Table 1 Characteristics of the included studies

Alqahtani et al., 2017 [55]

N = 29

mean 87,

sd 6

Frailty

Fried criteria

Upright balance, lower extremity muscle strength

Balance accelerometer; uni-axial load cell

Observational (Diagnostic accuracy study design)

Assessment

Validation of inexpensive measurements of strength and balance

Ambagtsheer et al., 2020 [56]

N = 592

median 88,

(IQR 9.0)

Frailty

Electronic Frailty Index

Clegg’s 36-items: activity limitation, chronic disease, falls, social isolation, cognition, mobility, polypharmacy, sleep quality and weight loss.

Machine Learning: K-Nearest Neighbours, Decision Tree, Support Vector Machines

Observational (Diagnostic accuracy study design)

Assessment

Identifying frailty from administrative records

Boumans et al., 2019 [62]

N = 42

mean 77.1,

sd 5.7

Frailty

Frailty Index

Time for completion of the questionnaires/robot–patient and nurse–patient interactions; percentage of robot–patient interactions completed autonomously

Social robot

Observational (Diagnostic accuracy study design)

Assessment

Effectiveness and acceptability of robot assistant assessment

Camicioli et al., 2015 [63]

N = 72

mean 74.97,

sd 1.44

Frailty

Fried criteria

Handwriting parameters: velocity, pressure, pauses.

Writing tablet with an instrumented pen for quantifying three-dimensional aspects of copying

Observational (Diagnostic accuracy study design)

Assessment

Studying relation between handwriting measures and frailty

Dupuy et al., 2017 [45]

N = 32

mean 81.63,

sd 1.57

Frailty

Fried criteria

Everyday activities; safety; social participation; interaction support; functional status; caregiver burden

Assisted-living platform: a set of wireless sensors and two touchscreen tablets

Interventional (Randomized controlled trial)

Intervention

Enhancing ADL authonomy, safety and sociality

Galan-Mercant & Cuesta-Vargas, 2013 [64]

N = 30

mean 76.98,

sd 4.85

Frailty

Fried criteria

Variability of the three-axes accelerations, angular velocity, and displacement of the trunk during the Si-St and St-Si transitions

iPhone 4 accelerometer

Observational (Diagnostic accuracy study design)

Assessment

Detecting frailty from Sit-to-Stand and Stand-to-Sit transition measures

Galan-Mercant & Cuesta-Vargas, 2014 [65]

N = 18

mean 79.95,

sd 5.37

Frailty

Fried criteria

Magnitude of accelerometry values

Trialxial gyroscope, accelerometer and a magnetometer in the iPhone 4 smartphone

Observational (Diagnostic accuracy study design)

Assessment

Improving the traditional assessment tools

Galan-Mercant & Cuesta-Vargas, 2015 [66]

N = 30

mean 76.98,

sd 4.85

Frailty

Fried criteria

ETUG test: sit-to-stand, gait go, turning, gait come, turn-to-stand-to-sit

Tri-axial gyroscope, an accelerometer and a magnetometer in the iPhone 4.

Observational (Diagnostic accuracy study design)

Assessment

Using intertial sensors embedded in a smartphone to measure kinematic variables in frail elderly

Garcia-Moreno et al., 2020 [67]

N = 79

mean 75

Frailty

Fried criteria

Fried criteria: “non-frail” 0 criteria, “pre-frail” 2 criteria, “frail” ≥3 criteria

Samsung Gear S3 wearable sensors; Microservices System Architecture; Frailty Status App; Cloud Server; Machine Learning algorithms

Observational (Diagnostic accuracy study design)

Assessment

To assess frailty status during the performance of IADLs

Gianaria et al., 2016 [68]

N = 30

mean 75.6,

sd 7.5

Frailty

Tillburg Frailty Indicator

Walking time/speed, covered distance, swing time, double support time, balance during walking, torso inclination angle

Microsoft Kinect sensors with skeleton tracking feature

Observational (Diagnostic accuracy study design)

Assessment

Detecting frailty from gait and posture features

Golkap et al., 2018 [69]

N = 36

mean 82,

sd 10

Frailty

Edmonton Frail scale

Arterial hemoglobin oxygen saturation; movement in a location; bed or chair occupancy

Home Monitoring Platform: sensors to acquire patient’s habits/clinical data; home gateway, a remote server to store patient data; clinician portal to view and manage patient data

Observational (Diagnostic accuracy study design)

Assessment

Studying an integrated care system to support independent living of frailty

Graňa et al., 2020 [76]

N = 645

mean 84.2

sd 6.76

Frailty

Fried criteria

Readmissions rates

Machine Learning - Linear discrimination analysis, Support vector machines, Multilayer perceptrons, K nearest neighbors, Random forests

Observational (Retrospective cohort study design - Predictive model)

Prediction

Studying frailty as a predictor of hospital readmissions

Hassler et al., 2019 [72]

N = 474

≥ 65

Frailty

Fried criteria

≥ 3 Fried criteria

Machine Learning – naïve Bayes classifier (NB), CART algorithm tree and bagging CART, C5.0 algorithm, Random Forest analysis, SVM, LDA

Observational (Retrospective cohort study design - Predictive model)

Prediction

Finding predictive factors for frailty

Held et al., 2017 [77]

N = 1686

≥ 70 years

Geriatric syndromes (Frailty; Cognitive impairment; Falls; Incontinence)

Fried criteria; Clinical assessment for cognition; The International Consultation of Incontinence Questionnaire; ≥ 2 falls in 12 months

Frequency of medication combinations

Machine Learning - Association Rule, Frequent-Set analysis

Observational (Cross-sectional study design)

Prevalence

Detect patterns of medication combinations according to geriatric syndrome status

Kubicki et al., 2014 [49]

N = 46

mean 81.87,

sd 5.9

Frailty

Fried criteria

Postural control, rapid arm movement

2D virtual reality-based program of motor telerehabilitation

Interventional (Randomized controlled trial)

Intervention

Enhancing postural control and balance

Kubicki, 2014 [57]

N = 37

mean 82.25,

sd 6.01

Frailty

Fried criteria

Gait speed; hand maximal velocity; timed up and go

Semi-immersive virtual reality with active motion-capture system based onvision technology

Observational (Diagnostic accuracy study design)

Assessment

improving detection of motor control efficiency

Lee et al., 2019 [51]

N = 65

≥ 65

Frailty

Custom questionnaire (based on Study of Osteoporotic Fractures index)

Health status, exercise, frailty, handgrip, body mass

Smart phone learning and balance/flexibility exercise

Interventional (Non-randomized trial)

Intervention

Reducing frailty

Martin-Lesende et al., 2016 [70]

N = 83

mean 81.3 (IQR: 77.1–85.4)

Multimorbidity

Presence of heart failure and/or chronic lung disease; ≥ 2 admission to hospital in the previous year

Mortality rate

Telemonitoring

Observational (Retrospective cohort study design)

Mortality rate

To assess mortality according to multimorbidity and telemonitoring status

Mateo-Abad et al., 2020 [52]

N = 856

mean 77.6,

sd 7.7

Multimorbidity

CIRS

Use of health care services, clinical control of the examined conditions, physical functional status, patient ́s satisfaction.

ICT-based platforms

Interventional (Non-randomized trial)

Intervention

Impact of an integrated care program on health resources use, clinical outcomes, and functional status

Merchant et al., 2020 [71]

N = 2.589

mean 73.1,

sd 6.5

Geriatric Syndromes (Frailty, Cognitive impairment, Sarcopenia, Anorexia of aging)

FRAIL questionnaire

Prevalence of frailty, sarcopenia, anorexia of aging

iPad mobile application for Rapid Geriatric Assessment

Observational (Cross-sectional study design)

Prevalence

Studying prevalence of frailty, sarcopenia and anorexia of aging

Orlandoni et al., 2016 [46]

N = 188

mean 85.47,

sd 7.03

Multimorbidity

CIRS

Incidence rates of complications, outpatient hospital visits, hospitalizations

Samsung Galaxy tablet for video consultation

Interventional (Randomized controlled trial)

Intervention

Enhancing home enteral nutrition management

Ozaki et al., 2017 [54]

N = 27

mean 73,

sd 6

Frailty

Fried criteria

Preferred and maximal gait speeds, tandem gait speeds, timed

up-and-go test, functional reach test, functional base of support, postural stability, muscle strength of the lower extremities, grip strength

Balance exercise assist robot

Interventional (Cross-over randomized controlled trial)

Intervention

Enhancing balance and walking

Paliokas et al., 2020 [58]

N = 80

mean 78.08,

sd 5.48

Frailty

Fried criteria

Errors related to the product types/number, payment errors, overall duration, selected item types/number, payment score, overall score

Non-immersive Virtual Reality Serious Game

Observational (Diagnostic accuracy study design)

Assessment

Detecting frailty from Virtual Reality Serious Game

Parvaneh et al., 2017 [59]

N = 120

mean 78,

sd 8

Frailty

Fried criteria

Daily postural transition

Unobtrusive shirt-embedded sensor with a three-axis accelerometer

Observational (Diagnostic accuracy study design)

Assessment

Identifying frailty from daily postural transitions

Peng et al., 2020 [73]

N = 86.133

mean 82.5

Frailty

Multimorbidity frailty index

All-cause mortality; unplanned hospitalizations; intensive care unit admissions.

Machine Learning - random forest method, Kaplan-Meier survival curve/log-rank test, Cox proportional hazard models

Observational (Retrospective cohort study design - Predictive model)

Predicition

Developing a machine learning–based multimorbidity frailty index

Persson et al., 2020 [53]

N = 94

mean 80,

sd 8

Comorbidity

CCI

Health-related quality of life; influence of healthcare dependency measures on HRQoL or vice versa

Telemonitoring system: digital pen technology supported by hospital-based home care

Interventional (Pre-post study design)

Intervention

Enhancing quality of life

Ritt et al., 2017 [60]

N = 123

mean 82.4,

sd 6.25

Frailty

Fried criteria; Frailty Index; Clinical Frailty Scale; Frailty index based on a comprehensive geriatric assessment

assessment

Spatio-temporal gait parameters

Electronic walkway; shoe-mounted inertial sensor-based mobile gait analysis system.

Observational (Diagnostic accuracy study design)

Assessment

Detecting frailty satus from gait analysis

Sargent et al., 2020 [74]

N = 1453

mean 79,

sd 0.54

Frailty

Fried criteria; MMSE score ≤ 23, TMT-A score ≥ 78, TMT-B score ≥ 106

Cognitive frailty: MMSE score ≤ 23, TMT-A score ≥ 78, TMT-B score ≥ 106; Physical frailty: ≥ 3 Fried criteria

Machine Learning - tree boosting approach model

Observational (Retrospective cohort study design - Predictive model)

Predicition

Studying biological mechanisms that relate physical frailty and cognitive impairment.

Schiltz et al., 2020 [75]

N = 6.617

≥ 65

Multimorbidity

Self-reported multimorbidity

30 day hospital readmission

Machine Learning - Random forest analysis, Classification and regression tree, Modified Poisson regression analysis, generalized estimating equation approach

Observational (Retrospective cohort study design - Predictive model)

Predicition

Studying IADL dependency as a predictor of hospital readmissions

Takahashi et al., 2012 [47]

N = 205

mean 80.3,

sd 8.2

Multimorbidity

CIRS

Hospitalization and emergency department visits

Telemonitoring device

Interventional (Randomized controlled trial)

Intervention

Reducing hospitalizations and emergency department visits

Tomita et al., 2007 [48]

N = 78

mean 73.8,

sd 4.7

Multimorbidity

CIRS

Functional status (ADL, IADL, MMSE, CHART)

Ambient assistive living: computer with internet access, X10-based smart home technology

Interventional (Randomized controlled trial)

Intervention

Studying feasibility and effectivenes of smart home technologies

Tsipouras et al., 2018 [61]

N = 73

mean 78.15,

sd 5.5

Frailty

Fried criteria

Number and durantion of transitions

Bluetooth localization system: sensorobluetooth beacons, smartphone andMaschine Learning for frailty level assessment: Naïve Bayes classifier, k-Nearest Neighbour, Neural Networks, Decision Trees algorithm, Random Forests

Observational (Diagnostic accuracy study design)

Assessment

Correlation between indoor activities and frailty status

Violán et al., 2019 [78]

N = 916.619

mean 75.4,

sd 7.4

Multimorbidity

> 1 of 60 chronic diseases

> 1of selected 60 chronic diseases; sociodemographics; number of invoiced drugs; use of health services

Machine Learning - fuzzy c-means clustering algorithm

Observational (Cross-sectional study design - Predictive model)

Predicition

Identifying multimorbidity patterns in the electronic health records

Volders et al., 2020 [50]

N = 585

mean 74.5,

sd 6.4

Multimorbidity

Self-reported multimorbidity

Physical activity

ActiGraph GT3X-BT accelerometer

Interventional (Randomized controlled trial)

Intervention

Studying the effect of a computer-tailored phisical activity intervention