Linking Fall-Risk Scores with Wearable Sensor Metrics in Older Adults from the DARE-FALLSPREDICT studies: Preliminary Findings

4680
Paola Di Florio1,M Colotti2,T Fiumana1,M Sicbaldi3,F Lamberti4,F Palmese2,F Gravina2,L Bettazzoni2,A Berardi1,L Palmerini3,P Palumbo3,G Ugolini5,6,A Rubboli1,7,P Cataleta2,S Tamberi5,8,A Buscaroli5,9,MG Piscaglia10,M Domenicali2,5,L Chiari3, A Silvani1
Dept of 1Biom and Neurom Sciences, 3Electrical, Electronic, and Information Eng, 4 Statistics, and 5 Med and Surg Sciences, University of Bologna; Units of 2 Internal Medicine 1,6 General Surgery, 7Cardiology, 8Oncology, 9Nephrology,AUSL Romagna,
Scientific Research
SP- Planned and ongoing trials
View DOI

Abstract

Falls are a leading cause of injury and disability in older adults, yet current predictors have limited accuracy. Continuous monitoring of activity, sleep, and heart rate (HR), combined with clinical assessments such as FRAT-up (Cattelani et al, JMIR, 2015) and the World Falls Guidelines (WFG) score (Montero-Odasso et al, Age Ageing, 2022), may improve personalized fall-risk identification.

DARE-FALLSPREDICT and DARE-FALLSPREDICT-GP are ongoing single-center studies recruiting older adults (> 65 y) in Italy, respectively at hospital discharge or from the general population. Participants wear a lower-back inertial unit and wrist accelerometer for one week; in DARE-FALLSPREDICT-GP, wrist photoplethysmography is also recorded for HR monitoring. Baseline and 6-month assessments include questionnaires (Short-FES-I, CES-D, PSQI, MMSE), lab walking test, and fall-risk classification via FRAT-up and WFG. Falls and related costs are tracked over 12 months via monthly phone calls. We performed an exploratory interim analysis of the first consecutive 331 participants (62.4% females, age 76.7 ± 6.9 y; N = 229 from DARE-FALLSPREDICT and 102 from DARE-FALLSPREDICT-GP). 

FRAT-up and WFG scores were strongly correlated in both cohorts (Spearman’s rho = 0.677; p < 0.001). 

In DARE-FALLSPREDICT, higher FRAT-up and WFG scores were significantly (all p < 0.001) associated with reduced stride length and walking speed, lower step count and MVPA, greater day-night rhythm disruption, and poorer subjective sleep quality.

In DARE-FALLSPREDICT GP, these results were largely replicated; moreover, higher FRAT-up scores significantly (all p < 0.05) correlated with lower HR variability during wakefulness, higher mean HR during sleep, and a more attenuated nocturnal HR dipping pattern. 

These preliminary findings suggest that metrics of real-life wake-sleep behaviour and cardiac autonomic control based on wearable sensor measurements carry information relevant to personalized fall-risk stratification in older subjects.

This research was funded by the Italian Complementary National Plan (PNC-1.1), DARE – DigitAl lifelong pREvention (PNC0000002; CUP B53C22006450001). 

This is a poster presented at a BGS conference. It constitutes research undertaken by the authors  and queries should be addressed to them. The BGS do not endorse or recommend any of the content. BGS do not make any warranties or guarantees in relation to the content.  If you wish to comment on the poster you will need to create a website account, sign in and submit for the authors to respond.