Components,design and effectiveness of digital physical rehabilitation interventions for older people: A systematic review

Abstract

Background

With the rapid advancement in digital technologies, the use of digital health applications is increasing day by day. Although a large number of digital applications have been developed for rehabilitation of older people, there has been no review of the evidence for effectiveness of these interventions.

Methods

The aim of our study was to review the evidence of digital rehabilitation interventions on outcomes including pain, function and quality of life in older people. We focused on digital interventions that are designed to improve and restore physical functioning. We searched six electronic bibliographic databases and included randomised controlled trials. Cochrane risk of bias tool and Cochrane’s Grading of Recommendations, Assessment, Development and Evaluation approach was used to evaluate the risk of bias and grade the evidence.

Results

Eight trials were included. The short-term effects of digital rehabilitation interventions on physical activity, quality of life, vertigo symptoms and falls are uncertain. Quality of trials was rated as very low to moderate evidence.

Conclusion

More research is needed to estimate effectiveness of these interventions.

Keywords

Older adults digital rehabilitation electronic health mobile health telehealth systematic review

Background

The world population is aging rapidly. By 2050, it is expected that the population over the age of 60 years will have increased by two billion people since the beginning of 21st century.¹ The number of people over 65 years old in the UK today is 18% of the total population.² By 2030, it is estimated that this number will rise to 21.8%² and there will be greater demands on health and social services³ as older people experience multiple health problems such as arthritis, diabetes, dementia and cancer,^2,3 and geriatric syndromes such as frailty, falls and immobility.^4,5

There is an increased need to develop strategies that promote healthy ageing and an emerging area of interest is digital health.⁶ Digital health interventions use digital-based technology to deliver accessible, usable, cost-effective and measurable interventions to improve health, healthcare services and quality of life of people or communities.⁷ These technologies include telehealth, electronic and mobile health applications (eHealth, mHealth), wearable devices and sensors, text messages, emails. They also potentially offer increased access to treatments from home, reducing the time, physical effort and travel costs of attending appointments.⁶

Technology-assisted healthcare systems generally focus on specific population groups, such as older people and patients with chronic diseases.⁸ Common interventions include self-monitoring and management of chronic diseases, patient education medication reviews, promotion of physical activity (PA) and exercise, healthy eating and cognitive behavioural therapies.⁹

In this review, we focus on digital physical rehabilitation interventions that are designed to improve and restore physical functioning in older people. There are different ways to deliver digital rehabilitation interventions and options include desktop computers, laptops, tablets, smartphones and their additional sensor systems.⁷

Taking part in regular PA and exercise is important for older people to maintain or improve the physical function needed to live independently and main good health.¹⁰ Home exercise programmes are often a fundamental part of successful rehabilitation for older people.¹¹ However, the adherence to home exercise programmes is often low¹¹ and older people are more likely to adhere to supervised exercise programmes.¹² Digital interventions have the potential to support older people by providing clinicians with a means of encouraging and motivating patients to undertake exercise and self-management strategies.^6,11

To date, there is only one systematic review of digital interventions in this area, and the study is limited to tele-health to support self-management in older people with chronic disease.¹³ However, there is no review available on the effectiveness of rehabilitation interventions delivered via digital routes in older people.

Objectives

The purpose of this review was to summarise the evidence on the benefits and harms of digital rehabilitation interventions on outcomes including pain, function and quality of life in older people.

Methods

Protocol and registration

The protocol is registered with the International Prospective Register of Systematic Reviews (PROSPERO). The registration number is CRD42018042471 and the protocol is available at https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=42471.¹⁴

Eligibility criteria

Studies: randomised controlled trials that evaluated the effectiveness of digital rehabilitation interventions specifically designed for older people. Qualitative studies that reported the patient experience of these interventions. Only peer-reviewed publications in English language were considered. Conference abstracts, dissertations and articles published in other languages were not included.

Population: men and women above 60 years of age with any chronic physical health condition including falls or mobility problems.

Intervention: self-directed digital rehabilitation interventions focusing on physical health delivered via Web-based, online platform, mobile applications were included. Digital rehabilitation interventions delivered by health professionals in clinical settings and those used solely for data collection or self-monitoring (e.g. physical activity data collected from a wearable sensor) were excluded.

Comparator: the control groups could be either receiving usual care, no treatment, a placebo (a digital rehabilitation intervention with limited features) or a non-digital rehabilitation intervention.

Outcomes: the main outcomes include pain, PA, function including mobility and fall-related outcomes, quality of life, adverse events and health resource use. Other outcomes of interest were psychological outcomes including self-efficacy, fear avoidance, anxiety and depression, and process outcomes such as intervention adherence rate and user perspectives.

Timing: outcomes were categorised into short-term (up to 3 months), medium-term (>3 to 11 months) and long-term (12 months and beyond).

Information sources

We searched the electronic databases of Ovid Medline (1946 to 6 November 2018), Ovid Embase (1974 to 6 November 2018), EBSCO CINAHL, the Cochrane Central Register of Controlled Trials (CENTRAL), and Physiotherapy Evidence Database (PEDro). We also searched the Journal of Medical Internet Research (JMIR) and its PubMed-indexed sister journals to identify additional relevant studies. We checked the World Health Organization (WHO) international clinical trials registry and clinicaltrial.gov trial databases for any ongoing trials.

Search strategy

We developed a search strategy (Supplemental material) in consultation with a health sciences librarian at the University of Oxford. The strategy was adapted for each bibliographic database. We used simple key words to search the JMIR and the trial databases.

Study selection

Two review authors (ET and CS) independently screened the titles and abstracts to identify potential studies from the database searches. ET and CS then screened the full-text publications of the potential studies based on the predefined eligibility criteria of the review. The reasons for excluding studies were documented. ET and CS consulted a third review author (EW) to resolve any disagreements during the study selection process.

Data collection process

The review authors ET and CS independently extracted the data from the eligible studies using the data extraction forms developed for the review. Data were collected on citation details, participants (age, gender, ethnicity and education level), inclusion-exclusion criteria, outcomes, interventions and the results. ET and CS cross-checked the data entries and resolved any discrepancies by consulting EW, the other review author. If date was insufficient for reporting from the included studies, ET attempted to contact the corresponding authors via email with up to three reminders.

Risk of bias in individual studies

The review authors ET and CS used the Cochrane risk of bias tool to evaluate the risk of bias of the included trials. The tool evaluates risk of bias arising from randomization (selection bias), effect of assignment to interventions (performance bias), measuring outcomes (detection bias) and missing data (reporting bias) and other sources of bias such as baseline variability between groups and small sample size. ET and CS summarised the overall risk of bias for each study as high, unclear and low as per the Cochrane guidelines.¹⁵ Studies are assessed at overall low risk of bias if all key domains were rated low risk; unclear risk if one or more key domains is rated unclear; and high risk if one or more key domains are rated at high risk. Any disagreements were resolved by consulting EW.

Summary measures

We proposed to calculate the treatment estimates as mean difference or standardised mean difference for continuous outcomes and risk ratio for dichotomous outcomes with 95% confidence interval at short, medium and long-term.

Synthesis of results

Before undertaking this review, it was unknown if it would be possible to carry out meta-analysis. Therefore, we specified statistical heterogeneity I² at 75% as cut-point to determine that meta-analysis was not approriate.¹⁵ The clinical heterogeneity was judged by the review authors based on the similarities and differences between participants, interventions and the outcome measures used in the included studies.

Grading of evidence

We used the Cochrane’s Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach¹⁶ to evaluate the quality of evidence as high, moderate, low or very low for the main outcomes of the review.

Results

Study selection

Figure 1 shows the study selection process. A total of 35,835 records were identified from the database searches (16,366 from Medline, 15,361 from EMBASE, 3623 from CINAHL, and 454 from PEDro). Additional records were identified by searching the Journal of Medical Internet Research (JMIR) journal and its PubMed indexed sister journals. After screening full-text articles, eight trials^10,17–23 were determined eligible for our review. One trial was reported in two publications.^18,19 Therefore, a total of seven trials were included.

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) 2009 flow diagram.

We did not identify any qualitative studies that explored older peoples’ experience of digital rehabilitation interventions.

Study characteristics

Of the seven trials, five focused on increasing PA.^{10,17–19,22,23} One trial included a vestibular rehabilitation programme for dizziness,²⁰ and the other evaluated a falls prevention programme.²¹ The interventions were delivered via a tablet in two trials,^17,21 Web-connected pedometer in two trials,^17,23 accelerometer in one trial.¹⁸ One trial used a combination of virtual animation coach and computer games with Kinect sensor.²⁰

The characteristics of the included trials are shown in Table 1. Three trials were from the USA:^10,17,23; two from Europe;^18,19,22 and one each from Australia²¹ and the UK.²⁰ The sample size ranged from 102–415. Three trials provided no intervention to those in the control arm.^10,22,23 Two trials delivered usual care/education^17,20,21 and one trial used a wait list control.^18,19 The duration of digital rehabilitation interventions ranged from 6 weeks to 4 months. The components of the interventions were mapped to the Behavioural Intervention Technology (BIT) model and are presented in Table 2.

Table 1.

Characteristics of the included trials.

First author and year Country	Randomised (n) Mean age in years Gender %	Key eligibility criteria	Outcomes relevant to the review	Control intervention	Digital rehabilitation intervention	Duration
Irvine 2013¹⁰ USA	40560.3 years69.4% females	Inclusion: aged above 55 years with no more 60 min/week moderate exercise.	Physical activity; quality of life; user satisfaction; barriers to exercise	No intervention	Active after 55: personalised physical activity programme. 11 weekly sessions, 10–15 min each.	3 months
Bickmore, 2013¹⁷USA	26371.3 years61.2% females	Inclusion: aged 65 year and over, inactive, stable medical condition. Exclusion: cognitive and depressive symptoms, high risk of fall risk.	Physical activity; adverse events; treatment satisfaction	Pedometers were worn and steps were monitored	Tablet computers connected with pedometers to deliver an embodied conversational agent (animated characters) based physical activity intervention.	2 months
Wijsman 2013¹⁸ andBroekhuizen 2016¹⁹The Netherlands	23565 years41% females	Inclusion: inactive, no diabetes, no glucose-lowering medication, no disability impeding physical activity.	Physical activity; quality of life	Waiting list	Direct Life physical activity programme: included an accelerometer-based activity monitor, a website, and a personal-coach to monitor and provide tips via email.	3 months
Geraghty 2017²⁰UK	29667 yearsv67% females	Inclusion: 50 years and above, dizziness over 2 years, have Internet access.Exclusion: medical contraindications, e.g. serious co-morbidity.	Dizziness related disability; depression	Usual care-education, reassurance, medication for nausea	Balance Retraining: fully automated online programme; 6 sessions with vestibular exercises and coping strategies.	1.5 months
Gschwind 2015²¹Australia	15374.7 years61.2% females	Inclusion: aged 65 years and above, able to walk 20 m unaided and watch television from 3 m.Exclusion: language problems, cognitive impairment, medical conditions limiting participation.	General health; depression; balance; adverse events; adherence; falls efficacy; user experience	Evidence-based educational booklet on healthy lifestyle and fall prevention	iStoppFalls: an information and communication based system for assessing fall risks and preventing falls in older people.	4 months
Mouton 2015²²Belgium	102Intervention: 61.2 yearsControl: 66.1 yearsIntervention: 39.6% malesControl: 38.3% males	Inclusion: 50 years and above, French language, access to Internet.	Physical activity; User Satisfaction	No intervention	Move More: The web-based programme covered benefits of physical activity, success stories, exercises, overcoming barriers, and tailored feedback.	3 months
Rowley 2017²³USA	108Intervention: 67.4 yearsControl: 66.1 yearsIntervention: 81.3% femalesControl: 78.6% females	Inclusion: inactive adults, no limitations to walking, access to computer.	Physical activity	No intervention	Tailored, Internet pedometer-based programme to reach 10,000 steps per day and maintain afterward.	3 months

Table 2.

The Behavioural Intervention Technology (BIT) Model mapped to digital rehabilitation interventions in older adults.

Digital rehabilitation intervention	Aims	Behavioural change strategies	Elements	Characteristics	Workflow
Active After 55: Internet based physical activity programme¹⁰	To promote function and mobility in older adults. Adopts a self-tailored approach based on theory of planned behaviour.	Goal setting; self-tailored exercise planning/ownership; barrier identification; review of progress; knowledge on exercise benefits; encouragement	Success stories; library for more articles; safety tips; disease specific recommendations	Medium: A narrator and personal coach presented text and video messagesComplexity: Grade 6-8 reading level; Bullet text points; New exercises addedAesthetics: Not reported.Personalisation: Self prescribed activities	The website was in browser format and participants navigate the website freely in non-linear pathway.
A computer and embodied conversational agent (ECA) based physical activity programme¹⁷	To promote health behaviour change of older adults with low education and health literacy levels.	Goal setting; review of progress; positive reinforcement; barrier identification; problem solving Self-monitoring.	Simulated face-to-face encounters with an embodied conversational agent (virtual coach); Animated characters demonstrating exercises; Step count charts; Daily exercise tips.	Medium: Interactive conversations with agent by choosing replies from multiple options; Images; Exercise demos.Complexity: Not reportedAesthetics: Intuitive and acceptablePersonalisation: Self-monitoring charts to track step counts.	Participants were given touch-screen tablet computers and taught how to use the conversational agent system and have daily conversations for two months. At two months, they were shown how to use a Kiosk computer (had same ECA conversations) when waiting for the scheduled clinical appointments.
An Internet-based interactive physical activity programme^18,19	To increase daily physical activity. Based on the stages of change and I-Change Model.	Goal setting; Regular feedback; Continuous contact with digital coaches.	Accelerometer-based activity monitor; a personal-website A personal e-coach	Medium: Interactive website; Coaches provided updates and advice via email.Complexity: Not reported Aesthetics: Not reported.Personalisation: Not reported.	Participants wore the activity monitor continuously through the day and personalised physical activity targets were set.
A fully automated online balance retraining programme²⁰	To facilitate central nervous system compensation mechanisms with specific vestibular rehabilitation exercises.	Goal setting; tailored feedback; education on dizziness, exercises and behavioural coping strategies; cognitive restructuring; problem solving.	Video demonstrations of exercises; success stories; Email reminders.	Medium: Text and exercise videos Complexity: Content in lay language; large font size; bullet points Aesthetics: Simple look and designPersonalisation: Feedback and exercise prescription based on symptoms	Participants logged in each week over 6 sessions. They undertook automatic tailored exercises based on their symptoms and balance capabilities and received symptom specific feedback each week.
iStoppFalls: An Information and Communication based system to predict and prevent falls²¹	To predict falls risk and provide tailored exercise programme for balance and feedback.	Education on falls prevention strategies and exercise safety; feedback; self-monitoring; social support (integrated social media platform to interact with others).	Exergames for balance, muscle strength and fall risk assessment; exercise activity tracker; monthly falls diary; printed guidelines on exercise safety.	Medium: Exer-gamesComplexity: Simple to Dual-tasking for balance; Increased repetitions, sets, difficulty level for strength exercisesAesthetics: Not reportedPersonalisation: Real-time feedback	Participants were provided with a computer, a set top box, and a Microsoft Kinect sensor, a Senior Mobility Monitor to monitor mobility and performance, and Android tablet to monitor the results. Participants interacted with the virtual Avatar to do the exergames.
Move More: Web-based physical activity programme²²	To improve self-efficacy and physical activity behaviours. Based on trans-theoretical model of health behaviour change and ecological model.	Goal setting; education on benefits of physical activity; overcoming barriers; action planning; feedback; social support.	Success stories; additional links; tips for physical activity; online forum; local physical activity opportunities.	Medium: Text, links Complexity: Not reported Aesthetics: Not reportedPersonalisation: Physical activity journal, automatic tailored monthly feedback	The website was in browser format and participants navigated the website in non-linear pathway.
An Internet-based physical activity programme²³	To understand the benefits of physical activity and to increase awareness of current physical activity level.	Goal setting; education; self-regulation; feedback; rewarding; barrier identification; ask the expert; social support.	Graphical feedback of daily steps; ask the expert; discussion forum; motivational messages; links to other physical activity sites.	Medium: Interactive messages Complexity: Not reported Aesthetics: Not reportedPersonalisation: Tailored messaging	Participants logged in once a week. To begin with, education and self-awareness of physical activity were provided. Participants then set goals to increase weekly steps by 10%.

The most commonly evaluated outcome was PA,^{10,17–19,22,23} followed by quality of life.^10,18,19 Other outcomes included disability or function (two trials), adverse events (two trials), fall risk (one trial), vestibular symptoms (one trial), psychological measures (two trials), user perspective (three trials) and intervention adherence (two trials). Seven trials reported short-term follow-up results.^10,17--21,23 The medium-term was reported in two trial^10,20 and long-term effects in two trials.^17,22

Methodological quality

The bias assessment of the included trials is shown in Table 3. The overall risk of bias was rated as unclear for two trials^20,22 and high risk for five trials.^{10,17–19,23} Trials were predominantly rated as at high risk of bias due to lack of blinding of participants and personnel delivering the interventions.Assessing outcomes were blinded in five of seven studies. While one study was conducted with over 400 participants, the number of participants ranged from 102–263 in other studies. Participants were recruited via online requirement strategies (g. social media, websites, emails, newsletters) in four studies.^{10,18,19,22,23} One study²¹ was carried out multicentre recruitment. One trial was single-centred²⁰ and other study required the adults from three outpatient clinics.¹⁷

Table 3.

Risk of bias assessments.

Trials	Random generation	Allocation concealment	Blinding of participants	Blinding of personnel	Blinding of assessment	Incomplete outcome data	Selective bias
Bickmore 2013¹⁷	Unclear	Unclear	High	High	Low	Low	Low
Geraghty 2017²⁰	Low	Low	Unclear	Low	Low	Low	Low
Gschwind 2015²¹	Low	Unclear	High	High	Low	Low	Low
Irvine 201¹⁰	Unclear	Unclear	High	Low	Low	Low	Low
Mouton 2015²²	Low	Unclear	Unclear	Unclear	Low	Low	Low
Rowley 2017²³	Unclear	Unclear	Unclear	Unclear	High	Low	Low
Wijsman 2013¹⁸ andBroekhuizen 2016¹⁹	Low	Unclear	High	High	High	Low	Low

Outcomes

We concluded that data could not be pooled due to the heterogeneous nature of the participants, interventions and outcome measures. Therefore, a narrative summary of the effects of digital rehabilitation interventions on the outcomes is presented (Tables 4 and 5).

Table 4.

Effects of digital rehabilitation interventions on physical activity in older adults.

Trials	Variables	Digital rehabilitation/control (n)	Assessment time points	Digital	Control	Statistical differencebetween groups
Bickmore 2013¹⁷	Step count	132/131	Baseline2 months (adjusted mean)12 months (adjusted mean)	Data not available 4041 steps3861 steps	Data not available 3499 steps3383 steps	Favours digital intervention.
Rowley 2017²³	Step count	57/51	Baseline (mean ± SD) 3 months (mean ± SD)	4688 ± 1475 steps10286 ± 3022 steps	4690 ± 1475 steps4654 ± 1447 steps	Favours digital intervention.
Wijsman 2013¹⁸Broekhuizen 2016¹⁹	AccelerometerMinutes/day	119/116	Baseline median (IQR) 3 month change	16.8 (18.6)Mean increase of 11.1 minutes	14.4 (23.8)Mean decrease of 0.1 minutes	Favours digital intervention.
Mouton 2015²²	METMinutes/week	52/50	Baseline (mean ± SD)12 month change	1215.3 ± 766.9Change of −28.21 to 220.47	1394.9 ± 836.34Data not available	No significant difference.
Irvine 2013¹⁰	Minutes/week Cardiovascular activitieStretching activitiesStrengthening activitiesBalanceactivities	200/205	Baseline (mean ± SD) 3 months (mean ± SD) 6 months (mean ± SD) Baseline (mean ± SD)3 months (mean ± SD) 6 months (mean ± SD) Baseline (mean ± SD)3 months (mean ± SD)6 months (mean ± SD)Baseline (mean ± SD)3 months (mean ± SD)6 months (mean ± SD)	53.7 ± 73.5108.6 ± 72.1122.1 ± 82.921 ± 39.649.1 ± 48.850.9 ± 54.522.5 ± 52.953.7 ± 53.751.5 ± 598.4 ± 20.835.6 ± 44.140.4 ± 57.6	47.5 ± 55.683.3 ± 71.189.8 ± 78.820.9 ± 32.129.3 ± 31.434.3 ± 39.215.6 ± 33.429.5 ± 50.631.6 ± 426.2 ± 19.917.4 ± 36.418.9 ± 39.4	Favours digital intervention.

IQR: inter-quartile range; MET: metabolic equivalent task; SD: standard deviation.

Table 5.

Effects of digital rehabilitation interventions on quality of life, vertigo, and risk of falls in older adults.

Trials	Variables	Digital rehabilitation/Control (n)	Assessment time points		Digital rehabilitation	Control	Statistical difference between groups
Gschwind 2015²¹	EQ-5D	78/75	Baseline (mean ± SD) 3 months (mean ± SD)		0.86 ± 0.110.86 ± 0.15	0.86 ± 0.130.87 ± 0.13	No significant difference.
Irvine 2013¹⁰	SF-12	200/205	SF-12 Physical log base 10 transformationBaseline3 months6 months		−0.01 ± 0.760.03 ± 0.660.04 ± 0.66	0 ± 0.82−0.07 ± 0.87−0.07 ± 0.84	Favours digital intervention.
Irvine 2013¹⁰	SF-12	200/205	SF-12 Mental log base 10 transformationBaseline3 months6 months		0.7 ± 0.80.88 ± 0.70.86 ± 0.72	0.8 ± 0.80.71 ± 0.80.7 ± 0.84	Favours digital intervention.
Broekhuizen 2016¹⁹	RAND-36	119/116	Physical function	Baseline (mean ± SD) 3 month-change	83.40 ± 14.981.84 (1.16)	84.61 ± 15.50.95 (1.01)	No significant difference.
			Mental health	Baseline (mean ± SD) 3 month-change	77.24 ± 15.522.52 (0.83)	77.31 ± 15.58−0.72 (1.19)	Favours digital intervention.
			Total RAND-36	Baseline (mean ± SD) 3 month-change	630.86 ± 120.126.28 (8.13)	639.68 ± 118.64−0.89 (7.57)	No significant difference.
Geragty 2017²⁰	VS-SF total score	160/136	Baseline median (IQR) 3 months median (IQR) 6 months median (IQR)		14 (8–22) 6 (3–12) 6 (3–14)	13 (7–22)9 (5–15)7 (4–17)	Favours digital intervention.
Gschwind 2015²¹	PPA	78/75	Baseline (mean ± SD)3 months (mean ± SD)		0.62 ± 0.890.41 ± 0.95	0.55 ± 0.900.39 ± 0.80	Favours digital intervention.

EQ-5D: European Quality of Life 5 Dimension; IQR: Inter-quartile range; PPA: Physiological Profile Assessment; RAND-36: Research and Development 36 item Health Survey; SD: standard deviation; SF-12: 12 Item Short-Form Health Survey; VSS-SF: Vertigo Symptoms Scale–Short Form.

PA

Pedometer, accelerometer and the International Physical Activity Questionnaire (IPAQ) were used to assess PA in five trials (Table 3). It appears that Web-based interventions do increase PA in the short/medium term with all studies reporting improvements compared to the control intervention. Two studies found that Web-based PA intervention is effective for increasing step count in the short-term.^17,23 Irvine et al.¹⁰ reported that Web-based PA interventions are effective in both the short and medium-term at increasing the time being physically active by evaluating the total minute PA for a week in cardiovascular activities, stretching activities, strengthening activities and balance.¹⁰

No difference was observed in outcomes at 12 months.

Quality of life

Three trials evaluated quality of life. Two of these were PA interventions.^10,19 One trial reported significant effects favouring digital rehabilitation interventions in the short and medium-term.¹⁰ Broekhuizen et al.¹⁹ only reported a significant improvement for emotional-mental subscale score of the Research and Development 36 item Health Survey (RAND 36).There were no significant difference in quality of life for digital vestibular training in the short-term.²¹

Vestibular and fall risk outcomes

Digital-based vestibular training was found effective in both short and medium-term for vertigo symptoms.²⁰ One trial focused on fall risk and reported that digital rehabilitation reduces the physiological fall risk.²¹

Disability

The impact of digital rehabilitation interventions on disability/function was reported in two trials. One trial that evaluated a balance retraining programme²⁰ found a significant reduction in dizziness-related disability using the Dizziness Handicap Inventory compare to control group. Another trial on a falls prevention programme²¹ measured general health (including mobility, activities, participation and self-care) using the World Health Organisation Disability Assessment Schedule (WHODAS) 2. No significant difference was found between the intervention and control groups.

Adverse events

Only two trials reported adverse events. The iStoppFall study reported that there were no adverse events in the study.²¹ Bickmore et al.¹⁷ reported 289 adverse events of which 10 were moderate-severe events that were likely be not related to digital intervention (eight in control, two in intervention group).

Pain

None of the included trials evaluated this outcome.

Health resource use

None of the included trials evaluated this outcome.

Psychological outcomes

Anxiety

Geraghty et al.²⁰ found a greater reduction in anxiety at 3 months in intervention group measured by the Hospital Anxiety and Depression Scale (HADS) compared to the control, but this difference was not sustained at 6 months.

Depression

Digital vestibular rehabilitation intervention had no significant effect on depression at 3 or 6 months compared to the control intervention.²⁰ Similarly, the iStoppFalls study reported no significant difference in a measure of depression between intervention and usual care groups study.²¹

Self-efficacy

None of the included trials evaluated this outcome.

Fear avoidance

None of the included trials evaluated this outcome.

Process outcomes

Intervention adherence

Only two trials reported data on intervention adherence which was determined by number of times they accessed the Web-based intervention or completion rates of the programme. Bickmore et al.¹⁷ reported that the embodied conversational agent-based PA intervention participants interacted with the virtual coach at an average of 35±19 times during the two-month intervention. Wijsman et al. determined that 91.2% of participants completed the web-based program.¹⁸ None of the studies used the adherence to digital intervention as a primary assessment measure and they did not evaluate the adherence by a patient -eported scale or questionnaires. In other words, the evaluations and results related to intervention adherence in the studies are insufficient. Therefore, it is not possible to conclude about intervention adherence with limited findings.

Intervention attrition

Two studies that compare the efficacy of an online PA intervention reported small dropout rates (Bickmore et al.=3%, Broekhuizen et al.=6.7%) for intervention attrition.^17,19 One study²¹ which assessed the effectiveness of a Web-based fall prevention programme had similar dropout rates between the intervention and control group (n=52, 15 dropouts from intervention group, 13 dropouts from control group). Two studies reported a high level of intervention attrition rates. Irvine et al.¹⁰ reported that 36.5% participants didn’t completed the all sessions. Besides, a study that compare the effectiveness of an online vestibular rehabilitation reported high attrition as 23%.²⁰

User perspectives

User satisfaction were evaluated in two trials.^10,17 Both trials used a Likert scale (1–7) to measure participants’ satisfaction, and both of them reported an average score of six (quite satisfied). In Irvine et al.,¹⁰ participants also rated the programme as very easy to use and very helpful, and, they would recommend the programme to friends or family (seven-point scale, mean=5.7±1.4).

Quality of evidence

The quality of evidence for the main outcomes of the review is presented in Table 6.

Table 6.

The quality of evidence for the effectiveness of digital rehabilitation interventions in older adults.

Outcomes	Comparisons	n	Direction of treatment effect	Quality of evidence
Physical activityShort-term	Web-based PA intervention vs no intervention²³Web-based PA intervention vs waiting list¹⁹Web-based PA intervention vs no intervention¹⁰	108235405	Favours digital intervention	⊕⊕⊕⊝Moderate¹
Short-term	PA intervention via animated virtual coach + pedometer vs pedometer intervention¹⁷	263	Favours digital intervention	⊕⊝⊝⊝Very low²
Medium-term	Web-based PA intervention vs no intervention¹⁰	405	Favours digital intervention	⊕⊕⊕⊝Moderate³
Long-term	Web-based PA intervention vs no intervention²²	102	No difference between groups.	⊕⊝⊝⊝Very low ²
Long-term	PA intervention via animated virtual coach + pedometer vs Pedometer intervention¹⁷	263	No difference between groups.	⊕⊝⊝⊝Very low²
Quality of life (QoL) QoL PhysicalShort-term	Web-based PA intervention vs waiting list¹⁹Web-based PA intervention vs no intervention¹⁰	235405	Inconsistent results. (No difference [19]; Favours digital intervention [10])	⊕⊝⊝⊝Very low⁴
QoL MentalShort-term	Web-based PA intervention vs waiting list¹⁹Web-based PA intervention vs no intervention¹⁰	235405	Favours digital intervention	⊕⊕⊝⊝Low⁵
QoL (Total score) Short-term	Web-based PA intervention vs waiting list¹⁹	235	No difference between groups.	⊕⊝⊝⊝Very low²
QoL (Total score) Short-term	Web-based Exergames vs falls prevention education²¹	153	No difference between groups.	⊕⊝⊝⊝Very low²
QoL (Physical & Mental) Medium-term	Web-based PA intervention vs no intervention¹⁰	405	Favours digital intervention	⊕⊕⊕⊝Moderate³
Vertigo symptomsShort & Medium-term	Web-based balance training vs usual care²⁰	296	Favours digital intervention	⊕⊝⊝⊝Very low²
FallsShort-term	Web-based Exergames vs falls prevention education²¹	153	Favours digital intervention	⊕⊝⊝⊝Very low²

¹Downgraded to one level for risk of bias; ²downgraded to three levels for risk of bias and being a single trial with <400 participants; ³downgraded to one level for risk of bias; ⁴downgraded to three levels for risk of bias, inconsistent results, and indirectness of evidence; ⁵downgraded to two levels for risk of bias and indirectness of evidence.

PA

In the short-term, digital rehabilitation interventions may improve PA in older adults compared to no intervention or waiting list but the evidence is of moderate quality. It is uncertain whether they are effective compared to a pedometer-only intervention as although results favour the digital intervention they are based on very low quality evidence.

In the medium-term, digital rehabilitation interventions probably improve PA compared to no intervention (moderate evidence).

In the long-term, it is uncertain whether digital rehabilitation interventions have no effect compared to no-intervention or a pedometer-based intervention as this is based on one study and the evidence is very low quality.

Quality of life

In the short-term, it is uncertain whether digital rehabilitation interventions have or do not have an effect on the quality of life-physical domain compared to no intervention or waiting list due to inconsistent findings based on very low quality evidence. They may slightly improve quality of life-mental domain compared to no intervention or waiting list (low quality evidence). It is uncertain whether digital rehabilitation interventions have no effect on the overall quality of life, compared to education only, due to very low quality evidence.

In the medium-term, digital rehabilitation interventions probably improve individual physical and mental domains compared to no intervention (moderate evidence).

Vertigo symptoms

In the short and medium-term, it is uncertain whether digital rehabilitation interventions are more effective in improving vertigo symptoms than the usual care as although results favour the digital intervention they are based on very low quality evidence.

Falls risk

In the short-term, it is uncertain whether digital rehabilitation interventions are more effective in reducing falls risk than a falls prevention education programme as although results favour the digital intervention they are based on very low quality evidence.

Discussion

This review evaluated the effectiveness and safety of digital rehabilitation interventions in people over 60 years of age. We included seven randomised controlled trials in this review. These trials compared digital rehabilitation interventions that focused on PA, falls prevention and vestibular retraining to a range of control interventions (usual care, education, no intervention or waiting list). The findings suggest that digital health interventions may improve PA and quality of life in the medium term (moderate evidence). However, there is inconclusive evidence for the short-term effects on PA, quality of life (physical and mental domains), vertigo symptoms and falls risk due to risk of bias, indirectness of evidence and small sample sizes. There was a lack of consistency on the effects on quality of life. The long-term effects on PA are unknown. Further research has the potential to change these findings. Only two studies included long-term follow-up and no difference between interventions was observed.

Secondary outcomes of interest that were studied included anxiety, depression, satisfaction, adherence and trial attrition. None of the included trials evaluated health resource use outcomes, pain and self-efficacy or fear-avoidance behaviour. There were very few adverse events reported that are likely to be related to the interventions but only two trials actually reported adverse events so the safety profile of this type of intervention is unclear.

A range of methods were used by the studies in this review. Only two trials measured satisfaction with the interventions but both suggested that these were acceptable interventions for older people and engagement was good. A recent systematic review concluded that tablet technology is acceptable and satisfying to older people, even if they have cognitive disorders.²⁴ Acceptability of digital rehabilitation interventions to older people is important if we want older people to access care in this way. Notably there were no trials in this review using the smartphone which is in very common use, although, we are aware of a feasibility trial that is underway in the UK to evaluate an intervention delivered via smartphone technology to support home exercises and prevent falls.²⁵

The majority of trials in this review focused on increasing PA.^{10,17,19,22,23} PA is a key target for health improvement or disease prevention in older people. Prior systematic reviews specified that population-based strategies with the use of eHealth to promote PA are effective.²⁶ Our findings would suggest there is also potential to improve PA as a rehabilitation strategy using digital rehabilitation interventions. All the interventions required participants to interact with the digital application on a daily or weekly basis. Mouton²² and Irvine¹⁰ noted that engagement with the programme was better when the programme was supervised.

The prevention of falls is another treatment target to improve health outcomes in older people. Nearly one in three older people aged over 65 years experience a fall at least once a year and this results in a large social and economic burden on individuals and health services. Exercise is an effective strategy for preventing falls and digital rehabilitation interventions have the potential to make this type of treatment accessible to large numbers of older people.⁵ There was only one trial included in this review that focused on balance training.²¹ It utilised game technology to deliver balance and strengthening exercises, a feature distinguishing it from other studies in this review. However, the system used in this trial contained a lot of additional technologies such as Kinect sensor systems, accelerometer, Google TV set and computers/tablets. The digital fall-stop intervention programme reduced the physiological fall risk for older people, but further research is needed including testing more simple technologies which would make the intervention more accessible. The final trial in this review focused on vestibular rehabilitation and resulted in reduced dizziness and disability compared to the control, demonstrating that this type of intervention can successfully be delivered using a digital approach overcoming economic barriers and increasing accessibility.²⁰ In all studies, the control groups were ‘usual care’ or ‘no treatment’. This had been noted as a bias for comparators.

Limitations

The review focused on physical rehabilitation interventions for older people delivered using digital platforms. This meant that we excluded studies that used tele-rehabilitation (telephone calls or messaging) which appeared to be more common in the literature and therefore resulted in a small number of studies. Although, five of the seven trials focused on PA, we were unable to perform a meta-analysis because of heterogeneity in the included studies. It was not possible to blind participants and personnel to the intervention received by participants so all studies were considered high risk for this element of the risk of bias assessment, making it impossible for studies to be considered at low risk of bias according to the Cochrane risk of bias tool.

Conclusions

Digital rehabilitation interventions seem to have potential to benefit older people in improving PA and quality of life in the medium term. However, there is uncertainty around the short-term effects on PA, quality of life (physical and mental domains), vertigo symptoms and falls. More research is needed to establish robust estimates of effectiveness including long-term outcomes. There is a need to conduct large trials that include evaluation of cost-effectiveness and safety of these interventions for older people.

Supplemental Material

sj-pdf-1-jtt-10.1177_1357633X20927587 - Supplemental material for Components, design and effectiveness of digital physical rehabilitation interventions for older people: A systematic review

Supplemental material, sj-pdf-1-jtt-10.1177_1357633X20927587 for Components, design and effectiveness of digital physical rehabilitation interventions for older people: A systematic review by Eda Tonga, Cynthia Srikesavan, Esther Williamson and Sarah E Lamb in Journal of Telemedicine and Telecare

Footnotes

Acknowledgements

We acknowledge the support of Elinor Harriss, an outreach Librarian at the Bodleian Health Care Libraries, University of Oxford, who developed and ran the search strategies. We thank Bethan Copsey, medical statistician at the Centre for Statistics in Medicine, University of Oxford for her guidance in data synthesis.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was funded by the SE2020 Translational Fund for Social Impact at the University of Oxford and the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care Oxford at Oxford Health NHS Foundation Trust, and supported by the NIHR Biomedical Research Unit, Oxford. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

ORCID iDs

Eda Tonga

Cynthia Srikesavan

Supplemental material

Supplemental material for this article is available online.

References

United Nations, Department of Economic and Social Affairs Report, Population Division (2017). World Population Ageing 2017 (ST/ESA/SER.A/408).

Later life in United Kingdom 2019, Fact-sheet of statistics, Age UK. https://www.ageuk.org.uk/globalassets/age-uk/documents/reports-and publications/later_life_suk_factsheet.pdf (accessed 20 June 2019).

Kingston

Comas-Herrera

Jagger

Forecasting the care needs of the older population in England over the next 20 years: Estimates from the Population Ageing and Care Simulation (PACSim) modelling study.

Lancet Public Health 2018; 3: e447–e455.

Inouye

Studenski

Tinetti

, et al. Geriatric syndromes: Clinical, research, and policy implications of a core geriatric concept. J Am Geriatr Soc 2007; 55: 780–791.

Sherrington

Michaleff

Fairhall

, et al. Exercise to prevent falls in older adults: An updated systematic review and meta-analysis. Br J Sports Med 2017; 51: 1750–1758.

Robbins

Lim Choi Keung

, et al. E-health for active ageing: A systematic review. Maturitas 2018; 114: 34–40.

WHO guideline: recommendations on digital interventions for health system strengthening. Geneva: World Health Organization; 2019. Licence: CC BY-NC-SA 3.0 IGO.

Stellefson

Chaney

Barry

, et al. Web 2.0 chronic disease self-management for older adults: A systematic review. J Med Internet Res 2013; 15: e35.

Gagnon

Ndiaye

Larouche

, et al. Optimising patient active role with a user-centred eHealth platform (CONCERTO+) in chronic diseases management: A study protocol for a pilot cluster randomised controlled trial. BMJ Open 2019; 9: e028554.

10.

Irvine

Gelatt

Seeley

, et al. Web-based intervention to promote physical activity by sedentary older adults: Randomized controlled trial. J Med Internet Res 2013; 15: e19.

11.

Argent

Daly

Caulfield

Patient involvement with home-based exercise programs: Can connected health interventions influence adherence?

JMIR Mhealth Uhealth 2018; 6: e47.

12.

Picorelli

Pereira

, et al. Adherence to exercise programs for older people is influenced by program characteristics and personal factors: A systematic review. J Physiother 2014; 60: 151–156.

13.

Guo

Albright

The effectiveness of telehealth on self-management for older adults with a chronic condition: A comprehensive narrative review of the literature. J Telemed Telecare 2018; 24: 392–403.

14.

Srikesavan

Tonga

Harriss

, et al. Components, design, and effectiveness of digital rehabilitation interventions for older adults: A mixed-method systematic review. PROSPERO 2018 CRD42018042471, https://www.crd.york.ac.uk/prospero/display_record.phpID=CRD42018042471 (2018, accessed 20 June 2019).

15.

Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (eds). Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019). Cochrane, 2019. http://www.training.cochrane.org/handbook; www.training.cochrane.org/handbook

16.

Ryan R and Hill S. How to GRADE the quality of the evidence. Cochrane Consumers and Communication Group. Version 3.0, December 2016. http://cccrg.cochrane.org/author-resources (2016, accessed 5 May 2019).

17.

Bickmore

Silliman

Nelson

, et al. A randomized controlled trial of an automated exercise coach for older adults. J Am Geriatr Soc 2013; 61: 1676–1683.

18.

Wijsman

Westendorp

RGJ

Verhagen

EALM

, et al. Effects of a Web-based intervention on physical activity and metabolism in older adults: Randomized controlled trial. Diabetes Technol Ther 2015; 17: S63–S64.

19.

Broekhuizen

de Gelder

Wijsman

, et al. An Internet-based physical activity intervention to improve quality of life of inactive older adults: A randomized controlled trial. J Med Internet Res 2016; 18: e74.

20.

Geraghty

Kirby

Essery

, et al. Internet-based vestibular rehabilitation for adults aged 50 years and over: A protocol for a randomised controlled trial. BMJ Open 2014; 4: e005871.

21.

Gschwind

Eichberg

Ejupi

, et al. ICT-based system to predict and prevent falls (iStoppFalls): Results from an international multicenter randomized controlled trial. Eur Rev Aging Phys Act 2015; 12: 10.

22.

Mouton

Cloes

Efficacy of a Web-based, center-based or combined physical activity intervention among older adults. Health Educ Res 2015; 30: 422–435.

23.

Rowley

Lenz

Swartz

, et al. Efficacy of an individually tailored, Internet-mediated physical activity intervention in older adults: A randomized controlled trial. J Appl Gerontol. Epub ahead of print 1 October 2017. DOI: 10.1177/0733464817735396.

24.

Ramprasad

Tamariz

Garcia-Barcena

, et al. The use of tablet technology by elderly in health care settings–is it effective and satisfying? A systematic review and meta analysis. J Gen Intern Med 2015; 2: S75–S76.

25.

Hawley

Tacconi

Mellone

, et al. Can smartphone technology be used to support an effective home exercise intervention to prevent falls amongst community dwelling older adults? The TOGETHER feasibility RCT study protocol. BMJ Open 2019; 9(9): e028100.

26.

Muellmann

Forberger

Mollers

, et al. Effectiveness of eHealth interventions for the promotion of physical activity in older adults: A systematic review. Prev Med 2018; 108: 93–110.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.39 MB