Research Programme 2: National Epidemiology
Dr Robert Lindsay, Dr Sarah Wild, Prof Helen Colhoun, Prof Andrew Morris
Aims
- To perform epidemiological studies on a national scale and use the information to estimate current and future health costs using diabetes as an exemplar.
Background
Scotland is unique in the UK in having developed over the last five years a nationwide clinical electronic system for the management of diabetes (SCI-DC). This system captures data daily from all general practices (n>1200) and all hospital clinics in Scotland. Key clinical processes and events related to diabetes care are recorded, including biochemistry, cardiovascular screening, digital retinal screening, all microvascular and macrovascular outcomes. All records have a CHI number and there are data for over 205,000 living people with diabetes on SCI-DC at present. Synergy across the SHIP programme will allow us to address the challenges of transferring clinically generated records into a high quality research tool using diabetes epidemiology as an exemplar.
Diabetes epidemiologists with established international reputations from Universities of Edinburgh, Glasgow and Dundee will lead analyses of SCI-DC data, linked anonymously with other routine health data under the auspices of the Scottish Diabetes Research Network (Chair Epidemiology Group: Robert Lindsay http://www.sdrn.org.uk). An important objective is to establish an efficient central mechanism for variable derivation and data extraction from the linked datasets for research users and to ensure research user input to augmenting meta-data.
Research Questions
Epidemiology and Pathogenesis of diabetes and diabetic complications
Developing risk engines for predicting cardiovascular disease (CVD) in type 1 and type 2 diabetes using routine clinical and laboratory data. These will be compared against other risk engines which could be used to determine at what age and with what risk profile absolute risk in type 1 DM typically reaches that where statin therapy is warranted.
Epidemiological analyses to inform care pathways
- analysis of retinopathy data from the current screening programme and variability in its uptake to model the effect of using different screening intervals for retinopathy
- analysis of the relationship of glycaemic control with current therapies to CVD morbidity and total mortality in different patient subgroups for example to establish whether the early stopping of the intensive glycaemic control arm of the ACCORD trial because of excess total mortality is relevant at a population level.
Diabetes and other diseases
It will be possible to explore novel associations such as the r positive association with psoriasis and the complex association with prostate cancer. In collaboration with other RPs we will also evaluate the extent to which any associations are modified by other factors including specific drugs or genetic factors.
Social dimensions of diabetes and diabetes care
Examining the magnitude of current geographic and socio-economic inequalities in the complications of diabetes and modelling to what extent given these are likely to be amenable to reduction with current intervention strategies.
Pharmacoepidemiology
We propose to use data on hospital admissions and biochemical results data to examine drug safety using data from people with diabetes as an exemplar.
Health Care Needs Assessment, Trend Analysis For Policy
We propose a programme of modelling current and future diabetes care costs: in collaboration with the internationally respected Health Economist Professor Andrew Briggs (University of Glasgow). This will include looking at how changes in health care practice (e.g. introduction of retinal screening) or risk factor control change patterns of diabetes rates and complications. For example, the effect of improvements in some areas of care (e.g. renal dialysis) on potential increases in burden of other complications (e.g. retinal outcomes).
Methods
SCI-DC data will be linked to data on hospital admissions, cancer registration, mortality, prescribing, other disease registers and anonymised prior to analyses to address specific research questions (see above). Tools developed in the RP3 programme will be used to detect unsuspected adverse drug reactions among people with diabetes.
Deliverables
- Publications on the epidemiology of diabetes and its complications of direct relevance to policy formulation. We envisage a Scottish equivalent of the Diabetes in America series published by NIH/NIDDK by the end of this project and multiple peer-reviewed publications.
- Health economic analysis of the costs of diabetes in Scotland.
- Improved availability of high quality linked data metadata.
- A core repository of data including a data dictionary and analysis code.
- An organisational and governance mechanism for using EPRs for collaborative research on other diseases.