Supporting People data linking project: update

The Welsh Government’s legacy Supporting People Programme (SPP) provided housing-related support. The programme aimed to help vulnerable people avoid homelessness and provided both long-term and short-term support (as outlined further below). This report outlines the analysis of the latest SPP data up until January 2020 from five local authorities in Wales (Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend) that have continued to provide data and builds on the findings from previous SPP data linking studies. It includes the demographics of SPP clients and an analysis of their healthcare usage.

• More younger people than older people received SPP support.
• Within the 15 to 29 and 30 to 44 age categories, more females than males were in receipt of SPP services.
• SPP clients were predominantly from areas with high levels of income deprivation.
• GP events, A&E visits and emergency hospital admissions were all notably higher for SPP clients than individuals in the matched control group.

SPP provided help to vulnerable people in Wales to live as independently as possible and was based on an approach focussing on prevention and early intervention. It aimed to help people who were at risk of being homeless, as well as assist people with other complex needs, such as those experiencing domestic violence, those with mental health needs and people with substance misuse problems.

It provided long-term maintenance support to help people keep or gain independence (and avoid the need for more costly interventions such as entering care). It also delivered short-term preventative services to those who were at risk of becoming homeless, by providing people with the help and skills they needed to live in their own homes, hostels, sheltered housing or other specialist housing.

From April 2019, SPP was replaced by the Housing Support Grant (HSG) which is an amalgamation of three grants including the SPP. The HSG’s overall purpose is to prevent homelessness and support people to have the capability, independence, skills, and confidence to access and/or maintain a stable and suitable home. As outlined in the Housing Support Grant Practice Guidance (2023 p3):

The HSG is an early intervention grant programme to support activity, which prevents people from becoming homeless, stabilises their housing situation, or helps potentially homeless people to find and keep accommodation’. Furthermore, ‘It supports vulnerable people to address the, sometimes multiple, problems they face, such as debt, employment, tenancy management, substance misuse, violence against women, domestic abuse and sexual violence, and mental health issues...

In 2018, the Supporting People data linking project: emerging findings report aimed to identify any reduction in NHS utilisation amongst people who have received SPP support. It outlined initial and provisional findings using linked administrative data analysis of SPP and health data in the Secure Anonymised Information Linkage (SAIL) databank and compared patterns of NHS use before and after SPP support across 19 local authorities in Wales with different comparison groups.  The SAIL databank at Swansea University is a trusted research environment through which approved researchers can access de-identified administrative data for approved research purposes.

Since then, more recent SPP data have been deposited in the SAIL databank for some local authorities. This report focusses on the data from January 2003 to January 2020, the vast majority of which is legacy SPP data. The last 10 months of data cover the early days of the HSG period, during which local authorities continued to collect data under the old SPP framework and deposit it in the SAIL databank.

Demographics

For the five local authorities in Wales that continued to provide data to the SAIL databank, the demographics of those receiving SPP support at the point of first contact were analysed. Figure 1 outlines that more younger people than older people received SPP support. Within the 15 to 29 and 30 to 44 age categories, more females than males were in receipt of SPP services. This indicates that the programme was primarily responding to the particular needs of a younger, female client base. The reasons for this could be investigated further in future research. Figure 2 outlines that the SPP client base was primarily within the first and second Welsh Index of Multiple Deprivation (WIMD) (2019) income quintiles, representing the most income deprived individuals between 2003 and 2020. WIMD is a measure of concentrations of deprivation at a small area level in Wales. 

Figure 1: Supporting People clients by gender and age group (from January 2003 to January 2020)

Image

Description of Figure 1: This bar chart shows the gender and age category of SPP clients from five local authorities in Wales. There was a greater proportion of younger (under 45) SPP clients than older (45 and over) SPP clients. There were more female than male SPP clients in in the 15 to 29, 30 to 44 and 75 and older age groups.

Source: Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend local authorities (combined).

Figure 2: Supporting People clients by gender and WIMD 2019 income quintile (from January 2003 to January 2020)

Image

Description of Figure 2: This bar chart shows the level of income deprivation and gender of SPP clients from five local authorities in Wales. There were more SPP clients in the most deprived income quintile than in any of the other quintiles. There were more females than males receiving SPP support in each of the income quintiles.

Source: Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend local authorities (combined).

Use of GP services

To assess the use of healthcare services, the number of visits to a GP can be used as a headline indicator. Within each GP visit, a number of ‘events’ can be recorded, for example representing activities such as measuring blood pressure, or prescribing medication. These are known as ‘GP events’.  Multiple GP events can occur during a single visit to the GP.

For this analysis, the number of days on which GP events occurred were aggregated and divided by the potential number of users of GP services to give the number of days on which GP events occur per user (as an indicator of healthcare use). This was calculated for both SPP clients and a comparator group who had not used SPP services (see under Methodology below for more detail on how this group was defined).

Figure 3 outlines the patterns of health service use for SPP clients in relation to the number of GP event days per user for each selected 30-day period before and after receipt of SPP support (which would have occurred between January 2003 to January 2020). A steady increase in GP event days per user can be observed before SPP support commences, followed by a sharp increase just afterwards. GP use remains relatively high up to 6 months afterwards. This is unsurprising given that the intention of the SPP and the HSG going forward, is to support vulnerable people and the complex and sometimes multiple problems they face.

Figure 3: Number of days on which GP events occurred per Supporting People client by time period (before and after the provision of support)

Image

Description of Figure 3: This bar chart shows the number of days GP events occurred per SPP client, from five local authorities in Wales, before and after SPP support was provided compared with the matched control group. The highest number of GP event days occurred one month after support began.

Source: Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend local authorities (combined).

A & E visits

Figure 4 shows A&E visits by SPP clients for each selected 30-day period before and after SPP support began, compared with the matched control group. This analysis indicates that A&E visits by SPP clients show an increase (compared with the control group) from 3 months before the start of support, followed by a notable drop just afterwards. A&E visits remain relatively high one and two months after SPP support was provided compared with the matched control group.

Figure 4: Number of A & E visits per 100 Supporting People clients by time period (before and after the provision of support)

Image

Description of Figure 4: This bar chart shows the number of times SPP clients from five local authorities in Wales visited A&E, before and after the provision of SPP support, compared with the matched control group. There were most visits to A&E one month before SPP support.

Source: Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend local authorities (combined).

Emergency hospital admission 

Hospital admissions can occur for any number of reasons and can be planned or unplanned. Emergency hospital admissions are defined here as unplanned, urgent and unexpected admissions to hospital.

Figure 5 shows emergency hospital admissions for SPP clients for each selected 30-day period, before and after SPP support (which would have occurred between January 2003 to January 2020), compared with the matched control group.

Emergency hospital admissions for SPP clients are highest 1 month before the start of support, then fall quickly 1 month after SPP support and continue to fall up to 12 months after SPP support. SPP clients are at least twice as likely to experience an emergency hospital admission than individuals in the matched control group for the selected periods. This indicates that SPP clients are more likely to be hospitalised and suggests that their health needs are more complex than other groups within the Wales population.

When considering emergency hospital admissions together with A&E visits for SPP clients, the overall trends appear similar, however emergency hospital admissions are less likely to occur than A&E visits for the selected periods.

Figure 5: Number of emergency hospital admissions per 100 Supporting People clients by time period (before and after the provision of support)

Image

Description of Figure 5: This bar chart shows the number of emergency hospital admissions of SPP clients from five local authorities in Wales, before and after the provision of SPP support, compared with the matched control group. There were most admissions 1 to 2 months before SPP support began.

Source: Newport, Torfaen, Blaenau Gwent, Cardiff and Bridgend local authorities (combined). 

GP events, A&E visits and emergency hospital admissions are all notably higher for SPP clients than individuals in the matched control group. This suggests that people receiving SPP support have a very particular set of health needs relative to the Wales adult population and that their healthcare utilisation is considerably different. This is the case even when age, gender and an area-based measure of deprivation are taken account of. As noted above, SPP and HSG clients are a vulnerable group, with complex and sometimes multiple problems and it is not unexpected that their health needs differ considerably from those with otherwise similar socio-demographic characteristics. The findings indicate that the programme is prioritising a vulnerable population with increased healthcare needs in Wales.

The demographic analysis shows that most SPP clients were from areas with high levels of income deprivation (WIMD income quintiles 1 and 2) and were from younger age groups (below 45), with a higher proportion of females in the youngest and oldest age categories. This is an indication that there are wider inequalities at play that impact SPP clients and potentially those in receipt of the current HSG.

This analysis highlights areas for potential further research. For example, the gender, age and socioeconomic dimension could be explored in more detail to understand inequalities. In addition, to investigate the full impact of the SPP and the HSG going forward, the repeat and long-term homelessness of those in receipt of support could be considered. The different health economic approaches could also be analysed as a means of evaluating the programme.

The basic method used for this analysis is essentially the same as that used for the Welsh Government Supporting People data linking feasibility study (2016) (the ‘feasibility study’). The feasibility study aimed to assess the feasibility of data linking to help evaluate SPP and understand the impact of the programme on health service use.  Data linking establishes links between administrative data sources so that de-identified information relating to the same person, family, place or event can be matched for research purposes. The feasibility study identified that there were data quality, management and acquisition challenges and considerable variation in data across different local authorities in Wales. It highlighted that creating a control group for analysis would permit the most credible way of assessing the impact of SPP and recommended undertaking a full data linking study of SPP across all local authorities in Wales.

Administrative records for over 88,000 users of SPP services across Wales were collected, anonymised and deposited in the SAIL Databank. For most local authorities, client data related to years 2015 onwards but in some cases record-level data were provided as far back as 2003 and up to 2021 in others. Following encryption, the SPP data were selected for those coming into contact with SPP services up to 1^st January 2020. The selected data were linked with healthcare datasets (GP data, A&E data, and hospital admissions data) using an Anonymised Linking Field (ALF). The number of GP visits, A&E visits, and emergency hospital admissions per selected SPP client were extracted based on the timing of health care events in relation to the SPP support provided.  Selected 30-day periods (12, 6, 3 and 1 months before and after the SPP intervention began) were used to understand the possible impact of the SPP as a policy intervention and its impact on use of health services.

To evaluate the impact of a policy intervention, it is necessary to estimate the counterfactual scenario to understand what would have happened without the intervention (by observing what happens to people who did not receive the intervention). A comparable group of people in Wales who had not been in contact with SPP services was therefore selected to compare with those who had. For each SPP client with linkable data, at least two other individuals were selected based on similar characteristics as defined by age, gender and area-based income deprivation (WIMD income quintile).  

Records for the matched control group were extracted and combined into a single dataset with those for SPP clients and processed together simultaneously. Both were linked to the same GP and hospital healthcare datasets by applying the same method as used in the previous reports. This updated analysis provides new insights as it presents figures for the matched control group compared with the five authorities who continued to provide SPP data to the SAIL databank.

Timeliness and punctuality

The aim of this evidence brief is to publish analysis as soon as possible using the available data. This research has been delayed due to the COVID-19 pandemic. In the meantime, however, local authorities have continued to provide SPP data to SAIL. It has been possible to incorporate this most recent data into the findings for this report, which are based on data just for those local authorities who provided SPP data to SAIL up until January 2020.

As noted above, the last 10 months of data cover the early days of the HSG period, during which local authorities continued to collect data under the old SPP framework and deposit it in the SAIL databank.

In 2021, the Welsh Government agreed that a new outcomes framework for the HSG should be developed to accurately capture the core purpose of the grant and its wider benefits. This new framework will be the successor to the legacy SPP outcomes framework and will capture the core purpose and all support services provided by the HSG.

Comparability and coherence

The SAIL Databank contains records of A&E visits and hospital admissions only for those individuals registered with a GP in Wales. However not all GP practices in Wales provide data to SAIL. Whilst the results for A&E and hospital admissions relate to all those registered with a GP in Wales, the results for GP events only relate to those registered at GP practices that provide data to SAIL.  As at March 2023 approximately 80% of GP practices had agreed to provide data to SAIL.

Limitations

This report has been developed using data recorded for those who have come into contact with SPP services for whom it is possible to link data with other datasets (for example healthcare). For some users of SPP services, it has not been possible to record adequate data. The healthcare utilisation patterns shown above are therefore representative only of those for whom it has been possible to collect the required information. The patterns may therefore not be representative of all SPP clients (as in some cases, insufficient information was recorded).

Given the SPP is intended to support people facing complex challenges, it is likely that the control group in this analysis will differ from SPP clients in many ways, so is perhaps not an ideal comparator. For this reason, when attempting to evaluate the impact of SPP, it is more instructive to compare trends in healthcare usage over time between the client group and the control group rather than the amount of healthcare usage.

Local authorities (not just those covered in this release), across Wales have collected SPP client data and uploaded it to the SAIL Databank. Their input is gratefully received, along with housing policy staff who have helped to develop this report.

ADR Wales brings together data science experts at Swansea University Medical School, staff from the Wales Institute of Social and Economic Research and Data (WISERD) at Cardiff University and specialist teams within the Welsh Government to develop new evidence which supports the Programme for Government by using the SAIL Databank at Swansea University, to link and analyse anonymised data. The ADRW Programme of Work 2022-2026 outlines the ten thematic areas that the ADR Wales team will focus their research on to help government address the most pressing issues facing society.

ADR Wales is part of ADR UK and funded by the Economic and Social Research Council (part of UK Research and Innovation).

Statistician: Tony Whiffen
Email: ADRWales@gov.wales

Media: 0300 025 8099

Research report number: 58/2023