How Effective Are Interventions to Increase Physical Activity Levels among Older Inpatients Receiving Rehabilitation, without Increasing the Amount of Therapy? A Systematic Review

Abstract

Purpose: We systematically reviewed the effectiveness of interventions to increase physical activity in older adults (aged ≥ 60 y), admitted for inpatient rehabilitation, without increasing the amount of therapy. Method: Five electronic databases were systematically searched to identify English-language articles reporting controlled trials of interventions to increase the physical activity (through participation or behavioural change) of older adults receiving inpatient rehabilitation. Trials were excluded if an intervention increased the intensity of usual care, either during the week or on the weekend. Two reviewers independently completed trial selection, quality assessment, and data extraction. Data were synthesized descriptively, and effect sizes with 95% CIs were calculated. Results: Of the 316 articles identified, 3 were included. Two were activity-based, and 1 was a behavioural change intervention. Physical activity was significantly improved in the behavioural change intervention trial that occurred during therapy (d = 0.27; 95% CI: 0.02, 0.52) and non-therapy time (d = 0.43; 95% CI: 0.19, 0.68). Participants in all trials were sedentary for the vast majority of the day. Conclusions: Older adults in inpatient rehabilitation have a high level of inactivity. Evidence is lacking that interventions increase physical activity in older adults admitted to inpatient rehabilitation without increasing the amount of time in therapy. Evidence from one trial indicates that behaviour-based strategies are effective for increasing physical activity levels in the inpatient rehabilitation setting.

Inpatient rehabilitation aims to increase the functional independence and mobility of older adults after an episode of acute ill health.¹ Causes of admission include, but are not limited to, fracture, a neurological event such as stroke, surgery, cardiopulmonary or respiratory illnesses, and infection.¹ Inherent in these periods of ill health is that patients reduce their level of physical activity as a result of surgical restrictions, an inability to organize and coordinate movement after brain trauma, body systems requiring a period of rest, severe fever, or sedation.

Muscular and aerobic deconditioning as a result of ill health adversely affects physical activity levels,² and the rehabilitation environment itself is also likely to play a role.³ Physiotherapists work with patients to improve their health status and build toward a therapeutically beneficial amount of daily physical activity. To achieve this, therapists often implement interventions such as exercise programs as a means to improve patient fitness and mobility⁴ so people can return to living at home and participating in their community. Ideally, inpatients admitted to rehabilitation have returned to their pre-morbid level of function and physical activity before discharge, although this is not always achieved.

Although inpatient rehabilitation aims to restore function and activity for older adults, this population has been reported to spend the majority of its time being sedentary.^5–7 Older adults completing inpatient rehabilitation after orthopedic interventions to the lower limb had a median step count of less than 400 steps per day, spent less than 10 minutes per day walking, and had a median sedentary time of 23 hours per day.⁵ This amount of activity contrasts with adults in the community aged 75 years or older, who on average spend 20 minutes per day doing physical activity and 5.5 hours standing, walking, or moving around in the upright position.^6,8

The level of physical activity of older adults in rehabilitation is also lower than that expected of older adults living with chronic illness in the community.⁹ Moreover, physical activity guidelines for older adults recommend that they complete at least 30 minutes of moderate exercise per day,¹⁰ a figure that equates to approximately 3,000 steps per day at a minimally moderate intensity,¹¹ in addition to the physical activity completed as a part of daily life.

Patients receiving inpatient rehabilitation are not meeting these benchmarks. Although older adults admitted to inpatient rehabilitation, often with multiple and chronic health issues, may not be at a stage at which meeting physical activity guidelines is feasible, their rehabilitation should involve an increase in function and incidental physical activity in preparation for their being discharged.¹⁰ However, the deconditioning associated with immobility or sedentary behaviour during inpatient rehabilitation can further reduce patients’ overall level of function.¹² This reduced level of function can make the transition to the community more difficult and potentially riskier, with a higher likelihood of increased community services or early hospital readmission.¹²

Evidence has shown that increasing the intensity and frequency of therapy during inpatient rehabilitation improves functional independence and reduces length of stay.¹³ Previous research with older people receiving rehabilitation has targeted periods when inpatients are most sedentary (e.g., weekends and outside regular therapy hours), resulting in significantly more physical activity and time spent upright.¹⁴ However, in rehabilitation settings, the frequency and intensity of therapy may be predominantly dictated by the use of resources, staffing levels in particular. This means that it can be difficult to implement additional therapy without additional funding. In these cases, it would be helpful for clinicians to know which interventions most effectively maximize physical activity in older adults admitted for inpatient rehabilitation.

Around the world, the percentage of people aged 65 years or older is projected to double by 2050,¹⁵ so the demand for inpatient rehabilitation is also likely to rise. A time-limited, in-demand health service such as inpatient rehabilitation makes the need to be efficient and effective increasingly important. Therefore, this systematic review investigated the effectiveness of interventions that aimed to increase physical activity in the population of older adults admitted for inpatient rehabilitation without providing additional therapy.

Methods

This systematic review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.¹⁶ A protocol for the review was registered with PROSPERO (ID CRD42018090718).

Identification and selection of trials

Articles were identified through searches of five electronic databases (PubMed, CINAHL, Embase, MEDLINE, and the Cochrane Library). All databases were searched from the inception of the database until February 2018, except for PubMed, which was searched only from August 2016 to March 2018 to capture recently published articles. The search strategy focused on four main concepts: older adult, rehabilitation, inpatient, and physical activity; it included synonyms for each concept. One reviewer (SQ) completed the searches. The search yields were then transferred to Endnote bibliographic software, version X7 (Clarivate Analytics, Philadelphia, PA). The database searches were supplemented by a manual search of the reference list of each included trial. In addition, citation tracking using Google Scholar (Google, Mountain View, CA) was completed for all included trials.

Inclusion and exclusion criteria

Articles were included if (1) they reported randomized, quasi-randomized, or non-randomized controlled trials; (2) participants were, on average, aged 60 years or older because we wanted to focus on older adults; (3) participants were admitted for inpatient rehabilitation (subacute services) including “slow stream” or less intense rehabilitation or rehabilitation in the home; (4) the intervention involved participation in physical activity (e.g., an exercise or walking program) or behavioural change strategies (environmental enrichment aimed at increasing physical activity or providing feedback on physical activity levels; the control group could receive usual care, no intervention, or an alternate physical activity intervention); (5) physical activity was reported as an outcome measure; and (6) the articles were full text articles published in English in a peer-reviewed journal.

Articles were excluded if (1) they included participants who were inpatients in an acute hospital setting; (2) the trial design was a pretest–posttest, case, or cohort study; or (3) the intervention was an increase in the intensity of usual-care therapy, either during the week or on the weekend.

Two reviewers (SQ, NS) applied the eligibility criteria independently of the search yield, initially excluding articles on the basis of title and abstract. Any disagreements between the reviewers were resolved by consensus. When an article could not be excluded on the basis of information in either the title or the abstract, the full text was obtained, and the eligibility criteria were reapplied. When there was uncertainty about an article, the author was contacted for clarification; we then assessed the article’s suitability for inclusion in the review. A k score and 95% CI were calculated using the QuickCalcs: Kappa online software (GraphPad Software Inc., San Diego, CA) to establish the level of agreement between the two reviewers.

Quality assessment of trials

All trials that met the eligibility criteria were critically appraised by two reviewers (SQ, NS), independently, using the PEDro scale to assess their quality.¹⁷ The PEDro scale consists of 11 items that assess eligibility criteria; random allocation; concealed allocation; similarity of groups at baseline; degree of participant, therapist, and assessor blinding; retention rate; intention-to-treat analysis; reported between-groups comparisons; point estimates; and variability among all trials. A score from ranging from 1 to 10 is determined (the eligibility criteria item is not scored), and trials with a score of less than 4 were deemed to be of a lower quality.¹⁸

Data extraction

A data extraction form was developed for use in this review. Two reviewers independently completed the data extraction (SQ, SC). The data extracted related to study design, participants’ characteristics (sample size, age, sex, BMI, primary diagnosis, comorbidities, living arrangements, pre-admission activity level, vocational status), as well as the characteristics of the trial for the intervention and control groups (type of intervention, frequency, intensity, duration, progression, individual and/or group, level of supervision, location, equipment used, attendance).

Data analysis

Data were synthesized descriptively. When appropriate data were available, standardized mean differences (or effect sizes) were calculated by dividing the mean differences by the pooled SD. We interpreted the standardized mean differences following Cohen’s interpretation,¹⁹ whereby 0.2 signifies a small effect, 0.5 a medium effect, and 0.8 a large effect. If data were unavailable or required clarification, the authors of the relevant trial were contacted to provide further information.

Results

A total of 316 articles were identified through database searches and citation tracking (Figure 1 presents a flow chart of the process). Of these, 309 were excluded on the basis of title and abstract. The criteria were then re-applied to the full text of the remaining 7 articles, and a further 4 articles were excluded, 3 because they did not include a measure of physical activity as an outcome and a fourth because the intervention involved increasing the intensity of usual-care physiotherapy.^20–22 Physical activity data were measured, although not reported, in one article;²³ we contacted the author and received these additional data. As a result of this process, three studies were included for review: two randomized controlled trials (RCTs) and one quasi-RCT.^24–26 There was good agreement among the reviewers (κ = 0.66; 95% CI: 0.38, 0.94) when applying the eligibility criteria.

Figure 1.

Preferred Reporting Items for Systematic Review and Meta-Analyses flow chart of included studies.

Quality assessment

Two of the included trials scored 7 and 8, respectively, on the PEDro scale, scores that were considered to be high quality.^24,25 One trial scored 3, which was considered to be low quality (see the Appendix and Table 1).²⁶ There was excellent agreement among the reviewers (κ = 0.868; 95% CI: 0.736, 1.000) when applying the PEDro scale to included articles. The nature of the interventions in the included trials meant that blinding the participants and therapists was not possible; therefore, the maximum possible score that these trials could attain on the PEDro scale was 8. Only two trials used assessors blinded to group allocation,^24,25 and one trial used intention-to-treat analysis.²⁵ Two of the three included trials had random, concealed allocation of participants to groups,^24,25 and all trials reported that groups were similar at baseline.

Table 1.

Study Characteristics

Study	PEDro score	Research design	N	Sex	Age, y, mean (SD)	BMI, mean (SD)	Primary diagnosis, n	Comorbidities, n*	Living arrangements, n	Pre-episode activity levels, n(%)
Healey & Falconer26	3	Quasi-RCT	32	22 women; 10 men	Con,75.1 (6.3); Exp, 79.6 (7.8)	Not reported	Orthopedic, 12; neurological, 8; medical-surgical, 12	Heart/vascular disease, 13; high blood pressure, 15; arthritis, 5; orthopedic impairment, 8; diabetes, 7; COPD-emphysema, 7; mental disorder, 4; cancer, 3; stroke, 4	Home alone, 19; home with family or friends, 13	Independent walking, indoors only, 7; independent walking, indoors-outdoors, 25
Peel ef al.25	8	RCT	255	148 women;, 107 men	Con, 82 (8); Exp, 81 (9)	Con, 24.5 (5.2); Exp, 25.6 (6.7)	Fractures, 88; infections, 47;. neurological, 29; cardiopulmonary, 25†	Con, mean = 8 (SD = 4); Exp, M = 8 (SD = 4)	Not reported	Walking (supervised): Con, 104 (82%); Exp, 111 (87%) Walking (no aids): Con, 11 (9); Exp, 4 (3)
van den Berg et al.24	7	RCT	58	36 women; 22 men	Con, 82(13); Exp, 78(10)	Not reported	Stroke, 3; other neurological,. 3; fracture, 14; other orthopaedic, 7; cardiac-pulmonary, 5; infection, 4; functional medical decline or reconditioning, 8; fall, 6; other, 6	Not reported	Not reported	Uses walking aid: Con, 28 (97); Exp, 26 (90)

^*Unless otherwise indicated.

^†Four most common; ns do not total 255.

^‡Diagnoses were provided for only 56 participants.

RCT = randomized controlled trial; Con = control group; Exp = experimental: group; C0PD = chronic obstructive pulmonary disorder.

Trial characteristics

The three trials included a total of 345 participants (206 women; see Table 1), with a mean age of 75–82 years, who were admitted for either geriatric or neurological inpatient rehabilitation. The interventions implemented were either for a set period of time (5–28 days) or for the duration of the patients’ rehabilitation admission.^24–26

The interventions were heterogeneous in nature. One trial involved a gait training intervention, one used technology to deliver video-based exercise, and one was a behavioural change intervention (see Tables 2 and 3).^24–26 Participants in the experimental group of the Healy and Falconer²⁶ trial took part in a gait training program that aimed to increase the percentage of therapy time spent walking, with a focus on improving endurance, working in functional environments outside the hospital space, and overall independence. The trial by van den Berg and colleagues evaluated the effect of video-assisted exercise versus usual care, with a focus on improving balance outcomes.²⁴ The intervention used commercially available game consoles and purpose-designed rehabilitation software to target participants’ individual goals and functional levels. The trial by Peel and colleagues aimed to increase physical activity by providing individual feedback on the previous day’s physical activity levels as measured by an accelerometer versus the physical activity target.²⁵

Table 2.

Intervention Dosage

Study	Intervention	Frequency	Intensity	Session duration	Progression	Individual or group
Healey & Falconer26	Gait training	2 sessions/d, weekdays, for duration of inpatient stay Length of stay, mean (SD): Con, 20.9 (4.0); Exp, 23.8	N/A	1 h (total of 2h/d)	N/A	Individual
Peel et al.25	Feedback on daily physical activity levels	Daily for 4 wk (28 d)	N/A	N/A	N/A	Individual
Van den Berg et al.24	Video assisted exercises	Daily for total of 5 sessions	Exercises prescribed according to functional level	1 h extra/day	Progression of mobility tasks reported in Box 1 of study	Circuit class format

Con. - control group; Exp. - experimental group; N/A- not applicable.

Table 3.

Intervention Description

Study	Control group	Intervention (type of exercise)	Location	Attendance	Adverse events	Supervision	Equipment used
Healey & Falconer26	Gait training activities made up 50%–74% of treatment time. Patients ambulated outdoors ≤1 time during stay. Transportation to and from therapy in wheelchairs. Close supervision of all walking during gait activities. Patients ambulate <10 min	Gait training activities make up ≥75% of total PT treatment time during stay. Patients walk outdoors ≥3 times during stay. Wheelchairs removed at least 2–3 d before discharge. Patients walk independently ≥3 times (with or without supervision of PT staff) during PT before discharge. Patients ambulate 10–15 min during PT at least 3 times before discharge.	On ward ancl outside	Mean no. of therapy visits: 28.2 (SD 9) per patient	None	By physiotherapist or independently	Journal used by therapist to record activity
Peel el al.25	Usual care, composed of setting mobility goals. Neither staff nor patients received data on walking aims to aid setting walking-time targets.	Accelerometer data downloaded daily. Feedback provided. Mobility goals set, including provisional targets for daily walking time. Goals reviewed weekly and modified on the basis of accelerometer data. Walking times during week summarized in chart form and made available at weekly case conference.	On ward, 24 h/d	Not reported	Not reported	Only for in-therapy walking time, not for non-therapy walking time	Accelerometer
Van den Berg et al.24	Usual care of assessment and management, by a multidisciplinary team.	Usual care and video- or computer-assisted exercises. Participants also wore an activity monitor and received feedback on activity levels.	Purposely designed video-and computer-based, interactive exercise space	39% attended >70% of sessions; 31% attended >80% of sessions; 19% attended >90% of sessions	None serious	1/physiotherapist, 1/ PT assistant	Nintendo Wii Fit; Xbox Kinect; HUMAC balance system; Fitbit Zip

PT - physiotherapy.

Effect on physical activity

The three trials measured physical activity in different ways (see Table 4): as minutes per day spent either walking or being sedentary or participating in moderate to vigorous physical activity.^24–26 Physical activity was the primary outcome measure in one trial and a secondary outcome in two trials.^24–26 Two of the three trials used accelerometers to measure physical activity, whereas Healey and Falconer relied on patients’ reported physical activity levels.^24–26 Studies have shown that when participants respond to self-report outcome measures, they can overestimate the amount of vigorous physical activity they do and underestimate their sedentary time.²⁷ Therefore, physical activity data from accelerometers can provide a more accurate measure of physical activity.^28–30

Table 4.

Post-Intervention Summary of Daily Physical Activity Levels

		Mean (SD)
Measurement (and unit)	Study	Control group	Experimental group
Moderate to vigorous physical activity (% of h awake)	van den Berg et al.24	0.1 (0.1)	0.3 (0.1)
Steps taken (per d)	van den Berg etal.24	1,230(1,010)	1,574(1,796)
Average walking time (x d, min)	Healey & Falconer26	15.4	19.0
	Peel et al.25	In therapy, 3.7 (2.6); outside therapy, 17.3	In therapy, 4.4 (2.6); outside therapy, 24.6
		(17.0)	(16.5)
Sedentary time (% of h awake)	van den Berg et al.24	86.6 (9.3)	84.8 (9.6)

In all three trials, participants in the experimental groups had very low levels of physical activity across the day; they spent the majority of their time being sedentary (Table 4). Participants in the experimental group of the van den Berg and colleagues trial were sedentary for 85% of the day on average and took just over 1,500 steps during their second week of rehabilitation.²⁴ Those in the experimental group of the Peel and colleagues trial averaged a total of just less than 30 minutes of walking per day.²⁵ Participants in the trial by Healey and Falconer averaged 19 minutes of walking a day when they were discharged from rehabilitation.²⁶

Differences between groups

A significant difference between the control and the experimental groups, favouring the latter, was seen in one high-quality trial for time spent walking in therapy sessions (standardized mean difference [SMD] = 0.27; 95% CI: 0.02, 0.52) and outside therapy sessions (SMD = 0.43; 95% CI: 0.19, 0.68; shown in Figure 2).²⁵ No differences were found between groups in the other two trials,^24,26 for any measure of physical activity, including time spent walking, time spent being sedentary, or time spent doing moderate- to vigorous-intensity physical activity.

Figure 2.

Standardized mean differences and 95% CIs of the effectiveness of the interventions to increase physical activity for older adults completing inpatient rehabilitation.

MVPA = moderate to vigorous physical activity.

Discussion

This systematic review has two main findings: (1) there is no evidence that interventions increase physical activity levels in older adults admitted to inpatient rehabilitation when the interventions do not involve increasing the amount or frequency of therapy, and (2) consistent with other literature, older adults admitted to inpatient rehabilitation have high levels of inactivity and spend only small periods of the day participating in physical activity.^5,7,31 Only one high-quality RCT – using a behavioural change intervention – was effective at increasing physical activity in this population. Until further evidence from rigorous RCTs is available, the results of this review may encourage clinicians to engage in behaviour-based techniques when aiming to increase physical activity levels in an inpatient rehabilitation setting. This recommendation is consistent with a recent meta-analysis that found significant increases in the physical activity levels of older adults who wore accelerometers and received feedback or counselling about their physical activity levels.³²

Increasing the physical activity levels of the participants was the primary aim in only one of the three included trials.²⁵ Two trials implemented physical activity interventions, but they were not designed to increase physical activity levels either during rehabilitation or in daily life after the patients were discharged. Rather, they aimed to improve the participants’ walking independence and balance. This may reflect the main aim of inpatient rehabilitation, which is to maximize patients’ safe discharge home. However, this goal is problematic: although inpatient rehabilitation aims to increase the health and function of patients, the long periods of time they spend being sedentary are metabolically deleterious and work against the goals of rehabilitation by having a negative impact on muscles and body systems,³³ resulting in a cycle of patient inactivity and illness. Despite being admitted to rehabilitation to regain functional strength and the ability to return to home and community life, older adults spend most of their time being sedentary, which can undermine their transition to community life.

There are many potential obstacles to increasing physical activity among older adults in a rehabilitation setting that relate to personal factors (a patient’s level of fatigue or fear of falling) and environmental factors (a lack of physical space or lack of staffing). One study reported examples of personal factors that affected physical activity levels; for example, elderly inpatients thought that they should be resting in their spare time and that their level of physical activity was adequate.³⁴ The length of an episode of acute ill health and subsequent recovery time can vary for patients with different diagnoses, and the available literature does not provide a clear picture of whether different diagnoses influence the level of physical therapy during an inpatient stay.

In addition to the rehabilitation environment,³ it is likely that the levels of physical activity that individuals typically complete before being admitted to hospital influence physical activity levels during rehabilitation. A study by Tudor-Locke and colleagues reported the levels of physical activity for people in the community with chronic diseases, showing a wide variation (1,200–8,000 steps per day) according to type of condition.⁹ As seen in the included trials, older adults completing inpatient rehabilitation may have multiple comorbidities, and this fact increases the complexity of their participation and may affect the expected levels of physical activity participation during inpatient rehabilitation. These patients in particular may benefit from individualized assessment and planning to increase their physical activity level and daily step count.¹¹ These issues should be investigated in future studies.

Physical activity guidelines across age groups have two components: (1) maximizing the amount of physical activity people do to maintain or improve their health and (2) minimizing sedentary behaviour.³⁵ For older adults living with chronic disease, including disease that limits mobility, the recommendation is to increase their current level of mobility and build toward 30 minutes of moderate-intensity exercise on most days.^10,36 What this physical activity consists of and how it takes place depends on individuals’ health conditions and the environmental setting in the context of inpatient rehabilitation, and it is likely to be planned with a health professional such as a physiotherapist.

Because of the outcome measures used to measure physical activity in the included trials, comparisons between trials and physical activity guidelines is difficult. Information provided in two trials gives some information on daily physical activity,^25,26 and van den Berg and colleagues presented a detailed description of physical activity patterns across the day.²⁴ Although meeting the physical activity guidelines may not be an appropriate goal for patients with long-term chronic illnesses that limit mobility, it provides a helpful benchmark. As described by Tudor-Locke and colleagues,¹¹ a daily step count of 4,600, a target that includes physical activity over and above performing daily tasks, may be appropriate for older adults with movement-limiting chronic health conditions. This daily step target contrasts starkly with the average daily step count of 1,400 for the participants in the trial by van den Berg and colleagues.²⁴

The one trial that successfully increased the physical activity levels of older people admitted for inpatient rehabilitation was the large, rigorously conducted trial completed by Peel and colleagues.²⁵ It implemented a behavioural intervention with the explicit aim of increasing physical activity levels, although because of the nature of the intervention, therapists and patients were not blinded to group allocation. The between-groups difference of 7 additional minutes of walking time per day was statistically significant, and although the clinical significance of this change is currently unclear, it is likely that an additional 49 minutes of walking per week could confer health benefits on the older adult population.

These data might also be interpreted in the context of convincing and consistent evidence that links prolonged sitting with adverse health outcomes, including premature mortality, and the growing body of research exploring the importance of breaking up sedentary (sitting) time among adults.^37–39 The advice to “move more, sit less, more often” complements the strategy of promoting regular moderate-intensity physical activity by encouraging individuals to reallocate sedentary time to light physical activity. This strategy of reducing and breaking up prolonged periods of sitting has shown the potential in laboratory-based RCTs for important health benefits, such as improved blood pressure control in adults.⁴⁰ This strategy may also be more acceptable to frail older adults than exercise training.⁴¹ However, further studies are required to determine how effective this strategy would be for older adults admitted for inpatient rehabilitation.

Future research may also help clarify whether physical activity guidelines specific to the older adult population during an inpatient rehabilitation admission are warranted. For clinicians who work in health services where additional funding for increasing the intensity of rehabilitation is unavailable but who want to maximize the opportunities for physical activity for individuals completing inpatient rehabilitation, this behavioural feedback on activity levels shows promise as an intervention.

This review has one main limitation: the lack of trials that were eligible for inclusion. Only two of the three included trials were of high quality; therefore, the data from which to draw conclusions are limited. Nevertheless, this review identified and thoroughly assessed the available literature on the physical activity levels of older people admitted for inpatient rehabilitation.

Conclusion

This systematic review revealed a dearth of evidence to help clinicians choose interventions that increase the physical activity level of older adults completing inpatient rehabilitation but that do not involve increasing the amount of time in therapy. This review adds to the body of work confirming that older adults admitted for inpatient rehabilitation spend most of their time being sedentary, a situation that poses risks to their health and recuperation. The results from one trial indicate that a behaviourally focused intervention may positively influence daily physical activity levels in this population. Future trials are needed to help determine effective interventions for increasing physical activity levels among older adults admitted for inpatient rehabilitation.

Key Messages

What is already known on this topic

Although restoring function and activity are the aims of inpatient rehabilitation for older adults, this population has been reported to spend the majority of its time being sedentary. The deconditioning associated with immobility or sedentary behaviour during inpatient rehabilitation can further reduce patients’ overall levels of function. In many workplaces, the frequency and intensity of therapy may be predominantly dictated by resources, including staffing levels, making it difficult to implement additional therapy without additional funding.

What this study adds

There is a lack of evidence for interventions that do not involve increasing the amount or frequency of therapy to increase physical activity levels in older adults admitted to inpatient rehabilitation. Rehabilitation professionals might consider addressing the high levels of inactivity of older adults admitted to inpatient rehabilitation using behavioural techniques. Until further evidence is available, this review may provide encouragement for clinicians engaging in behaviour-based strategies when aiming to increase inpatients’ physical activity levels in a rehabilitation setting.

Appendix: PEDro Scores of INCLUDED TRIALS

Study	Random allocation	Concealed allocation	Groups had similar baseline	Blinding of participant	Blinding of therapist	Blinding of assessor	<15% dropouts	Intention-to-treat analysis	Reported between-groups difference	Point estimate and variability reported	Score (out of 10)
Healey & Falconer26	N	N	Y	N	N	N	Y	N	N	Y	3
Peel et al.25	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
van den Berg et al.24	Y	Y	Y	N	N	Y	Y	N	Y	Y	7

N = no; Y = yes.