Diabetes is a condition affecting an increasing number of people and which can lead to serious complications such as heart disease, nerve damage and amputations. Managing diabetes therefore places a significant burden on the NHS, which is expected to rise over time. We aim in this study to estimate the financial cost to the English NHS of treating and caring for people with diabetes over the five years from 2013 to 2017. Using data from the Clinical Practice Research Datalink (CPRD) and linked Hospital Episode Statistics (HES), we will count and cost the use of GP surgery services, prescriptions, outpatient attendances and inpatient stays and compare these between patients with type 1 and type 2 and no diabetes, between males and females, and between ten-year age bands. This will provide valuable information about the effects of diabetes and different patient characteristics on healthcare burden and help to inform the planning of future services.
The aim of this study is to characterise the primary and secondary healthcare costs of treating people with diabetes.
Patients in CPRD whose data are eligible for linkage to HES and have a recorded Read or ICD-10 code indicative of diabetes mellitus and/or a prescription for a glucose-lowering therapy and no diagnosis indicative of secondary diabetes will be selected. People will be classified as having type 1 and type 2 diabetes using a previously published algorithm1 based on the diagnostic code recorded, prescribed glucose-lowering therapy, age at onset and body mass index (BMI). NHS resource use (primary care consultations, prescriptions, outpatient appointments and inpatient admissions) and associated costs will be characterised across five years (2013 to 2017), chosen to coincide with the most up-to-date linked secondary care data. The linked Hospital Episode Statistics (HES) inpatient and outpatient data will be used to categorise inpatient admissions and outpatient appointments and to provide the information needed to cost these resources using standard NHS costing methods. Healthcare costs will be stratified by sex and 10-year age band and compared between patients with type 1 and type 2 diabetes and the general population without diabetes. Healthcare resource use and costs will also be compared between patients with type 1 and type 2 diabetes and their matched controls with no diabetes. We predict that costs will not have a normal distribution, in which case a non-parametric Mann-Whitney U test will be used to compare costs between those with diabetes and the general population.
Health Outcomes to be Measured:
Healthcare resource use and associated healthcare costs.
HES Admitted;HES Outpatient