Exploring variation in population health and social care outcomes: implications for performance assessment in the NHS and social care outcomes frameworks (Responsive mode)

The Outcomes Frameworks are to be a key method to assess the performance of the health and social care system. The strength of using outcome indicators, such as health-related quality of life, for this purpose is that they get directly to the heart of what the health and social care systems aim to achieve. The challenge is finding a way to identify the impact of the activity of the care system on the observed changes in these outcome indicators at population level. The difficulty is that a wide range of other factors also drives changes in these indicators. For example, important mitigating factors are long-term trends in chronic conditions including co-morbidities, and environmental and economic factors.

This project group is mainly concerned with addressing this attribution challenge and therefore helping to cast light on the performance of relevant organisations. There are analogies with our ‘using outcomes information to improve decision-making about care and support’ project. That work focuses on using outcomes information at the service or intervention level, while this project is about the use of outcomes information to inform performance at the system level. In particular, the NHS Outcome Framework is mainly focused on questions about the progress of the NHS overall, and by organisational units such as CCGs. Similarly, the Adult Social Care Outcome Framework (ASCOF) mainly concerns activity at the local authority level.

The intention is to develop methods for attribution so that we can measure the impact of NHS and social care activity on the headline outcome indicators in the NHS OF and the ASCOF.

Aims and methods

Strand 1: NHS Outcomes Framework

We have recently begun a new QORU variation-to-contract project concerning the developing of a performance assessment methodology around Domain 2 of the NHS Outcomes Framework.
This work strand was initiated as a responsive mode request.

We will (continue) to work with the DH on indicators and impact methodologies regarding Domain 2 of the NHS Outcomes Framework. The aim is to find ways to benchmark NHS performance by estimating a ‘no change’ future profile for the (main) Domain 2 indicator (currently EQ-5D) and comparing actual patient average scores in the future. We are using data from the GP patients survey for this purpose and estimating the extent to which the above factors, such as LTC conditions/co-morbidities, account for changes in EQ-5D scores.

The intention is to undertake developmental work to investigate and populate a more sophisticated benchmarking framework that allows better attribution methods to isolate the impact of NHS (and social care) performance changes (improvements) on the relevant Outcomes framework indicators (e.g. Domain 2). This work will include both a consideration of the methods that can be used for attribution and development of the relevant information and analysis.

The main methods would be conceptual and statistical analyses. In particular, we will explore the variation in health-related quality of life (EQ-5D) of patients in the GP patient survey. Most of the data linkage occurs using small area approaches.

Strand 2: Adult Social Care Outcomes Framework (ASCOF)

The aim of this study is to develop a methodology to adjust the outcome indicators collected in the Adult Social Care Outcome Framework (ASCOF) so they more closely estimate the contribution made by local authorities to improve the wellbeing of social care service users. More specifically the method should allow comparisons between LAs and within LAs over time to be made on a more comparable basis.

The focus of this work will be on the Adult Social Care Survey (ASCS) outcome indicators, most notably social care-related quality of life as measured using the Adult Social Care Outcomes Toolkit (ASCOT). Statistical analysis will be used to ‘risk adjust’ the ASCOT score so that comparisons between LAs are not affected by the different mix of service users that are eligible for care support.

Various methods, such as the residual method and the production function method, will be used to analyse the ASCS data. We will explore the variation in social care-related quality of life of service users in the ASCS.

Timing and outputs

The main outputs would be models of the relationship between outcome change and mitigating factors. These relationships could be used to inform analysis of the NHS and social care performance. They would also help us to better understand the key inter-dependencies between LTC conditions, such as depression, disability levels and other factors.

We expect to produce papers in the methodologies that could be used for attribution or risk adjustment. We will then provide adjustment factors, potentially in toolkits (e.g. spreadsheet tools).

As regards strand 2, we aim to provide support and advice to the Social Services User Survey Group (SSUSG) in the further development of the ASCS.

This work started in phase I and will be ongoing through phase II.