Costs and complications of diabetes - investigation of a comprehensive national diabetes register

The number of people with diabetes is increasing. As new treatments for diabetes become available (for example to reduce body weight, lower blood sugar or lower blood pressure) it is important that these are assessed not only to provide real clinical benefit to patients, but also available at the best value for the health service to be achieved. The main aim of this research will be creating methods to determine the costs and benefits of different, novel treatments in the management of diabetes.

We have already successfully assessed inpatient costs and working is ongoing in assessing the costs of prescriptions for people with diabetes. This project would allow modelling of the impact of new treatments in the NHS. This is critical to health service planning and to support the development of an equitable service for people with diabetes. My fellowship training will allow me to develop more sophisticated methods in handling large diabetes datasets. Currently in Scotland such datasets already exist. I will take advantage of the broad range of existing datasets from the NHS including: recording routine inpatient care of people with diabetes, demographic characteristics (such as age, sex), clinical information (such as blood pressure, cholesterol) and drug prescription.

As a first step, using existing datasets, I will be able to determine which risk factors (such as cholesterol and blood pressure) are associated with admission and cost of admission to hospital for complications. The novel component of this project would be examining how these risk factors change over time and how these affect the timing of events. Combining the information we will be able to determine how cost-effective novel treatments for the management of diabetes will be and how these treatments will affect the risk of complications and admission.

This existing dataset is most advanced at present and this research has the potential to enhance the quality of health research in this area and improve health policy decision making. The methods developed in this project will be applicable to other disease areas allowing us to understand the costs and benefits of treatments for different diseases within the NHS.

Aim
This fellowship will extend my existing prevalence-based models in diabetes and develop more complex incidence-based models. I plan to take advantage of the unique data source available and develop new incidence-based models to estimate lifetime costs of people with diabetes and investigate the impact of novel treatments for diabetes.
Objectives
1. Review existing incidence-based diabetes models in order to create my own model. This will involve 3 stages:
a. Develop risk equations to predict first complications requiring inpatient care
b. Develop equations to predict the time trends for modifiable risk factors
c. Develop cost equations for predicting cost of complications
2. Combine all the equations created in 1) to develop a comprehensive predictive model
3. Use the model created in 2) to evaluate the impact of novel treatments on risk and cost of complications
Methods
Analysis of diabetes population and outcomes are based on the comprehensive national register of people with diabetes in Scotland based around the Scottish Care Information-Diabetes registry with linked outcomes from centralised Scottish Morbidity Records on hospital admissions and mortality records held by the General Register Office for Scotland. A major component of incidence-based models involves the timing of events in a survival analysis framework. I will apply more complex survival methods to create the equations that will be used to develop a comprehensive predictive model. This model will be used to assess the impact of novel treatments on risk of complications and admission and form a policy model to help plan future health care for patients with diabetes.
Opportunities
The maturing linked dataset offers an unrivalled opportunity to examine the health impact of diabetes. This project offers an excellent training opportunity because of the importance of the questions addressed and potential for interaction with several institutions working on different aspects of the data.

Direct beneficiaries of this research will include medical statisticians, epidemiologists, health economists and health policy makers. As expanded upon in our Case for Support, I believe incidence-based modelling are ideally suited when making decisions on alternative treatment programs by allowing analysis of long-term cost and effectiveness. Unfortunately, these models are not consistently used to due to limitations of data available for analyses, but they have the enormous potential to identify the benefits and costs of different treatments and can be used to estimate the impact of prioritising different groups in terms of health inequalities. My research has the potential to improve the current methodologies of incidence-based models by employing various topical techniques such as multi-state modelling, the results of which can help decision makers decide which treatments will be cost effective. Furthermore, it will inform policy makers of the associations between modifiable risk factors, treatments and complications of those with diabetes in the Scottish population. Moreover, this research will provide more realistic and accurate estimates of the costs and effectiveness when evaluating the impact of novel treatments in the management of diabetes.

I intend to disseminate my research at various conferences and meetings to an audience with a wide range of expertise, and in the form of publications in high-quality, peer-reviewed journals. Diabetes medical conferences in Europe and America will give me the chance to disseminate my findings to those working directly with people with diabetes. I will follow-up my presentations with publications in diabetes journals. Methodological conferences such as a health economics conference will allow health economists in a similar area of research to discuss my methodologies. Similarly, the Mount Hood Challenges will allow me to discuss my work with experts in the area of diabetes modelling. I will also play an active part in local seminars and presentations at University of Glasgow and during my training visits to McMaster University and University of Melbourne to discuss my results and findings with local experts.

Indirect beneficiaries of my research will be patients with diabetes and the wider general public. The Scottish Diabetes Research Network (SDRN) actively encourage public engagement including involvement of lay members on the SDRN executive committee, liaison with charities including Diabetes UK and presentation at national meetings with a strong lay and service user element. The Scottish Health Informatics Programme (SHIP) also involves members of the lay public at their annual retreats and conferences. Both SDRN and SHIP produce newsletters made available to the wider research group and lay members. Additionally, improving the methodological base to modelling the complications and impact of future treatments in people with diabetes will have obvious benefits to the people with diabetes and the wider general public, ensuring public funds are spent in an appropriate manner.