Predictive RIsk Stratification Models: AssessmenT of Implementation Consequences (PRISMATIC 2)

Contact details

Barbara Gomes (study manager), Swansea University

Principal / Lead Investigator
Co-Investigators / Research Team
Type of study

We will use mixed methods to investigate effects, mechanisms and patient perspectives on the implementation of emergency admission predictive risk software.


We know that when patients are admitted to hospital as an emergency, they may stay in hospital for a long time and are at risk of losing mobility and independence. UK policy has supported initiatives to reduce emergency admissions, where possible. One widely implemented initiative in primary care has been the installation of software that predicts each patient’s risk of emergency hospital admission. The software allocates a risk score to each patient that allows GPs to target management, such as medication review, to those at highest risk. This software has been introduced without knowing whether it brings benefits to patients or to the NHS. However, our evaluation of the introduction of predictive risk software in 32 general practices in the South Wales area (PRISMATIC trial) found that emergency admissions to hospital, Emergency Department (ED) attendances and days spent in hospital all increased when the software was available to GPs. 

We propose to undertake PRISMATIC 2 to build on these findings to:

  1. Assess effects of introducing emergency admission risk prediction software across England
  2. Investigate whether GPs change how they manage risk and make decisions about admitting patients to hospital when the software is introduced
  3. Understand patients’ views about communication of risk scores by their GP – do they want to know if they are classified as high risk of emergency admission to hospital?

Using national data, we will look at changes in trends in emergency admission and other healthcare use that occurred when the software was made available. This approach is suited to this study with over 200 health commissioners - Clinical Commissioning Groups (CCGs) - across England and a range of dates when risk prediction software was implemented. We will look at anonymised data on emergency admissions, Emergency Department attendances and days spent in hospital in each CCG before and after the software was available so that we can understand whether there were any effects when it was introduced.

We will explore whether GPs changed their decision making when the risk prediction software became available. For instance, did GPs lower their threshold for deciding to admit patients? Or did they prioritise patients in different ways once their risk scores were available to them? We will interview 48 members of staff at 16 GP practices to find out whether they think practice changed with implementation of the software.

We will carry out focus groups (x2) and interviews (x16) with patients to ask them whether they think hearing their own risk score may affect them, and if so, how.

Public and patient involvement

We are strongly committed to the involvement of patients and the public in PRISMATIC 2, and in line with good practice. The UK Standards for Public Involvement will be followed throughout the study.

Patient/public contributors were involved in the development of the study design, including two group discussions.

Two patient/public contributors are co-applicants and sit on the Research Management Group - Jan Davies and Rashmi Kumar. Both have personal health and care experience relevant to the topic of study. In addition, both are experienced in contributing to research relating to primary and emergency care. Our PPI partners feed into several networks and gain others’ perspectives and input to supplement their contributions and widen perspectives on use of primary and emergency care services. Links include to the SUPER (Service Users for Primary and Emergency care Research) group, which is coordinated by co-applicant Bridie Evans, and co-applicant Jeremy Dale is a member; and to the Patients and Public Participation Groups (PPGs) Network in South East England, of which co-applicant Rashmi Kumar is a trustee.

A further two patient/public contributors sit on an independent steering committee – Samina Begum and Ray Harris.

How can this research potentially benefit patients?

From the point of view of patients, it is important to reduce emergency admissions to hospital. Emergency admissions are generally unwelcome to the patient; they can be associated with adverse outcomes including death, frailty and difficulties regaining independence; and are challenging to manage in terms of quality and safety (e.g. exposure to hospital acquired infections). Although predictive risk stratification has been advocated as one tool to help support reductions, its impact and worth as a policy option remains unclear. This research study will investigate effects (including those unexpected and unintended), on emergency admissions to hospital, Emergency Department attendances and days spent in hospital, associated with the introduction of predictive risk stratification software.

We anticipate that this study will produce definitive evidence to enable policy makers and commissioners to understand the impact of investment in predictive risk stratification tools, and so make appropriate investment decisions.

We will communicate findings to:

  • Patients to inform them about the use of risk prediction software and allow them to participate more fully in decisions about their own healthcare
  • NHS and social care staff and policy makers to ensure that the best evidence about the effects of risk prediction software can be used to inform practice and policy.

NIHR (NIHR150717)

Total grant value

£ 828,912

Amount to Wales


Start date

August 2022

End date

July 2025

Outputs generated
Related links
  • To follow