Assessing keywords relevant to the severity of slip-and-fall injuries by text mining method on nursing records

Abstract

To investigate risk factors relevant to the severity of accidental slip-and-fall injuries, we assessed keywords by comparing the words extracted from nursing reports using a text mining method in groups that were classified according to the level of injury severity. A total of 462 cases that were reported in 2013 in our incident reporting system were classified into four groups: cases without injury, cases with injury requiring no treatment, cases with mild injury requiring minor treatment, and cases with severe injury requiring major treatment. Text mining was carried out using the package “RMeCab” that enables the tool for morphological analysis “MeCab” to be added to the software environment for statistical computing and graphics “R”. The results showed that there were substantial descriptions about excretion and physical activity in the nursing reports for cases with severe injury, and suggested that problems of excretion and poor physical activity were risk factors for these cases. On the other hand, there were fewer descriptions about symptoms such as pain and fever, as well as using a wheelchair, in the nursing reports for cases with severe injury. A defensive approach may enable risky actions to be avoided. These observations suggest that descriptions in nursing reports may provide clues to reduce future severe slip-and-fall injuries.