Multicomponent rehabilitation to improve independence and functioning in elderly patients with common age-associated diseases: a scoping review

Abstract

Objective

Multicomponent rehabilitation (MR) could restore functioning in elderly patients after hospitalisation, even beyond geriatrics, but specific evidence seems lacking. This review mapped the evidence on MR in elderly patients following hospitalisation for age-related conditions regarding functioning-related outcomes.

Design

Scoping review.

Data sources

PubMed, Cochrane Library, International Clinical Trials Registry Platform and ClinicalTrials.gov (searched through 24 June 2024).

Eligibility criteria

We included randomised controlled trials (RCT) and controlled cohort studies (CCS) comparing centre-based MR with usual care (medical care excluding exercise training) in patients ≥75 years, hospitalised for age-related cardiac, neurological, oncological and orthopaedic diseases. MR was defined as exercise training and at least one additional component (eg, nutritional counselling), starting within 3 months after hospital discharge. RCTs and CCS were included from inception, without language restriction. Care dependency, physical function, health-related quality of life (HRQL) and activities of daily living (ADL) after ≥6 months follow-up were the outcomes of interest.

Data extraction and synthesis

Four reviewers independently screened titles, abstracts and full texts for inclusion and extracted data. MR components and the typology of outcome assessments used were mapped at the final data synthesis level.

Results

Out of 20 409 records, nine studies were investigated in the final data synthesis. Throughout these studies, disease education was the most frequent MR component besides exercise training, while physical function, HRQL and ADL were commonly assessed outcomes. One RCT (cardiac rehabilitation, 80±0.3 years, MR/usual care n=24/23) fully met the inclusion criteria and reported improvements in physical function (2 months) and in HRQL (2, 8, 14 months post intervention) in MR patients.

Conclusions

Evidence on MR regarding functioning-related outcomes in ≥75-year-old patients is sparse beyond geriatrics. There is an essential need for studies investigating the capabilities of MR in this growing and under-represented patient population.

Trial registration number

OSF (https://doi.org/10.17605/OSF.IO/GFK5C).

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The evidence of multicomponent rehabilitation (MR) in patients ≥75 years old was systematically mapped in the main somatic specialties of orthopaedics, oncology, cardiology and neurology beyond geriatrics.

• Functioning-related outcomes such as care dependency, physical function, health-related quality of life and activities of daily living were considered.

• Only one study from the reviewed literature fully met the eligibility criteria, highlighting the fragmentary evidence for MR in patients aged ≥75 years.

• Implications for future clinical trials addressing the research gap regarding MR in the growing population of elderly patients are discussed.

Introduction

Population ageing is a consolidated phenomenon affecting high-income countries, where one in six people will be aged 60 or older by 2030 and the number of people aged 80 or older is expected to triple by 2050.¹ For international and national health systems, this represents one of the greatest challenges of the 21st century, as older people are subjected to age-related health burdens.² The age of 65 years is often used to define older adults,³ but the differences between individuals aged 75 or older and those of younger age (eg, 65–74 years) can be far-reaching and range across physical, cognitive and psychosocial domains.⁴ In adults over 70 years of age, the prevalence of age-related conditions such as cardiovascular diseases, metabolic disorders or chronic respiratory diseases is up to 80%⁵ and comorbidity is common in this population.^{6 7} Thus, hospitalisation rates due to age-related events such as hip replacement, stroke, heart valve surgery or cancer therapy are high among older adults.⁸

Hospitalisation can be a traumatic event for patients ≥75 years, leading to an acute decline in functioning⁹ as a central expression of the individual ability to participate in social life. Functioning comprises the individual’s physical and mental capacities, environmental factors of a person and the interaction between the two. Implications of medical procedures (eg, surgery, interventions) such as scars, polypharmacy or prolonged immobilisation can temporarily impair both physical and mental capacities.10,12 Following hospital discharge, impairments in physiological functions (eg, breathing, circulation) or deficits in body structures (eg, organs, limbs) can lead to muscle weakness, pain, reduced range of motion or fatigue.^{10 13} Mild cognitive impairment and emotional distress (eg, depression) may also manifest. As a consequence, activities of daily living and social participation may be restricted, leading to a higher risk of care dependency and disability if functional health is not regained after discharge from the hospital.1013,15

The ability to regain functioning is directly related to individual mental and physiological reserves and environmental factors that can mitigate deficits (eg, strong social networks, easy access to medical and social care) as defined by the International Classification of Functioning, Disability and Health (ICF) model.¹⁶ Thus, the population of older patients after hospitalisation is characterised by nuances of functional impairments.¹⁷ In the majority of patients, the severity or complexity of impairments does not indicate geriatric intervention, which primarily aims to recover age-appropriate mobility, support self-sufficiency and avoid long-term care in patients affected by age-specific multimorbidity and clinical states like frailty.¹⁸ In any case, geriatric rehabilitation is not standardised and available throughout Europe, much less globally.¹⁹ Regardless of the intervention and care facility, the therapeutic regime of older patients needs to be adapted to and aware of the risks (permanent functional impairment), but more importantly the opportunities (restoring functional health) at a critical time point in the lives of this highly heterogenic population.20,22

Multicomponent rehabilitation (MR) is a suitable approach recommended by the WHO for restoring functioning in older patients.^{2 21 23} Multiple medical professions should be involved in the delivery of MR, ideally including exercise, disease education, psychotherapy and social therapy as well as dietary counselling following a patient-centred, integrated rehabilitation strategy based on the concept of the ICF.1623,25

Nevertheless, appropriate designs and characteristics of MR (beyond geriatrics) to regain functioning and corresponding outcome measures are still uncertain, particularly in patients aged 75 and older, following hospitalisation for an age-associated event or disease. Therefore, we performed the REHabilitation in OLDer patients (REHOLD) scoping review to map the available information concerning MR characteristics, functioning-related outcomes and their assessments in studies aiming to improve the functioning in elderly orthopaedic, cardiac, neurological or cancer patients. We decided to chart the MR components and the typology of the outcome assessments used in order to derive implications for future MR development, evaluation and implementation research projects.

Methods

The REHOLD scoping review was designed according to the Preferred Reporting Items for Systematic review and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR)²⁶ recommendations and based on the guidelines set by the Joanna Briggs Institute (JBI Manual for Evidence Synthesis).²⁷ A comprehensive description of the methodology including eligibility criteria and parameters of the search strategy was previously published.²⁸

Eligibility criteria

We defined a priori eligibility criteria in terms of population, concept, context and types of evidence sources. The population of interest was patients with an age of ≥75 years (mean), hospitalised for acute events (eg, surgery, acute decompensation, stroke) due to age-related diseases (eg, cancer, cardiovascular diseases, osteoarthritis) in the specialties of orthopaedics, oncology, cardiology or neurology. Although it was not specified in the study protocol,²⁸ we decided to set an average age of the study population of ≥75 years instead of a minimum age as an inclusion criterion, in order to cover all relevant studies in the field that would have been excluded when using the stricter age threshold.

We screened randomised controlled trials (RCT) and controlled cohort studies (CCS) with prospective CCS or retrospective CCS designs that examined the effects of MR in comparison to usual care. MR was defined as any type of exercise training and at least one additional component (eg, nutritional counselling, occupational therapy), provided centre-based and starting within 3 months after hospital discharge. Other intervention characteristics, such as duration or frequency, were not specified. Usual care was defined as medical care provided by a general practitioner or specialist (eg, pharmacotherapy), not including exercise training. Investigations assessing the functioning-related outcomes care dependency, physical function, health-related quality of life (HRQL) or activities of daily living (ADL) after ≥6-month follow-up after MR completion were considered.

Studies were checked for subgroup analysis. If a subgroup met the inclusion criteria and outcomes were clustered accordingly, the study was considered eligible and the outcome data were extracted. Exclusion criteria included: patients not hospitalised or hospitalised for diseases in fields other than the indicated specialties, rehabilitation intervention that did not meet our MR definition, studies with a different design (eg, case series) or assessing outcomes other than the predefined.

Search strategy and screening process

The major biomedical electronic databases PubMed (including MEDLINE)²⁹ and the Cochrane Library (CENTRAL), which include trials contained in EMBASE and CINAHL databases, were searched on the 14 June 2022, the 02 November 2023 and the 24 June 2024. In addition, the search was supplemented by exploring the ICTRP Search Platform (WHO) and the public register ClinicalTrials.gov (https://clinicaltrials.gov/). Literature saturation was ensured by manually screening the reference lists of included studies and by retrieving ‘grey literature’ using Google Scholar (https://scholar.google.com/), which were considered on a case-by-case basis.

The REHOLD search strategy adhered to population, concept, context and type of evidence descriptors. Thus, four search components, specialty (A), rehabilitation (B), interventions (C) and population (D), connected by the Boolean term ‘AND’ were created as combinations of Medical Subject Heading-indexations-terms (MeSH). This search strategy was repeated for four specialties: ‘oncology’ (A1), ‘orthopaedics’ (A2), ‘cardiology’ (A3) and ‘neurology’ (A4).²⁸

At each stage of the review process, studies were screened for eligibility by at least two reviewers (OB, TT, HS or MLD). Their decisions were blinded to each other and disagreements were solved by the study guarantor (AS). The reviewers followed a standardised screening procedure for all studies. At every screening stage, first the eligibility of the population, followed by the intervention (MR characteristics), the control group (usual care) and the outcomes based on the predefined criteria were evaluated. In the full-text screening stage, key sections of the publications (eg, baseline demographic for population, methods section for intervention) were screened, while studies in the full-text analysis were investigated thoroughly including supplementary information when available. The software package Citavi (V.6.10; Swiss Academic Software, Switzerland) and the internet-based review production software Covidence (www.covidence.org; Melbourne, Australia) were used for study management and screening. The reasons for exclusion and the number of excluded records were specified according to the PRISMA-ScR recommendations.²⁶

Outcomes

The primary outcome was care dependency following the definition of Boggatz et al³⁰ as ‘a subjective, secondary need for support in the domain of care to compensate for a self-care deficit’.³⁰ It was operationalised by the number of nurse visits at home, admission to a nursing home or by the results of specific care dependency scales such as the Care Dependency Scale for Rehabilitation or by the Northwick Park Dependency Score.^{31 32}

Further, functioning-related outcomes such as physical function, HRQL, ADL, rehospitalisation and mortality were defined as secondary outcomes.^{16 28} HRQL captures the physical and mental health perception of the patients^{33 34} and can be evaluated with a general (eg, Short Form Questionnaire 36-items)³⁵ or disease-specific questionnaire (eg, Chronic Heart Failure Questionnaire).³⁶ Physical function is associated with the ICF dimension of body functions and structures,¹⁶ which commonly has been assessed in terms of physical performance such as mobility (eg, Timed Up-and-Go test), muscle strength (eg, handgrip) or gait speed (eg, 6-minute walk test (6MWT)).³⁷ ADL include basic self-care (eg, bathing, toilet hygiene) and instrumental tasks (eg, preparing meals) of daily living,²¹ which can be evaluated with questionnaires such as the Barthel Index³⁸ or the Lawton Index.³⁹ The primary and/or secondary outcomes should have been assessed after at least 6 months of follow-up subsequent to MR discharge.²⁸

Data charting

Information regarding the study (design, author’s name, year of publication, continent/region, medical specialty), analysed population (total number of participants and of each subgroup, age, sex, referral diagnosis, inpatients/outpatients/mixed), characteristics of the MR programme (eg, number and type of components, duration, number of sessions per week), usual care (eg, type of care, duration, frequency) and outcomes (care dependency, HRQL, physical function, ADL, rehospitalisation, mortality) were extracted.

Data synthesis

Essential information from all studies in the full-text analysis was synthesised narratively into a descriptive overview, although this was not prespecified in the study protocol.²⁸ We found studies reaching the last stage of standardised screening procedure despite minor deviations in the eligibility criteria relevant in terms of the subject of the REHOLD review. Therefore, we conducted this additional step to outline characteristics, particularly MR components and outcome measures, of studies dealing with MR in older patient populations stratified by medical specialties comprehensively. Characteristics of the MR components were listed according to items of the Template for Intervention Description and Replication checklist (TIDieR)⁴⁰ and stratified for medical specialties. The types of assessment tools used for outcome measurement were charted and stratified into clinician-assessed or patient-reported outcome measures. Study characteristics, MR components and outcome measurements were summarised and structured into tables.

Study quality assessment

The risk of bias was assessed using the Cochrane Risk-of-Bias tool⁴¹ for RCTs and the Risk Of Bias In Non-randomised Studies of Interventions⁴² for controlled cohort studies. The overall risk of bias of the studies eligible for final data synthesis is displayed in a summary table as recommended in the assessment manuals of the tools.

Patient and public involvement

None.

Results

Search results

The search generated a total of 33 557 records (figure 1), which combined 27 276 hits from databases and 6281 from study registers. After the removal of duplicates, the title and abstract of 20 409 studies were screened for eligibility. Of these, 19 184 studies were excluded and the remaining 1225 underwent full-text screening. Most of the studies involved oncological (406 studies) and neurological (366 studies) patients, followed by the medical specialties of orthopaedics (235 studies) and cardiology (213 studies).

Figure 1. Preferred Reporting Items for Systematic review and Meta-Analyses study selection flowchart. Green boxes indicate studies included in the final data synthesis (eligibility criteria predominantly or entirely met). MR=multicomponent rehabilitation.

During the full-text screening, the majority of the records (n=1216) were excluded due to an ineligible patient population (822 records), intervention (92 records), control group (43 records), outcomes (21 records), other reasons (eg, study design; 49 records) or were study protocols (189 records).

In the final data synthesis of nine studies,43,51 only the study of Marchionni et al⁴³ fully met the inclusion criteria (table 1). This RCT enrolled 270 patients (67% men) after myocardial infarction, who were randomised into three groups: hospital-MR, home-MR and usual care, clustered by age (middle-aged 45–65 years, old 66–75 years, very old >75 years). Physical function and HRQL were assessed at admission to MR, with follow-up measurements at 8 and 14 months.⁴³ Of the remaining eight studies, three^{44 45 47} investigated unmonitored rehabilitation programmes, three studies^{48 50 51} had a follow-up time of less than 6 months, one study⁴⁹ investigated a patient population aged <75 years (mean) and in one study⁴⁶ usual care included physical exercises (table 1). Detailed study characteristics are displayed in the extended table in the online supplemental appendix.

Table 1. Main characteristics of the studies eligible for the final data synthesis.

Studya: Author (year)b: Countryc: Medical specialtyd: Design	Patientsa: Diagnosisb: N. of included patients (n)c: Age mean±SD (range) in yearsd: Sex n (%)	MR (supervised exercise training+additional components)a: N. of patients (n)b: Components characteristicc: Setting	Control groupa: N. of patients (n)b: Characteristic	Outcomesa: Follow-up periodb: Care dependencyc: Physical functiond: HRQLe: ADL	Results (N. of patients included)benefits of MR (significant improvement and/or difference)
a: Marchionni et al43 (2003)b: Italyc: Cardiologyd: RCT	a: Myocardial Infarctionb: n=270c: Group ‘very old’ (>75 years)=80±0.3d: Female Group ‘very old’ n=10 (40) total cohortn=86 (32)	a: IG n=24 (Hosp-CR group)b: Exercise training (endurance, stretching), counselling (cardiovascular risk factors management), support group meetingsc: Hospital	a: CG n=23(no-CR group)b: One education session; referred to family physician	a: 8 and 14 monthsc: TWCd: SIP	∆ baseline - 8 and 14 months (IG n=24; CG n=23)TWC: improved (p<0.001) in IG at the end of MR (2 months), but went back to baseline values at follow-up (14 months).SIP: improved significantly in IG (p<0.01) over the entire study duration (2, 8, 14 months).Effect estimates or CI not reported.
a: Witham et al44 (2012)b: UKc: Cardiologyd: RCT	a: Heart failureb: n=107c: 80.4±5.8d: Femalen=35 (33)	a: n=53b: (I) Exercise sessions, (II) counselling and educationc: Not supervised	a: n=54b: Educational booklet	a: 6 monthsc: 6MWT, Sit-to-Stand, Activity counts, Strength, FLPd: EQ-5D, MLWHF, HADS	∆ baseline - 6 months (IG n=43; CG n=44)Benefits (significant difference) in IG compared to CG only for Sit-to-Stand (−6.4; 95% CI −12.2, –0.6 s; p=0.03 adjusted*).
a: Xueyu et al45 (2017)b: Chinac: Cardiologyd: RCT	a: CHF (NYHA II/III)b: n=78c: IG=78±3; CG=76±4d: FemaleIG n=10 (25)CG n=13 (33)	a: n=40b: (I) Exercise training programme, (II) transitional care programmec: Not supervised	a: n=38b: Counselling on physical activities	a: 12weeksc: 6MWT, TUGd: MLWHF	∆ absolute values at baseline and 12 weeks (IG n=40; CGn=38)Benefits (significant improvement) in IG for 6MWT, TUG, MLWHF.
a: Thingstadt et al46 (2016)b: Norwayc: Orthopaedicsd: RCT	a: Hip fractureb: n=397c: IG=83.4±5.5, CG=83.2±6.4d: FemaleIG n=145 (73)CG n=148 (74)	a: n=198b: (I) Early mobilisation, (II) education, multidimensional assessmentc: Hospital	a: n=199b: Mobilisation, physiotherapy, exercise, written information	a: 12 monthsc: Gait characteristics (eg, speed, postural control)e: Barthel Index, NEADL	∆ IG (n=139) - CG (n=143) at 12monthsBenefits (significant difference) in IG compared with CG only for gait speed, postural control, NEADL mobility subscale.
a: Tseng et al47 (2021)b: Taiwanc: Neurology, orthopaedicsd: RCT	a: Hip fracture and cognitive impairmentb: n=152c: 82±7.1d: Femalen=118 (78)	a: n=76b: (I) Exercise programme, (II) geriatric assessment, educationc: Home training	a: n=76b: In-hospital education	a: 6 and 12 monthsd: SF-36, EQ-5De: Self-care ability (Barthel Index, IADL)	∆ absolute values at 6 and 12 months (IG n=76; CG n=76)Benefits (significant improvement) in IG only for EQ-5D.
a: Olsson et al51 (2007)b: Swedenc: Orthopaedicsd: Prospective cohort	a: Hip fractureb: n=112c: 84±7 (65–97)d: Femalen=83 (74)	a: n=56b: (I) ADL training, early post-operative ambulation, (II) educationc: Hospital	a: n=56b: Usual care*	a: 11–46 dayse: ADL Katz Index, walking aids in/dependence	∆ absolute values at baseline (prefracture) - 46 days (IG n=53; CG n=56)Benefits (significant improvement) in IG only for ADL Katz Index (rank sum comparison), independence from walking aid (∆ between groups).
a: Miller et al48 (2006)b: Australiac: Orthopaedicsd: RCT (four groups)	a: Fall-related lower limb fracture, malnutritionb: n=100c: 83.5d: Femalen=79 (79)	a: n=24b: (I) Exercise, (II) nutritional supplementationc: Hospital	a: n=26b: Education (eg, nutrition, exercise advice)	a: 12weeksb: Admission rate to higher level carec: Gait speed, strengthd: SF-12e: Higher level of care	∆ baseline - 12 weeks (IG n=24; CG n=26)No benefits (no significant differences) for IG compared with CG in all parameters.
a: Bulthuis et al49 (2007)b: Netherlandsc: Orthopaedicsd: RCT	a: Hip (n=52)/knee (n=35) others (n=11)b: n=98c: 68±11d: FemaleIG n=31 (78)CG n=46 (79)	a: n=58b: (I) Physiotherapy, (II) educationc: Resort with professional care	a: n=40b: Referred to physician, physical therapy or temporary admission to nursing homes	a: 6 and 12 monthsc: EPM-ROMd: RAND-36e: HAQ, MACTAR	∆ absolute values baseline, 6 and 12 months (IG n=58; CG n=40)Benefits (significant improvement) for IG only for EPM-ROM and MACTAR total score.
a: Taricco et al50 (2014)b: Italyc: Neurologyd: Prospective cohort	a: Strokeb: n=229c: IG=71.8±10.5 CG=70.1±10.7d: FemaleIG n=41 (33)CG n=41 (40)	a: n=126b: (I) Adapted physical activity (II) therapeutic patient educationc: Hospital	a: n=103b: Written education; two follow-up visits with a physician	a: 4monthsb: CSIc: 6MWT, SPPB, BBS and Motricity Indexd: SF-12e: Modified Barthel Index	∆ baseline - 4 months (IG n=99; CG n=100)Benefits (significant improvement) in IG and compared with (significant difference) CG for 6MWT (gait endurance, gait velocity) and SPPB summary score.Benefits (significant improvement) in IG for BBS score, SF-12 (physical and mental components), no difference between groups.

Colour scheme: grey=basic study information; green=fully met eligibility criteria; red=deviation from inclusion criteria (bold=specification of deviation); blue=study results.; # further specified in the study;

(Full data available for each parameter are provided in the extended table in the online supplemental appendix).

^*Not further specified in the study.

ADL, activity of daily living; BADL, basic activities of daily living; BBS, Berg Balance Scale; CG, control group; CHFChronic Heart Failure QuestionnaireCSICaregiver Strain IndexEPM-ROM, Escola Paulista de Medicina–Range of Motion scale; EQ-5D, EuroQoL-5D; FLP, Functional Limitations Profile; HADS, Hospital Anxiety and Depression Score; HAQ, Health Assessment Questionnaire; HRQL, health-related quality of life; IADL, instrumental activities of daily living; ICF, International Classification of Functioning; IG, intervention group; MACTAR, McMaster Toronto Arthritis Patient Preference Disability Questionnaire; MLWHF, Minnesota Living With Heart Failure Questionnaire; MRmulticomponent rehabilitation6MWT, 6-minute walk test; NEADL, Nottingham Extended Activities of Daily Living Scale; NYHA, New York Heart Association; PA, physical activity; RAND-36, RAND 36-Item Health Survey 1RCT, randomised controlled trial; SF-12, Short Form Questionnaire 12-items; SF-36, Short Form Questionnaire 36-items; SIP, Sickness Impact Profile; SPPB, Short Physical Performance Battery; TUG, Timed Up-and-Go test; TWC, total work capacity

Out of the nine studies, seven were RCTs43,49 and two were observational investigations^{50 51} with a prospective design. The patient cohorts were mainly from European regions (Italy,^{43 50} Norway,⁴⁶ Sweden,⁵¹ Netherlands,⁴⁹ UK⁴⁴), Asia (Taiwan,⁴⁷ China⁴⁵) and Australia.⁴⁸ Orthopaedic patients were included most frequently (five studies), mainly after fall fractures and hip or knee replacements.46,4951 Patients with myocardial infarction,⁴³ heart failure,⁴⁴ chronic heart failure⁴⁵ or stroke⁵⁰ were included in four studies, respectively. The studies were characterised by a moderate to high overall concern of bias (online supplemental appendix).

Multicomponent rehabilitation characteristics

The MR programmes lasted between 2 and 24 weeks with a median duration of 10 weeks. Besides exercise training (which was set as an inclusion criterion), disease education was the most common rehabilitative component included in the MR programmes of seven studies43,4648 49 51 (table 2). Endurance and resistance exercises were the predominant forms of training. In the study of Marchionni et al,⁴³ endurance training, stretching as well as flexibility exercises in combination with cardiovascular risk factor management and emotional support (as a group discussion) were the components provided to the hospital-MR group.

Table 2. Overview of the rehabilitation components of the studies in final data synthesis by medical specialty.

Medical specialty	Rehabilitation components
	Exercise training	Physiotherapy	Disease education	Emotional support	ADL-training	Nutritional counselling
Cardiology (n=3)43,45	n=343,45		n=343,45	n=343,45
Orthopaedics (n=5)46,4951	n=546,4951	n=347 49 51	n=347 49 51		n=151	n=148
Neurology (n=1)50	n=150		n=150

Colour scheme: green=rehabilitation component was implemented; red=rehabilitation component was not implemented.

Rehabilitation components definition: Eexercise training=interventions involving physical activity and respecting training principles; physiotherapy=physical therapy measures besides exercise training; Ddisease education=all formats of information on the disease/therapy given by members of the rehabilitation team; Eemotional support=caregiver/family involvement, group discussions; ADL-training=training of activities of daily living (eg, grocery shopping); Nnutritional counselling=counselling including supplementation.

The frequency of training sessions ranged from several times daily in orthopaedic patients (knee arthroplasty or arthritic flare-up,⁴⁹ hip fracture^{47 51}) to two times per week in patients with cardiovascular or neurological diseases (heart failure,⁴⁴ stroke⁵⁰). The duration of a single exercise session varied between 0.5 hours^{43 45 47 48} and 1.5 hours.⁴⁴ Studies implementing endurance training reported its intensity in different ways: ‘mild to moderate’,⁴⁴ 70–85% of the maximum heart frequency (MaxHF) achieved at the baseline exercise test,⁴³ 60% of the estimated MaxHF⁴⁹ and 10 bpm above the resting heart rate or, respectively, 11–13 out of 20 points on the Borg Scale.⁴⁵ Studies that implemented strength or resistance training failed to provide information on their targeted intensities.4447,50

Physiotherapy, such as early postoperative ambulation or mobilisation, was implemented in three^{46 49 51} of the five orthopaedic studies in the final synthesis. Furthermore, one study in orthopaedics⁵¹ used ADL training (eg, grocery shopping simulation) to regain patient’s independence, another⁴⁸ applied a nutritional oral supplementation (1.5 kcal/mL, 16% protein, 35% fat and 49% carbohydrate).

Outcomes of multicomponent rehabilitation

Care dependency, defined as admission to a higher level of care, was considered in only one⁴⁸ study in the final data synthesis (table 3), while physical function, HRQL and ADL outcomes were commonly captured. Physical function was assessed predominantly according to clinician-reported outcome measurements (CROMs), generally evaluating gait (eg, 6-metre walk test; 6MWT),^{44 45 50} strength (quadriceps strength),^{44 48} endurance (total work capacity; TWC),⁴³ balance (Berg Balance Scale; BBS)⁵⁰ or complex motoric-tasks (Timed Up-and-Go; TUG).^{45 50} HRQL was exclusively assessed using patient-reported outcome measures (PROMs) like the Short-Form Questionnaire 12-Items^{48 50} or the Minnesota Living With Heart Failure Questionnaire.⁴⁵ ADL were evaluated by clinician judgement using the Barthel, Katz or Lawton indices,^{46 47 50} although patient-administered surveys like the Nottingham Extended Activities of Daily Living Scale⁴⁶ or the Health Assessment Questionnaire⁴⁹ were also conducted.

Table 3. Functioning-related outcome assessments used in the studies in the final data synthesis.

	Factual data or clinician-assessed outcome measures	Patient-reported outcome measures
Care dependency	Admission rate to higher levels of care48	CSI50
Physical function	6MWT44 45 50Quadriceps strength44 48Total work capacity43TUG45Motricity Index50Sit-to-Stand Test44Gait characteristics46SPPB50BBS50PA level44EPM-ROM49	FLP44
Activities of daily living	Barthel-Index (modified)46 47 50BADL (Katz Index)51IADL (Lawton Index)47	NEADL46HAQ50MACTAR50
Health-related quality of life		SF-1248 50SF-3647MLWHF44 45EQ-5D44SIP43HADS44RAND-3649
Rehospitalisation	Total number of rehospitalisations44 48 51
Mortality	Total number of deaths44 46 47 51

ADL, activities of daily living; BADL, basic activities of daily living; BBS, Berg Balance Scale; CSI, Caregiver Strain Index; EPM-ROM, Escola Paulista de Medicina–Range of Motion scale; EQ-5D, EuroQoL-5D; FLP, Functional Limitations Profile; HADS, Hospital Anxiety and Depression Score; HAQ, Health Assessment Questionnaire; HRQL, health-related quality of life; IADL, Instrumental Activities of Daily Living; MACTAR, McMaster Toronto Arthritis Patient Preference Disability Questionnaire; MLWHF, Minnesota Living With Heart Failure Questionnaire; 6MWT, 6-minute walk test; NEADL, Nottingham Extended Activities of Daily Living Scale; PA, physical activity; RAND-36, Research and Development 36-Item Health SurveySF-12, Short Form Questionnaire 12-items; SF-36, Short Form Questionnaire 36-items; SIP, Sickness Impact Profile; SPPB, Short Physical Performance Battery; TUG, Timed Up-and-Go test

No effects of MR on care dependency were determined, with similar rates of admission to a higher level of care in the MR and control groups (table 1).⁴⁸ Marchionni et al⁴³ reported an increased physical function defined by TWC in the intervention group at the end of MR (baseline vs 2 months). This effect was not sustained toward the follow-up period, while HRQL (Sickness Impact Profile) improved over the entire study duration (baseline vs 8 and 14 months). However, the authors did not report effect estimates or CIs. In general, outcomes of physical function improved in patients who participated in MR, with significant changes in gait parameters (eg, 6MWT), strength, balance (eg, BBS) and mobility (eg, TUG).44,4649 Contradictory effects were observed for HRQL: Three studies^{44 48 49} revealed no differences between MR and the control group, two^{45 50} found beneficial effects of MR on general and disease-specific HRQL. Two studies^{49 51} reported a decrease in ADL impairment in MR-patients, while one study⁴⁷ found no difference compared with a control group.

Discussion

In the REHOLD scoping review, we mapped the available evidence of MR on independence and other functioning-related outcomes of patients aged 75 years and older after hospitalisation for age-associated diseases in four main medical specialties. Out of more than 1200 full-text screened studies, we reviewed nine in detail. We identified one study—a randomised trial reporting short-term but not lasting beneficial effects of MR for physical function and suggesting an improved HRQL in post-myocardial infarction patients—that fully met our inclusion criteria regarding population, the content of MR or comparator intervention and aspects of study design.⁴³ In the narrative synthesis of the nine studies in the full-text analysis, disease education was the most common MR component alongside exercise training, a predefined eligibility criterion. Physiotherapy, ADL training and nutritional counselling were applied only in orthopaedics, while emotional support interventions were specific to studies in cardiology. Both clinician-assessed outcomes and patient-reported measures were used to assess functioning. Physical function, HRQL and ADL were most frequently assessed, though a wide range of tests was used (eg, 12 different tools for physical function).

Under-representation of older patients: age-related issues

Based on the results of our screening process, the majority of the reviewed studies were excluded due to an ineligible population. Generally, the population of older patients aged 75 years and above remains under-represented in interventional studies, for example, due to excessively strict exclusion criteria, even in those investigating rehabilitation programmes in patient groups with age-associated disease.⁵² Setting age cut-offs in clinical trials and reviews can and should be discussed since functioning is individual and cannot be directly attributed to age. With the help of age cut-offs in a study, researchers can ensure that previously under-represented age groups are included and investigated. This mandatory representation of certain age groups can be necessary since study populations at higher risk like elderly or minors are often excluded from clinical studies. For example, a recent Cochrane review assessing the impact of exercise-based cardiac rehabilitation in adults after valvular surgery found that almost all included trials recruited highly selected populations of younger participants with low to moderate risk.⁵³ He et al⁵⁴ investigated in a systematic review the exclusion rates in about 300 RCTs on treatments (including rehabilitation) for physical conditions such as cardiovascular disease, diabetes and cancer. The authors observed an exclusion quota of over 75% of patients due to strict exclusion criteria such as age limit or multimorbidity.⁵⁴ On both sides, researcher and participant, several identified barriers lead to the frequent exclusion of elderly patients. Psychological factors (eg, fear of physical exertion) or transportation issues may represent obstacles in elderly patients, particularly after hospitalisation when their functional capacity might be temporarily compromised.^{52 55 56} From the researcher’s perspective, the presence of comorbidities, the bias of ageism towards elderly patients, the lack of adequate insurance, the additional time and cost expenses or factors like the knowledge about higher risk of adverse events and drop-out rates frequently leads to the decision to exclude elderly patients.^{51 52 55}

On the other hand, age cut-offs can be seen critically as they do not discriminate regarding impairments. The age group of patients >65 years is characterised as very heterogenic concerning their functional status,^{18 57} ranging from completely independent to highly impaired and care-dependent. One useful approach might be to define the eligible population through impairment or (in-) dependence scores, such as the ADL Katz Index, combined with a subgroup analysis of study results stratified for different age groups.

However, this review referred to the age aspect defining a corresponding inclusion criterion rather than depicting the population by impairment. In line with previous reviews on elderly patients in clinical trials,^{51 54} REHOLD’s findings clearly show that elderly patients are understudied in rehabilitation research and that there is a lack of information on the implementation of MR to improve their functioning.⁵⁸

Content of multicomponent rehabilitation

Our definition of MR follows the WHO recommendations of MR that should be provided by a multidisciplinary workforce.⁵⁹ We considered exercise training a key component of MR due to its positive effects on physical performance (eg, gait speed, maximal strength), cognitive capacity and emotional distress, which contributes to restoring functioning in older adults.5360,62 In all the studies analysed in the final data synthesis, exercise training was mainly based on several endurance and resistance training sessions per week performed according to standardised protocols. The overall reporting quality regarding exercise intervention parameters such as method (eg, differentiation of hypertrophy or strength endurance training), targeted intensity (eg, % of one-repetition maximum) or equipment was poor, with studies providing fragmentary information or failing to report any details other than the type of exercise (eg, ‘endurance training’). Beside exercise, counselling with a focus on risk-factor management was predominantly implemented within cardiology, orthopaedics and neurology. The counselling offered ranged from a simple informational booklet to individual sessions and group educational seminars. Positive effects of counselling on disease-related anxiety, depression and general well-being in elderly patients with rheumatoid arthritis, stroke, osteoporosis or cardiovascular diseases were confirmed in a systematic review.⁶³ Additionally, physiotherapy (mostly passive mobilisation), ADL training or nutritional counselling including nutritional supplements were also implemented in the reviewed MR protocols.

Overall, our observations illustrate a heterogeneous understanding of MR between the screened studies across and even within the individual medical specialties considered. The differences in the implementation of MR interventions are likely attributable to the different healthcare systems of countries, policies of health service providers and insurance companies, study-related aspects or lack of specific standards such as medical guidelines. This review found heterogeneous MR components used in studies within the same medical specialty (eg, orthopaedics). The heterogeneity of the MR components could result from differences in the organisation of rehabilitation services in the context of the national healthcare system in the individual countries where the studies were conducted (eg, Norway, Sweden, the Netherlands). Health policy regulations may affect the rehabilitation delivery or the multiprofessional or interprofessional nature of MR services. For example, in many European countries, rehabilitation is predominantly carried out on an outpatient basis and rehabilitation facilities are not very specialised for specific indications.⁶⁴ The use of a standardised rehabilitation definition, as provided by Cochrane Rehabilitation,⁶⁵ is highly recommended to develop and evaluate MR measures for tailored care of patients aged 75 years or older and to generate sufficient evidence.

Reported outcome measurements

We observed that aspects of functioning such as physical function, ADL and HRQL were most frequently assessed after MR, where the outcome measurements could be differentiated into PROMs or CROMs. The used tools are predominantly validated, reliable and well-established throughout rehabilitation research in older patients.⁶⁶ A broad set of CROMs of physical function addressing specific performance capacities such as endurance (eg, 6MWT, TWC), strength or mobility (eg, TUG, Short Physical Performance Battery (SPPB)) was carried out in the reviewed studies. For example, the 6MWT was used to investigate gait characteristics (eg, gait speed) in two studies, quadriceps strength was also measured in two studies as well as TUG or SPPB assessed mobility and balance, respectively. The interpretation of these findings needs to recognise the different focus of the four medical specialties in relation to physical function (eg, endurance in cardiology, movement quality in orthopaedics).

Assessing care dependency or the ability to perform ADL represent another valid operationalisation of functional health in the elderly.⁶⁷ Yet, the study meeting inclusion criteria did not assess care dependency or ADL performance. Only four out of nine studies in the final data synthesis investigated ADL, each using a different measurement tool (Katz Index, original or modified Barthel Index, Nottingham Extended ADL Scale). Indeed, there is no consensus on the measurement of functioning in older hospitalised patients.^{67 68}

Implications for research

Our results show a deficiency of information about MR in previously hospitalised ≥75-year-old patients throughout all four investigated specialties (orthopaedics, oncology, cardiology or neurology). Well-designed controlled studies are needed to close this research gap. As the intervention itself, concomitant research should be based on the WHO principles of a patient-centred, integrated rehabilitation strategy grounded in the ICF concept of functioning.^{16 65} The characteristics of the population to be included, the rehabilitation intervention, and the outcomes to be assessed should be carefully predefined. Figure 2 provides a simple scheme for introductory considerations of key study design characteristics following the ICF.

Figure 2. ICF-compliant key design characteristics of studies on multicomponent rehabilitation in elderly patients. Based on the International Classification of Functioning, Disability and Health (ICF), rehabilitation provides a holistic approach focusing on patients’ functioning, as should be reflected in appropriate study characteristics. ADL, activities of daily living; HRQL, health-related quality of life.

In order to gain insights for a specific population like patients aged 75 or older, it is mandatory to investigate this population group specifically. Researchers may consider various approaches to represent target age groups, for example, using age as a stratification factor in randomised and non-randomised studies in elderly patients (≥65) with planned appropriate subgroup analyses. People over the age of 75 often suffer from age-related syndromes (eg, sarcopenia), have multiple conditions and polypharmacy with changing metabolism and further physiological systems, leading to altered effectiveness of therapeutic interventions.⁵ At the same time, in the context of functioning, the goals of rehabilitation may evolve with the age of the recipients. Hence, outcome parameters (eg, physical capacity vs care dependency) and their operationalisations (eg, bicycle ergometry vs gait speed testing) should be chosen accordingly.^{69 70} Rehabilitation programmes should be based on a holistic patient-centred approach and informed by a pertinent theoretical framework (eg, the Medical Research Council’s Guide to the Development and Evaluation of Complex Interventions or the Cochrane definition of rehabilitation).^{65 71} Involving different professions such as sports scientists in the research process and using standardised reporting tools like the TIDieR checklist⁴⁰ might improve the study and reporting quality. Also, adverse events (AE) should be reported more consistently (eg, Consolidated Standards of Reporting Trials (CONSORT) Harms extension,⁷² Good Clinical Practice guidelines⁷³) to help minimise researcher’s bias towards excluding older patients. Overall, the reporting level on AE was insufficient in the investigated studies. Out of nine studies, four failed to report any information on AE.4347,49 Five studies did report on AEs44,4650 51 but only three used the correct term44,46 and none of the studies reported in a standardised way (eg, following CONSORT Harms checklist⁷²). A follow-up of 6 months or longer should be preferred, as one of the WHO’s principles of MR emphasises the long-term recovery of functional health, allowing older patients to re-engage in meaningful life roles.²⁴

Limitations

The restrictive nature of the inclusion criteria may have served as a limiting factor for study inclusion. In particular, setting the population age of interest to a mean of 75 years or older might have limited the number of eligible studies. In detail, 822 out of the total 1216 study exclusions during the full-text screening stage were accounted to an ineligible population, mostly due to the age criterion. Although subgroup analyses were incorporated, the intended population was a crucial factor, leading to an extensive exclusion of studies. However, while patients are commonly classified as old from the age of 65,³ this chronological definition does not reflect the functional health deterioration that increases with age. A higher prevalence of multimorbidity, cognitive decline, functional limitations and pathophysiological deterioration is observed in patients aged 75 or older.⁷⁴ Consistently, an increased need for care occurs by age ≥75 compared with the 65–74-year-old,^{3 75} implying different demands on rehabilitation to restore functional health for the age groups described. Based on these considerations, we have defined the age criterion in this review.

Conclusion

Evidence on MR (beyond geriatrics) regarding functioning-related outcomes such as care dependency, physical function, ADL or HRQL in patients ≥75 years old is sparse. The lack of evidence makes it impossible to draw any conclusion on the effects of MR in this population. With the REHOLD scoping review, we note an urgent need for research focusing on the impact of MR on the functioning of older patients with rehabilitation needs for common age-associated neurological, oncological, cardiac or orthopaedic diseases.

Data availability statement

Data sharing is not applicable to this article as no new data beyond the presented were generated in this study.