Individually randomised controlled multi-centre trial to determine the clinical and cost effectiveness of a home-based exercise intervention for older people with frailty as extended rehabilitation following acute illness or injury, including embedded process evaluation

DESIGN: Pragmatic, multi-centre (10 sites) RCT with two-level, partially nested hierarchical design, internal pilot and embedded process evaluation. SETTING: Recruitment from elderly medicine, trauma & orthopaedics wards in 10 UK hospitals across 2 hubs (Yorkshire/South West) and from linked intermediate care services. Participants randomised to the intervention (n=400) will complete a 24 week home-based exercise programme supported by a therapist trained in the HOPE manual. POPULATION: Older people with frailty admitted to hospital following acute illness or injury, and discharged home from hospital or from intermediate care. INCLUSION CRITERIA: Age>65; frailty (Clinical Frailty Scale (CFS) 5-7); mobility (complete Timed Up-and-Go Test, TUGT); capacity to give consent, or named consultee for assent willing to support intervention delivery; able to communicate via telephone. EXCLUSION CRITERIA: Permanent care home residents; moderate/severe dementia; severe stroke; unstable angina/recent MI; another household member in trial; severe frailty (CFS 8); terminally ill (CFS 9); palliative care; referral for disease specific rehabilitation; unable to complete TUGT. INTERVENTION: 24 week home-based exercise intervention (HOPE programme) delivered by community rehabilitation staff, + usual care. The HOPE intervention consists of 12 week HOPE programme (5 home visits & 7 telephone sessions) plus 12 further weekly telephone sessions. CONTROL: Unrestricted usual care provided by primary, community and social services. RANDOMISATION: Individually randomised via CTRU in a 1.25:1 allocation ratio, stratified by site; discharge setting; level of HOPE programme (level 1, 2, or 3); reason for admission (acute illness or injury). COSTS/OUTCOMES: Baseline assessment; follow-up at six & 12 months; routine GP, HES & ONS data. PRIMARY: SF36 physical component summary (PCS) at 12 months. SECONDARY: SF36 mental component summary (MCS); Barthel index; Nottingham Extended Activities of Daily Living (NEADL) index; EQ5D-5L; mortality; falls; new care home placement; hospital readmission; health/social care resource use; cost effectiveness; intervention adherence; patient experience. SAMPLE SIZE: 718 patients (318 control, 400 intervention) will provide 90% power, 5% significance, to detect minimum clinically important difference (MCD) of 3 points on SF36 PCS (SD 9.47), accounting for 25% loss to follow-up, and clustering in the intervention arm (20 cluster size, 0.03 ICC). ANALYSIS: Intention to treat, reported according to CONSORT. Primary analysis will compare mean SF36 PCS scores between groups using a random-effects heteroscedastic model. Cost effectiveness analyses will report differences in cost of service use between groups and ICERs using QALYs derived from i) EQ5D-5L and ii) SF36/SF6D (sensitivity analysis). INTERNAL PILOT: In 4 sites with progression criteria assessed at 6 months (provision & recruitment); 9 months (acceptability); 12 months (follow-up). Intervention provision assessed through participants receiving 1st home visit within 3 weeks; acceptability through retention of participants; follow-up through completion of primary outcome. Criteria for all: green = >80%; amber = or equal to 65%; red = or equal to 2 pts/month/site; red = <2pts/month/site. RECRUITMENT: The recruitment target is 718 patients across 10 sites over a total of 23 months, with recruitment staggered to accommodate the internal pilot. TIMETABLE: Set-up 1-9m; recruitment 10-32m; internal pilot progression criteria assessed at 16m (provision/recruitment), 19m (acceptability), 22m (follow-up); follow-up 16-44m; data cleaning, analysis & reporting 45-51m. A 15 mth recruitment extension was approved in September 2019, and the recruitment period increased from 23 mths to 38 mths. EXPERTISE: Clinical, community rehabilitation, trial expertise, statistics, health economics, qualitative, PPI Important changes to methods after trial commencement: The sample size has increased from 718 to 742 (325 control, 417 intervention) in September 2019. The revised sample size calculations incorporate updated assumptions, based on data collected at the time of a recruitment extension, predicting final numbers of trained therapists and numbers of patients per therapist and ensuring loss to follow-up is minimised as far as possible. No change was made to the original ICC estimate of 0.03.