Theme 3: Effective use of data

Key findings

As was flagged under the social prescribing theme, data-linkage initiatives combining NHS health data, income, housing and fuel poverty data are vital to effectively target those experiencing fuel poverty and related health conditions.
There are clear examples of innovation in this space, with the UrbanTide uZero platform demonstrating what is possible when relationships are established, datasets linked, and insightful and user-friendly dashboards developed.
However, the NHS can be hard to access as an outsider, with high levels of data protection and systems under pressure. The Cheshire and Merseyside project worked well because it came from within the NHS, but such success may be harder to achieve if initiated externally.
Cheshire and Merseyside’s provision of a toolkit and blueprint guide is a useful way to support other areas to develop their own data dashboards and subsequent projects.
In the cases below, fuel poverty data is available at a granular level. Consideration should be given to how the Scottish fuel poverty definition could be used to target support in a similar way.
Where the Scottish definition cannot be operationalised, it is important to consider what data proxies are appropriate. The types of data used by UrbanTide to develop their fuel poverty metric are a good starting point here.

Case study 15: Cheshire and Merseyside fuel poverty project

Key details: linked dataset utilised to create trailblazer fuel poverty projects for target adult and child groups in Cheshire and Merseyside. First trailblazer commenced in 2023.

Organisation(s): NHS Cheshire and Merseyside, Cheshire and Merseyside Integrated Care Board (ICB), Health Innovation West Coast, Optum UK, Graphnet

Locations(s): Cheshire and Merseyside, specifically trailblazer projects in St. Helens, Knowsley, Warrington, The Wirral.

Duration: Ongoing, first trailblazer project in St. Helens commenced in February 2023

Who is supports: The dataset discussed below identified two target cohorts:

An adult group with a COPD diagnosis who lived in neighbourhoods with the highest rates of fuel poverty and carried a 50% or higher risk of an emergency admission.
A children’s group aged 0-4 years prescribed a salbutamol inhaler in the past 12 months, and with a 5% or higher chance of emergency admission.

What is delivered: Project brings together the NHS, voluntary and community sector and local authority partners. NHS Cheshire and Merseyside and Health Innovation North West worked in conjunction with Optum UK and Graphnet to create a linked dataset which combines GP data with secondary care, mental health, social care and other socio-economic datasets. Specifically, the approach brought together the following, as defined in the scheme’s blueprint guide:[35]

Analysis of fuel poverty hotspots: drawing on official government data, this identified the Lower Layer Super Output Areas (LSOAs) of around 1,500 people or 650 households with high levels of fuel poverty across the ICS.
Predicted and actual health care utilisation data: using the Johns Hopkins Adjusted Clinical Groups (ACG) algorithms, the Integrated Care Board (ICB) population was also divided up into Resource Utilisation Bands (RUB) reflecting expected and actual usage of health care in acute and primary care settings.
Further stratification of fuel poor populations: local CIPHA data[2] also enables teams to assess the vulnerability of its fuel poor population using diagnosed condition, risk of admission, mortality risk and other factors such as living alone.

The data showed that 1.54m people across Cheshire and Merseyside have a medical condition that would seriously increase their risk of harm if their home was poorly heated. 490,000 of those are known to live in a fuel poverty ‘hotspot’, defined as being in the top 20% of the most fuel poor areas in England.

Optum then facilitated workshop sessions, effectively utilising multi-agency working to involve clinicians and other professionals across NHS, local authority and voluntary sector organisations to interrogate the data and work together to develop targeted actions to support people at risk. Trailblazer projects were developed off of the back of this work. These projects each deliver tailored clinical and non-clinical interventions, such as medicines optimisation, pulmonary rehabilitation referrals, support for smoking cessation, and signposting to financial assistance.

Health Innovation North West Coast has produced a fuel poverty toolkit. This brings together assets developed with the teams across Cheshire and Merseyside to identify, engage and support those at greatest risk. This includes guidance on how to access the CIPHA dashboard, a blueprint guide on how the project was developed, and an asset map of all partners engaged with as a part of the project.

Evaluation and evidence: Graphnet undertook an evaluation of the project’s fuel poverty dashboard in 2023[36]. Two groups were targeted, members of the Fuel Poverty Steering Group for the project and Fuel Poverty Dashboard Users (clinical and non-clinical). This was a very small scale and targeted evaluation, utilising a survey (n=11 of a possible 21) and interviews (n=8 of a possible 15). In terms of the Fuel Poverty Steering Group, key findings include:

100% agreed that the fuel poverty dashboard provides an opportunity to improve the quality of care to high-risk patient experiencing fuel poverty.
91% agreed that data analytics helped draw meaningful insight to address fuel poverty.
91% agreed that the partnerships established within the fuel poverty project have contributed to achieving fuel poverty goals and objectives, and increased access to resource for COPD and Asthma patients living in fuel poverty.

The interviews with fuel poverty dashboard users found that:

The fuel poverty dashboard provides actionable insights to support the design of preventative interventions.
Promotes understanding of social determinants of health and the delivery of holistic health and care services[37].

There is not yet evaluation available for all the trailblazer projects, however an outcome evaluation 12 months following intervention has been published for the St. Helens and Knowsley project[38]. The purpose of the evaluation was to investigate whether the fuel poverty pilot in St. Helens and Knowsley place led to a reduction in primary and or secondary care activity in the first 12 months following a patient intervention. A differences-in-differences analysis was applied to the cohort comparing outcomes in the 12 months (52 weeks) after intervention between people who had received an intervention within the pilot and patients who did not. It should be noted that the cohort size for this intervention was small (254 patients in each of the intervention and control groups).

The fuel poverty pilot was associated with a 9.8% reduction in GP appointments (400 appointments avoided) in Knowsley and St. Helens place compared to what would otherwise have been expected. Within the scope of activity for Knowsley and St Helens, for the 254 patients analysed, this is estimated to have avoided approximately £14,800 from primary care in the first-year post intervention (based on indicative GP consultation costing £37). Scaled to 1000 patients, which would have avoided approximately 1,575 GP appointments or approximately £58,275.

No impact was found on emergency admissions or A&E attendance due to the intervention, nor was there any impact found on primary care prescriptions[39].

Case study 16: UrbanTide collaboration with Sunderland City Council

Key details: a 2022 pilot project with Sunderland City Council to identify ageing households most at risk of fuel poverty and cold and damp homes. The platform integrated smart meter data with health, housing and demographic data to better target care interventions.

Organisation(s): UrbanTide, Sunderland City Council (SCC)

Location: Sunderland

Duration: Pilot collaboration with SCC took place in 2022, but work to develop UrbanTide’s AI platform uSmart began in 2015

Who it supported: Ageing households most at risk of fuel poverty and cold and damp homes

What was delivered: UrbanTide is a small company which supports government, local authorities, and private companies and organisations to unlock data and utilise AI to solve global challenges. Between 2021 and 2022, UrbanTide led a UKRI-funded project to develop a method for identifying customers who are in, or at risk of fuel poverty. The project team merged a deep understanding of vulnerability and fuel poverty with expertise in artificial intelligence (AI) and data science, supported by partners including the Data Communications Company, University College London and the Connected Places Catapult. In parallel, UrbanTide also contributed AI capabilities to a related project led by UK Power Networks and Energy Systems Catapult (“Social Connect”), which similarly used smart meter, socio-economic and geographic data to identify households in need of support. This combination of expertise was essential for identifying and analysing relevant data, and for developing a predictive tool capable of estimating, at a street-by-street level, the prevalence of customers who are in or at risk of fuel poverty. This platform has been called uZero and incorporates smart meter data provided by the Data Communications Company, EPC data as a proxy for the energy efficiency of homes, Index of Multiple Deprivation/Scottish Index of Multiple Deprivation data, data on benefits claimants, and income data in order to produce a fuel poverty metric. The smart meter data is key to the platform, and contains information such as when emergency credit is activated, or a prepayment meter is topped up. The selected datasets have been integrated into a digital tool designed to analyse and present insights on customers estimated to be at risk. A user-friendly interface was also developed, enabling users to explore and interrogate the data[40].

In 2022, UrbanTide worked on a pilot project with Sunderland City Council to support them to identify ageing households most at risk of fuel poverty and cold and damp homes. The adapted platform integrates smart meter data with health, housing and demographic data to provide reliable insights that help SCC efficiently plan support services and better target care interventions[41]. Datasets utilised included GP data related to obesity, asthma, and other health conditions linked to deprivation. The result was 1 billion cross-sector data points used to identify and target support services.

Evaluation and evidence: none available.

Case study 17: Warm Homes Programme, Bedford

Key details: In December 2022- July 2023, 53 households received an energy efficiency installation following identification as having a chronic health condition exacerbated by living in a cold damp home and as living in areas of high deprivation and fuel poverty.

Organisation(s): Bedford Borough Council, NHS Arden and Greater East Midlands Commissioning Support Unit (CSU), The National Energy Foundation

Location: Bedford

Duration: December 2022 – July 2023

Who it supported: Residents living in areas of high deprivation and fuel poverty who had chronic health conditions which could be exacerbated by living in a cold and/or damp home.

What was delivered: The Warm Homes programme was implemented against a backdrop of concerns about rapidly rising fuel costs, the cost of living generally and a particularly cold winter. Bedford Borough Council worked with the CSU’s Business Intelligence Team to identify patients with specified health conditions living in an area with high levels of deprivation and fuel poverty. These patients were flagged on local GP systems and sent letters on behalf of the GP practice inviting them to participate in the Warm Homes programme. The programme was also promoted through partner organisations who could signpost people to self-refer to the programme[42].

The Warm Homes programme offered eligible residents improvements which could make their homes warmer and/or less damp. The main improvements offered were the installation of replacement gas boilers, loft insulation and/or cavity wall insulation.

An additional intended benefit of the programme was to reduce carbon emissions, to help the Council meet its targets towards addressing climate change.

This was a fairly small scale project. 1,635 people were invited to participate via their GP surgery. 112 people contacted the programme to find out more and an additional 7 people came forward after learning about the programme through wider communications or partner organisations. After assessment of their eligibility and requirements, 53 people received an installation.

Evaluation and evidence: Evaluation of the project was carried out in two parts. An initial mixed-methods evaluation produced in September 2023 which reviewed wider issues affecting householders’ ability to keep their homes warm and dry; analysed household and programme activity data, and analyses of health outcomes data (numbers of GP appointments, A&E attendances and inpatient admissions) for those invited to participate in the programme who did not receive an installation; those who received installations as part of the programme, and a control group in a neighbouring local authority. In autumn 2024, the health outcomes analysis was repeated to reflect the potential longer term impacts of the interventions. This further analysis used difference-in differences regression modelling.

In the initial evaluation in September 2023, of people who had received an installation (n=53):

81% reported a positive change in their rating with regards to the warmth of their home.
The proportion of people who felt warm or very warm rose from 26% to 70%.
45% showed a positive change with regards to how worried they felt about paying their energy bills.
54% believed that their home has less of a negative impact on their health after receiving an installation.

The overall average cost of the programme per household was £2,514. The mean lifetime carbon saving for households who received an installation was approximately 14,000 kg CO2 (based on estimated lifetime Carbon savings of each installation).

The further analysis in autumn 2024 found that:

Rates of A&E attendances and hospital admissions were similar across the programme and control cohorts. However, at this level of granularity, the majority of patients have no contact with NHS hospitals.
There was a material decrease in primary care activity in January-March 2024 for patients who received an installation, falling by 1.2 appointments per patient.
Given an average number of four appointments per quarter, this decrease is also practically significant from the patients’ perspective[43].

When interpreting these evaluation findings, it is again worth noting the relatively small number of installation recipients, as well as the fact that any additional actions that people may have taken to keep their homes warm or look after their health are not accounted for.

[2] CIPHA is a linked data platform which brings together health and social care data from across Cheshire and Merseyside. Originally established to support the region’s COVID-19 response, CIPHA now helps NHS and partner organisations across the ICS to understand and map population health data to support service design and improvement.