Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Validity of a smartphone-based activity monitoring and fall detection application for older persons

 
: Vermeulen, J.

The gerontologist 54 (2014), Supplement 2, S.208
ISSN: 0016-9013
ISSN: 1758-5341
Gerontological Society of America (Annual Scientific Meeting) <67, 2014, Washington/DC>
Englisch
Abstract
Fraunhofer AICOS ()

Abstract
Background: The GoLivePhone® (GLP) is an easy-to-use smartphone-based application with various functionalities (e.g. activity monitoring, fall detection, medication management support, etc.) that can assist community-dwelling older persons in independent living. The aim of this study was to evaluate the validity of the activity monitoring and fall detection functionality of the GLP.
Methods: For the validation study of the activity monitoring functionality 15 participants (6 men, age 18-70) were included. Participants walked on a treadmill for 10 minutes starting at 3 km/h accelerating their walking speed with 1 km/h each minute, while wearing a GLP in their trouser pocket and an Actigraph GT3X and a PAM accelerometer on their hip. Pearson Correlations were calculated for intensity of activity registered per minute (in METs). For the validation study of the fall detection functionality, 8 healthy adults (4 men, age 18-24) were included. Participants simulated in total 400 falls, 200 falls with recovery, and 440 daily activities while wearing the GLP in their trouser pocket. Sensitivity and specificity were calculated.
Results: Pearson correlation between the GLP activity monitoring application and the Actigraph GT3X and PAM accelerometer were .77 and .84 respectively (p <0.01). Sensitivity and specificity of the fall detection application were .89 and .89 respectively.
Conclusion: Smartphone-based applications, such as the GLP, can support older persons in independent living by providing valid feedback to them regarding their level of physical activity. Furthermore, valid fall alarms can be generated but it is recommended that sensitivity and specificity are increased before implementation in practice.

: http://publica.fraunhofer.de/dokumente/N-332101.html