The Telephysiotherapy for Older People (TOP‐UP) program for improving mobility in people receiving aged care: a hybrid type 1 effectiveness–implementation randomised controlled trial

Older people receiving aged care often have multiple health problems, including dependence on others for basic needs and restricted mobility.1 In the United States, mobility disability affects 35% of people aged 70 years or older. The World Health Organization recommends that all adults undertake 150–300 minutes of moderate intensity activity per week, including balance and strength training, to improve mobility.2 The 2018 Australian Sunbeam trial found that a six‐month physiotherapist‐led moderate intensity strength and balance exercise program improved mobility and reduced falls among people in aged care.3 However, an updated Cochrane literature review found that falls prevention in residential aged care requires ongoing, tailored strength and balance exercise delivered at moderate intensity for at least one hour each week.4,5

The Australian Royal Commission into Aged Care highlighted the need for interventions that improve mobility and reduce falls in aged care.6,7 Physiotherapists have the expertise to deliver tailored exercise programs.3 Despite the limited resources of allied health services such as physiotherapy,8 recent residential aged care funding reforms provide an opportunity for optimising use of staff time, enhancing care quality, and introducing innovative solutions, such as telehealth, particularly in regional areas.9

Telehealth uses information and communication technology to deliver health care services remotely.10 The coronavirus disease 2019 (COVID‐19) pandemic accelerated infrastructure development and the adoption of telehealth by older people and health care professionals, making it a viable option in aged care.11 However, evidence supporting the value of telehealth for delivering physiotherapy (telephysiotherapy) is limited. Our recent systematic review found that few randomised controlled trials had specifically evaluated the effect of telephysiotherapy on outcomes such as mobility, falls, and quality of life for people in aged care.12

We therefore assessed the implementation of the Telephysiotherapy for Older People (TOP‐UP) program and its effectiveness for improving mobility (sit‐to‐stand performance, balance, gait speed), physical function (fall rate, fall risk, pain, physical activity), mobility goal attainment, and quality of life among adults aged 65 years or older who were receiving at‐home or residential aged care. We also examined implementation outcomes (adoption, reach, fidelity, and dose delivered) to provide information that could inform health care policy, clinical practice, and the design of telephysiotherapy services as part of aged care.

Methods

The TOP‐UP trial was a two‐arm, pragmatic, hybrid type 1 effectiveness–implementation randomised controlled trial, conducted from 1 September 2021 to 30 November 2023.13 Trial and protocol reporting were guided by the CONsolidated Standards Of Reporting Trials (CONSORT) statement,14 the Template for Intervention Description and Replication (TIDieR) checklist,15 and the Consensus on Exercise Reporting (CERT) guidelines.16 The trial was registered prospectively with the Australian New Zealand Clinical Trials Registry (ACTRN12621000734864; 10 June 2021).

Participants

The TOP‐UP trial was conducted in collaboration with aged care providers who offer home or residential aged care services in metropolitan, regional, and remote areas of New South Wales, Victoria, and Western Australia. Aged care providers were recruited through research networks and direct outreach, including calling a broad range of providers in different regions to ensure geographic and service diversity. Providers assisted with participant recruitment, using a standardised screening form. The eligibility criteria for participants were the willingness to use a tablet computer for video consultations and to commit to two hours of weekly exercise, and adequate sensory, neurological, cognitive, and English language skills for participation. Exclusion criteria were severe cognitive impairment (Modified Telephone Interview for Cognitive Status score of 10 or less17), inability to walk ten metres, current participation in a physiotherapist‐led exercise program, and life expectancy less than six months. All participants provided written consent, together with an enduring power of attorney if required by aged care service management. Participant agency was a priority, and we ensured that the study was clearly explained and that the willingness to participate was confirmed before consent was requested.

Randomisation and blinding

Baseline data were collected during a Zoom meeting prior to randomisation by research team physiotherapists, with the support of aged care workers. Participants were randomly assigned to the TOP‐UP program or to the waiting list control group using a centralised web system with random permutation blocks of two and four, stratified by aged care type. A statistician computer‐generated the allocation sequence, independent of recruitment; enrolment staff did not have access to the sequence. The nature of the intervention precluded blinding of participants or care providers to allocation; outcomes assessors and data analysts were blinded through role separation and participant instructions. Cognitive impairment was assessed with the Modified Telephone Interview for Cognitive Status (range, 0–41; severe: score of 10 or less; moderate: 11–20; mild: 21–30; no impairment: 31–4017). Readiness to adopt new technology was assessed with the Technology Readiness Index (low scores = low readiness because of perceived complexity or lack of usefulness).18

The TOP‐UP intervention

Experienced aged care physiotherapists, employed or contracted by our aged care partners and independent of the research team members who conducted follow‐up assessments, delivered ten Zoom sessions over six months, during weeks 1, 3, 5, 7, 9, 12, 16, 20, 22, and 24 of the trial. The sessions described tailored balance and strength exercises in the Otago Exercise Program.19 Participants were encouraged to independently undertake moderate intensity exercise (Borg scale, 12–14 of 20)20 for two hours each week. Participants had access to further information in the TOP‐UP website and booklet (https://www.topupexercise.com.au); aged care support workers assisted with telephysiotherapy appointments and supervised weekly 30‐miunte exercise sessions. Physiotherapists prescribed exercise selections for independent and supervised sessions of gradually increasing intensity, for safety reasons. To maximise adherence, physiotherapists and support workers received one hour of training (via Zoom) in health coaching and behaviour change strategies from the research team.21,22 Adherence to the program was tracked in participant self‐report diaries.

Control group

Participants in the control group received usual care, including seated exercise, non‐strength/balance exercise programs, and massage and heat therapies. They were offered a three‐month TOP‐UP program after completion of the study.

Outcomes

All outcomes were assessed at baseline and six months after randomisation. The primary outcome was mobility as indicated by the combined score from the 12‐point Short Physical Performance Battery (SPPB), which assesses standing balance, gait, and timed sit‐to‐stand. The SPPB score is a predictor of mortality, care admission, and disability.23

The secondary outcomes were:

• fall rate (falls per person) and falls risk (proportion of people who experienced at least one fall), based on incident reports (residential care) or self‐report diaries (home care);
• SPPB components (five times sit‐to‐stand, four‐step balance, gait speed);
• pain, assessed with a visual analogue scale (1–10 scale);24
• mobility goal attainment, assessed with the 5‐point Goal Attainment Scale (+2 = much better than expected, +1 = somewhat better than expected, 0 = expected level, –1 = somewhat less, –2 = much less);25
• physical activity, assessed with the Incidental and Planned Exercise Questionnaire (IPEQ);26 and
• quality of life, assessed with four measures: the EuroQol 5‐dimension 5‐level (EQ‐5D‐5L) visual analogue scale (range, 0–100);27 the question “Compared to six months ago, how would you rate your quality of life?” (–2 = much worse, 0 = no change, +2 = much better); reported balance self‐confidence (1 = excellent to 5 = poor); and reported fear of falling (1 = not at all to 5 = extremely).28

Implementation outcomes were assessed to inform future program design, delivery modifications, and expansion:29

• adoption: proportion of invited aged care services that participated in the trial;
• reach: proportion of aged care service users who were screened, consented to participation, and participated in the intervention;
• fidelity: the number and duration of physiotherapy and support worker‐assisted sessions completed by participants; and
• dose delivered: total amount of exercise undertaken by participants.

To evaluate the generalisability of the intervention, feasibility factors, such as the proportion and representativeness of people still participating at six months, were considered throughout the study.

We also examined program acceptability, barriers, facilitators, and uptake factors in surveys and interviews with participants, physiotherapists, support workers, and service managers. Our thematic analysis of interview data has been reported elsewhere;30 in this article, we discuss only the trial participant survey responses. All adverse events, including falls, injuries, and deaths, were documented in incident reports (residential care) or participant diaries (home care). Two investigators independently reviewed each event, using medical records and care documentation to assess whether it was related to the intervention and its impact on study viability. We established a data monitoring committee to review serious unexpected events related to the intervention.

Sample size

We determined that 240 participants (120 per group) were required to provide 80% statistical power to detect a 0.9‐point between‐group difference in 12‐point SPPB scores at six months, assuming a standard deviation (SD) of 2.8, α = 0.05, and 20% attrition. We defined a 0.5‐point between‐group difference in SPPB as being clinically significant.31 This sample size was also expected to be adequate for detecting 10–15% between‐group differences in the secondary outcomes.

Statistical analysis

We registered our statistical analysis plan with the Open Science Framework prior to data analysis (https://osf.io/qyt5a; 29 April 2024). We used a modified intention‐to‐treat approach; we analysed data for all participants for whom follow‐up data were available as part of their original randomised group, regardless of intervention adherence. Participants lost to follow‐up were excluded from outcomes analyses; missing data were not imputed. The statistical significance of between‐group differences in mobility (SPPB scores) and other continuous outcomes at six months was assessed in linear regression models adjusted for baseline scores; we report adjusted mean differences with 95% confidence intervals (CIs). The statistical significance of between‐group differences in ordinal outcomes (including SPPB components, Goal Attainment Scale scores, quality of life) was assessed in ordinal regression models adjusted for baseline; we report adjusted odds ratios (aORs) with 95% CIs. The statistical significance of the difference in fall rate was assessed using negative binomial regression (with follow‐up time as offset), adjusted for baseline; we report the adjusted incidence rate ratio (IRR) with 95% CI. We report fall risk, analysed using log‐binomial regression, as estimate relative risk (RR) with 95% CI. Other dichotomous outcomes (pain improvement, mobility goal met, self‐reported improvement in quality of life) were analysed using logistic regression; we report adjusted odds ratios (aORs) with 95% CIs. We also undertook pre‐specified subgroup analyses by gender, care type (home care, residential aged care), and cognitive impairment status that included interaction terms. No post hoc analyses were performed, and no interim analyses planned. Stopping guidelines were developed prior to trial commencement. Implementation outcomes are summarised numerically and descriptively.

Ethics approval

The human research ethics committee of the Sydney Local Health District (Concord) approved the trial (CH62/6/2021‐009).

Results

During 1 September 2020 – 31 March 2023, 1348 people aged 65 years or older were screened at 27 sites, of whom 242 people were recruited for the trial: 120 were randomised to the intervention group (residential aged care, 51; home care, 69), 122 to the control group (residential aged care, 52; home care, 70) (Box 1). The mean age of the 242 participants was 83 years (standard deviation [SD], 8 years; range, 65–100 years). Cognitive impairment was detected in 108 participants (45%; mild impairment in 76 and moderate impairment in 32 people). Fifty‐one participants (21%) had prior telehealth experience; the mean Technology Readiness Index score was 2.1 (SD, 0.8; range, 1.0–4.6). The baseline demographic characteristics of the two groups were similar (Box 2).

A total of 192 participants (79%) completed the six‐month mobility assessment (primary outcome). The mean age at baseline of the 50 people who withdrew from the study was similar to that for the 192 available for the six‐month follow‐up (82 [SD, 9] v 82 [SD, 8] years), as was the baseline mobility score (5.1 [SD, 2.9] v 5.5 [SD, 2.9]), and the proportion of women (33 of 50, 66% v 127 of 192, 66%). The mean pain visual analogue score was higher for the participants who withdrew (4.6 [SD, 2.9] v 3.9 [SD, 3.0]) and the proportion with cognitive impairment was larger (26 of 50, 52% v 82 of 192, 43%).

One minor adverse event was related to the intervention, a non‐injuring fall, but there were no serious adverse events. The most frequent reasons for withdrawing from the trial were poor health, including COVID‐19 (22 participants, 9%) and deaths unrelated to the intervention (sixteen participants, 7%: cerebrovascular disease, six; chronic heart disease, five; cancer, one; cause unknown, four people). As no concerns were reported by staffs at the trial sites, these deaths were deemed unrelated to the intervention, consistent with our safety monitoring procedures.

Primary outcome

After adjusting for baseline mobility scores, the mean difference in mobility score at six months (intervention v control group) was 2.1 (95% confidence interval [CI], 1.4–2.7) points (Box 3, Box 4); the estimated mean difference was similar after also adjusting for study site (2.0 points; 95% CI, 1.4–2.7 points).

Secondary outcomes

Sit‐to‐stand performance was more likely to improve in intervention participants than in control participants (aOR, 2.7; 95% CI, 1.3–4.3); in comparison with the control group, they needed less hand support to stand up, their gait was more rapid over 2.4 m (adjusted mean difference, 0.9 [95% CI, –1.6 to –0.1] seconds), and they were more likely to meet their mobility goals (aOR, 14.1; 95% CI, 4.1–48.5). Pain visual analogue scores (adjusted mean difference, –1.1 points; 95% CI, –1.8 to –0.3 points) and planned exercise participation (adjusted mean difference, 1.7 [95% CI, 1.1–2.3] hours/week) were better for the intervention than the control group (Box 3). Intervention group outcomes were also better than for the control group with respect to the sit‐to‐stand, balance, and gait speed SPPB components (raw data: Supporting Information, table 1).

Mobility goals were met or exceeded by 48 of 92 intervention group participants (53%) and seven of 100 control group participants (7%) (Supporting Information, table 2). At six months, intervention participants reported greater quality of life (EQ‐5D‐5L visual analogue scale: adjusted mean difference, 6.2; 95% CI, 1.8–10.7); they were more likely to report improved quality of life (aOR, 7.0; 95% CI, 3.4–14.2), and 47 intervention group participants (51%) reported that their quality of life had improved between baseline and follow‐up, compared with thirteen control participants (13%) (Box 5). Intervention group participants also reported greater self‐confidence about balance (adjusted mean difference, –0.3 [95% CI, –0.2 to –0.5] points) and of falling (adjusted mean difference, –0.5 [95% CI, –0.3 to –0.5] points).

Over six months, 58 falls were recorded in the intervention group and 96 for the control group, but the difference in incidence was not statistically significant (incidence rate ratio, 0.60; 95% CI, 0.35–1.01) (Box 6). By six months, six of 92 participants in the intervention group (10%) had experienced two or more falls, and 19 of 100 participants in the control group (19%). The mean surveillance period was slightly longer in the intervention group (196 [SD, 37] days) than in the control group (180 [SD, 17] days). The mean number of falls per participant over six months was 0.52 (SD, 1.45) per person in the intervention group and 0.96 (SD, 1.81) per person in the control group.

Subgroup analyses

Mobility improvement was greater for intervention participants who reported at least 45 minutes of baseline physical activity per week (adjusted mean difference, 2.7 [95% CI, 1.8–3.5] points) than for those who reported less (1.3 [95% CI, 0.3–2.3] points; P for interaction = 0.044). Sex, cognitive impairment, care type (home or residential aged care), and technology readiness did not influence differences between the intervention and control groups in the primary outcome (Box 7).

Implementation outcomes

Nine of twelve invited aged care service providers (75%) agreed to participate (adoption); 242 of 1348 people who received aged care from the participating service providers (18%) were screened, provided consent for participation, and participated in the trial, and 92 of 120 (77%) completed the six‐month program (reach). Intervention fidelity was good: intervention participants attended a mean 7.4 telephysiotherapy sessions (SD, 2.3 sessions) and 23 (SD, 4.9) of 26 support worker sessions. Based on exercise diaries, total mean exercise dose was 32.8 [SD, 17.7] hours (mean, 1.3 hours per week). One non‐injuring fall was reported (Box 8).

In post‐intervention surveys completed by 66 intervention participants (72%), the mean rating (scale, 1–10) for improvement in walking was 7.4 (SD, 1.7), in balance 7.5 (SD, 1.6), and fall risk reduction 7.0 (SD, 1.7). The mean rating for physiotherapy advice was 8.7 (SD, 1.3), the Zoom experience 8.1 (SD, 1.8), exercise videos 8.1 (SD, 1.3), and support worker engagement 9.2 (SD, 0.9). The overall rating for the TOP‐UP intervention was 7.8 (SD, 1.7) (Box 9); 62 people (94%) would recommend telephysiotherapy for other people.

Discussion

We report the first pragmatic randomised trial to evaluate a locally supported telehealth physiotherapist‐led program for improving mobility and preventing falls of people receiving aged care. TOP‐UP safely improved mobility in older people receiving aged care; its adoption, feasibility, and adherence were all good. Fewer intervention participants experienced falls during the program than control group participants. Sit‐to‐stand performance, balance, gait speed, mobility goal attainment, and quality of life also improved for intervention participants, and they reported reduced pain. Most intervention group participants would recommend the program to other people.

Mobility gains did not differ significantly by sex, cognitive impairment status, or care type (home or residential aged care), but were greater for participants who reported more baseline physical activity; however, clinically meaningful improvements were also achieved by participants who reported less baseline physical activity. The overall mean improvement in mobility score (2.1 points; 95% CI, 1.4–2.7 points) exceeded the thresholds for both the minimal clinically important difference (0.5 point) and for substantial change (1.0 point).31 The level of improvement achieved is associated with reduced risk of disability, institutionalisation, and mortality.34,35 Improved mobility may support functional recovery, including safer transfers, such as moving between bed and chair or between chairs, and greater independence in daily activities.23

Our findings are consistent with those of two effective in‐person, physiotherapist‐led group fall prevention programs for people in aged care in Australia: the Sunbeam trial3 and a dementia‐focused program.36 In the Sunbeam study, supervised strength and balance training using gym equipment was associated with a 55% reduction in the fall rate;3 the simpler, more readily expandable dementia‐focused program found a significant between–group difference in the falls rate during the intervention period.36 In the TOP‐UP program, participants received a mean 1.7 hours of exercise per week, similar to the other two programs, and this level may be sufficient for people receiving aged care. An update of the 2018 Cochrane review found that the fall reduction effects of physiotherapy programs are lost after interventions conclude,37 indicating that sustained exercise is needed. Telephysiotherapy could facilitate such exercise by providing ongoing remote access and care worker engagement.

The positive effects of TOP‐UP should be considered in the context of the broader telehealth literature. Physical benefits and cost savings are associated with telerehabilitation after knee replacement.38 Asynchronous telehealth is effective for the neurological rehabilitation of people with mobility limitations when adequate support is provided.39 Real‐time video‐based telerehabilitation achieves greater adherence and satisfaction than in‐person therapy, and is viable for older adults with mobility disabilities;40 it also fosters motivation and therapeutic relationships.41 Our qualitative findings suggest that the high acceptability of TOP‐UP is attributable to its accessibility, personalised care, and local support, which enhance motivation, self‐efficacy, and exercise independence.30

Limitations

The TOP‐UP trial was co‐designed with clinicians, older adults, and carers, and proved safe, effective, and acceptable, even for people with cognitive impairment and multiple medical conditions. However, the COVID‐19 pandemic slowed recruitment, and lockdowns may have influenced physical activity levels. Excluding people with severe cognitive impairment limits the generalisability of our findings. Relying on self‐reported participant information introduces recall bias. Support worker engagement probably varied, which could have affected the fidelity of intervention implementation. Recruitment of participants through aged care services may have favoured the participation of more engaged or better functioning older people. As intervention effects beyond six months were not measured, we cannot assess the longer term impact of our intervention. Digital literacy and technology‐related barriers were not comprehensively assessed, which may affect the generalisability of our findings to older people who have limited resources or are less comfortable with digital technology. We did not investigate the sustainability of the intervention program, workforce capacity, or system readiness.

Conclusion

Our TOP‐UP trial findings indicate that telephysiotherapy is a safe, effective, and expandable approach to improving the mobility of people receiving aged care, reducing falls risk and pain, and enhancing quality of life. The participants achieved clinically meaningful gains with high adherence to the program. The combination of telehealth, physiotherapy, and local support was effective, including for people with mild to moderate cognitive impairment. Our findings indicate that telephysiotherapy is a viable alternative that could be integrated into aged care when in‐person care is limited.

Box 1 – Selection of participants for the Telephysiotherapy for Older People (TOP‐UP) trial, 1 September 2021 – 30 November 2023

---

COVID‐19 = coronavirus disease 2019.

Box 4 – Mobility (Short Performance Physical Battery, SPPB) scores at baseline and six‐month follow‐up: mean values (with standard deviations), by participant study group

---