Developing anautomatic scale to delirium detection based on the electronic medical record system / 医学研究生学报

ABSTRACT

ObjectiveAt present, there are many bedside tools for delirium, but these manual tools are time-consuming and poor feasible. The aim of this study was to establish a delirium screening scale, automatically extracting keywords from electronic medical records (EMR).MethodsWe selected electronic medical records of 779 elderly hospitalized patients in West China Hospital of Sichuan University from 2015 to 2017. Then, R software was used to automatically extract keywords to form a database undercritical ration, correlation coefficient and different analysis methods. Finally, the Delphi method and Analytic Hierarchy Process Weight were carried out to the construct weight coefficient, so as to form the formal scale.ResultsIn the study, we developed a formal scale consisting of 59 items and 11 dimensions. The score of the scale ranged from 0 to 53.4, with a mean value of 6.64, skewness of 2.6 and kurtosis of 8.2.ConclusionThe delirium screening scale based on the EMR can improve the recognition rate of delirium through intelligent and automatic warning, so as to early diagnosis and timely intervention of delirium.