Rehabilitation after critical illness (RaCI) enhanced physiotherapy input following critical care discharge: A quality improvement project

Introduction

Patients surviving critical care often suffer physical and psychological morbidity following critical illness (Salisbury et al. 2010). Multi-organ failure, prolonged mechanical ventilation, and neuromuscular dysfunction are all factors associated with both increased critical care and hospital length of stay (McWilliams et al. 2015, 2019). These patients often experience short and long-term reduction in health-related quality of life which has profound consequences for the individual and their families (McWilliams et al. 2015; McWilliams et al. 2019).

The National Institute for Clinical Excellence outlined the key principles of care in the recovery of critically ill patients in their clinical guideline entitled ‘Rehabilitation after critical illness’ (NICE 2009). Structured rehabilitation within intensive care is recommended and increasingly recognised (NICE 2009, 2017; FICM 2019, 2020); with early, daily physiotherapy reported to be safe, feasible and established as standard care in many units (McWilliams et al. 2015).

A substantial body of research has explored the benefits of rehabilitation after critical care discharge; however, there is currently no evidence to conclude benefit from such interventions (Connolly 2016). The ‘RECOVER’ study (Walsh et al. 2015) attempted to evaluate the effect of increasing physical and nutritional rehabilitation post critical care discharge using rehabilitation practitioners. The intervention group received a two to three-fold increase in the frequency of mobilisation, increased dietetic involvement, and individualised goal setting. However, there was no improvement in physical recovery or health related quality of life.

Patients discharged from critical care often have multi-factorial rehabilitation needs which require input from allied health professionals and nursing staff (Salisbury 2010; van der Schaaf 2008; Vollam 2021; Silveira 2019). Those discharged to specialities dispersed across the hospital (for example, respiratory medicine, gastrointestinal surgery) commonly experience uncoordinated journeys, where critical care associated problems are poorly understood (Salisbury 2010; FICM 2020). Furthermore, the optimum timing, frequency, duration and components of rehabilitation post critical care is uncertain (Walsh 2015). Conversely, studies completed on already established rehabilitation pathways for example, stroke and ortho-geriatric services suggest that a co-ordinated approach to physical rehabilitation improves outcome for patients and can reduce hospital length of stay (Wu 2019; Stucki 2005). These findings suggest that there may be potential for improving outcomes within the critical care population through a pathway approach, although this is yet to be demonstrated within the literature.

A recent evaluation within our hospital has shown that the phase immediately following critical care often presents with a plateau or reduction in functional ability. This deterioration is often linked with an increase to patients’ length of stay, which can have significant cost implications to the health service (Cuthbertson 2007).

In summary, despite recent interest in rehabilitation during and after critical care, there remains a clear lack of clarity on its effectiveness and cost. Whilst intervention during critical care appears to have some short-term benefit, there is very little evidence for its effectiveness post critical care discharge.

Based on the above, this quality improvement project aimed to improve patient’s functional recovery and reduce post-critical care length of stay through the provision of enhanced physiotherapy input. This enhanced physiotherapy input was provided in addition to existing services and was delivered by the critical care physiotherapy team for the first 14 days post-critical care discharge.

Methodology

Context

This quality improvement project was completed in a 1000-bed University Teaching Hospital within South Wales (U.K.). The hospital has a 32-bed, mixed-dependency critical care unit, admitting more than 1500 patients per year from all major specialities including general medicine, liver, trauma, neuro-critical care, and complex upper gastrointestinal surgery. On discharge from critical care patients are transferred to the most appropriate ward for their clinical presentation and medical speciality.

Prior to this quality improvement project, all patients discharged from critical care would be transferred to the ward-based physiotherapy team. The frequency of physiotherapy intervention received was dependent on the demand and prioritisation of physiotherapy case-load. Patients often only received two to three physiotherapy treatments per week of varying durations (on average between 20 and 30 minutes). No weekend or public holiday physiotherapy input was provided unless the patient required urgent ‘respiratory’ physiotherapy intervention.

Patient selection

Patients deemed ‘at risk of physical morbidity’ (patients scoring >3 of the following on first assessment) were included in the quality improvement project during December 2019 and March 2020.

1 Unable to get out of bed independently.

2 Anticipated invasive ventilation >72 hours.

3 Obvious physical/neurological injury.

4 Lack of cognitive function to exercise independently.

5 Unable to mobilise short distances independently.

6 Unable to ventilate with <35% oxygen.

6 Pre-morbid respiratory disease.

7 Pre-morbid mobility problems.

Patients were not eligible if they had: a contraindication to mobilisation (for example, unstable fractures), an already established rehabilitation pathway, or a profound acquired neurological deficit, where it was thought a short-term enhancement in therapy input was unlikely to influence recovery time.

Comparator group

To assess the impact of the QI project, data from a comparator group was also collected, and was based on the same four-month period from the previous year. The same eligibility criteria were utilised, specifically only considering patients at ‘risk of physical morbidity’, and the exclusion of those on pre-existing rehabilitation pathways and patients with profound neurological deficit. The data collated only routinely collected data that was readily available.

Interventions

For the four-month quality improvement project, funding was gained to support an additional physiotherapist within the critical care physiotherapy team with the purpose of providing enhanced physiotherapy input to eligible patients discharged from critical care. The aim was for all patients to receive daily physiotherapy input (from the critical care physiotherapy team) in addition to existing ward-based physiotherapy services. This additional input was provided on weekdays for the first two weeks post critical care discharge (unless discharged from hospital within two weeks). There were no limitations on the duration of individual physiotherapy sessions and no guidance was provided as to the content of the physiotherapy intervention other that it should be targeted at the patient’s maximum functional ability. The patients’ general management remained the responsibility of the ward-based physiotherapy team and regular liaison between teams was encouraged.

Study of interventions

The quality improvement project was designed to reduce post critical care length of stay and increase patient’s functional recovery. A thorough review of the existing physiotherapy model of care was completed (including the physiotherapy records of a the comparator group), which suggested a patient’s recovery either slowed or plateaued in the early post critical care period. Based on this review, and discussion with the ward based physiotherapy teams, the quality improvement project focused on increasing physiotherapy involvement for the two-week period post critical care discharge. Once designed, the improvement programme was re-discussed with all in-patient physiotherapy teams to ensure awareness of the project and to answer any queries or concerns. Following the two-week post critical care discharge period, all physiotherapy services returned to baseline with no involvement from the critical care physiotherapy team.

Measures

The primary outcome was post critical care length of stay (LOS) compared to the comparator group. Secondary outcomes explored patients’ functional ability, the frequency and duration of physiotherapy intervention provided, the frequency of ‘unmet need’ (non-completion of planned physiotherapy sessions, for example, identified as requiring physiotherapy input but no received due to prioritisation of case-load) and physiotherapy/community resource team (care and therapy input) requirements on hospital discharge. Functional ability was measured using the Chelsea critical care assessment tool (CPAx) (which consists of 10 commonly assessed components of physical ability) and, the ICU mobility scale (IMS) (which is a 11-point scale used to record a patient’s level of mobilisation); the higher the score the greater the mobility and functional ability. These outcome measures were already in use within critical care unit and familiar to the staff involved. Comparator data was only available for length of stay and physiotherapy/community resource requirements required on discharge.

Analysis

A customised data collection tool was used to aid analysis of this project. Data were analysed using Microsoft Excel. Simple descriptive data are presented using means (standard deviation) or median (inter-quartile range) depending on the nature of the data. No statistical testing was completed.

Ethical considerations

This project constituted an improvement in the standard care delivery with no randomisation and thus met the definition of a quality improvement project under the NHS Health research authority guidelines. This was confirmed with Health and Care Research Wales and as such ethical approval was not required. The project was registered as a quality improvement project (QI project 48) within the host organisation and underwent local peer review as per standard and complied with local governance processes. All data was stored electronically on password protected NHS computers in accordance with data protection requirements.

Results

Demographics

Between December 2019 and March 2020, all eligible patients (n = 44) were involved in this single-site service improvement project which looked at enhanced therapy after critical illness. Participants comprised of 34 males and 10 females with an average (SD) age of 61.9 (15.6) years. Patients median (IQR) critical care length of stay was 20 (12.0–25.5). The comparator group consisted of 38 patients who were slightly younger with an average age of 56.3 (17.6) years, and a shorter length of stay (17.7 days (IQR 8.5–26.1)). Further demographic information is shown in Table 2.

Post critical care length of stay

A median 11-day reduction in LOS between the project group and comparator group was observed, which must be considered as highly clinically significant. Further comparison details are shown in Table 1.

Table 1: Comparison of project and comparator group.

	Project Group (2019–2020) (n = 44)	Comparator Group (2018–2019) (n = 38)
Age (Mean, SD)	61.9 (15.6)	56.3 (17.6)
Male:female	34:10	21:17
Speciality Cardiology Cardiothoracic surgery General surgery General medicine Haematology Neurology Neurosurgery Oro-maxillofacial Renal Spinal Thoracic medicine Trauma	3 1 18 6 1 3 2 2 1 1 6 0	2 1 7 10 0 1 1 1 0 4 7 4
Critical care LOS (Median, IQR)	20 (9.0–32.3)	17.7 (8.5–26.1)
Post Critical Care LOS (Median, IQR)	15 (8.3–31.5)	26 (8.0–44.5)

Physical function

An improvement in physical function and mobility (measured using the IMS and CPAx) between critical care discharge and hospital discharge was demonstrated in all but one participant, with the rate of improvement most noticeable within the first 14 days (see Figure 1). Most patients leaving critical care had stood but not stepped (IMS 4 (IQR 4–5)) whereas by day 14 most patients were mobilising with assistance of one person (IMS 8 (IQR 7–9)). The comparator group had a similar mobility status at critical care discharge (IMS 4 (IQR 3–5)), however, no comparison data is available for physical function at either 14-days post critical care discharge or at discharge home from hospital as this was not routinely collected data, nor was it possible to complete these outcome measures retrospectively.

Demonstrates median changes in outcome measures at each time point.

See Figure 1: Displays improvements in CPAx score in the intervention group at each time point with superimposed median scores (black line).

Table 2: Demonstrates median changes in outcome measures at each time point.

	Critical care discharge	14-days post critical care discharge	Hospital discharge
CPAx	28 (23–32)	42 (37–45)	45 (42–46)
IMS	4 (4–5)	8 (7–9)	9 (9–9)

Data shown as median (inter-quartile range).

Support required on hospital discharge

79.5% (n = 35) patients were discharged home, with nine being discharged prior to day 14. This is compared to 71.1% (n = 27) in the comparison group. Those in the QI group not discharged home were transferred for either specialist rehabilitation or repatriation to their local hospital (20.4% and 18.4% respectively), with no onward data available. Table 3 outlines differences in physiotherapy/community resource requirements.

Table 3: Physiotherapy and community resource requirements on hospital discharge.

	Project group (2019–2020) (n = 35)	Comparator group (2018–2019) (n = 27)
No support	37.1% (n = 13)	37% (n = 10)
Community resource team (care and therapy input)	34.2% (n = 12)	18.4% (n = 5)
Community physiotherapy	20.0% (n = 7)	25.9% (n = 7)
Musculo-skeletal physiotherapy	9.0% (n = 3)	18.5% (n = 5)

Physiotherapy activity

The quality improvement project aimed to enhance physiotherapy input for the first 14 days post critical care discharge. Patients included in the QI programme (n = 44) received a mean (SD) of 8.6 (4.4) physiotherapy sessions in those 14 days. A mean of 5.9 (3.2) sessions were delivered by the critical care physiotherapy team, and 2.6 (2.0) delivered by ward team. However, this included patients who were discharged within the 14 days. When these were removed from the analysis, the remaining (n = 30) received an average of 10.6 (3.3) physiotherapy sessions (average 7.4 [2.3] and 3.6 [1.9] from the critical care and ward-based physiotherapy teams respectively). During the first 14 days, only 2% of planned sessions (for example, 2% ‘unmet need’) by the critical care physiotherapy team were not completed compared to 42% of the sessions planned by the ward physiotherapy team (for example, identified by ward team as requiring input on a specific day and recorded on a register but intervention not delivered). There was little variation in the duration of physiotherapy sessions with both the critical care and ward-based physiotherapy team sessions lasting an average of 26 minutes.

No comparison data is available for physiotherapy sessions delivered in the 14-days post critical care discharge prior to the initiation of the quality improvement project, nor was it possible to collect this retrospectively.

Discussion

In this quality improvement project, the critical care physiotherapy team continued to provide input to patients, in additional to ward-based physiotherapy input for the first 14-days post critical care discharge. The project included 44 patients deemed at high risk of developing physical morbidity post critical illness with our results demonstrating a 11-day reduction in median hospital LOS compared to a similar group from a year prior. Whilst the results were statistically non-significant, clinically this reduction implies significant health and cost implications to the health service.

Previous research exploring rehabilitation after critical care has failed to consistently show significant difference in outcomes (Connolly et al. 2016), with suggestions that therapy provision three times higher than standard practise may be needed to have significant benefits (Wu et al. 2019). In a different sample population, Atkins et al. (2019) reported similar results to the current QI project, with a 14-day reduction in hospital LOS following the introduction of more consistent rehabilitation on an acute medical ward. Despite acknowledgement of further research warranted to confirm findings in other clinical areas, it suggests that enhanced therapy may influence patient outcome and patient flow.

Patients included in this project demonstrated improvements in physical function at each time-point and was most notable within the first 14-days following critical care discharge. The speed of improvement may have been a reflection on the additional input received by the critical care physiotherapy team, but the absence of comparator data makes this only a suggestion. What remains unclear is the speed in which natural recovery may have occurred, for example, these patients would have continued to improve with or without input. However, our previous unpublished service evaluations have suggested a plateau in physical recovery in the immediate post-critical care period.

Usual standards of rehabilitation are important to consider when analysing and comparing project results (Wu et al. 2019); our unit aims to complete therapy twice a day for those at ‘high risk at risk of physical morbidity’, whilst resources for rehabilitation on acute wards are limited, and as research suggests, can be as little as two to three times per week (Cuthbertson et al. 2007; Salisbury et al. 2010; Wu et al. 2019). This is highlighted in the current QI project by the significant difference in percentage of intended sessions completed by the therapists; 42% of the intended physiotherapy sessions were not completed by the ward team in comparison to 2% by the critical care physiotherapy team. This ‘unmet need’ reflects local prioritisation tools and clinical case-load affecting the ability to provide physiotherapy intervention as frequently as planned.

Additionally, within our project, the number of treatment sessions provided (not in relation to unmet sessions) were less than expected over 14-days (average 7.4 [2.3] and 3.6 [1.9] for critical care physiotherapy team and ward-based physiotherapy teams respectively). Reasons for this include the lack of physiotherapy input at the weekend within the host organisation. Additional reasons include patients being unable to tolerate two treatment sessions a day, that they were otherwise engaged or down to individual opinion resulting in re-prioritisation of ward case-load to meet the demand of those patients not seen.

Within this limited QI project, a higher percentage of patients were discharged directly home following when compared to comparison group. This is reflected in a study by Denehy et al. (2017) that showed a significantly higher number of patients being discharged home in their control group compared to usual care. However, in our QI project, whilst more patients were discharged home, there was an increase in requirement for community support. The reasons for this are not clear especially as physical function data from the comparator group was not available. Potentially, the ability to discharge home more frequently, and in a timelier manner, was only achieved through greater reliance on community services. This cannot be confirmed based on this QI project however would need consideration in future studies.

This study aimed to improve patient’s functional recovery and reduce post-critical care length of stay through the provision of enhanced physiotherapy input. However, based on our findings it is difficult to interpret what elements of ‘enhanced’ input the patients received. Due to clear differences in the number of treatments delivered by the ward and critical care physiotherapy team, and the lack of baseline data, it is unclear whether any improvements were a result of the amount of physiotherapy input delivered, or the continuity of physiotherapy staff involved from critical care into the wards, or a combination of both. However, despite that lack of clarity, the results are suggestive of a clinically significant reduction in length of stay and therefore require further exploration and research.

Limitations

As expected for a small, local quality improvement project there are several limitations affecting the ability to generalise our findings to the wider critical care population. Firstly, the limited sample size of both the QI and comparator groups is insufficiently powered for reliable statistical analysis. However, it should be noted that all patients that were eligible were included in the QI project and therefore the sample is an accurate reflection of the critical care discharges during those 4-months. The single-site nature of the study also limits the generalisation of the results and recognition of limitations and exclusion criteria must be considered when applying to all patients within a critical care population.

The use of a historical comparison group for data is also a significant limitation. Whilst the two groups were matched in terms of both being patients ‘at risk of physical morbidity’, it was based on local models of risk assessment. There were no significant differences between the groups in terms of age or length of critical care, but there was no assessment of severity of illness and so on. Furthermore, the absence of routinely collected data regarding patients’ functional abilities reduced comparisons. Based on these, the results of this project should be considered with caution and likely only a suggestion for future research and discussion.

The potential ceiling effect of the outcome measures also requires discussion. Figure 1 suggests a plateau in patient recovery after day 14. The exact reasons for this are unknown. A plausible explanation is that the patients were at, or close to pre-morbid level of function. It is given that many critical care survivors have chronic diseases in addition to their presenting diagnoses therefore have a lower pre-morbid function Denehy (2013). However, without baseline mobility data it is impossible to state for certain. The sensitivity of the outcome measures used need to be considered. Despite both being validated and shown to be reliable in measuring functional ability they are designed as tools to be used within the critical care setting and not for higher functioning patients (Corner 2013; Tipping 2016).

It must also be highlighted that the end of the project coincided with the start of the COVID-19 pandemic which brought significant changes to the NHS. It is unclear whether patients within the QI group experienced shorter lengths of stay directly because of bed-capacity pressures. This is unlikely to be the case given most patients were discharged prior to the 1st wave of the pandemic but must be considered as a potential factor.

Further research

The importance of early rehabilitation on critical care step down on influencing patients’ hospital stay clearly warrants further investigation. Previous research has failed to show significant improvements in outcomes through post-critical care rehabilitation although quality improvement projects have shown potential. Prospective projects need to continue to explore the most suitable methods of delivery of rehabilitation, whilst also considering which professions must be involved, the timing of the intervention and the most appropriate outcomes for use.

Conclusion

In this single site quality improvement project, the provision of 14-days input from the critical care physiotherapy team following discharge from critical care was associated with a median 11-day reduction in hospital length of stay for patients at high risk of morbidity following critical care. Given the significant limitations to this study, and the findings of larger randomised control trials, further research is required into the most appropriate structure, timing, and frequency of rehabilitation in the early post critical care period.

Data availability

Raw data available by direct request to the corresponding author.

Ethics approval

This project constituted an improvement in the standard care delivery with no randomisation and thus met the definition of a quality improvement project under the NHS Health research authority guidelines. As such ethical approval was not required. Consent for involvement was gained as would be for normal therapy treatment sessions.

Declarations

The authors declare no declarations of interest. No material has been used from other sources.

Funding

Additional ‘winter pressures’ funding supported the recruitment of a 1.0 WTE band five physiotherapist to support this quality improvement project. Funding was provided by the critical care department, Cardiff and Vale University Health Board.

Atkins, G., Finch, L., Rawson, B., Hawker, R., Wells, J., & McWilliams, D. (2019). A quality improvement project looking at the impact of structured rehabilitation on medical wards Physiotherapy, 105(1). https://doi.org/10.1016/j.physio.2018.11.110.

Cuthbertson, B. H., Rattray, J., Johnston, M., Wildsmith, J. A., Wilson, E., Hernendez, R., Ramsey, C., Hull, A. M., Norrie, J., & Campbell, M. (2007). A pragmatic randomised, controlled trial of intensive care follow up programmes in improving longer-term outcomes from critical illness. The PRACTICAL study. BMC Health Services Research, 7, 116. https://doi.org/10.1186/1472-6963-7-116.

Connolly, B., O’Neill, B., Salisbury, L., Blackwood, B., & Enhanced Recovery After Critical Illness Programme Group. (2016). Physical rehabilitation interventions for adult patients during critical illness: An overview of systematic reviews. Thorax, 71(10), 881–890. https://doi.org/10.1136/thoraxjnl-2015-208273.

Corner, E. J., Wood, H., Englebretsen, C., Thomas, A., Grant, R. L., Nikoletou, D., & Soni, N. (2013). The Chelsea critical care physical assessment tool (CPAx): validation of an innovative new tool to measure physical morbidity in the general adult critical care population; an observational proof-of-concept pilot study. Physiotherapy, 99(1), 33–41. https://doi.org/10.1016/j.physio.2012.01.003.

Denehy, L., Skinner, E. H., Edbrooke, L., Haines, K., Warrillow, S., Hawthorne, G., Gough, K., Hoorn, S. V., Morris, M. E., & Berney, S. (2013). Exercise rehabilitation for patients with critical illness: A randomised controlled trial with 12 months of follow-up. Critical Care, 17(4). https://doi.org/10.1186/cc12835.

FICM. (2019). Guidelines for the provision of intensive care services – version 2. https://www.ficm.ac.uk/sites/default/files/gpics-v2.pdf.

FICM. (2020). Recovery and rehabilitation for patients following the pandemic. https://www.ficm.ac.uk/sites/default/files/ficm_rehab_provisional_guidance.pdf.

NICE. (2009). Rehabilitation after critical illness in adults. https://www.nice.org.uk/guidance/cg83.

NICE. (2017). Rehabilitation after critical illness in adults – quality standard 158. https://www.nice.org.uk/guidance/qs158/resources/rehabilitation-after-critical-illness-in-adults-pdf-75545546693317.

McWilliams, D., Weblin, J., Atkins, G., Bion, J., Williams, J., Elliott, C., Whitehouse, T., & Snelson, C. (2015). Enhancing rehabilitation of mechanically ventilated patients in the intensive care unit: A quality improvement project. Journal of Critical Care, 30(1), 13–18. https://doi.org/10.1016/j.jcrc.2014.09.018.

McWilliams, D., Snelson, C., Goddard, H., & Attwood, B. (2019). Introducing early and structured rehabilitation in critical care: A quality improvement project. Intensive & Critical Care Nursing, 53, 79–83. https://doi.org/10.1016/j.iccn.2019.04.006.

Salisbury, L. G., Merriweather, J. L., & Walsh, T. S. (2010). Rehabilitation after critical illness: could a ward-based generic rehabilitation assistant promote recovery? Nursing in Critical Care, 15(2), 57–65. https://doi.org/10.1111/j.1478-5153.2010.00382.x.

Silveira, L., Silva, J., Tanaka, C., & Fu, C. (2019). Decline in functional status after intensive care unit discharge is associated with ICU readmission: A prospective cohort study. Physiotherapy, 105(3), 321–327. https://doi.org/10.1016/j.physio.2018.07.010.

Stucki, G., Stier-Jarmer, M., Grill, E., & Melvin, J. (2005). Rationale and principles of early rehabilitation care after an acute injury or illness. Disability and Rehabilitation, 27(7–8), 353–359. https://doi.org/10.1080/09638280400014105.

Tipping, C. J., Bailey, M. J., Bellomo, R., Berney, S., Buhr, H., Denehy, L., Harrold, M., Holland, A., Higgins, A. M., Iwashyna, T. J., Needham, D., Presneill, J., Saxena, M., Skinner, E. H., Webb, S., Young, P., Zanni, J., & Hodgson, C. L. (2016). The ICU mobility scale has construct and predictive validity and is responsive. A multicentre observational study. Annals of the American Thoracic Society, 13(6), 887–893. https://doi.org/10.1513/AnnalsATS.201510-717OC.

van der Schaaf, M., Dettling, D. S., Beelen, A., Lucas, C., Dongelmans, D. A., & Nollet, F. (2008). Poor functional status immediately after discharge from an intensive care unit. Disability and Rehabilitation, 30(23), 1812–1818. https://doi.org/10.1080/09638280701673559.

Vollam, S., Gustafson, O., Young, J. D., Attwood, B., Keating, L., & Watkinson, P. (2021). Problems in care and avoidability of death after discharge from intensive care: A multi-centre retrospective case record review study. Critical Care, 25(1), 10. https://doi.org/10.1186/s13054-020-03420-5.

Walsh, T. S., Salisbury, L. G., Merriweather, J. L., Boyd, J. A., Griffith, D. M., Huby, G., Kean, S., Mackenzie, S. J., Krishan, A., Lewis, S. C., Murray, G. D., Forbes, J. F., Smith, J., Rattray, J. E., Hull, A. M., Ramsay, P., & RECOVER Investigators (2015). Increased hospital-based physical rehabilitation and information provision after intensive care unit discharge: The RECOVER randomised clinical trial. JAMA Internal Medicine, 175(6), 901–910. https://doi.org/10.1001/jamainternmed.2015.0822.

Wu, J., Vratsistas-Curto, A., Shiner, C. T., Faux, S. G., Harris, I., & Poulos, C. J. (2019). Can in-reach multidisciplinary rehabilitation in the acute ward improve outcomes for critical care survivors? A pilot randomised controlled trial. Journal of Rehabilitation Medicine, 51(8), 598–606. https://doi.org/10.2340/16501977-2579.