The multimorbidity (having one or more additional diseases to a main chronic condition) burden of people with type-2 diabetes (T2DM) and Coronary Heart Disease (CHD) is an important consideration in disease management. Most people with these conditions are found to have at least one other medical condition. There are various ways of measuring multimorbidity, the simplest being counting the total number of present chronic conditions. Multimorbidity has also been found to affect people differently, due to variations in patient demographics such as gender, age and socio-economic background.
Using anonymised data from the Clinical Practice Research Datalink, we will analyse the trends of 16 other existing diseases in patients with two chronic conditions: T2DM and CHD. We will estimate the current burden of these comorbidities and forecast the trends of each condition in 2020 and 2030.We will analyse the multimorbidity trends using different multimorbidity measures and identify the most commonly co-occurring groups of comorbidities and groups of similar conditions.
This will enable us to develop a tool to identify patients at a high risk of hospitalisation and death, while taking into account patients characteristics. We will determine the most important factors in predicting hospitalisation and death in these groups of patients.
Most patients with type-2 diabetes mellitus (T2DM) and coronary heart disease (CHD) have comorbidities; however, research has predominantly focused on single conditions, rather than multimorbidity. There are various multimorbidity measures available, all capturing different aspect of multimorbidity burden. Using primary care data from Clinical Practice Research Datalink, we aim to identify patterns of multimorbidity in patients with T2DM and CHD as measured by different available multimorbidity scores including the 1) total count of 16 chronic conditions included in the Quality and Outcome Framework (QOF), 2) Charlson Comorbidity Index (CCI) 3) electronic Frailty Index (eFI) 4) a Diabetes Severity Score, currently under development.
We will identify patients diagnosed with T2DM or CHD between 2007 and 2017. Based on Read/ICD10 codes and prescription data, we will identify the presence of individual conditions and calculate the prevalence rates and multimorbidity measures in patients stratified by gender, age and deprivation. We will estimate how patients of different characteristics perform on these measures. Furthermore, we will use machine learning methods to 1) Identify groups of similar conditions and 2) Create an algorithm predicting death and hospitalisation of T2DM/CHD patients and identify variables most important in predicting these outcomes.
Percentage of T2DM/CHD patients with individual comorbidities and pairs of comorbidities at the time of T2DM/CHD diagnosis and after the follow-up period.
Expected prevalence for 2020 and 2030
Percentage of patients diagnosed with T2DM/CHD in a given year between 2007 and 2017 who also had comorbidity
All-cause and index disease related death within 12 month of the T2DM/CHD diagnosis
All-cause and index disease-related hospitalisation within 12 months of the T2DM/CHD diagnosis
Evangelos Kontopantelis - Chief Investigator - University of Manchester
Magdalena Nowokowska - Corresponding Applicant - Farr Institute of Health Informatics Research
Christian Mallen - Collaborator - Keele University
Darren Ashcroft - Collaborator - University of Manchester
Mamas Mamas - Collaborator - Keele University
Salwa Zghebi - Collaborator - University of Manchester
Stephen Weng - Collaborator - University of Nottingham
HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation