Identifying Outliers in Data from Patient Record.
Journal:
Studies in health technology and informatics
Published Date:
Jan 1, 2016
Abstract
It is important for health services to be able to identify potential outliers with minimal effort as part of their daily evaluation of care data from patient record. This study evaluates the suitability of three statistical methods for identifying nursing outliers. The results show that by using methods implemented in the nursing workload measurement system "LEP" with reference to real data, unusual LEP minute profiles (movement, nutrition and so on) can be identified with little effort and therefore seem promising for application to the health services' daily evaluation process. The lessons learned are used to create requirement criteria for the further development of software solutions. It is recommended that the methods for identifying outliers in the daily evaluation process should be standardized in order to increase the efficiency of secondary use of care data from patient record.