively discussed with a third (TK) in order to agree a final categorisation for error type and severity.
