Police Service Modelling Associates (PSMA) Pilot Programme

Research Institution / Organisation

University of Exeter

In Collaboration With

Alexis Poole (Performance and Analysis, Devon and Cornwall Police); Dr Iain Lang (NIHR Applied Research Collaboration (ARC) South West Peninsula, University of Exeter)

Principal Researcher

Dr Daniel Chalk

Level of Research

Professional / Work-based

Project Start Date

January 2020

Research Context

In 2016, PenCLAHRC, now the NIHR Applied Research Collaboration South West Peninsula (PenARC) launched the Health Service Modelling Associates (HSMA) Programme. In this programme, NHS staff are released from their usual role for a day a week for one year to undertake an Operational Research (computer modelling to inform decision making) or data science project that addresses a question of importance for their organisation.

The associates are provided with training from academics who are experts in modelling and data science, and are taught a range of methods including an overview of the modelling process, Discrete Event Simulation for modelling pathways and queuing problems, Geographic Modelling to optimise the location of services and Problem Structuring Methods to help structure focused research questions from complex real world problems. Associates are trained in how to program in Python and R in order to develop Free and Open Source models and solutions that can be freely shared and require no software investment costs.

In addition to receiving extensive training, associates are also allocated a mentor – a member of the PenARC modelling team – who provides ongoing advice and guidance to the associate about their project. The associate is additionally supported by a Workplace Supervisor – a senior member of the associate’s organisation who helps to facilitate the implementation of the project and ensures that the work of the HSMA becomes "knitted in" to the analytical function of the organisation.

Previous HSMA projects have looked at using artificial intelligence to make better surgery cancellation decisions and determine when it is safe to deprioritise ambulance callouts, computer simulation to model the resources required to staff new mental health and urgent care facilities and geographic modelling to determine the optimum location for new health services. The programme has led to significant impact for NHS organisations and patients, including over £13 million of investment in new mental health facilities,  support for business cases to establish new additional stroke beds in Cornwall, and the creation of brand new Operational Research roles in NHS organisations across Devon and Cornwall.

The fundamental aims of HSMA are to train analysts and service improvement staff to use advanced modelling and data science methods to make better use of their organisation’s data and improve decision making and to help build a sustainable Operational Research and Data Science function within the associates’ organisations. We believe that these aims are also relevant to the work of the Police Service, and that the programme could be successfully applied into this sector.  This project proposes to run a small-scale pilot of the programme with Devon and Cornwall Police, Avon and Somerset Police, Wiltshire Police, Dorset Police and the College of Policing to a) explore the potential for the programme to assist in making better use of data in policing, b) provide training and increased awareness of Operational Research and Data Science methods to police staff, and c) support policing projects that have the potential to lead to real impact for the service and / or the public.

Research Methodology

The proposal is to pilot the programme as a Police Service Modelling Associates (PSMA) programme. This would consist of a three month Phase 1 training and project development phase with a cohort of up to 10 associates, followed by a 9 month Phase 2 in which two projects will be selected to be taken forward. Multiple associates from Phase 1 may choose to work collaboratively on the two projects chosen for Phase 2 to continue their engagement with the programme, but each project will have a project lead (one of the associates) who will take ownership of the project and see it through to its completion.

In Phase 1, 17 full day training sessions will teach analytical staff in Devon and Cornwall Police, Avon and Somerset Police, Wiltshire Police, Dorset Police and the College of Policing the key principles and methods in Operational Research and Data Science, and explore with them potential projects to which these methods could be applied to address key problems for the police service. The 17 sessions covered 13 modules :

  1. An Introduction to Operational Research and Data Science
  2. Principles of Programming
  3. Programming in Python
  4. Programming in R
  5. Discrete Event Simulation
  6. Geographic Modelling
  7. System Dynamics
  8. Network Analysis
  9. Agent Based Simulation
  10. Machine Learning
  11. Natural Language Processing
  12. Forecasting
  13. Deep Reinforcement Learning

Staff admitted onto the programme should have some experience in analytical and / or basic modelling, with at least basic mathematical and IT skills. Previous experience in programming would be desirable but not essential. Due to these requirements, data analysts and intelligence staff are the most likely candidates, but the programme would be open to wider roles as long as analytical and / or basic modelling skills can be demonstrated.

At the end of Phase 1, the PSMA Programme Lead (Dan Chalk) will work with Alexis Poole and Iain Lang, as well as representatives from the PenCHORD team, to decide on two projects to be taken forward into Phase 2. Projects will be prioritised based on their potential impact and their practicability given the reduced 6 month time frame for Phase 2 compared to the standard 9 months for HSMA. Practicability considerations will include complexity of the project, skills of the associates to undertake the project and data availability to support the project. The two projects selected for Phase 2 will each have a single associate allocated as the project lead. Other associates from Phase 1 will be offered the opportunity to join the selected projects to work collaboratively and share skills.

In Phase 2, the projects will be undertaken, driven by the project leads. Associates will work on the projects for at least one day per week. Each project will be allocated a mentor from the PenCHORD team who will provide advice and guidance, and a workplace supervisor within DCP who is a senior member of staff able to help facilitate the delivery and implementation of the project. Associates will come together once a month for a four hour hackathon session, along with their mentors, to discuss progress, share ideas moving forward and have time to work on their projects with modelling experts on hand to assist.

At the end of the programme, associates will be asked to present their projects to an audience of DCP staff and local academics.

Interim reports and publications

​Not available

Date due for completion

October 2021
Return to Research Map