Healthcare
NHS Gloucestershire
IMPROVING PATIENT CARE WITH A DIGITAL TWIN
CLIENT
REQUIREMENT
NHS Gloucestershire needed a digital tool that would support them in decision-making and optimising hospital procedures and resources to offer the best care possible in uncertain and changing situations.
CHALLENGE
The key challenges of this project were modeling the queue times and the outliers.
SOLUTION
A dynamic simulation model that helps to optimise the daily challenges of the hospital operations and find the answers of how to resolve the situation faced.
A digital solution for optimising hospitals to offer the best patient care
Gloucestershire Hospitals NHS Foundation Trust manages three hospitals across Gloucestershire, employing around 8,000 people. Our project for the Trust was about building a digital twin simulation tool for two of their hospital sites: Cheltenham General Hospital and Gloucestershire Royal Hospital. NHS Gloucestershire data science team wanted to be able investigate what-if scenarios at a hospital level. Running these scenarios will support optimising hospital procedures and resources to offer the best care possible in dynamic situations.
Individual patient journey and medical outliers
There were two main things we focused on in this project. The first was queue times. In the modelling, we work at the patient level to understand what happens to the individual: what does each patient experience in hospital and how much of their time is spent queuing? The second was modelling medical outliers. These are patients who get placed into wards that have a different speciality than the patient’s actual medical problem, for example a nephrology patient in a cardiology ward. Caused by limited space capacity of the wards, they can have major knock-on effects in terms of treatment inefficiencies, queue building up and increased waiting times for individuals. Our model can identify the medical outliers and then track their movements. It can then be used to find ways of working that could reduce the number of medical outliers by changing the capacities of certain wards – some wards potentially need to be bigger than others due having a larger number of patients that require their treatment. Our model can help provide insight into these issues and how they can be addressed.
This is a complex problem with potentially very large supply-demand deficits. Because of the long planning horizon, the uncertainties in climate, population and legislative drivers are significant. There is a need to optimise against this uncertain future to produce resilient plans. Previously, the main focus had been on identifying the least cost plan that met demand – a relatively straightforward optimisation problem. But being able to produce plans that are resilient and adaptable to uncertain events decades in the future and also incorporate other performance metrics (environmental benefit, customer preference and more) is a major technical challenge. How can we plan 75 years ahead for such a complex problem and with so many uncertainties?
A detailed simulation model ready to support hospitals in real-world dynamic situations
Our simulation model provides a wealth of data. It can give hourly statistics and output results on the length of queues. It provides data on utilisation of personnel and beds. It outputs metrics such as time to triage, time to treat and then to discharge. It provides all these results as a summary, but the user can also drill right down to the individual patient level. They can choose the patients of interest to visualise their personalised journey, providing a clear and focused output. The model can also calculate key performance indicators such as the percentage of patients who pass through A&E within four hours. The model uses realistic scenarios to figure out the best planning strategies.
An extensive documentation for the highly detailed hospital digital twins
We worked closely with the NHS Gloucester team to develop the model, holding regular meetings to ensure that our model lined up with their expectations. In between these meetings we kept a detailed list of questions to ask during the next meeting. This approach helped the project to progress smoothly and ensured that the process logic was accurate and covered all possible scenarios. As a team we effectively managed tasks so that multiple developers could work on different parts of the model simultaneously and kept a high level of communication to make sure this did not cause any conflicts. By keeping a high level of communication within the team, we were also more effective at identifying issues and points within the logic that needed to be clarified with the client. In the closing stages of the project, we delivered an extensive documentation to our client and provided a demo of the model. This enabled us to effectively share all necessary knowledge of the model with the client and allowed them to evaluate it and confirm that it meets their requirements.
Decision intelligence tool ready for any NHS hospital
The model we built can serve as a simulation tool to help optimise any hospital. It can be used on an on-going daily basis by the data analysts working for the NHS as a decision support tool, and provide them with answers and recommendations to give to key hospital stakeholders – the people involved in management, administration, recruitment, etc. The model can consider different what-if scenarios and figure out how to resolve the situation faced, such as answering: How well can the current setup cope the demands of the coming winter? Where are the queueing bottlenecks? What will the implications on the number of staff be? Should we make new recruitments to support the expected turnover of patients?
Let’s collaborate
Send us an email, to discuss a new project.
We’re a team of innovators who are excited about unique ideas and help companies to create amazing solutions.