FRADE: Pervasive Platform for Fall Risk Assessment,

Prevention and Fall Detection

Joana Silva

, Nuno Cardoso

, Jorge Ribeiro

, Alberto Carvalho

, Mariana Pereira

Fernando Ricaldoni

, Carlos Resende

and Jo

ao Oliveira

Fraunhofer Portugal Research Center for Assistive Information and Communication Solutions,

Rua Alfredo Allen 455/461, 4200-135 Porto, Portugal

Keywords:

Fall Detection, Fall Risk Assessment, Fall Prevention, Otago Program, Inertial Sensors, Wearable Device.

Abstract:

The ageing of the global population has an impact on the elderly quality of life, as the reduced mobility and

balance contribute to the increasing of falls. Fall detection solutions can trigger emergency alerts and reduce

the negative effect of falls. Fall risk assessment strategies can help to early identify fall risk factors and tailor

strategies to revert those risk factors, by means of fall prevention exercises. However, most technological solu-

tions do not simultaneously address these three aspects of the falls management cycle. FRADE platform will

allow to pervasively detect falls using a wearable device, that can also be used to monitor fall risk assessment

tests and recommended individual exercises, that can be performed at home with a tablet and two wearables.

1 INTRODUCTION

The worldwide population aged over 65 is rapidly

growing and the consequences are simultaneously so-

cial, health-related and economic. The process of

ageing impacts mobility, muscle strength and balance

control which contributes to the increase of falls oc-

currence in this population. Currently, there are a va-

riety of solutions to address only speciﬁc stages of

the fall management lifecycle: assessing multiple fall

risk factors, detecting falls automatically, and provid-

ing strategies for falls prevention that focus on atten-

uating speciﬁc fall risk factors (Rajagopalan et al.,

2017). However, most technological solutions do not

allow to close the falls management loop by simul-

taneously addressing fall detection (FD), fall risk as-

sessment (FRA) and prevention (FP).

Among the elderly population, falls are one of the

major causes of death and injury. More than 30% of

people over 65 falls each year and the prevalence in-

https://orcid.org/0000-0002-6214-5868

https://orcid.org/0000-0001-5736-7995

https://orcid.org/0000-0001-5532-2574

https://orcid.org/0000-0002-1458-687X

https://orcid.org/0000-0002-3278-1230

https://orcid.org/0000-0003-2836-8782

https://orcid.org/0000-0002-1834-0420

https://orcid.org/0000-0003-3667-4958

creases for people above 80 (Bergen et al., 2016). Be-

sides social and personal consequences, falls also play

an important role in healthcare costs. Centers for Dis-

ease Control and Prevention estimate approximately

645 fall-related emergency visits for every 10000 el-

derly. In the USA in 2015, the direct costs for fatal

and nonfatal fall injuries were 637,5 million and 31,3

billion dollars, respectively (Burns et al., 2016). The

Personal Emergency Response Systems (PERS) mar-

ket is valued at 6245 million in 2018 and expected

to reach 9452 million dollars by 2025 (Bergen et al.,

2019).

Even a minor fall can severely affect the physi-

cal and mental health of an elder and increase the

fear of falling again. Thus, the elderly quality of

life and of their caregivers can be severally affected.

For community-dwelling elderly, the problem affects

mostly the elder and his family. However, given the

high demand for specialized care due to aging and

falls, most of the older adults need to be institution-

alized, and thus the problems also affect daycare cen-

ters, retirement homes, nursing homes, and healthcare

facilities. Fall detection solutions can trigger emer-

gency alerts and reduce the negative effect of falls.

FRA strategies can help to early identify fall risk and

tailor strategies to revert those risk factors.

Currently, there are a variety of solutions to ad-

dress only speciﬁc stages of the fall management cy-

cle: assessing multiple fall risk factors, detecting falls

320

Silva, J., Cardoso, N., Ribeiro, J., Carvalho, A., Pereira, M., Ricaldoni, F., Resende, C. and Oliveira, J.

FRADE: Pervasive Platform for Fall Risk Assessment, Prevention and Fall Detection.

DOI: 10.5220/0010328103200326

In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 4: BIOSIGNALS, pages 320-326

ISBN: 978-989-758-490-9

automatically, and providing strategies for falls pre-

vention that focus on attenuating speciﬁc fall risk fac-

tors. However, current solutions fail to provide a com-

plete system for fall detection, fall risk assessment,

and fall prevention, which would enable value-based

healthcare. Most commercial solutions available in

the market target only one speciﬁc functionality.

Personal emergency response systems (PERS), or

medical alert services, provide prompt access to emer-

gency help for elderly who fall. PERS are usually

in the form of a pendant or wrist band, and enable

the users to press a button, to transmits a signal to

a call center or caregiver, who contacts emergency

help. Most PERS are not automatic, instead, the user

must press a button. PERS are stigmatizing since they

should be attached to a speciﬁc on-body position, and

most of them do not allow the user to select differ-

ent levels of sensitivity for FD. The main competitors

are Philips Lifeline

, Medical Guardian

, AlertOne

Services

, Tunstall

and Life Alert Emergency Re-

sponse

. These companies provide PERS with an

emergency button, but few detect falls automatically.

FRA solutions fail to address multiple risk fac-

tors and only provide an estimation of FRA based

on a limited number of risk factors evaluated in spe-

ciﬁc periods. There are several commercial solu-

tions for FRA that can use wearable devices or pres-

sure sensors to evaluate standard tests, such as QTUG

from Kinesis

, Go from GaitUp

and MoveTest from

McRoberts

FP solutions do not combine the assessment of

fall risk factors with personalization of FP strate-

gies. There are a variety of solutions that provide

technological support to gamify physiotherapy exer-

cises, either with wearable devices, pressure sensors

or Kinect, such as Sword Health

, Biosensics

and

SilverFit

Considering that there are no existing solutions

that close the falls management loop, FRADE will

be unique in that perspective. Such level of integra-

tion comes with beneﬁts in the acquisition and main-

https://www.lifeline.philips.com/medical-alert-systems/

fall-detection.html

https://www.medicalguardian.com/

https://www.alert-1.com/

https://www.tunstall.co.uk/resources/product-datasheets/

vibby-fall-detector/

http://www.lifealert.com/

https://www.kinesis.ie/qtug/

https://clinical.gaitup.com/gait-up-go/

https://www.mcroberts.nl/products/movetest/

https://swordhealth.com/

https://biosensics.com/

https://silverﬁt.com/en/

tenance costs, which can be cheaper than acquiring

individual solutions. It also brings a broader view of

the elder status, where all the information is available

in a single platform, which should improve the efﬁ-

ciency of healthcare for the elderly population. The

FD system is a standalone wearable device that sup-

ports multiple body positions. The FRA is composed

of a multifactorial set of questionnaires and instru-

mented functional tests, which allow recommending

individual exercises.

2 FRADE PLATFORM

FRADE is composed of a bundle of components

(wearable sensor, desktop application, Android appli-

cation, data storage, and data visualization web in-

terface) to perform FD, FRA, and FP. The elder will

use the wearable sensor to monitor falls automatically,

based on movement analysis. It sends an alarm to the

backend server and an SMS to a caregiver whenever

a fall is detected. The wearable device will also be

used to monitor the elder’s movements and communi-

cate with the desktop and Android application while

he performs FRA tests and FP exergames. All the

monitored data is stored in the data storage and can

be accessed through a web interface, by a caregiver or

an healthcare provider.

2.1 Fall Detection Wearable Device

Kallisto is an hardware module with sensing, com-

munication and power management capabilities. It

includes a set of inertial and ambient sensors, Blue-

tooth Low Energy (BLE) and NFC radios, as well as

USB and Qi inductive charging systems. It is supplied

as a module and can be used as standalone or as part

of a more complex device. To build the wearable de-

vice, the COMM mainboard was used that comprises

Kallisto as its core block and complements the sys-

tem by providing extra sensing and communication

functionalities. It includes GNSS, Narrow Band-IOT

(NB-IoT) and RFID radios

The wearable device casing was conceived from a

user-centered perspective in a way of making it easy

to use whilst accommodating the HW platform. The

design approach was to make the device almost part

of the user clothes, as a simple shirt button for exam-

ple, as represented in Figure 2. When attached to the

clothes the device will provide automatic fall detec-

tion, and occasionally, when attached on an accessory

to the users’ thighs or feet, it will provide data stream

for movement analysis.

https://demo.sensry.net/

FRADE: Pervasive Platform for Fall Risk Assessment, Prevention and Fall Detection

321

Figure 1: Hardware and software components of the FRADE platform.

Figure 2: Wearable device Kallisto placed in the clothes as

a button.

This wearable device integrates an automatic fall

detection algorithm and offers features such as alarm

sending, through the wearable device, which can be

used discreetly by the user in the belt, pocket or in the

chest. The fall detection algorithm is embedded in the

ﬁrmware of the wearable device. This algorithm anal-

yses the inertial sensors data and detect whenever a

fall incident occurs, as previously described in (Alves

et al., 2019). In case of detection of a fall event, a noti-

ﬁcation will be sent to a data storage that in turn sends

an SMS and e-mail to a set of predeﬁned emergency

contacts. The communication between the device and

the data storage is made through NB-IoT. The de-

vice also triggers an audible alarm to attract the at-

tention of people nearby, featuring an emergency but-

ton that allows the cancellation of false alarms. The

device works independently of a smartphone or other

resource and only needs to be charged via an induc-

tion charger, which is made available with the device.

Figure 3: Elderly performing the fall risk screening using

the Clinical App.

2.2 Fall Risk Assessment Clinical

Application

The Clinical app is a Windows desktop application

that was designed for the healthcare providers to al-

low the creation of questionnaires and functional tests

to screen the elderly regarding the fall risk. The exer-

cise monitoring algorithms are based on the analysis

of movement and balance (Martins et al., 2018) using

sensor data from two wearable devices and a Phys-

ioSensing pressure platform, as can be seen in Figure

3. This application also allows healthcare providers

to prescribe exercises for the home application. The

application was developed in Unity 2019.3.3 and re-

quires a laptop with BLE to connect to the wearable

devices and a USB port to connect to the pressure

platform

The Clinical App extracts information about the

movement and balance of the user while he is per-

https://www.physiosensing.net/

BIOSIGNALS 2021 - 14th International Conference on Bio-inspired Systems and Signal Processing

322

forming the functional tests Sit-to-Stand Test (Cho

et al., 2012), Timed-Up and Go Test (Kojima et al.,

2015), and 4 Stage Balance Test (Thomas et al.,

2014), as shown in the main screen of Figure 4. The

application also allows the insertion of the personal

proﬁle of each participant and the answer to several

questionnaires for the assessment of several risk fac-

tors for falls, such as activities of daily living (Ara

ujo

et al., 2008), fear of falling (Figueiredo and Santos,

2017), and home hazards (Watson et al., 2014) ques-

tionnaires. In addition, the application allows the pre-

scription of Otago Exercise Program (OEP) exercises

and their schedule for each participant through a ded-

icated exercise prescription interface. This exercise

prescription will be sent automatically to the Home

application.

Figure 4: Screening exercises in the Clinical App.

2.3 Fall Prevention Home Application

The Home App is an Android tablet-based application

for the elderly to perform fall prevention exercises

based on the Otago Programme, an evidence-based

program for reducing falls (Campbell et al., 1997).

The application provides interfaces with instructions

on how to perform the exercises, as in Figure 5, and

interfaces with feedback during the exercises, which

are based on the movement analysis algorithms pro-

vided by the inertial sensors of a wearable device, as

previously described in (Silva et al., 2018).

The interactive Home application, based on the

Otago exercise program, aims to improve physical

functionality. The application features 8 exercises

from the program, i.e., knee ﬂexion, knee extension,

hip abduction, knee bending, toe raises, calf raises, sit

to stand, and one leg standing, that are static, easy and

well accepted exercises with interactive feedback on

the execution of the movement, as depicted in Figure

6. The application could also be a source of moti-

vation for the participants who perform the exercises

at home. The use of the application requires an An-

droid Tablet, a support for the tablet, two wearable

inertial devices, and the respective chargers. The ap-

plication only requires internet connectivity once at

the app setup and occasionally to synchronize the ex-

ercises metrics with the cloud (on follow up appoint-

ments with the clinician). The application is compat-

ible with any tablet with Android version above An-

droid 4.4 (API 19). The user will be guided through

a weekly exercise plan, which he/she will be able to

select through the tablet interface where the instruc-

tions for executing each exercise will be presented, as

well as an interface with visual feedback during the

execution of each exercise.

Figure 5: Example of one interface of the Otago exercises

using the Home App.

Figure 6: Execution of one exercise using the Home App.

2.4 Data Storage

All the information collected through the aforemen-

tioned components is stored on Firebase Cloud Fire-

store. Cloud Firestore is a non-relational database that

enable developers to safely store and sync data across

multiple devices and applications. Firebase is certi-

ﬁed under major privacy and security standards and

follows the GDPR rules. All the data in Firebase is

also encrypted on the tablet side. Cloud Firestore is

also used to manage the authentication of users with

the home and clinical applications. Only registered

users can have access to the data, i.e., the healthcare

FRADE: Pervasive Platform for Fall Risk Assessment, Prevention and Fall Detection

323

providers or caregivers. Data from the clinical app,

home app, as well as from the wearable device are

stored on the database to be made available to the

other applications in the system (Fig. 1). For in-

stance, plans generated on the clinical app by health-

care providers are available for the elderly to follow

in the mobile app. Cloud Firestore stores the proﬁle

data, such as height or clinical condition, the answers

to the questionnaires for risk assessment, plans cre-

ated by the healthcare providers, and session and ex-

ercise data detailing the measurements of individual

exercises. Besides allowing developers to synchro-

nize data across applications, it enable us to monitor

the elderly during the trials, and to keep a detailed log

of participants measurements for further analysis of

the trials results.

2.5 Trial Monitor

Trial Monitor is a web application created with the

purpose of supporting researchers monitoring partici-

pants remotely during ﬁeld trials (Vasconcelos et al.,

2019). The tool enables researchers to create per-

sonalized data visualization dashboards with the data

generated by participants during the trials (Fig. 7).

The web application connects to the data storage (i.e.

Cloud Firestore) to retrieve users data, and generates

visualizations for health providers to follow the el-

derly remotely.

The platform enables nursers to understand how

participants are progressing during the trial, by pro-

viding a system for monitoring the results of the

prescribed exercises and how they evolve over time.

Healthcare providers can use the platform to visual-

ize and analyse the results of the Otago exercises, or

the fall risk assessment questionnaires. The platform

displays the number of sessions completed and the re-

sults from individual exercises (e.g. number of repe-

titions, range of motion). Moreover, the platform al-

lows researchers to easily export data (i.e. CSV ﬁle)

from the system for further analysis after the end of

the trial.

3 VALIDATION TRIALS

We have produced 20 units of system to be used dur-

ing the validation trials. The participants are being

recruited. The trials to be conducted will allow to

evaluate the performance of the FD and to assess the

fall risk of the elder population in the North region of

Portugal. The trials will also provide insights on the

effectiveness of the FP exercises using this technolog-

ical platform.

The study will bring the objective measure of fall

risk factors and movement-based metrics extracted

during the Otago exercises. The technological plat-

form allows the centralization of all relevant variables

in a uniﬁed and secured database, that is accessed

through a web portal, that will be available for the

healthcare providers that will supervise the study. Be-

sides the aforementioned variables retrieved by the

Clinical app, i.e., personal proﬁle, medical conditions,

medication, answer to the questionnaires, and scores

of the three functional tests, the Home app will allow

to measure range of motion, number of repetitions and

duration of ascending and descending movements for

eight exercises of the OEP, i.e., knee ﬂexion, knee ex-

tension, hip abduction, knee bending, toe raises, calf

raises, sit to stand, and one leg standing.

4 CONCLUSION

The use of technological devices for early detection

and prevention of falls is a key strategy to minimize

the consequences of this event, namely the perma-

nence on the ground for long periods after the fall and

hypothermia, which often leads to the death of the el-

derly. This reality is veriﬁed not only in the elder’s

home but also in nursing homes. To respond to this,

the scientiﬁc production in this area has been prolifer-

ating, both in the technological sector and in the aca-

demic and clinical context. The synergies created by

the joint work of these areas, allow this project to con-

tribute to the development of more effective strategies

for prevention, reduction of falls and their physical,

psychological and economic consequences.

In technological terms, we consider that the tech-

nological literacy of each participant could impact the

way each one will use the platform and we are aware

that some limitations may raise during the course of

the eight weeks of intervention, considering that in

some of the sessions the elderly will perform the ex-

ercises alone at home. To try to overcome these

limitations, we foresee frequent contacts between the

healthcare providers and the elderly, as well as remote

guidance whenever possible.

At the end of the validation study, we will have

a system for fall management composed of a wear-

able device and a bundle of applications. The system

can be used to i) pervasively detect falls and assess

fall risk factors in daily life; ii) monitor movements

during fall risk assessment tests to estimate a risk of

falling using a desktop application, with the supervi-

sion of a healthcare provider; iii) monitor movements

during the execution of fall prevention exercises us-

ing by a tablet application. The trials conducted will

BIOSIGNALS 2021 - 14th International Conference on Bio-inspired Systems and Signal Processing

324

Figure 7: Trial monitor web platform displaying the metrics of the Otago exercises.

allow us to evaluate the performance of the FD, and

to screen part of the institutionalized population in the

North region of Portugal. The FD, FRA will provide

insights on the effectiveness of the fall prevention ex-

ercises.

This validation study aims to evaluate the effec-

tiveness of the technological solution for detecting

falls and the identiﬁcation of the risk of falling in the

elderly at their homes, as well as validating a tech-

nological solution for fall prevention. The primary

economic buyer will be companies that sell healthcare

equipment, mainly to nursing homes and daycare cen-

ters. The secondary economic buyer will be the nurs-

ing homes and healthcare institutions and the tertiary

economic buyer will be the ﬁnal consumer, the elderly

since the solution can also ﬁt a home scenario.

The main obstacles for our solution to reach the

market are mainly related to the user and market ac-

ceptance. The technological solutions that will be in-

tegrated into this system were previously developed

based on user-centered design and all the prototypes

were tested with the elderly in usability tests using a

network of end-users and institutions. The fact that

the same wearable can be used for FD, FRA and FP,

will require less devices to be bought and maintained.

The same applies to the monthly fees normally associ-

ated with these services, there will be a single fee for

a system that covers all components, instead of a fee

for each component. We can also mitigate the risks

of poor market acceptance by providing only parts of

the system.

ACKNOWLEDGEMENTS

The authors would like to thank all the participants

involved in the usability tests, and the healthcare

providers that collaborated in the deﬁnition of the re-

quirements for the technological solution. Authors

would like to thank the ﬁnancial support of Project

RIS-1001-8273 of EIT Health RIS Innovation Call

2020; EIT Health is supported by the EIT, a body of

the European Union.

REFERENCES

Alves, J., Silva, J., Grifo, E., Resende, C., and Sousa, I.

(2019). Wearable embedded intelligence for detection

of falls independently of on-body location. Sensors,

19(11).

Ara

ujo, F., Ribeiro, J., Oliveira, A., Pinto, C., and Martins,

T. (2008). Validac¸

ao da escala de Lawton & Brody

numa amostra de idosos n

ao institucionalizados. Actas

do 7

Congresso Nacional de Psicologia da Sa

ude,

pages 217–220.

Bergen, G., Stevens, M., and Burns, E. (2016). Falls and fall

injuries among adults aged > 65 years - United States,

2014. MMWR. Morbidity and mortality weekly report,

65:993–998.

Bergen, G., Stevens, M., and Burns, E. (2019). Global

personal emergency response systems (PERS) market:

Global market size, trends, competitive, historical &

forecast analysis, 2019-2025. BMRC 854,, Medical

Devices and Consumables:202.

Burns, E. R., Stevens, J., and Lee, R. (2016). The direct

costs of fatal and non-fatal falls among older adults -

United States. Journal of safety research, 58:99–103.

Campbell, A. et al. (1997). Randomised controlled trial of

a general practice programme of home based exercise

to prevent falls in elderly women. BMJ (Clinical re-

search ed.), 315:1065–9.

Cho, k. H., Bok, S., Kim, Y.-J., and Hwang, S. (2012). Ef-

fect of lower limb strength on falls and balance of the

elderly. Annals of rehabilitation medicine, 36:386–93.

Figueiredo, D. and Santos, S. (2017). Cross-cultural val-

idation of the falls efﬁcacy scale-international (FES-

FRADE: Pervasive Platform for Fall Risk Assessment, Prevention and Fall Detection

325

I) in portuguese community-dwelling older adults.

Archives of gerontology and geriatrics, 68:168–173.

Kojima, G., Masud, T., Kendrick, D., Morris, R., Gawler,

S., Treml, J., and Iliffe, S. (2015). Does the timed

up and go test predict future falls among british

community-dwelling older people? prospective co-

hort study nested within a randomised controlled trial.

BMC Geriatrics, 15.

Martins, A., Moreira, J., Silva, C., Silva, J., Tonelo, C., Bal-

tazar, D., Rocha, C., Pereira, T., and Sousa, I. (2018).

Multifactorial screening tool for determining fall risk

in community-dwelling adults aged 50 years or over

(fallsensing): Protocol for a prospective study. JMIR

Research Protocols, 7.

Rajagopalan, R., Litvan, I., and Jung, T.-P. (2017). Fall

prediction and prevention systems: Recent trends,

challenges, and future research directions. Sensors,

17(11).

Silva, J., Moreira, D., Madureira, J., Pereira, E., Dias, A.,

and Sousa, I. (2018). A technological solution for sup-

porting fall prevention exercises at the physiotherapy

clinic. 2018 IEEE International Symposium on Medi-

cal Measurements and Applications (MeMeA).

Thomas, J. C., Odonkor, C., Grifﬁth, L., Holt, N., Percac-

Lima, S., Leveille, S., Ni, P., Latham, N. K., Jette,

A. M., and Bean, J. F. (2014). Reconceptualizing bal-

ance: attributes associated with balance performance.

Experimental Gerontology, 57:218 – 223.

Vasconcelos, A., Lopes, I., Ribeiro, J., and Barros, A. C.

(2019). Challenges and lessons learned from imple-

menting longitudinal studies for self-care technology

assessment. LDC 2019: Workshop on Longitudi-

nal Mobile, Wearable and Ubiquitous Data Collection

from Human Subject Studies, pages 893–898.

Watson, M., Benford, P., Coupland, C., Clacy, R., Hind-

march, P., Majsak-Newman, G., Deave, T., and

Kendrick, D. (2014). Validation of a home safety

questionnaire used in a series of case-control studies.

Injury Prevention: Journal of the International Soci-

ety for Child and Adolescent Injury Prevention, 20.

BIOSIGNALS 2021 - 14th International Conference on Bio-inspired Systems and Signal Processing

326