Using Census 2021 data to improve diabetes prevalence estimates
Analysts in the Clinical Epidemiology team at the Office for Health Improvement and Disparities (OHID) have used data from the Office for National Statistics (ONS) to update sub-national estimates of type 2 diabetes prevalence in England. The data was used within an ONS Trusted Research Environment (TRE).
The team used record-level data from the 2021 Census to calculate the number of usual residents aged 16 and over, stratified by:
- sex
- age group
- ethnic group
- area deprivation quintile
These population statistics were produced for selected local authorities, administrative regions, Integrated Care Boards, NHS regions, and for England as a whole.
This approach enabled more robust estimation of both diagnosed and undiagnosed prevalence of type 2 diabetes across local areas. Previous models relied on older Census 2011 data and less reliable administrative sources, which limited the accuracy of demographic stratification.
The updated prevalence estimates are available through the Fingertips public health profiles and are widely used across the health system. For example:
- Diabetes UK has used the estimates to calculate the economic effect of undiagnosed diabetes and advocate for improved care recovery post-pandemic
- NHS England and NHS RightCare use the data to identify areas with possible high undiagnosed prevalence and populations at greatest risk
- Integrated Care Boards have begun incorporating estimates into their business planning
- local authorities rely on these estimates to support service planning, particularly where no other data exists at this level of geography for type 2 diabetes
The team plans to extend this work by modelling prevalence for years beyond 2021 and for other conditions, such as non-diabetic hyperglycaemia.
This project demonstrates how high-quality data from ONS can support better public health analysis and decision-making.
Nicholas Brown