Work Package 1

A quantitative work package focused on measuring how often older care home residents need emergency treatment in hospital in England.

Work Package Aims & Objectives

The aim of this work package is to measure variation in hospital attendances and admissions of care home residents across England.

This will be achieved through the following objectives:

Create a nationally representative cohort of care home residents clustered by postcode districts containing one or more care home and integrated care system (ICS) (years 2022-2024).

Use multi-level regression models (adjusting for resident and care home factors) to report the difference between observed and predicted rates of Emergency Department (ED) attendance/admission for each postcode district and integrated care system.

Planned Investigation

We are using healthcare data collected by GPs (between 2022 and 2024) and a database of care homes in England. The healthcare data used in this study is from the Optimum Patient Care Research Data (OPCRD). More details about OPCRD can be found at the link to the right. The database of care homes is the Care Quality Commission (CQC) register. All of the data we use is anonymised so people cannot be identified.

Using this data, we will use a multi-level regression model to measure how often older residents are treated in hospital as an emergency in different parts of England.

Residents will be grouped by postcode district and districts will be mapped to Integrated Care Systems (ICS) across England, as shown in the figure below:

The model will aim to adjust for resident case mix and care home factors, allowing an estimate of remaining unexplained variation at the level of postcode district and ICS, potentially reflecting differences in Enhanced Health in Care Homes services.

OPCRD | Real-world Patient Data for Evidence-based ResearchGains insights with Optimum Patient Care Research Database (OPCRD). Access de-identified patient data, analyse real-world information, and drive evidence-based research for healthcare outcomes.

From the model, we will be able to create a type of graph called a funnel plot. The figure to the right shows an example of a funnel plot:

The funnel plots will show care home containing postcode districts and ICSs with observed-versus-expected ED attendances and inpatient admissions. This will help us identify which postcode districts and ICSs have higher-than-expected, average and lower-than-expected rates of emergency department attendances and inpatient admissions of older care home residents.

Where there are differences in how often residents are treated in hospital in different areas it could be due to the NHS providing better or worse enhanced services. For example, if in one area care home residents go to hospital less than other areas, it could mean NHS services in that area are working well. Alternatively, it could mean that care home residents are not being transferred to hospital when they need to be.

Anticipated Outputs & Impact

The findings will be disseminated to the Care Quality Commission’s Data analytic team and South Yorkshire Integrated Care home systems lead to assess how the statistical modelling developed in WP 1 could be used by commissioners and regulators to monitor and plan services.

Approvals

Ethical Approval from the University of Sheffield Research Ethics Committee: Reference no. 065432, 13/11/2024

ADEPT Approval: ADEPT0325, 29/01/2025

Report abuse