PROJECT NIHR206126.02
Administrative data for evaluating the social care system
BACKGROUND
Social care policy evaluation is limited by a lack of quantitative evidence, and in particular of linked, individual-level data about the characteristics of individuals with care needs and their carers, the support they receive (type, amount and quality, provider characteristics) and the care outcomes generated.
AIMS
This project aims to continue ASCRU’s assessment of different sources of administrative social care data to generate new evidence about how the social care system in England works and to understand the potential and limitations of using administrative evidence for policy evaluation.
The study will use different types of administrative data to generate new findings about key policy questions regarding the commissioning and targeting of social care support, and of their impact on the care system’s costs and outcomes. While carrying out these evaluations, the project will explore the strengths and limitations of the different sources of administrative data used in the estimations.
The project will examine two broad sets of policy questions:
-
What are the patterns of social care use in England and their consequences on costs and outcomes? The analysis will explore how local authority-brokered social care support is targeted on different individuals, how this targeting varies across areas, and the consequences of different support models on local performance (e.g. patterns of institutionalisation; costs; substitution and complementarity between services). The project will combine cross-sectional and longitudinal analyses, to explore how lifetime patterns of service use vary across individuals and areas at a point in time, and to estimate the lifetime cumulative patterns of service use (e.g. to derive lifetime care cost estimates). Subject to data availability, the project will also examine the interdependencies between health and social care use.
-
How do social care markets operate? The second set of questions relate to the way in which individual-level, care package unit costs and service quality vary with the characteristics of the care recipient, of the care provider, of local area characteristics, and of local commissioning arrangements. We will focus these analyses on home care, residential care and nursing care services, for which we will have access to individual-level data from two regions in England.
METHODS
The project will use data from:
-
Local authority social care records. These records will include structured data on care needs from assessments and reviews; and continuous data about the content of care packages through time (e.g. type of service, cost and provider if available);
-
Care quality management systems (e.g. local authority collected, provider level local minimum standards; CQC ratings);
-
Market capacity and supply, from Capacity Tracker and CQC.
These data will be collected for all adult social care user groups. They will be matched with other data about local characteristics (e.g. from Census and other ONS datasets) to control for the effect of place on the care system.
These data will be analysed using a range of statistical and data science methods, including data visualisation, multivariate regression and simulation methods.