CPM (Clinical Prediction Models) Networks

Project overview

This 3 year project looks at the stats methodology which underpins CPM’s. CPM’s are models or algorithms, which take a collection of a patients characteristics (e.g age, blood pressure, family disease history) and produces an output of the likelihood of an event happening, for example the risk of developing Type 2 diabetes. Current models available are only designed to calculate a patients risk for one particular disease. This project aims to model a whole patient pathway and predict the likelihood of coronary heart disease, atrial fibrillation, stroke, chronic kidney disease and Type 2 diabetes developing for that individual, using one CPM.

Intended outcomes

– The development of stats algebra to support this type of multi-disease risk model.

– Creating a CPM for coronary heart disease, atrial fibrillation, stroke, chronic kidney disease and
Type 2 diabetes.

Start: MAY 2020

End: MAY 2023

Funded by:

Medical Research Centre

Data Source

– Routinely collected primary care data.

– Main source of data is Clinical Practice Research Datalink (CPRD), which collects data from
certain GP practices in the UK

Benefits

This research is beneficial to any patient who is at high risk of developing the 5 conditions mentioned
above. The new CPM is hoping help enable early intervention strategies and effective prevention of disease development for high risk individuals.

Hopes beyond the project

The researchers hope that this project will ultimately lead to individuals being able to go to their GP, have their data imputed into this CPM, and then be able to receive an accurate risk likelihood for the 5 diseases

Manchester Researchers Involved

Glen Martin
Niels Peek
Tjeerd van Staa
Jamie Sergean