BACKGROUND

​

Large numbers of people require social care support following their hospital discharge. Hospital discharge arrangements vary significantly across hospitals and local authorities, but in recent years a key policy and practice drive has been to discharge patients as soon as their needs no longer meet the “criteria to reside” in hospital, using Home First and Discharge to Assess (D2A) discharge “models”.

 
An important element of the discharge process involves the assessment of care needs, both whilst the person is in hospital and once they are in the community. These assessments are crucial for understanding the person’s short and long-term care and support needs. How these assessments are carried out therefore drives the support that individuals receive, and the associated discharge costs and outcomes (wellbeing, destination, re-admissions etc.). 

Although there is growing evidence about the use and perceptions of discharge to assess arrangements, our understanding of their impact is still limited. ASCRU colleagues are already looking at ‘discharge to assess’ and exploring local configuration of D2A schemes and the impact of different D2A arrangements on outcomes across London Boroughs and NHS stakeholders for ESHCRUII. There have also been (mainly descriptive) studies by the AHSNs. 

​

AIMS

​​​

This project will build on existing analyses carried out by ASCRU colleagues of the way in which D2A arrangements are implemented and their impact on health and social care systems. Building on this work, it will analyse:

 

• Processes used for assessing the care needs of older individuals discharged from hospital (e.g. assessment styles and documentation, professionals involved, information recording systems). This analysis will explore the assessment of whether the person meets the criteria to reside in hospital, their care needs assessment and associated expected care plan whilst in hospital, the assessment of their care needs and care support immediately post discharge and the Care Act assessment to identify a possible long-term support plan.

• Impact of different care management arrangements on patterns of care and care outcomes for different hospital patients. The study will aim to quantify the relationship between models of assessment and length of stay in hospital, type of support provided post-discharge, costs through time and available care outcomes (e.g. rates of long-term care home admission, readmission rates, death rates). 

​

METHODS

 

This project will use mixed methods to evaluate the hospital discharge process and how decisions about which support is provided post discharge are taken in the hospital and in the community. 

​

A process evaluation will be carried out in five areas across different regions across England. Key professionals involved in the discharge process will be invited to participate in an interview to describe local processes and to explore their perceptions of the strengths and limitations of their local systems. This element of the work will develop local process maps and identify key care management arrangements, which will be used to interpret the findings from the quantitative analysis. This component of the study will explore factors perceived by stakeholders to hinder the implementation of effective support packages post discharge (e.g. supply constraints, challenges with information sharing, caseloads).

​

The quantitative analysis will use administrative data from health and social care departments.  This analysis will cover the five local areas included in the process evaluation, but we will also seek to include other authorities (e.g. in London and the North East) in which we have access to suitable administrative care records. The aim will be to use linked, individual-level administrative care records to examine variations in models of support post discharge, and their consequences on costs and outcomes for different types of hospital discharges. We will seek to interpret variations in post-discharge support models to different care management arrangements. We expect to use a range of statistical modelling techniques, including survival models and multivariate regression models.

​