Fujifilm and Juntendo Hospital in Japan have developed an AI that can accurately predict the risk of falls among outpatients using hospital data.
Their research team collated over 500 features associated with falls, including age and prescription history, from hospital data collected in Fujifilm’s CITA Clinical Finder. The data platform is used to centrally manage data across various hospital departments, including EMRs, radiology, and endoscopy.
An AI model was then trained on these features to predict an individual’s fall risk. The technology displays predicted fall risk in percentage and those features that are potential risk factors.
The researchers later tested the effectiveness of their AI on data from approximately 70,000 outpatients at Juntendo Hospital. Based on findings from their latest study, the AI scored 96% accuracy in predicting and generating fall risks.
According to a media release, Fujifilm and Juntendo will continue testing their AI and seeking its early clinical application.
WHY IT MATTERS
Hospitals in Japan are said to be challenged in identifying outpatients susceptible to falls given their limited time in their facilities. They also realised that there are more outpatients to assess for fall risks than inpatients. An efficient and accurate solution to prevent falls in medical settings has been increasingly raised and demanded.
MARKET SNAPSHOT
There are existing studies in Japan that have also utilised AI to help prevent falls in hospitals and nursing homes. Fujitsu and Wakayama Medical University, for one, had worked on trials of a combined sensor and AI technology to detect falls while preserving patient privacy.
Meanwhile, Fujifilm’s latest AI innovation adds to its growing elderly care technology portfolio. In 2022, it revealed its AI for predicting Alzheimer’s disease progression among patients with mild cognitive impairment.