Welcome to the Midlands Analyst Network

Thank you for being part of our growing community!

If you are an analyst who works in health and care or the wider public sector, then join the Midlands Analyst Network!

The Network has almost 900 members from a range of health and care and other organisations and with representation from all parts of the Midlands and many other areas too. Established in April 2020 the Network provides a space for analysts across the Midlands and beyond to share information, ideas and resources, as well as seek advice and guidance from one another.

Your Network Lead: Rachel Caswell

Rachel is a Senior Healthcare Analyst in the Strategy Unit who is working to broaden the scope of the Network. Rachel is also the Operational Lead for the Training & Development.

Fortnightly Huddles

Rachel hosts a fortnightly online Huddle on MS Teams (see link below). The Huddles cover a large range of topics and you are very welcome to join one at any time – for quick access to a Huddle, just follow the link below. You don’t have to be a member of the Midlands Analyst Network to come to a Huddle but if you’d like to be kept up to date with all future Huddles and training opportunities then just contact Rachel and we’ll take it from there.

We try to keep them as informal but as informative as possible and welcome contributions from Network members as well from academia, subject specialists, national and occasionally international presenters.

If you have any queries about the Huddles, how to access them, if you want to find out what’s gone before, or you want to present then please do contact Rachel.

Next Huddle: 17th October, 2024 (11-12): Modelling Long-Term Changes in Population Health State and Associated Healthcare Resource Requirements with Luke Shaw (NHS Bristol, North Somerset and South Gloucestershire ICB)

Join this Huddle to find out more about the mathematical model, which has already influenced financial decisions of resource allocation for long-term commissioning needs within the Bristol, North Somerset, South Gloucestershire (BNSSG) ICS.

Healthcare policy makers face regular challenges on how to allocate healthcare resources with limited budgets, both in the short and longer term. Mathematical and computer modelling tools can capture, subject to assumptions and simplifications, these interacting factors in estimating the long-term trajectory as well as the implications of different mitigatory measures. We have developed a mathematical model to support decisions around long-term commissioning needs within the Bristol, North Somerset, South Gloucestershire (BNSSG) ICS.

The model is a finite horizon discrete-time Markov chain where the state space representing the health state of individuals within the population is based on segmentation using Cambridge Multimorbidity Score. Essentially, the model accounts for the life-course of individuals as they age and (typically) advance through the states with declining health. Such movements are extrapolated from the observed transition rates within the system, anchored on demographic projections (births, deaths, and migration) from the Office for National Statistics (ONS).

Our results indicate notably different growth rates for different healthcare settings, such as maternity and A&E attendances, which helps to demonstrate the added value of data science techniques – especially over common approaches based simply on ONS percentage uplifts. Ultimately, while the population is expected to increase by 14% over the 20-year horizon, the total cost is expected to increase by 41%, indicating the scale of the challenge ahead.

The model has already influenced financial decisions of resource allocation within the system, and we will present examples of the model being used in real life situations. Finally, we will review outputs against two other externally developed models.

This work follows naturally on from the Population Segmentation talk at HACA2023 by Nick Hassey, introducing the segmentation approach being used in BNSSG, and now as a foundational building block in our dynamic model.