Diabetes is a progressive disease defined by high blood sugar levels; it is a major risk factor for heart attacks, strokes, visual problems and kidney disease, which lead to significant illness, work absence, and early death.
Being sedentary or overweight increase the chance of developing diabetes, and addressing these with lifestyle programmes can prevent or delay the condition. In order to offer these, we need to identify those at highest risk. Various diabetes risk scores estimate risk by combining information such as family history of diabetes, medical conditions, weight and ethnicity. Information on these risk factors is often contained in a persons electronic health record.
Our study aims to assess the performance of a recently published diabetes risk score, and to use it to produce diabetes risk maps across the UK at regional level to inform policymakers and allocate resources by identifying key areas with the most at risk populations. These risk maps are visually dramatic, highlighting hot spots of high risk in deprived multi-ethnic areas and thus providing a talking point for local policymakers and citizens.
We have already produced maps of diabetes risk in several London boroughs using data held on GPs computer systems. For various reasons, we cannot use GP computer systems to generate regional or national level maps so we would like to see if it is possible using clinical research data sets.
Effective approaches to diabetes prevention are essential to reduce the human and economic costs of this increasingly common and serious condition. Identifying sub-populations at the highest risk of diabetes could improve targeting of preventative interventions. Validated diabetes risk scores are a simple and scalable way of estimating and comparing risk across different populations; some can be calculated from routinely collected data available on GP clinical computer systems and/or research databases.
The aim of this project, which builds on previous work undertaken by TGs team in East London, is to externally validate the recently derived type-2 diabetes risk score, and use it to estimate the 10-year risk of developing diabetes across the UK. We will use the Clinical Practice Research Datalink (CPRD) to obtain patient level risk factor data and estimate diabetes risk using a the QDiabetes-2018 risk score. Traditionally, CPRD has only included data from practices using Vision computer systems. However, CPRD have recently added data from practices using EMIS Web computer systems. Using both data sets, individual risk estimates will be aggregated to form regional summaries and presented as a heat-map of the UK. The availability of data from both Vision and EMIS Web practices means that any heat maps produced will be more representative of the risk in each health region than was previously possible.
Risk of type 2 diabetes.
Benjamin Feakins - Chief Investigator - University of Oxford
Benjamin Feakins - Corresponding Applicant - University of Oxford
Dianna Smith - Collaborator - University of Southampton
Evangelos Kontopantelis - Collaborator - University of Manchester
Richard Stevens - Collaborator - University of Oxford
Sarah Lay-Flurrie - Collaborator - University of Oxford
Tim Holt - Collaborator - University of Oxford
Trisha Greenhalgh - Collaborator - University of Oxford
Patient Level Townsend Score