Format

Send to

Choose Destination
Int J Qual Health Care. 2016 Apr;28(2):175-82. doi: 10.1093/intqhc/mzv122. Epub 2016 Feb 6.

Development and evaluation of an automated fall risk assessment system.

Author information

1
Seoul St Mary's Hospital, The Catholic University of Korea, Seoul, Korea.
2
College of Nursing, The Catholic University of Korea, Seoul, Korea.

Abstract

BACKGROUND AND OBJECTIVE:

Fall risk assessment is the first step toward prevention, and a risk assessment tool with high validity should be used. This study aimed to develop and validate an automated fall risk assessment system (Auto-FallRAS) to assess fall risks based on electronic medical records (EMRs) without additional data collected or entered by nurses.

METHODS:

This study was conducted in a 1335-bed university hospital in Seoul, South Korea. The Auto-FallRAS was developed using 4211 fall-related clinical data extracted from EMRs. Participants included fall patients and non-fall patients (868 and 3472 for the development study; 752 and 3008 for the validation study; and 58 and 232 for validation after clinical application, respectively). The system was evaluated for predictive validity and concurrent validity.

RESULTS:

The final 10 predictors were included in the logistic regression model for the risk-scoring algorithm. The results of the Auto-FallRAS were shown as high/moderate/low risk on the EMR screen. The predictive validity analyzed after clinical application of the Auto-FallRAS was as follows: sensitivity = 0.95, NPV = 0.97 and Youden index = 0.44. The validity of the Morse Fall Scale assessed by nurses was as follows: sensitivity = 0.68, NPV = 0.88 and Youden index = 0.28.

CONCLUSION:

This study found that the Auto-FallRAS results were better than were the nurses' predictions. The advantage of the Auto-FallRAS is that it automatically analyzes information and shows patients' fall risk assessment results without requiring additional time from nurses.

KEYWORDS:

accidental falls; prevention; risk assessment; validity

PMID:
26851379
DOI:
10.1093/intqhc/mzv122
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Silverchair Information Systems
Loading ...
Support Center