An invention that predicts which elderly people are most likely to fall, developed by the Catharina Hospital, has won prizes at a congress on elderly care in Den Bosch.
The hospital has been researching how to predict falls in patients for years. As it happens, preventing falls has quite an impact on the workload of nursing staff.
Here’s the thing: almost every patient in the geriatrics department currently wears a motion sensor. This sounds an alarm if someone falls. But the system also often false alarms if, for example, a blanket moves or the patient drops his handkerchief.
Alarm tired
As a result, nurses hear that alarm so often that at some point they become ‘alarm tired’ and react less alertly. Moreover, it takes a lot of time because they do have to check what is going on with a patient after such an alarm goes off.
Smart alarm
To reduce the number of false alarms, the hospital used AI to predict which people are at risk and which are not. The smart prediction model, developed with Catharina Hospital’s AI Expertise Centre, is based on several words that nurses write down in the free text of patient records, such as ‘discharge’, ‘normal’, and ‘spontaneous’.
Based partly on those words, the artificial intelligence can estimate the likelihood of a fall. The patient’s age and the use of different types of medication have also been added to the model.
The invention, on which TU/e also collaborated, received an incentive prize of 2,500 euros. This is intended to further develop the system.
Source: Studio040
For Eindhoven news: Chaitali Sengupta
