Counting the Cost of Choice and Control: Evidence for the costs of self-directed support in Scotland

2 OVERVIEW OF STUDY

2.1 In this chapter we give details of the aims, design and methods used in this study. Further details, particularly of the sampling frame, data sources, and the ways in which we analysed the data and tested for the validity and reliability of our findings, as well as ethical considerations, can be found in Appendix 1.

Aims and design of the study

2.2 The overarching aim of this project was to establish the current costs of SDS in three Local Authorities in Scotland, and from that baseline estimate the future costs of giving users more choice and control using SDS options. The most robust evidence available to date is around DPs and 'standard' community care packages rather than the fuller range of options envisaged in the forthcoming legislation. We have therefore based our methods on using the DP evidence base. This project gathered macro-level financial and economic evidence on the actual and potential costs, benefits and impacts of an increase in the uptake of DPs and DP-type SDS services in Scotland.

2.3 The study aimed to:

• establish the baseline and current position of DP services in three Local Authorities;
• compare the costs for a group of DP users with a matched group of users receiving standard services for a stratified sample of users controlled for user group and need level; and
• use any differences between levels of resources allocated to DP users compared to matched recipients of arranged services revealed by this data to model and predict the future costs associated with an increase in uptake of DP-type services within those Local Authorities.

Methods

2.4 We used a flexible design and drew on methods which had been used and validated in previous studies, in particular interrogating a range of existing health and social care-related datasets to inform a microsimulation model (Bowes and Bell, 2007) and using questions developed and validated in related studies (Glendinning et al, 2000; Glendinning et al 2008) as the basis for establishing the costs of SDS for users. The data collected via these different methods contributed to our understanding of the present and future costs of SDS in Scotland, although we drew primarily on the qualitative data from interviews with Local Authority/service commissioners and service providers, and on both qualitative and quantitative data from the comparison of data on DP users and users of standard services, the survey of users and other data in the public domain.

2.5 The following outlines the methods and the information collected in each element of the study:

Reference Group

2.6 We appointed a reference group of SDS users, carers and advisors who were not based in the sample Local Authorities to provide assistance in the design of the costing methodology, and to check the external validity and reliability of findings.

Local Authorities

2.7 Three Local Authorities were selected to take part in the study based on key criteria (range and volume of DP users, non-overlap with SDS test sites). It was important that participating Local Authorities would be able to provide data on sufficient numbers of DP users across the range of user groups for use in the microsimulation modelling process. SDS test sites were not approached as it was felt that these might not be reflective of the national position in relation to the implementation of SDS.

2.8 All three selected had a high volume of DP users across several user groups, and together represented a mix of urban and rural provision. In advance of confirming their inclusion in the study it was established, as far as possible, that the data required for the study was available and accessible. All three Local Authorities agreed to take part in the study.

2.9 Data on the profile of DP users in the three Local Authorities and the costs of their packages was collected using a pro forma. The pro forma provided an indication of the required fields for data extraction for the purposes of the research. In order to minimise the additional burden on the participating Local Authorities and facilitate the data extraction process the pro forma was based on selected elements of the Scottish Government Health Directorate's 2011 data specifications for annual returns related to Self-directed Support and Home Care respectively.

2.10 The overhead costs associated with DPs in each of the Local Authorities were examined. This included costs of central services, transport, accommodation etc. The data was collected through a combination of email and telephone conversations, structured preliminary meetings and structured interviews with relevant Local Authority officers. In total 14 Local Authority officers were consulted. Participants in preliminary meetings (n=10) included Heads of Social Work and Health and senior staff in contract management, business performance, research and information services, and financial teams supporting social work. Interviewees' (n=5) held senior roles in social work development, assessment and care management services (2), contracts management and direct payments team management. One person was both a meeting participant and an interviewee.

Service providers

2.11 A focus group with 8 participants was conducted with a service provider with experience of providing services across different Local Authorities. The participants included the CEO, Finance Director, Contracts and Commissioning Officer, Director of Client-related Services and a number of Regional Managers. We asked about: experiences of SDS implementation across different Local Authorities and understandings of the experiences of their clients; experiences of Local Authority contract models and anticipated changes; costing and provision of services and whether this would change with different numbers of SDS users; perceptions around support required by and provided for SDS users; workforce requirements in anticipation of changes in SDS use; and threats and opportunities around the promotion and implementation of SDS.

2.12 We identified and interviewed service providers (with experience of providing services to DP users) in the private and third sectors operating in different geographical areas and providing services to a mix of different client groups. The interviews aimed to estimate the providers' setup and running costs for SDS packages (two interviews with single interviewees and two group interviews with 2 and 4 interviewees respectively). We also asked them about perceived opportunities for and obstacles to expansion should demand increase, including both labour and finance constraints.

2.13 We then estimated set up and brokerage costs for providers using data provided by service provider interviewees.

2.14 An interview with representatives of a DP support organisation (2 participants) was also conducted to provide data on costs of support services for DP users, to discuss the implications for support organisations should numbers of SDS users increase, and to get support organisation perspectives cost-related issues raised by interviewees from other stakeholder groups.

Service users

2.15 The three participating Local Authorities were asked to provide the anonymised records of 'matched' samples of SDS (DP) users and 'standard' users (receiving care services arranged by the Local Authority). Users included in these samples were matched in terms of type of disability² and basic demographic details such as sex and age. Local Authorities were requested to stratify the samples in terms of disability type but to select user records randomly within each type.

2.16 The sample sizes varied between participating Local Authorities, based on numbers of SDS users and the authority's resources available to access the required information. Data requested as part of the anonymised records included: age; gender; marital status; household structure (if known); service user type; risk band; date current SDS / arranged package started; details of current SDS / arranged package in terms of numbers of hours and type of support; sources of funds included in package (DP, ILF etc); date of assessment; date of care plan; and details of previous package(s).

2.17 In sampling the non-DP users for comparability, we matched user groups, service users and levels of need as far as possible. The sample of DP users for the user-derived costs was stratified according to levels of need using Local Authority data. Additional data was also collected on need levels from users themselves, using a structured questionnaire and validated measures such as the GHQ IADL scores to ensure our findings were valid and reliable (see para 2.18).

2.18 Supplying the requested data provided a challenge for all of the participating authorities, as individuals' records often spanned different computer and paper-based systems. In total 265 usable anonymised records were returned.

User survey

2.19 A postal survey of the sample of DP users provided by the Local Authorities was conducted to gain structured information about levels of need (using GHQ IADL³ and other validated measures). The questionnaire was initially tested with the reference group, advisory group and with the three Local Authorities. The wording on a number of questions was subsequently modified to try to provide greater clarity of purpose and to reflect variations in terminology and differences in ranges of services included in DP packages across the participating Local Authorities.

2.20 The questionnaire sought information on the size of the DP package, the length of time the user had been receiving the DP, the support services purchased using the DP, other services purchased and demographic data. Respondents were also asked why they chose a DP and whether they were satisfied with the value of the DP. Data on costs including set up costs and hidden costs borne by users and carers was also collected. We distributed 210 questionnaires to DP users and received 59 usable responses.

Modelling

2.21 Desk based analysis was undertaken to simulate different rates of increase in SDS provision to estimate future demand and costs, using information on the prevalence of various types of disability and population estimates by Local Authority from GROS.

Validation of findings

2.22 We tested the reliability and validity of our findings by checking them with our reference group and by carrying out a workshop at the University of Stirling to which we invited all the research participants, together with a range of other stakeholders.

The participating Local Authorities

2.23 Table 2.1 details the characteristics of the three Local Authorities which took part in this study with respect to uptake and value of DPs.

Table 2.1 Characteristics of participating Local Authorities

Authority A was a primarily urban Local Authority. It was the largest of the participating authorities in terms of the population that it served, but the smallest in terms of area. Authority A had a rate of receipt of DP per 10,000 population higher than the Scottish average, and an estimated median value of DP package higher than the Scottish median value.

Authority B was a primarily rural Local Authority. It had the smallest population of the three participating Local Authorities but was responsible for the largest area. Authority B had a rate of receipt of DP per 10,000 population higher than the Scottish average and an estimated median value of DP package lower than the Scottish median value. Authority B has a relatively larger proportion of older DP users compared to Authorities A and C.

Authority C was a mixed urban-rural Local Authority. It had a population smaller than Authority A but larger than Authority B, and covered an area larger than that of Authority A but smaller than that of Authority B. Authority C had a rate of receipt of DP per 10,000 population lower than the Scottish average and an estimated median value of DP package higher than the Scottish median value.

Costings data

2.24 Data gathered from Local Authorities was likely to miss some components and under-report the full costs of packages of care (particularly of DPs). Data gathered from individual users therefore gave a fuller picture of the range of costs to users and carers. However, there were problems of recall or lack of data about full amounts, frequency and type of services and timings, and users could not be expected to know details of the costs of 'standard' services. We therefore used both data sources to allow for the strengths and weaknesses of each approach, relying on Local Authority data and a modified version of the costings used by Curtis (2009), adjusted for inflation and the local context, and on the results of the postal questionnaire of users.

2.25 For a fuller discussion on reliability and validity of instruments and results and limitations of the data please see Appendix 1, paragraphs 8.3 to 8.7.