Document resource
Introduction There is growing pressure to offer ‘Hospital at Home’ (H@H) as an alternative to in-patient care for worsening heart failure (HF). NHS England published guidance on H@H, based primarily on observational data, soon after the publication of our UK pilot randomised controlled trial (RCT) that showed H@H increased days alive out of hospital but there is concern over a higher rate of hospitalisation by 60 days.Objective The Safe@Home-HF Survey aims to assess if there is enough uncertainty regarding safety to justify a multicentre RCT.Methods An online survey was distributed via the NIHR Regional Research Delivery Network Study Support Service, Safe@Home-HF investigators and British Society for HF to ≥55 healthcare organisations managing patients with HF, between January 2020 and January 2026. Questions addressed H@H service availability, and if there is enough uncertainty regarding safety to justify a randomised trial of H@H care. Participants were asked to consider examples of equipoise (uncertainty regarding safe discharge), rating the likelihood of patients being safely managed at home. The clinical scenarios were scored on a five-point Likert scale (1=very likely unsafe, 3=equipoise, 5=very likely safe).Analysis: Likert scale responses are presented as median [interquartile range] to assess the degree of clinical uncertainty.Results Consultants and nurses from 40 sites responded by January 2026. Twenty-eight (70%) offer H@H, delivering IV diuretics to an estimated 2932 of 23316 patients annually (12.6%). More sites offer H@H compared to our 2022 survey when 9.7% of patients with HF were estimated to be treated by H@H.Safety perceptions varied, with clinical scenario scores ranging from ‘very likely unsafe’ to ‘very likely safe’ (1-5) for all but two cases which scored 1-4 (very likely unsafe to likely safe). [Figure 1] This variability highlights the lack of consensus on H@H safety. [Tables 1 and 2]73% (29/40) of respondents reported sufficient uncertainty regarding safety to justify an RCT. 35 of 40 centres (88%) expressed interest in participating in an RCT. It is feasible to randomise 2 patients per-month per site. In addition, 5 patients can be enrolled into a registry for patients who meet eligibility criteria but choose not to be randomised or cannot be randomised due to organisation factors (e.g. no bed). Figure 2 shows where H@H services are based.Conclusion There is enough uncertainty regarding safety of outpatient delivery of IV diuretics for HF to justify an RCT according to 73% of respondents in this survey. There is a wide spectrum of opinion regarding whether H@H is a safe management strategy.Abstract 346 Figure 1Abstract 346 Table 1Perception of uncertainty regarding safety in patients who would have been excluded in the only UK based pilot trial comparing home IV diuretic treatment vs standard inpatient care (1 is very unlikely safe, 3 is equipoise, 5 is very likely safe)Examples of patients that would have been excluded in the pilot trialMedian score [IQR]Acute kidney injury stage 33 [2,3]Blood Potassium >6.5mmol/L1 [1,2]Blood Potassium <3.0mmol/L2 [2,3]Anaemia (haemoglobin 70g/l) with no acute active bleeding3 [2,3]Presyncope with postural hypotension -20mmHg drop in systolic BP on standing to 89 systolic, but normal aortic valve and no arrhythmia in last 48 hours3 [2,3]Active infection - fever, white count, and CRP improving with antibiotics3 [2,4]Tachypnoea at rest-respiratory rate 20 per minute2 [2,3]Hypoxaemia at rest breathing air (oxygen saturation 86%)1 [1,2]Abstract 346 Figure 2Abstract 346 Table 2Perception of safety (1 is very unlikely safe, 3 is equipoise, 5 is very likely safe) of IV diuretic treatment at home in other grey case scenariosGrey case scenariosMedian score [IQR]Sodium 128mmol/L (120 yesterday)4 [3,4]Breathless on minimal exertion (from chair to exam couch), basal crackles with oedema to thighs; not hypoxic4 [3,4]NYHA 4, orthopnoea and paroxysmal nocturnal dyspnoea in last 3 days with crackles to midzone but not hypoxic.3 [2,3]Three falls with syncope in last month but no postural hypotension or aortic stenosis or arrhythmia in last 48 hours2 [2,3]Mild troponin elevation with dyspnoea but no chest pain or ECG changes4 [3,4]NT-proBNP over 5000 in 65-year-old woman4 [2,5]Frail 80-year-old man with dementia and multiple comorbidities (stable except DHF); lives with carer3 [2,4]65 years old woman who lives alone (no family and social support)4 [3,5]