29th November 2024
The Lung Health Check programme aims to find lung cancer earlier, before symptoms start to show, and will therefore enable earlier diagnosis and treatment.
By 2029, all patients with a smoking history will be invited for a Lung Health Check. This will generate a significant increase in demand for diagnostic and treatment services. The NHS Transformation Unit (TU) built a model for West Midlands Cancer Alliance (WMCA) to predict the demand for these services over the next five years for a range of different scenarios.
The Challenge
As it rolled-out its targeted Lung Health Check programme, West Midlands Cancer Alliance (WMCA) needed to understand what additional demand there would be at each stage of the Lung Cancer pathway due to identifying a greater number of people with suspected lung cancer. WMCA wanted to be able to model demand for various diagnostic and treatment modalities over the next five years. They required a model that was flexible and could capture inputs and assumptions relating to smoking rates, conversion rates and treatment requirements. This would help the screening programme to plan for and accommodate varying population needs, healthcare resources and environmental factors. Considering the population needs of patients, this model will support broader outreach, better outcomes and the use of efficient resource.
Our Approach
We created a flexible model to project future demand at key points across the Lung Cancer pathway over the next five years. To achieve this, we:
- Mapped the pathway for patients who are offered a Lung Health Check. This enabled us to identify key patient interactions along the pathway, and where patients moved to following that interaction including to further diagnostic tests or treatment.
- Translated the pathway model into a set of assumptions that impact how a patient will progress on their journey. We outlined the structure of the code that needed to be created within the model to capture information about key events and demand. We created a user-friendly spreadsheet for the user to record and input all the assumptions required to run the model.
- Built a model using the open-source programming language R. This model used Discrete Event Simulation (DES) methodologies to randomise patients to an event at each stage of the pathway based on the input assumptions. The codebase for the model was developed collaboratively and is available on GitHub. GitHub is an online platform for storing, sharing and collaborating on code using robust version control. It helped our team to track changes, collaborate on the project, and manage different versions of our work efficiently. Our use of open-source tools enables transparency, shares learning across the analytics and modelling community and better enables easier, further development of the model in the future.This will enable WMCA to effectively optimise their resources, and therefore improve health outcomes and ensure long-term operational viability.
We delivered the final model to West Midlands Cancer Alliance alongside guidance on how the model functions and how to develop new scenarios. We also created a report for the model scenarios detailing the projected demand at each stage of the pathway.
The Outcome The Cancer Alliance are now able to run a range of scenarios across different geographies to determine the additional demand for services. The analytics team at WMCA have used the GitHub repository to replicate the model across several of the different ICBs that WMCA are partnered with. This is possible due to the modular design of the model, allowing different teams to reflect not only their varied populations but also the different assumptions and experiences of their respective clinicians