Scaling Hospital at Home Through Smarter Patient Identification
Context
South Eastern Health and Social Care Trust (SET) worked with Akeso to assess, develop and operationalise MetiSense, an Akeso-developed decision-support tool that identifies suitable Hospital at Home patients earlier, and provides an evidence-led basis for sustainable service expansion.
SET’s Hospital at Home (HaH) service provides an alternative to hospital admission for clinically suitable patients. The service was already highly valued by patients and plays an important role in supporting community-based hospital avoidance.
However, the Trust had limited visibility of potentially suitable patients within Emergency Department (ED) and inpatient pathways. Identification relied heavily on clinician time, awareness and referral behaviours, rather than a consistent approach across the system. This created a risk that people who could have received hospital-level care at home were not identified or referred.
The Trust wanted to understand the scale of this unmet demand, whether more patients could transfer home earlier and what this would mean for the future capacity of Hospital at Home. The work also supported a wider ambition to reduce pressure on acute services while providing care in a setting better suited to the needs of older people living with frailty.
Challenge
The central challenge was not whether HaH could provide clinically effective care, it was how the Trust could identify suitable patients earlier and more consistently and ensure the service had sufficient capacity to support them.
Within SET’s busy ED and inpatient environments, manual identification placed additional pressure on clinical teams and could not provide a systematic view of potential demand. The Trust also needed to understand what scaling the service would require in practice, including the operating model, workforce, capacity and investment needed to convert identified demand into safe and sustainable delivery.
A data-led solution had to do more than produce an analytical result. It needed to reflect real clinical decision-making, support rather than replace professional judgment, and provide practical information that teams could use within existing workflows.
Our Approach
Working alongside HaH, ED and Frailty-at-the-Front Door teams, Akeso combined clinical engagement, pathway review, advanced analytics and machine learning.
The work began with a rapid six-week discovery phase. Akeso mapped how patients were identified and referred, reviewed ED and inpatient pathways, and worked with clinical teams to understand the factors influencing referral and acceptance decisions.
Our approach combined:
Pathway Review
Mapping how patients were identified, assessed and referred across ED and inpatient settings
Clinical Engagement
Working with frontline teams to understand decision-making and test whether model outputs aligned with clinical practice
Machine Learning
Developing a predictive model using historic patient and referral data
Clinical Validation
Reviewing model results with clinicians and incorporating additional variables to improve alignment with professional judgement
Operational Testing
Developing and testing a clinician-facing interface that generated prioritised patient lists for review and referral decisions
Service Modelling
Translating identified demand into capacity, workforce, financial and bed requirements
To move from analysis to operational use, Akeso developed and tested a clinician-facing interface through which teams could review prioritised patient lists. The refined model was then used to quantify future demand and inform a HM Treasury Green Book-aligned business case and phased implementation roadmap for scaling HaH.
Impact
The analysis gave the Trust a clearer view of unmet demand for Hospital at Home and quantified the potential impact of expanding the service.
of patients were identified as potentially clinically suitable for Hospital at Home but were not accessing the service
were eligible and referred but were not admitted because of capacity constraints. Initial modelling indicated that admitting all suitable patients to the Hospital at Home service could potentially release:
equivalent to 10% of the Trust’s inpatient bed base, significantly improving flow. The service and financial modelling also demonstrated the potential to release:
acute bed days released annually
Estimated net present value over 10 years
These are modelled benefits rather than realised savings. The business case and roadmap provided SET with a structured basis for implementation and informed decisions about the future scale of Hospital at Home.
The outputs moved the Trust from a broad ambition to expand Hospital at Home towards a clearer understanding of who could benefit, how those patients could be identified and what sustainable expansion would require.
Wider Impact
MetiSense demonstrates how machine learning can support clinical teams to identify suitable patients earlier, while keeping professional judgement at the centre of decision-making.
The work also showed that predictive accuracy alone is not enough. Digital tools must align with clinical pathways and be supported by realistic workforce, capacity and investment plans.
By combining machine learning, clinical insight and operational planning, the programme created stronger foundations for moving appropriate care closer to home and reducing pressure on acute services. The next step is to embed MetiSense within clinical workflows, test its impact through live delivery and use that evidence to guide sustainable HaH expansion.
Contact Our Experts
Sarah Pinto-Duschinsky
Managing Partner
sarah.pinto-duschinsky@akeso.co.uk