Background

In March 2020, a global pandemic was declared with the emergence of COVID-19, an infectious disease, viral by aetiology and caused by the SARS-CoV-2 virus [1]. As of the 17th of March 2023, the World Health Organisation (WHO) reported 760,360,956 confirmed cases and 6,873,477 deaths globally [2]. Common symptoms of COVID-19 include fever, dry cough, and fatigue; less commonly people experience headache, dizziness, abdominal pain, nausea, and vomiting [3].

Older age and male gender place people at higher risk for disease severity [3,4,5,6] and a poorer prognosis [4, 7]. Those with other underlying health conditions namely cancer, obesity, chronic kidney disease, chronic lung disease, cystic fibrosis, dementia, diabetes, people with disabilities, heart conditions, HIV infection, and those who are immunocompromised are also at greater risk of severe illness [8,9,10]. Given that over 50% of those aged over 65 have two or more chronic health conditions [11], it would suggest that older persons are at significant risk for COVID-19 disease severity.

Between 13.9 and 43% of patients infected with COVID-19 develop long term symptoms, with fatigue and memory difficulties or brain fog amongst the most common [12, 13]. Additionally, the quality of life (QOL) of those post COVID-19 is significantly impacted regardless of the time since discharge or recovery and older age and co-morbidities are among the most frequently reported factors associated with low levels of QOL post COVID-19 [14]. Worse mobility and functional outcomes have also been identified in older adults admitted to hospital due to COVID-19 [15] and in older adults with mild to moderate COVID-19 who did not require hospitalisation [16].

In the early months of the pandemic there was a dearth of literature describing the rehabilitation needs of people recovering from COVID-19 and the efficacy of interventions [17]. Since this time, the body of evidence has grown significantly to include longitudinal studies exploring clinical progression, symptoms, and rehabilitation recommendations [18,19,20].

The WHO’s living guideline on the clinical management of COVID-19 recommends screening for rehabilitation needs throughout the recovery process [21] and both the WHO guidelines and the National Institute for Health and Care Excellence (NICE) guidance document for the management of Long Covid [22] recommend multidisciplinary input given the virus’ impact on several body structures and functions, and its long-term sequelae. These guidelines reflect other COVID-19 rehabilitation guidelines developed for clinicians of specific disciplines [23,24,25,26,27] and the European Geriatric Medicine Society (EuGMS) guidance [28]. The use of comprehensive geriatric assessment, long term follow-up and ongoing monitoring of patients following discharge from rehabilitation for COVID-19 is also advised by EuGMS, with suggested time points of 6 weeks and 6 months [28].

Studies have evaluated multidisciplinary team (MDT) rehabilitation for various groups including adults with severe-to-critical illness in intensive care units [29] and those adults post intensive care [30]. Older adults face increased risks for COVID-19 severity and poorer prognosis. While the literature supports multidisciplinary rehabilitation for adults hospitalised with COVID-19, little is known yet about how MDT rehabilitation in this group impacts outcomes. To date, there are no randomised controlled trials or analytical cohort studies published exploring the effect of MDT rehabilitation on older adult outcomes following hospitalisation for COVID-19. However, several observational studies have described the rehabilitation outcomes of older adults with COVID-19 following MDT intervention. There is a need to profile the clinical characteristics, functional and process outcomes of older adults who have undergone MDT rehabilitation in the acute or post-acute inpatient hospital setting to inform the development and response of services in the future and to guide the development of trial studies. This systematic review aims to explore and synthesise the totality of evidence regarding the outcomes of older adults with COVID-19 who have undergone MDT intervention in the acute or post-acute inpatient setting. The author hypothesises that older adults with COVID-19 will have improved function following completion of MDT rehabilitation.

Methods

Study design

The conduct and reporting of this systematic review of observational studies is in accordance with the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) guidelines [31], see Additional file 1. The protocol for this systematic review has been registered on the PROSPERO register (PROSPERO ID = CRD42022341365).

Search strategy

The searches were conducted on the 1st of June 2022 of the following databases: Cochrane library, EMBASE, Cinahl and Medline (via EBSCO), PubMed, and Web of Science by the first author. Reference lists of eligible studies were also checked. Literature was limited to publications from March 2020 to the date of search completion and limited to English language full text. The search was completed by AMC, Master of research candidate at the University of Limerick. A repeat search was conducted on the 17th of March 2023 to identify additional papers published between the initial and repeat search date.

The following MeSH terms and associated keywords covering three concepts were used;

  • COVID-19

  • Multidisciplinary rehabilitation

  • Hospital setting

Appropriate synonyms were compiled to identify all appropriate studies. See Additional file 2 for search terms and synonyms.

Eligibility criteria

Studies meeting the following criteria were included:

  • Population: Older adults (with mean or median age of 65 or greater) with a diagnosis of COVID-19.

  • Study design: Prospective and retrospective descriptive cohort studies, comparison groups of experimental studies including randomized controlled trials, quasi randomized studies or controlled before after studies, case series (with more than 1 participant), and the ‘cases’ in case control studies.

  • Intervention: multidisciplinary (MDT) rehabilitation provided by two or more Health and Social Care Practitioners (HSCP) including but not limited to the following disciplines in the inpatient setting: Occupational Therapy, Physiotherapy, Speech and Language Therapy, Human Nutrition and Dietetics, Psychology and/or Medical Social Work.

Outcomes

The primary outcome for the study was any validated measure of functional ability that reflect activity limitations and participation restrictions in keeping with the International Classification of Functioning e.g., Barthel Index, or Functional Independence Measure.

Secondary outcomes included:

  • Discharge disposition e.g., discharge directly home, long term care, transitional care, and/or to the care of a family member

  • Hospital length of stay (LOS)

  • Mortality

  • Primary/Community and secondary healthcare utilisation (unplanned ED return, unscheduled hospital admission)

  • Long term effects of COVID-19 i.e. signs and symptoms reported during the post COVID-19 phase for example fatigue, headache, attention disorder, hair loss and dyspnoea [32].

Studies were excluded if they met any of the following criteria:

  • Population: Persons with COVID-19 with mean or median age of < 65 years.

  • Study design: The control arm of experimental or analytical observational studies where MDT intervention has not been implemented, and cross-sectional studies.

  • Intervention: Studies reporting outcomes following uni-disciplinary interventions, pulmonary rehabilitation only or papers describing medical interventions only. Studies reporting outcomes following rehabilitation only in the Intensive Care Unit were also excluded.

Data extraction

Studies obtained through the search strategy were reviewed and duplicates removed in Endnote. Remaining studies were then exported to Rayyan for initial screening by the first author (AMC). Rayyan is a web-based platform that facilitates the methodical and efficient screening of search results by title and abstract. It allows researchers to allocate labels to explain reasons for exclusion facilitating transparency in the systematic review process [33]. One third of included articles were independently reviewed by another author (RG). Following the initial screening, full text articles were obtained and screened for eligibility by two members of the research team (AMC & RG). Disagreement was resolved through review by a third review team member (KR). Where information relating to inclusion and exclusion criteria was ambiguous or not reported in an article, the authors were contacted by email to screen for eligibility.

Data were extracted from included studies by one reviewer (AMC) using a custom template. The following data were extracted: Author, year of publication, country, methodology/ study design, population (including patient demographics and baseline characteristics where applicable), interventions received, and outcomes measured. A quality check of 20% of the data extraction was completed by a second independent reviewer (RG).

Quality assessment

The methodological quality of included studies was assessed independently and in duplicate by two reviewers (AMC, RG). The CASP critical appraisal tool for cohort studies [34] and the JBI critical appraisal tool for case series [35] were applied as appropriate. Disagreements regarding bias were resolved by a third reviewer (KR). GRADE analysis was applied to the primary outcome of functional ability to evaluate the quality of evidence [36].

Statistical analysis

Statistical analysis was performed using Review Manager Software (version 5.4) for meta-analysis. For the primary outcome of functional status, the mean and standard deviation values for the MDT group were extracted at baseline and post MDT rehabilitation. In instances where the mean and standard deviation (SD) were not available, the median was used as a proxy for the mean and a multiple of 0.75 times the interquartile range (IRQ) or 0.25 the difference in the range [37]. In studies that assessed the same construct but used a different validated outcome measure to report the construct, the exposure (MDT rehabilitation) effect was determined using the standardised mean difference (SMD). In studies that measured the same outcome using the same scales, the mean difference (MD) was used. The standard error (SE) was calculated using the SD divided by the square root of the number of values in the data set (n). For all outcomes, the denominator in each group was considered as the number of participants allocated to that group at baseline.

We assessed clinical variation across the studies by exploring the characteristics of participants, the content and duration of the MDT intervention, outcome measures administered and timing of outcome assessments. Statistical heterogeneity was examined by visual inspection of the forest plots and using the Chi2 statistic and the I2 test. As strict thresholds for interpreting I2 are not recommended, we interpreted the I2 statistic using the approximate guide by Deeks and colleagues [38]. Furthermore, to explore potential explanations of heterogeneity, moderator analysis was conducted where sufficient data was available. For example, random effects meta-regression was conducted when ≥ 10 studies reported a continuous moderator variable—age, gender, length of stay, and number of health and social care professional disciplines. In instances where there was considerable variation in the results or where there was not enough data available to conduct a meta-analysis, we opted for a narrative summary of the outcomes of interest.

Results

Flow of studies in the review

Figure 1 displays the flow of studies in the review. A total of 10,515 studies were identified across the database searches, 9168 were excluded on the basis of title/abstract screening and 195 full text articles were reviewed. Ultimately 12 articles were deemed eligible for inclusion.

Fig. 1
figure 1

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram of included studies

Full size image

Study and patient characteristics

Twelve studies met the criteria for this systematic review. Four studies were conducted in the United States [39,40,41,42], two in Italy [43, 44], two in Switzerland [45, 46], and one study in Canada [47], Romania [48], Taiwan [49] and France [50]. Six studies in the review were published in 2021 [39, 40, 43, 44, 47, 50]. Four were published in 2022 [41, 45, 48, 49]. Two papers were published in 2023 [42, 46]. The total number of participants from included studies was 570. Nine out of 12 studies reported the age of the cohort as a mean (65 to 85.33 years) [39,40,41,42, 44,45,46, 48, 49]. The remaining three studies reported a median age of 65 to 75 [43, 47, 50]. Older adults required ICU admission in seven out of 12 studies ranging from 23 to 100% of their total cohort [41, 43, 44, 46, 47, 49, 50].

Rehabilitation programme

Each paper described MDT rehabilitation which included 2 or more HSCP disciplines including Physiotherapy (PT), Occupational Therapy (OT), Speech and Language Therapy (SLT), Psychology, Social Work, Clinical Nutrition and Dietetics and Pharmacy [39,40,41,42,43,44,45,46,47,48,49,50]. See Table 1 for summary of disciplines provided by study. All 12 papers reported intervention from a PT [39,40,41,42,43,44,45,46,47,48,49,50]. Nine papers described intervention from an OT [39,40,41,42,43, 47,48,49,50]. Eight papers described intervention from an SLT [39,40,41, 44, 46, 47, 49, 50]. Seven papers reported patients received psychological interventions when needed as part of the MDT intervention [40, 41, 43, 45, 46, 49, 50]. In four out of seven studies, this intervention was provided by either a Neuropsychologist [40, 41, 43] or Psychologist [50]. Three of the studies did not report the specific discipline of psychology providing the service [45, 46, 49]. In two studies, Social Workers were part of the MDT [47, 50]. Clinical Nutrition and Dietetics and pharmacy were part of the MDT in only one study [47]. Five out of 12 papers reported input from a physician alongside the rehabilitation programme [40, 41, 44, 47, 50] including a rehabilitation physician, medical doctor, hospitalist or physiatrist and specialists such as geriatricians and liaison psychiatrists.

Table 1 Disciplines provided by study
Full size table

Intensity of multidisciplinary rehabilitation was not reported in any study. Eight studies presented detailed information on the nature of rehabilitation intervention [41, 43,44,45,46, 48,49,50]. The description of rehabilitation programmes was heterogenous however domains reported include respiratory/pulmonary rehabilitation, motor and strengthening interventions, training in activities of daily living, energy conservation techniques, advice regarding the home environment and practice of functional mobility and transfers. Please see Table 2 for characteristics of included studies for additional information. The remaining four studies reported only the disciplines that were involved in the MDT intervention or the assessment domains [39, 40, 42, 47].

Table 2 Study and patient characteristics
Full size table

Methodological quality

Table 3 details results of the CASP checklist for cohort studies and Table 4 details results of the JBI critical appraisal tool applied to the case studies included in this systematic review.

Table 3 CASP checklist
Full size table
Table 4 JBI critical appraisal tool
Full size table

Cohort studies

All cohort studies in the review addressed a clearly focused question [39,40,41,42, 44,45,46,47, 50]. In eight out of nine studies, the exposure was accurately measured to minimise bias [39,40,41,42, 44,45,46, 50]. In eight out of nine studies the follow up of patients was deemed adequate [40,41,42, 44,45,46,47, 50]. However, all nine studies recruited a convenience sample of patients. Four of the studies identified all important confounding factors for results[41, 45, 46, 50] and five studies took these factors into account when designing the methods or completing analysis [41, 42, 45, 46, 50]. In addition, no study followed up patients for long enough evidenced by the absence of follow up beyond the point of discharge.

Case study/Case report

There were two case report/case series of high quality in this systematic review satisfactorily meeting all criteria in the JBI checklist(43, 48). The complete and consecutive inclusion of participants by Chuang and colleagues was unclear however it met all other quality criteria in the JBI checklist(49). Of relevance to our secondary outcomes, Di Pietro and colleagues documented an intent to follow up patients at eight to 10 months [43] however the results of this review have not been published to the authors knowledge. Table 4 details results of the JBI critical appraisal tool.

Table 5 details results of GRADE analysis for the primary outcome of functional ability. Analysis discovered very low certainty for quality across studies meaning the true effect is probably markedly different from the estimated effect.

Table 5 GRADE assessment of outcome: functional ability
Full size table

Primary outcome

Functional ability

Functional ability was assessed pre and post MDT intervention in all studies. The validated measures used in eleven of the 12 studies for meta-analysis were the Barthel Index (BI) [43, 44, 48,49,50], the Modified Barthel Index (m BI) [39], the Functional Independence Measure (FIM) [45,46,47], the Boston AM-PAC “6 Clicks” Daily Activity Inpatient Short Form [42] and the US Centres for Medicare and Medicaid Services mandated section GG Functional Abilities score [41]. Figure 2 demonstrates that there was a statistically significant improvement in functional ability among older adults with COVID-19 who received multidisciplinary rehabilitation (REM, SMD = 1.46, 95% CI 0.94 to 1.98). Heterogeneity across the studies was significant and considerable (p < 0.00001, I2 = 91%). However, random effects meta-regression showed age (p = 0.747), gender—% males (p = 0.314), and number of disciplines (p = 0.784) did not moderate functional outcome post-MDT or explain sources of heterogeneity. See Table 6 for results of meta-regression. In the study by Maltser et al., authors reported a statistically significant change in functional ability measures following their described rehabilitation protocol [40]. This change was measured using the US Centres for Medicare and Medicaid Services mandated section GG Functional Abilities and Goals of the Improving Post-Acute Care Transformation Act. GG scores measure changes related to self-care (GG0130) and mobility (GG0170).

Fig. 2
figure 2

Functional ability pre and post MDT rehabilitation in the acute setting

Full size image
Table 6 Random Effects Meta-Regression
Full size table

Secondary outcomes

Rehabilitation length of stay

Rehabilitation length of stay was measured across 12 studies. The mean length of stay for older adults in rehabilitation units was 19 days (95%CI, 15.88–21.79 days). Heterogeneity was substantial across the pooled studies (p < 0.00001, I2 = 95%). See Fig. 3. Data from 10 of these studies could be pooled to examine the moderating effect of rehab length of stay on functional outcomes. Meta-regression showed length of stay did not significantly predict functional outcome post-MDT, (p = 0.299).

Fig. 3
figure 3

Rehabilitation length of stay among older adults with COVID-19

Full size image

Acute hospital length of stay

Acute hospital length of stay was measured across six studies comprising eight cohorts. The mean acute hospital length of stay for older adults was 18 days (95%CI, 13.35- 23.13 days). Heterogeneity was significant (p < 0.00001, I2 = 97%). See Fig. 4. Insufficient number of studies were available to analyse acute hospital length of stay as a moderator on functional outcomes post-MDT.

Fig. 4
figure 4

Acute hospital length of stay among older adults with COVID-19

Full size image

Discharge disposition

Seven studies reported discharge disposition of older adults. The proportion of older adults who were discharged directly home from the acute setting ranged from 62 to 97% [40,41,42, 44, 47, 49, 50]. Other discharge destinations included a relative’s home, COVID-19 free rehabilitation unit, sub-acute rehabilitation units, skilled nursing facilities and return to acute care.

Mortality

Two studies reported 2% mortality of older persons [42, 50], during rehabilitative care. Piquet and colleagues’ patient cohort had a mean length of stay in the acute hospital of 14.4 days and 9.8 days in rehabilitation and 23% required intensive care unit care [50]. Coakley and colleagues had a median length of stay of 6 days in the acute setting with 0% admission to ICU [42].

Primary/Community and secondary healthcare utilisation

No studies reported primary and secondary healthcare utilisation, including unplanned Emergency Department return, or unscheduled hospital admission after discharge from rehabilitation units.

None of the studies reported on long-term effects of COVID-19 at discharge from rehabilitation units or at agreed follow-up points in time. Some studies did describe patients need for supplementary oxygen [41, 44, 50] on discharge, reporting prevalence of 4 and 58%. In addition, Bertolucci also reported persisting symptoms at the time of discharge. The author reports that tracheostomies were removed in 97.44% of patients on discharge from rehabilitation (28.22% of patients had a tracheostomy on admission), 100% of patients achieved complete oral alimentation, zero patients presented with mental confusion, 10.2% of patients were continuing to be prescribed corticosteroids and 12.8% were continuing to be prescribed antipsychotics.

Discussion

This review aimed to describe the clinical characteristics, functional and process outcomes of older adults with COVID-19 who received MDT rehabilitation in the inpatient acute or post-acute hospital setting. There was heterogeneity across the 12 included studies with regards to study design, MDT intervention provided, and outcomes measured. There was a significant improvement in functional ability among older adults with COVID-19 who received MDT rehabilitation, but only two studies had a comparator group [40, 42]. The proportion of older adults who were discharged directly home from the acute setting ranged from 62 to 97%. No studies followed up patients after discharge or reported on long term effects of COVID-19 on discharge from rehabilitation units.

The key finding of our review is that MDT rehabilitation for older adults with COVID-19 in acute or post-acute inpatient hospital setting resulted in statistically significant improvement in function. Moreover, this improvement in functioning was not moderated by length of rehabilitation stay. Our primary outcome, function, aligns with the WHO agenda for healthy ageing globally [51] which recognises society’s duty to facilitate the rights of the older adult to healthy ageing. Our findings support guidelines by the European Geriatric Medicine Society (EuGMS) [28] and the WHO [21] which recommend MDT rehabilitation for older adults hospitalised with COVID-19.

This review found that older adults stayed in hospital for an average of 18 days (95%CI, 13.35- 23.13 days) and in rehabilitation units for 19 days (95%CI, 15.88–21.79 days). Mortality was not routinely reported across studies, but the incidence was low (2%). Rehabilitation length of stay following COVID-19 has already been reported in the literature however in a younger cohort of patients, where length of stays ranged from 11 days to 44.96 days [30, 52,53,54,55,56,57,58,59]. Most of this evidence represents patients of high illness acuity with patients described as having critical illness or severe illness or requiring intensive care unit treatment [30, 52, 55, 57,58,59]. This is comparable to the evidence presented in this review, where older adults required ICU admission in seven out of 12 studies [41, 43, 44, 46, 47, 49, 50]. In a study by O’Kelly and colleagues, authors reported patients had a median length of stay of 9 days, with 17% requiring ICU admission, however again patients were younger, with a median age of 45 years old [60] and the extent of rehabilitation services provided, if any, was not reported.

The long-term sequelae of COVID-19 are well documented [61,62,63] however we found that none of the included studies followed up participants after the point of discharge and none of the studies reported on residual COVID-19 symptoms at the point of discharge or follow up. The long-term effect of multidisciplinary rehabilitation is unclear and remains to be investigated rigorously. Existing research in the older adult population indicates decline in function, increases in frailty and a reduction in quality of life over time following COVID-19 [64, 65]. It would be valuable to determine through robust experimental research if MDT rehabilitation can impact functional deterioration and worsening frailty over time in older adults with COVID-19 as it has been shown to benefit these outcomes with other older adult populations [66, 67].

This review included no studies reporting healthcare utilisation following MDT rehabilitation at the point of discharge or at follow up. It is important that intervention studies assess older adults’ healthcare use on discharge from acute or post-acute hospital settings for COVID-19 as people discharged from hospital following treatment for COVID-19 are at significantly higher risk for readmission to hospital when compared to demographically matched controls and people discharged from hospital following treatment for influenza, suggesting a significant burden to healthcare services for the cohort [68].

The 12 included studies in this review consisted of seven descriptive cohort studies, one pre-post intervention cohort, one quasi experimental study and three case series highlighting a dearth of robust experimental studies or analytical cohort studies describing the effect of multidisciplinary rehabilitation on the outcomes of older adults in the acute or post-acute setting following COVID-19 to facilitate systematic review and meta-analysis. A quasi-experimental study by Rodriguez and colleagues aimed to describe the effects of a multimodal rehabilitation programme in patients with COVID-19 admitted to the ICU [69] however this study was ineligible for inclusion in our review as the average age of the intervention cohort was 56.5 years and it was unclear if the intervention was multidisciplinary in nature. A large number of descriptive cohort studies and case series were not included in this review reporting outcomes following MDT rehabilitation following COVID-19 as their focus was on a younger population [30, 52,53,54,55,56,57,58,59, 70, 71]. GRADE analysis of included studies showed very low certainty of evidence which limits the applicability of results and highlights the importance of future trial studies to determine the effect of rehabilitation for the cohort.

Three studies included in this review excluded patients with a diagnosis of delirium or dementia [43, 45, 46]. Older adults with COVID-19 commonly present with delirium on admission or during the course of their acute illness in hospital [72,73,74]. Additionally, older adults with an underlying cognitive impairment or dementia pathology are at higher risk of delirium incidence [75]. Existing evidence from studies with older adults not specific to COVID-19 supports the assertion that older adults with cognitive impairment can benefit from rehabilitation [76, 77]. Exclusion of those with cognitive impairment in rehabilitation research, limits the applicability of outcomes to a significant cohort of older adults seeking acute medical care for COVID-19.

The results of this review must be considered in the context of the global progress with the roll out of COVID-19 vaccination programmes. The European Centre for Disease Prevention and Control (ECDC) reports a total of 966,099,169 vaccination doses administered as of the 14th of December 2022 [78]. The total number of people who have been vaccinated with at least one dose in the European Union is reported as 342,182,404 in the total population, representing 75.5% of the population [78]. It is established that mRNA COVID-19 vaccination greatly reduces the risk of mortality, disease progression, death and mechanical ventilation [79]. Our review included studies in which patients were recruited between March 2020 and December 2021 and therefore not all patients could have been vaccinated. Three studies were carried out during a time when vaccinations were available to older adults [46, 48, 49]. It is possible to deduce that as more people are vaccinated worldwide that fewer adults and older adults will require hospitalisation and rehabilitation. However, there are cases of unvaccinated cohorts internationally due to inequity in vaccine roll out with the WHO reporting only 25% of older adults have had a complete series of vaccines in lower income countries [80]. It has also been reported that COVID-19 patients infected with the Omicron variant have a lower risk of hospitalisation compared with patients infected with the Delta variant [81, 82]. It is possible that new variants will emerge with unknown associated admission rates to hospital.

Geriatric rehabilitation programmes for patients with COVID-19 require additional consideration for the physical environment, equipment, resources and staffing in order to minimise the impact of infection control measures on patient experience and outcomes [28]. The multi-organ involvement of COVID-19 requires an interdisciplinary approach to address the numerous complications associated with COVID-19 infection [83] provided by an interdisciplinary team including, Physicians, Nurses, Physiotherapists, Occupational Therapists, Dietitians, Speech and Language Therapists, Psychologists and Social Workers [28]. In this systematic review, each study met the criteria for MDT rehabilitation however team composition varied. PT, OT, SLT and Psychology were the most prevalent disciplines. Few studies reported Dietitians as part of the MDT despite the high prevalence of malnutrition in COVID-19 hospitalised patients [84, 85]. Heterogeneity of rehabilitation programmes and limited reporting of rehabilitation programmes were evident in this systematic review however seven papers described their rehabilitation programme in sufficient detail [41, 43,44,45,46, 49, 50]. It is recommended that geriatric rehabilitation for COVID-19 should address frailty, malnutrition, cognition, activities of daily living and participation, mood, pain and symptom management, retraining of mobility, strengthening exercises, psychological disturbances, and speech and swallow impairments with discharge planning to facilitate follow up to the appropriate primary care or specialist outpatient care setting [28]. None of the studies included in this review described a rehabilitation programme that addressed all of these domains.

Strengths and limitations

The conduct and reporting of this systematic review was in accordance with the MOOSE guidelines [31]. The identification of suitable papers was completed with a standardised and reproducible search strategy and with clear inclusion and exclusion criteria. A PRISMA flow diagram was used to map the flow of information through the different phases of the review. Critical appraisal of included papers was completed using the CASP checklist for cohort studies and the JBI Critical Appraisal Tool to assess bias. GRADE analysis also assessed the quality of evidence.

A limitation of this review is the heterogeneity of rehabilitation programmes with limited reporting of the frequency, intensity, time and type of interventions. No trial studies were included in this review and critical appraisal of the studies included highlight quality deficits which limits the internal and external validity of the findings.

Clinical and research Implications

This review highlights the need for experimental studies exploring the effect of multidisciplinary rehabilitation on older adults with COVID-19. The ethical challenge this poses to the research community must be considered however as experimental studies would place patients into control and experimental groups.

This review highlights the need for greater attention to long term follow up in studies with older adults post COVID-19 to assess function, ongoing symptoms, and healthcare utilisation to determine the long-term effect of multidisciplinary rehabilitation. Long term outcomes and ongoing symptoms should be explored objectively by measures designed for the population and health states in question such as the COVID-19 Yorkshire Rehabilitation Scale (C-19 YRS) [86] which is recommended by the United Kingdom’s National Health Service [87] and the National Institute for Health and Care Excellence [22].

Given the heterogeneity of rehabilitation programmes in this review, future experimental research should describe a defined and reproducible rehabilitation programme using the TIDieR checklist [88]. An economic evaluation of multidisciplinary rehabilitation in this population could explore the financial implications to our health care systems. It is estimated that COVID-19 rehabilitation costs twice that of other rehabilitation units due to the complexity of its presentation, the heterogenous complications and the infection control measures required [89] however exact figures do not exist.

Conclusion

This review demonstrates that multidisciplinary rehabilitation may result in improved functional outcomes on discharge from acute or post-acute hospital settings for older adults with COVID-19. There is a need for robust and experimental research into the long-term effect of rehabilitation for older adults following COVID-19. Future research should comprehensively describe MDT rehabilitation in terms of disciplines involved and the intervention provided using a standardised method of reporting.