Research and analysis

Methodology note on coronary heart disease prevalence by lower layer super output area

Published 25 September 2025

Applies to England

Background

This is a supplementary release, showing detailed tables underpinning the charts published in Health trends and variation in England 2025: a Chief Medical Officer report.

This methodology note explains how published general practice (GP) level health data has been attributed to local areas called lower super output areas (LSOAs). A LSOA is a small neighbourhood of around 1,500 residents. GPs tend to cover multiple LSOAs and LSOA populations can be served by multiple GPs. This data uses 2021 LSOAs.

This methodology has been developed by Alex Gibson at the University of Plymouth. The code to attribute from GP level to LSOA level, and an explainer of how to do so, are available on GitHub. This note focuses on the specific use case of this method to calculating coronary heart disease (CHD) prevalence at LSOA level.

This data is intended to inform changes in clinical practice by highlighting areas where prevalence varies. Variation in GP recorded diagnoses may indicate differences in case ascertainment or alternatively differences in population risk, such as age structure, as the prevalence of CHD increases with age.

Data sources

The methodology to get to GP level data used in the report draws on 3 publicly available datasets from government sources.

Office for Health Improvement and Disparities (OHID) - Fingertips API

Fingertips provides GP level public health indicator data, including measures from the Quality and Outcomes Framework (QOF). The QOF is a voluntary annual reward and incentive programme for all GP practices in England, detailing performance across clinical and public health domains (such as management of long-term conditions and screening uptake).

The indicator used in this analysis was indicator 273 - CHD: QOF prevalence. We use the average for financial year end 2023 to financial year end 2024.

NHS Digital - Patients Registered at a GP Practice

NHS Digital - Patients Registered at a GP Practice provides the number of patients from each LSOA registered at each GP practice. It is used to weight GP level values based on where patients live.

For financial year end 2023 we use the April 2022 dataset. For financial year end 2024 we use the April 2023 dataset.

Office for National Statistics - lower layer super output areas (December 2021)

This data supplies the official list of 2021 LSOAs in England. It is used to ensure all areas are consistently represented in the analysis.

Aggregating from GP level to LSOA level

Quality and Outcomes Framework data is not routinely published at LSOA level. To produce estimated LSOA prevalence, we aggregate up from published GP practice data. However, the areas in which patients live for each practice do not necessarily fall into the same LSOA as the practice. To produce estimates that reflect health outcomes in LSOAs, this method attributes values from each GP to the LSOAs in which their patients live.

This section describes the steps used to complete this population weighted attribution process.

GP level indicator data

Indicator data for each GP practice is retrieved from the Fingertips public health data API. This data represents the population registered at GP practices in England that are recorded as having a diagnosis of CHD.

Obtain patient distribution data

Each GP’s patient population is mapped to the LSOAs where those patients live using NHS Digital’s Patients Registered at a GP Practice data. This dataset provides a count of registered patients broken down by practice and patient LSOA.

For each GP, the proportion of patients residing in each LSOA is calculated. These proportions provide the attribution weights used in the next step.

Attribute GP values to LSOAs

Each GP’s indicator value is distributed across the LSOAs where their patients live. The amount assigned to each LSOA is proportional to the number of that GP’s patients who live there. For example, if 12% of a GP’s patients live in a specific LSOA, then 12% of that GP’s count is attributed to that LSOA.

This process is repeated for all GP practices in the country.

Aggregate to the LSOA Level

Once each GP’s contribution to an LSOA is calculated, all contributions to the same LSOA are summed to produce an overall estimated value for that area.

If multiple GPs serve patients in the same LSOA, those contributions are combined.

Calculate percentages

To produce estimated percentages of CHD prevalence, the aggregated count is divided by the total LSOA population and then multiplied by 100 to get the percentage.

This step ensures the resulting figures are comparable across areas of different population sizes.

Data quality

Not all practices participate in QOF and if they are also not listed in the Organisation Data Service they are not captured by this collection. In financial year end 2024 this was 528 practices. Some practices are not included in Fingertips QOF indicators because they have a registered population (list size) of less than 750. In financial year end 2024 this was 16 practices. Because GPs vary between time periods, some don’t appear in both years. Where this is the case, we assume that the same patients still appear in the other year at a different practice.

The data reported here is a crude percentage of CHD prevalence. It does not adjust for differences in age distribution of the population of LSOAs to allow for a comparison of prevalence independently of the age structure of the area.

As QOF data is only available for practice populations as a whole, attributing practice-level CHD prevalence percentages to LSOAs must assume that all cohorts attributed to LSOAs from each GP have the same practice-level risk of CHD. The reliability of the attribution for specific LSOAs will be affected if this is not the case. For instance, if an LSOA with a younger population is served by several GPs serving wider areas with predominately older populations, and which will therefore tend to have higher percentages of CHD, then it is likely that percentages will be overestimated for the LSOA with the younger population. Attributed QOF CHD percentages for individual LSOAs are estimates and should be treated with caution, though broader spatial patterns are likely to be informative.

If you have any questions in relation to these statistics, please contact statistics@dhsc.gov.uk.