Publication

• Title: Remote Multicomponent Rehabilitation in Intensive Care Unit Survivors: A Randomized Clinical Trial
• Acronym: iRehab
• Year: 2026
• Journal published in: JAMA
• Citation: O’Neill B, Bradley JM, Connolly B, Bruce J, Ratna M, Lall R, et al; iRehab Trial Investigators. Remote multicomponent rehabilitation in intensive care unit survivors: a randomized clinical trial. JAMA. Published online May 18, 2026.

Context & Rationale

• Background

  • Survivors of critical illness frequently experience persistent physical deconditioning, fatigue, impaired daily activities, dysphagia or respiratory symptoms, anxiety, depression, cognitive problems, and reduced health-related quality of life after ICU discharge.¹
  • Evidence for rehabilitation after hospital discharge has been inconsistent, partly because previous trials varied in population, timing, intensity, intervention components, and outcome measurement.²
  • Patients and families have repeatedly described recovery needs as multifactorial and individualised, not limited to strength or exercise tolerance alone.
  • Remote delivery offered a plausible route to overcome geography, travel burden, post-ICU fatigue, and unequal access to specialist rehabilitation.
• Research Question/Hypothesis

  • Does a 6-week, remotely delivered, multicomponent, individualised rehabilitation programme improve health-related quality of life at 8 weeks after randomisation in adult ICU survivors who required invasive mechanical ventilation for at least 48 hours, compared with standard care?
  • The protocol hypothesis was that remote multicomponent rehabilitation would improve quality of life, physical function, fatigue, mood, and other health-related outcomes compared with standard care.³
• Why This Matters

  • Post-ICU survivorship is common, disabling, and resource-intensive.
  • Face-to-face ICU recovery pathways are patchy, resource dependent, and difficult for fatigued or mobility-limited patients to attend.
  • iRehab tested a scalable, centralised, theory-informed rehabilitation model that could plausibly be adopted across a national health system if effective.

Design & Methods

• Research Question: In adult survivors of critical illness discharged home after an ICU admission requiring at least 48 hours of invasive mechanical ventilation, does a 6-week remote multicomponent rehabilitation programme improve EQ-5D-5L utility score at 8 weeks compared with standard care?
• Study Type: Pragmatic, multicentre, open-label, assessor-blinded, randomised clinical trial conducted at 52 NHS hospitals in the UK from December 2022 to November 2025.
• Population:
  • Adults aged at least 18 years.
  • Required continuous invasive mechanical ventilation for at least 48 hours during ICU admission.
  • Within 12 weeks of discharge home from hospital at consent.
  • Required ability to participate in trial procedures, including telephone or computer use.
  • Required spoken English understanding or access to a family member, friend, or other person able to translate trial materials.
  • Key exclusions were inability or refusal to consent, previous randomisation in iRehab, participation in another rehabilitation or self-management trial, contraindication to exercise, severe mental health problems precluding group participation, discharge to a rehabilitation unit or care facility, and prisoner status.
• Intervention:
  • A 6-week, remotely delivered, patient-centred, structured, individually tailored multicomponent rehabilitation programme.
  • Weekly 1-to-1 remote sessions with a trained iRehab intervention specialist using BEAM, Zoom, Microsoft Teams, or telephone.
  • Core components were weekly symptom management, targeted exercise and physical activity, psychological wellbeing support, and peer support/information.
  • Participants received printed manuals, an exercise manual, access to pre-recorded exercise or wellbeing resources, and were invited to weekly group exercise and peer-support café sessions.
  • The intervention team was supported by a multiprofessional group including dietetics, occupational therapy, speech and language therapy, clinical psychology, physiotherapy, and critical care expertise.⁴
• Comparison:
  • Standard NHS care at the recruiting site.
  • Sites were ineligible if they provided an ongoing structured rehabilitation programme similar to iRehab.
  • Standard care was documented and monitored during the trial; 39 of 52 sites had some ICU follow-up clinic or service, but content, eligibility, frequency, and staffing were variable and generally not equivalent to structured iRehab.
• Blinding: Participants and interventionists were not blinded, but trial staff collecting outcomes were blinded to allocation. This is appropriate for a rehabilitation trial, but the self-reported primary outcome and acceptability measures remain vulnerable to expectation and contact effects.
• Statistics: A total of 428 participants was required to detect a 0.08 between-group difference in EQ-5D-5L utility score, assuming SD 0.20, 90% power, two-sided α=.05, and 30% loss to follow-up; the planned allocation was 231 intervention and 197 standard care participants. The primary analysis followed the intention-to-treat principle using adjusted linear regression for baseline score, stratification variables, and prespecified covariates. Secondary outcomes were exploratory without multiplicity correction.
• Follow-Up Period: Primary outcome at 8 weeks after randomisation, with further outcome assessment at 6 months.

Key Results

This trial was not stopped early. It completed planned recruitment: 3705 patients were screened, 563 registered, and 429 were randomised.

• The primary result was clinically and statistically borderline but negative: adjusted MD 0.04 was half the prespecified MCID of 0.08, and the 95% CI crossed zero.
• Secondary outcomes showed a coherent signal for physical function, fatigue, anxiety, acceptability, and 6-month EQ-5D-5L, but these analyses were exploratory and unadjusted for multiplicity.
• The prespecified ventilation-duration subgroup suggested that iRehab may be more useful in shorter-stay, less persistently disabled ICU survivors, but this requires confirmation.

Internal Validity

• Randomisation and Allocation: Central, 24-hour web-based randomisation with minimisation by hospital site and duration of invasive ventilation provided strong allocation concealment. Baseline measures were collected before randomisation.
• Dropout and Exclusions: Post-randomisation outcome loss was material: primary outcome data were available for 178/231 (77.1%) in iRehab and 167/198 (84.3%) in standard care. Loss to follow-up participants were younger and had higher baseline anxiety and depression scores, raising a plausible risk that missingness was not completely random.⁴
• Performance and Detection Bias: Blinding participants and interventionists was not feasible. Blinded outcome assessment mitigated detection bias, but the primary outcome was self-reported EQ-5D-5L and therefore susceptible to expectation, therapeutic contact, and reporting effects.
• Protocol Adherence: Delivery was strong for the main 1-to-1 intervention: 205/231 (88.7%) participants attended at least one 1-to-1 appointment and 169/231 (73.2%) attended at least 5 of 6 appointments. Group components were less used: 67/205 (32.7%) attended at least one group exercise session and 60/205 (29.3%) attended at least one peer-support café.
• Fidelity: Interventionists required certification at ≥90% before independent delivery and maintained ≥80% competency thresholds for drift monitoring. Fidelity remained high, with no evidence of systematic drift.
• Baseline Characteristics: Groups were broadly comparable: mean age 55 years in both groups, mean APACHE II 18.4 in both groups, ICU length of stay 18.7 vs 18.9 days, ventilation duration 11.3 vs 11.4 days, and baseline sit-to-stand 7.5 vs 7.6 repetitions. There were some numerical comorbidity imbalances, including diabetes at 26.0% vs 15.2%.
• Heterogeneity: Heterogeneity was central rather than incidental. Participants varied by recovery trajectory, mechanical ventilation duration, comorbidity, mental health symptoms, and baseline function. The ventilation-duration interaction indicates that a single undifferentiated post-ICU rehabilitation pathway may be too crude.
• Timing: Mean time from hospital discharge to randomisation was 57.4 days in iRehab and 57.9 days in standard care. This tested an early post-discharge programme, but not an immediate ICU-to-home transitional model. Patients with rapidly evolving recovery or very prolonged critical illness may have required different timing.
• Dose: The delivered 1-to-1 dose was substantial for a remote model, but the programme lasted only 6 weeks. For participants ventilated longer than 7 days, this may have been insufficient for persistent weakness, complex multimorbidity, and prolonged recovery needs.
• Separation of the Variable of Interest: Separation was credible. iRehab participants had structured weekly specialist contact, action planning, symptom management, exercise, psychological support, and optional group sessions. Standard care was variable; although 39/52 sites had some follow-up service, it was generally ad hoc and not equivalent to iRehab.
• Crossover: Crossover to the intervention was not a major issue. The more important issue was non-adherence within the intervention arm: 26/231 (11.2%) had zero 1-to-1 appointments.
• Adjunctive Therapy Use: Usual care included variable follow-up, signposting, and local services; co-enrolment occurred in 145/429 (34%) participants across other trials. This improves pragmatic realism but may add background noise to recovery trajectories.
• Outcome Assessment: EQ-5D-5L is standardised and policy-relevant, but generic. The 30-second sit-to-stand added a valuable performance-based outcome, although remote assessment and missingness reduce certainty.
• Statistical Rigor: The planned adjusted intention-to-treat analysis was appropriate. Sensitivity analyses were less supportive than the headline borderline result: multiple imputation gave adjusted MD 0.04; 95% CI −0.003 to 0.09; P=.07, and alternative EQ-5D crosswalk scoring was also non-significant. Secondary outcomes were not multiplicity adjusted.

Conclusion on Internal Validity: Internal validity is moderate to strong. Randomisation, allocation concealment, blinded outcome assessment, intervention fidelity, and prespecified analysis were strengths; open-label delivery, self-reported outcomes, differential missingness, heterogeneity, and exploratory secondary/subgroup findings temper confidence.

External Validity

• Population Representativeness: The trial was geographically broad across 52 NHS hospitals and included patients from deprived areas, but the enrolled cohort was highly selected. Only 429 of 3705 screened patients were randomised, and 1109 people declined participation.
• Home-Discharged Survivors: The findings apply best to ICU survivors discharged home after at least 48 hours of invasive ventilation. They do not apply well to patients discharged to care homes, rehabilitation units, those unable to engage with trial procedures, those with severe mental health problems precluding group participation, or patients with exercise contraindications.
• Ethnicity and Digital Access: Participants were predominantly White (88%), and trial participation required telephone or computer access. The intervention may be less generalisable to populations with digital exclusion, language barriers, cognitive impairment, unstable housing, or limited social support.
• Healthcare System Applicability: A centralised remote team could plausibly translate to other high-income and some middle-income systems with telehealth infrastructure, but it requires trained rehabilitation specialists, governance for remote exercise safety, and pathways for escalation.
• Severity Spectrum: The lack of benefit in participants ventilated for more than 7 days limits generalisability to the most persistently impaired ICU survivors, who may require longer, more intensive, or facility-based rehabilitation.

Conclusion on External Validity: External validity is moderate for home-discharged, digitally reachable ICU survivors in NHS-like systems. It is limited for patients with prolonged critical illness, major cognitive or mental health barriers, institutional discharge, or inadequate access to remote care.

Strengths & Limitations

• Strengths:
  • Large, pragmatic, national multicentre post-ICU rehabilitation trial.
  • Central randomisation with allocation concealment and blinded outcome assessment.
  • Theory-informed, patient-centred, multicomponent intervention co-developed with ICU survivors and multiprofessional stakeholders.
  • High 1-to-1 intervention adherence and high treatment fidelity.
  • Careful characterisation and monitoring of standard care, reducing comparator ambiguity.
  • Use of both patient-reported and performance-based outcomes.
  • Important process and fidelity data that make the trial unusually informative even though the primary outcome was negative.
• Limitations:
  • Primary outcome did not demonstrate a clinically meaningful or clearly statistically significant effect at 8 weeks.
  • Open-label rehabilitation and self-reported primary outcome increase risk of expectation and contact effects.
  • High non-participation and differential follow-up reduce representativeness and may bias estimates.
  • EQ-5D-5L may be insufficiently responsive to the multidimensional recovery targets of a complex rehabilitation intervention.
  • Secondary outcomes and subgroup analyses are hypothesis-generating because they were exploratory and not multiplicity adjusted.
  • The intervention’s multiple components make it difficult to identify which element, if any, drove benefit.
  • Six weeks may be too short for those with prolonged mechanical ventilation or persistent critical illness.
  • Economic evaluation and full process evaluation were not reported with the main clinical results.

Interpretation & Why It Matters

• Primary message

  iRehab does not support routine, universal provision of this 6-week remote multicomponent programme to all home-discharged ICU survivors with the expectation of improving EQ-5D-5L by 8 weeks.
• Clinical signal

  The intervention was feasible, acceptable, safe, and associated with improved sit-to-stand performance, fatigue, and anxiety, suggesting that it may address some patient-important domains even when a generic health utility measure is unchanged.
• Targeting matters

  The ventilation-duration subgroup suggests that rehabilitation may need to be stratified by recovery phenotype: shorter-ventilation survivors may benefit from remote self-management and exercise support, whereas longer-ventilation survivors may need more intensive, longer, or facility-supported rehabilitation.
• Practice implication

  Remote rehabilitation should be viewed as one modality in a broader post-ICU recovery pathway, not as a standalone replacement for structured in-person or facility-based rehabilitation in patients with major disability.

Controversies & Other Evidence

• Borderline primary result: The adjusted primary estimate, MD 0.04, was below the 0.08 MCID and the CI crossed zero. Treating this as positive would overstate the finding; treating it as null without recognising the 6-month and secondary-domain signals would understate the therapeutic signal.
• Single primary outcome problem: The accompanying editorial argued that a single generic primary outcome may not fully capture a complex survivorship intervention targeting physical, psychological, social, and self-management domains. It also highlighted the severe persisting functional impairment: sit-to-stand values around 8 repetitions at baseline, around 10 at 8 weeks, and around 12 at 6 months remained far below age-based norms and below commonly cited thresholds for community independence.⁵
• EQ-5D-5L is defensible but imperfect: EQ-5D-5L is recommended in critical illness rehabilitation core outcome work, but it may be insensitive to granular changes in strength, fatigue, anxiety, participation, and self-efficacy.⁶
• Complex intervention interpretation: iRehab was a multicomponent intervention with interacting components and behavioural mechanisms. The MRC framework warns that such interventions require interpretation across mechanisms, implementation, context, and outcome patterns rather than a single endpoint alone.⁷
• Uptake is a major implementation barrier: Among those who either registered or declined, 1109 of 1672 declined participation. Systematic review evidence on ICU follow-up models shows that feasibility, timing, flexibility, travel burden, and perceived need are major determinants of engagement.⁸
• Consistent with other recent neutral post-ICU support trials: Recent trials of a mobile critical care recovery programme, a post-ICU telehealth care model, and face-to-face multidisciplinary post-ICU consultations did not establish clear patient-centred benefit, reinforcing that post-ICU recovery interventions are difficult to make broadly effective without better targeting and dose selection.⁹¹⁰¹¹
• Current rehabilitation standards are broader than iRehab: Contemporary rehabilitation reviews and guidance emphasise a continuum from ICU to ward to community, with individualised, progressive, multiprofessional rehabilitation, detailed dose reporting, and implementation strategies. iRehab should be interpreted as evidence about one remote post-discharge model, not as evidence against rehabilitation after critical illness.¹²

Summary

• iRehab randomised 429 home-discharged ICU survivors who had required at least 48 hours of invasive ventilation to a 6-week remote multicomponent rehabilitation programme or standard care.
• The primary 8-week EQ-5D-5L result was negative: 0.69 vs 0.67; adjusted MD 0.04; 95% CI −0.001 to 0.09; P=.05, below the prespecified 0.08 MCID.
• Secondary outcomes favoured iRehab for sit-to-stand performance, fatigue, anxiety, acceptability, and 6-month EQ-5D-5L, but these findings are exploratory.
• The intervention was feasible and delivered with high fidelity, but uptake into the trial and follow-up completion were important limitations.
• The most plausible future role is targeted remote rehabilitation for selected recovery phenotypes, not universal deployment as a single standard post-ICU pathway.

Overall Takeaway

iRehab is an important trial in which the primary endpoint was not met, rather than a practice-changing demonstration of short-term health-utility benefit. It shows that a well-designed, acceptable, remotely delivered post-ICU rehabilitation programme can be implemented at scale, but that modest domain-specific improvements do not necessarily translate into clinically meaningful short-term health utility gains across a heterogeneous ICU survivor population.

Bibliography

• 1.Herridge MS, Azoulay É. Outcomes after critical illness. N Engl J Med. 2023;388(10):913-924.
• 2.Cazeta BBR, de Queiroz RS, Nacimento TS, et al. Effects of exercise interventions on functioning and health-related quality of life following hospital discharge for recovery from critical illness: a systematic review and meta-analysis of randomized trials. Clin Rehabil. 2024;38(7):898-909.
• 3.O’Neill B, Bradley JM, Connolly B, et al. Remote multicomponent rehabilitation compared to standard care for survivors of critical illness after hospital discharge (iRehab): a protocol for a randomised controlled assessor-blind clinical and cost-effectiveness trial. NIHR Open Res. 2025;5:29.
• 4.O’Neill B, Bradley JM, Connolly B, et al; iRehab Trial Investigators. Supplemental content to: Remote multicomponent rehabilitation in intensive care unit survivors: a randomized clinical trial. JAMA. Published online May 18, 2026.
• 5.Moncion K, Smith-Turchyn J, Kho ME. The future of survivorship in critical illness. JAMA. Published online May 18, 2026.
• 6.Connolly BA, Barclay M, Davies C, et al. PRACTICE: development of a core outcome set for trials of physical rehabilitation in critical illness. Ann Am Thorac Soc. 2024;21(12):1742-1750.
• 7.Skivington K, Matthews L, Simpson SA, et al. A new framework for developing and evaluating complex interventions: update of Medical Research Council guidance. BMJ. 2021;374:n2061.
• 8.Dimopoulos S, Leggett NE, Deane AM, Haines KJ, Abdelhamid YA. Models of intensive care unit follow-up care and feasibility of intervention delivery: a systematic review. Aust Crit Care. 2024;37(3):508-516.
• 9.Khan BA, Perkins AJ, Khan SH, et al. Mobile critical care recovery program for survivors of acute respiratory failure: a randomized clinical trial. JAMA Netw Open. 2024;7(1):e2353158.
• 10.Bakhru RN, Flores L, Cain JM, et al. A randomized controlled trial of a post-ICU telehealth care model (WFIT). Am J Respir Crit Care Med. 2025;211(9):1662-1670.
• 11.Sharshar T, Grimaldi-Bensouda L, Siami S, et al; Suivi-Rea Investigators. A randomized clinical trial to evaluate the effect of post-intensive care multidisciplinary consultations on mortality and the quality of life at 1 year. Intensive Care Med. 2024;50(5):665-677.
• 12.Hodgson CL, Ayre E, Broadley T, et al. Standard of care for rehabilitation in critical illness. Intensive Care Med. 2026.