In this blog, our Head of Data Science - Chris Beeley – proudly launches an open-source model, with support from the New Hospitals Programme, designed to help plan NHS services for many decades into the future. The model represents significant progress in our commitments to transparency, collaboration, and efficiency in healthcare planning.  

Planning effectively for healthcare infrastructure is paramount when faced with the evolving needs of healthcare. Imagine that you are responsible for planning a new hospital. You face multiple challenging tasks: one of which is working out how big it needs to be.  

There is a responsibility to ensure that hospitals are designed with long-term functionality at the forefront of everything you do. Too small and the capacity will quickly be overwhelmed, too big and it will be seen as a waste of valuable public money.  

So how would you figure this out? How would you navigate the swirling and beguiling range of factors in play? Projecting population growth might be easy enough, but how healthy might people be? Will the population of 2040 be in better shape than that of 2025? And will care in the future look like care does today? What about medical breakthroughs and technological advance? How might that change things?  

Whilst this remains an abstract concept for most of us, it remains a challenge for those tasked with building the healthcare infrastructure of the future. Those leading hospital building schemes in the New Hospital Programme (NHP) need to model their designs based on an assessment of potential future needs and how services might look to accommodate these changes in the years to come. They need models. And they need models knowing that existing models for projecting future needs have suffered from:

  • Overlooking some big-ticket items. For example, not accounting for the likely future health status of the population.  
  • False precision. Giving single point estimates, rather than probabilistic ranges (which would, paradoxically, be more accurate).  
  • Not being clear about how results are produced. Many existing models are, in effect, a ‘black box’ with hidden methods and assumptions.  

To overcome these problems, the NHP asked the Strategy Unit to develop and build a demand and capacity model. A model to provide consistency, reduce duplication, and increase scope for learning, as well as being available to every scheme in the programme.  

As part of our commitment to developing in the open, the NHP and the Strategy Unit are thrilled to announce that the model is now being shared for the benefit of the wider healthcare system.

Hospital planners are now able to access a high quality model for assessing plans that combines standard elements with local features. And our enthusiasm runs far deeper.  

The NHP model is entirely open source. Anyone can freely examine the inner workings of the model. They can explore its source code on GitHub. They can test and critique it. They can use it and share it. And they can do all of this for free. The model will never be ‘finished’. We have created something that can be updated and improved over time.  

In some ways, this might sound like a niche concern: something for us technical types to get excited about. And yet the implications run wider than that, because:

Transparency matters

Sharing the source code is an important way of building trust and confidence, allowing stakeholders to transparently understand and verify the model – something not possible with traditional ‘black box’ models .  

Quality matters

Our team can attest – first hand– that sharing code means you are especially careful. Knowing that anyone could examine the inner workings of our model definitely means that we give all our attention to sharpening quality. It also ensures we give ourselves no room for obscure methods or overstating expertise. We believe these principles enhance credibility and set a standard for accountability in healthcare analytics.    

Public value matters

The NHP model was built with public money; we are sharing it openly for public benefit, circumventing poor commercial incentives that may stunt growth.  

Open source is the future

In many ways, open source is not a new idea. It has been around since the birth of computing and was popularised particularly with the birth of the free software movement in the 1980s. Its value in analytics has been increasingly recognised – thanks in large part to the efforts of Ben Goldacre and others. We are now at a point where the value of open source modelling is required by policy. The Data Saves Lives policy, for example, states that:  

“Public services are built with public money, and so the code they are based on should be made available across the health and care system, and those working with it, to reuse and build on.”

A major benefit of the open-source model is that whilst every new hospital will have to be built brick by brick, nobody using a model to plan a hospital will ever have to start from scratch. Because it is open source, the NHP model can be developed, iterated and improved upon for many years to come.  

Building on the example set by this model, openness should become standard practice for all NHS analytical tools. Sharing code transparently not only builds trust and ensures quality but also sets a benchmark for how such tools can deliver maximum public value.

The NHS will always need new hospitals; the NHP model will help us get better and better at ensuring they are fit for the future.