Executive Summary
Study objectives and approach
This report considers the extent, nature, causes and implications of variations in the unit costs of services provided under the Supporting People programme. A series of key objectives are addressed:
- To establish whether meaningful unit cost variations exist for comparable services in England and Scotland, and to test a set of possible explanations.
- To identify unit costs for key services across Scotland and to analyse the extent and causes of variation in unit costs between local authorities and providers.
- To explore the extent to which unit variations reflect variations in:
- providers costs
- the scope of the services eligible for SP funding
- commissioning strategies
- To identify and quantify factors or drivers of variations in unit costs.
- To consider whether opportunities exist to improve value for money, and to make recommendations.
- To consider whether it is practical to promote convergence of costs for comparable services, and to identify services where such convergence may secure value for money.
- To provide advice on action which the lead officers' network might take to benchmark costs and to promote best value in service delivery.
The study comprised two key elements:
- Statistical analysis to identify and explore relationships within the secondary data. Data were available from England, enabling the study to consider the cost distributions across authorities and services, both within Scotland, and between Scotland and England.
- A series of case studies - eight in Scotland, three in England. The Scottish case studies each focused on one of the "super groups" (amalgamations of client groups which have been used as the basis for analysis because of data problems 1). The English case studies focused on the older people and learning disabilities groups. The case studies collected information on the services provided, costs, staffing, budget and commissioning.
In addition, a brief consultation meeting was held with the SP lead officers. This involved participating in the first meeting of the Lead Officers' Group, Benchmarking Sub-group, held in March.
Key findings
The analysis of the secondary data demonstrates a clear link between the use that is made of the Supporting People funds (in terms of the client groups that are being provided for) and the unit cost of provision of services. The link is identified in both Scotland and England. In general the analysis suggests where the mix of clients is skewed towards older people, the unit cost is lower overall; where the mix of clients is skewed towards groups such as learning disabilities or mental health problems, the unit costs tend to be higher.
This result is, no doubt, unsurprising - some service cost most than others, and the local authorities that deliver disproportionately more of the higher cost services will have higher overall costs. However, it is worth noting that the analysis finds that the relationship is more straightforward in England than it is in Scotland. Put technically - when these principles were used to specify a multiple regression model to explain the variation in unit costs, the best fit model derived from the English data had significantly more explanatory power than the model derived from the Scottish data.
There is, therefore, a substantial amount of variation between Scottish authorities that is not explained by the client-mix relationship, and which must come down to other - possibly local - factors. The statistical analysis therefore highlighted a number of key issues to be explored in the case study work, including:
- The reasons why there are wide variations in unit costs between local authorities for the same client group across Scotland.
- the reasons why the cost of provision for specific services in Scotland are strikingly different from unit costs in England, particularly for services such as older people and learning disabilities.
The case studies also found marked variations in the intensity of support provided to different client groups - with fewest support hours provided to the older people client group and most support hours to the leaning disabilities client group. There was much less variation in the support cost per hour - indeed, the analysis suggests that per hour costs for the learning disability client group is only 5% higher than that for the older people client group. Therefore, the key determinant of cost variation is the number of hours of support.
Overall the study concludes that the average costs for older people in England are substantially affected by the high proportion of "alarm based" projects. These projects are very cheap to run on a unit cost basis - especially where on a large scale. If such projects are removed, the cost differences between Scotland and England narrow, but still remain.
Cost variations between Scotland and England for the older people group are moderate but still remain when allowance is made for the extensive use of alarm systems in England. The case studies do not provide a clear conclusion as to whether higher costs in Scotland reflect higher contact hours or higher hourly costs. While our expectation is that higher contact hours are the main factor, we also note that some costs included in Scotland (e.g. building maintenance, capital costs of alarm services) are not included in English costs (covered by rents/ HB).
Rather different conclusions emerge for the learning disabilities group in England and Scotland. Unit costs in Scotland are about 50% higher than in England according to the secondary data analysis. The case study work pointed clearly to the conclusion that this difference is due to differences in the hours of care provided. Scottish projects examined had between twice and three times as many hours per client per week as English projects. There seems to be a clear policy difference, with some English authorities "capping" housing support hours per week, or not providing housing support when large care packages are already in place - both measures are designed to ensure that resources are targeted towards people who are most likely to move towards/live as independently as possible and ensure that housing support is not replacing care. The strong impression is that Scottish authorities are less strict about enforcing this boundary issue and that housing support to this group in Scotland at times contains an element of "care".
There appears to be more unit cost variation within Scotland than within England at the service level - suggesting greater variety of provision of services.
Within Scotland unit cost variation is driven both by hours of support per week and by hourly rates but to varying degrees. For older people the main driver of cost variation was hours of contact - though high costs were also associated with high hourly costs. It is less clear why hourly direct staff costs should vary for floating support services to older people since issues such as shift working or unsocial hours payments are less prevalent than in accommodation based projects.
For people with learning disabilities the picture is somewhat clearer. As in the comparison with England, high costs are primarily associated with high numbers of hours of support. There is some evidence in the case studies that higher hours also produce a higher hourly rate. We consider that "high hours" projects almost certainly involve support to persons with considerable needs and provision of support which borders on care.
Variation in unit costs for mental health projects appear to be related to hours provided and to the balance between floating and mixed support. Hourly and staff related costs for mental health projects are notably high and we consider that this is likely to reflect the specialist skills needed to work with this client group.
Homeless projects comprise accommodation-based and other projects. Accommodation-based projects are relatively expensive because they usually involve some element of 24 hour/7 days a week cover. Variations in costs for accommodation-based projects reflect the intensity of the required "supervision" and the extent to which the project is seeking to engage with clients. Other homeless projects involve very low intensity contact with large numbers of clients. Staff costs per hour may be higher than in accommodation based projects because staff require different qualifications. Variation in unit cost for these projects will be related to the number of clients per staff member and the intensity of contact.
We do not consider that there is any evidence to indicate that higher costs in Scotland than in England - insofar as they exist - necessarily demonstrate poorer Value for Money - rather they reflect different choices over the nature and intensity of services provided. The clearest difference between Scotland and England is in relation to support to people with Learning Disabilities. As we have seen, the intensity of support is, on average, far higher in Scotland. This is a policy issue rather than a value for money issue.