Choosing the best glucose-lowering drug is often a major clinical dilemma for the four million people in the UK with Type 2 diabetes. This project will use routine and trial data and state-of-the-art statistical methods to develop a treatment selection model to select the optimal glucose-lowering drug for people with Type 2 diabetes based on their clinical characteristics. This model to be developed has the potential to be used as a decision aid in primary care to improve outcomes for people with Type 2 diabetes.
Explaining the science
Precision medicine can be defined as the targeting of treatment according to the biological or risk characteristics of individuals. Type 2 diabetes glucose-lowering therapy is an excellent candidate for a precision medicine approach to prescribing for the following reasons:
1. Each of the four drug classes recommended after metformin differ markedly in mechanism of action and side-effect profile but have the same principal aim: to lower blood glucose.
2. There is a clear knowledge gap as current treatment guidelines do not provide information which
medication is best for lowering glucose or will be best tolerated, for which individuals.
3. Although average glucose-lowering of the four second-line therapies is similar, at the individual level glucose-lowering response and susceptibility to side-effects vary greatly.
4. There is major heterogeneity in type 2 diabetes. Individuals with different underlying pathophysiology are likely to vary in response to the different drug classes, and in susceptibility to side-effects, depending on the mechanism of action of the drug.
Type 2 diabetes (T2D) is a progressive, multifactorial condition characterised by chronically raised blood glucose levels. It affects over 4 million people in the UK and accounts for over 10% of NHS expenditure as it is a major risk factor for cardiovascular disease, renal disease, loss of vision, reduced quality of life, and increased mortality.
The majority of people with type 2 diabetes will require glucose-lowering medication to reduce their symptoms and risk of developing complications. Metformin is the recommended initial glucose-lowering drug and is prescribed first-line to over 90% of individuals with Type 2 diabetes. Choice of second-line therapy, however, is a major clinical dilemma with current NICE UK guidelines recommending four medication classes as options for second-line (DPP4-inhibitors, sulfonylureas, thiazolidinediones, and SGLT2-inhibitors), with little guidance on which treatments will be beneficial to which individuals. This has resulted in marked and arbitrary variation in prescribing practice by geography across the UK.
The aim of this project is to develop and evaluate a decision support tool to select the likely best medication class for individuals with type 2 diabetes, with optimisation based on likely glucose-lowering, tolerance and cardiovascular outcomes.
This work will lead to a model to optimise glucose-lowering that can be integrated within primary care electronic health record systems. Use of the model could improve patient defined outcomes by:
● Reducing cardiovascular disease and death, and microvascular complications.
● Reducing/delaying requirement for additional glucose-lowering therapy/insulin in the future.
● Increasing tolerance of therapy and reducing side-effects.
Researchers and collaborators
Bilal MateenClinical Data Science Fellow
Professor Spiros DenaxasProfessor of Biomedical Informatics, University College London
John Dennis - [email protected]