Calculating the date of your death is not your average school maths lesson, but when Health Economics Unit Director Andi Orlowski was invited to give a talk to London students, the curriculum quickly became a matter of life and death…
When I was a boy science and maths were seen as the preserve of geeks and nerds – subjects to be quietly studied in the corner of a library, not excitedly discussed at breaktime alongside the latest football results or TV shows. So I was delighted and happily surprised by the sea of attentive faces and barrage of enthusiastic questions with which I was greeted when I visited William Patten School, in Stoke Newington, to talk all things science, technology, engineering an maths (STEM).
My aim was to explore some of the analytical methods we use at the Health Economics Unit (HEU) and the theory behind it in a relatable and engaging way, so my talk was titled “How You Can Predict the Future with Maths”. Below are some of the areas which I discussed with the students.
Obviously most of our analytical work at the HEU is focused around health and care, but did you know that you can also use data modelling to create chart-topping hits? Analysts at the University of Bristol have used mathematical models to predict hit songs by analysing variables including time signature, tempo and even volume variability.
Similar predictive modelling was used by the Department of Health and Social Care to predict hospital readmissions. By using variables such as age, postcode, and previous admissions, these models can predict whether someone will return to the hospital within 30 days, allowing commissioners to better plan care and resource allocation.
It was great to be able to demonstrate these links between the theoretical maths which the students learn in the classroom, and real world applications which are actively improving their health and wellbeing.
We talked about how we use statistical models, like linear regression and logistic regression, to see patterns in the data and how they help in making critical decisions in health economics. For example, predicting the cost of healthcare services, or determining the likelihood of a patient having a particular disease or progressing to multimorbidity based on symptoms and test results.
We also delved into the world of machine learning, explaining concepts like K-Nearest Neighbours (KNN) and deep learning. These theories sound complex, but they drive applications which many of the pupils use every day, such as recommending movies or recognising faces in photos. The students showed a particular interest in Monte Carlo simulations, which use randomness to predict outcomes. We even simulated how hospitals manage resources using Discrete Event Simulation (DES), helping them understand the complex systems that keep healthcare running smoothly.
It was a privilege to be able to engage with young minds in a school setting, and when I invited questions I was overjoyed to be confronted with a sea of raised hands and some very thoughtful and inciteful challenges. Questions included: ‘Do your predictions ever go wrong?’ ‘Do your predictions ever cause harm?’ and the ultimate analytical challenge ‘Can you predict when we’ll die?’
While speaking at a school as a representative of the HEU builds stronger ties between the next generation and the NHS, it does more than that – it plants the seeds of curiosity and enthusiasm for STEM subjects. Seeds which will hopefully blossom into brilliant minds and innovations which will one day join us in shaping the future of health and care.