Completeness and representativeness of small area socioeconomic data linked with the UK Clinical Practice Research Datalink (CPRD)

Study type
Protocol
Date of Approval
Study reference ID
21_000382
Lay Summary

The Clinical Practice Research Datalink (CPRD) is a repository of primary care electronic healthcare records in the United Kingdom (UK). CPRD collects anonymised patient data from a network of general practices across the UK. Primary care data from CPRD are linked to many other datasets, including socioeconomic (SES) measures, to provide a fuller picture of health in the UK.

This study is designed to describe SES and Rural-Urban Classification (RUC) data that can be linked with CPRD. We will assess the proportion of patients and practices in CPRD that have SES/RUC data. We will assess the proportion of patients whose individual-level SES/RUC classification matches that of their GP. This is important as not all patients have individual measures available and researchers may approximate by using the classification of their GP. We will assess the agreement between the different measures to enable us to better advise on the similarities and differences between these indices. We will assess the representativeness of the CPRD patient population in terms of SES, RUC, age, and sex by comparing the distribution of patient and/or practice classifications in CPRD to the SES, RUC, age, and sex distributions of the general populations of the UK, Great Britain, England, Scotland, Wales, and Northern Ireland.

This project will provide a better understanding of the small area SES/RUC linkages available in CPRD, and the representativeness of these data, and will serve to improve interpretation of research results of studies using CPRD data where SES has an important contextual impact on health.

Technical Summary

This study will be a retrospective cohort study of completeness and representativeness the area level socioeconomic (SES) and Rural-Urban Classification (RUC) data linked with the Clinical Practice Research Datalink (CPRD) GOLD and Aurum databases, amongst acceptable patients and currently registered patients eligible for patient postcode linkage. SES/RUC data linked with CPRD can be provided at the practice-level across the UK, and at the patient-level for England only. Currently, CPRD offers linkage with the Indices of Multiple Deprivation (IMD), the Townsend Deprivation Index, the Carstairs Index, and RUC, which are commonly used in health research as a proxy for individual level SES in analyses.

Firstly, we will assess the completeness of SES/RUC data linked to CPRD primary care data by determining the proportion of patients and practices with SES data available and the characteristics related to SES/RUC recording. Secondly, we will assess the agreement and correlation of patient- and practice-level SES/RUC measures to explore whether the practice-level measures can be used as a proxy for the patient level measure where this is missing. Thirdly, we will assess the correlation between the various patient-level SES metrics for patients in England , to ascertain the likely impact of choice of area-based SES measure on interpretation. Finally, we will assess the representativeness of each SES measure in CPRD compared with the general populations of the England (at patient- and practice-level), and the United Kingdom, Great Britain, Scotland, Wales, and Northern Ireland (practice-level only), as well as assessing the representativeness of the CPRD populations by age and sex.

This project will serve patients and researchers by providing up-to-date information on these important demographic variables including their completeness, usability, limitations, and representativeness of SES linkages in CPRD to inform choice of data sources, interpretation of results, and translation of those results into practice for patients.

Health Outcomes to be Measured

- Completeness of patient-level SES/RUC measures in CPRD for England and English regions, and completeness of practice-level SES/RUC measures in CPRD across the UK
- Agreement and correlation of patient- and practice-level SES/RUC in CPRD for England and English regions
- Correlation of different SES measures in CPRD at the patient-level for England and English regions, and at the practice-level across the UK
- Agreement of SES, age and sex distribution with the general UK, GB, and country populations

Collaborators

Eleanor Axson - Chief Investigator - CPRD
Eleanor Axson - Corresponding Applicant - CPRD
Helen Booth - Collaborator - CPRD
Karen Cuenco - Collaborator - The Gates Foundation
Mia Harley - Collaborator - CPRD
Preveina Mahadevan - Collaborator - CPRD
Rebecca Ghosh - Collaborator - CPRD
Susan Hodgson - Collaborator - CPRD

Former Collaborators

Mia Harley - Collaborator - CPRD

Linkages

2011 Rural-Urban Classification at LSOA level;2011 Rural-Urban Classification at LSOA level;Patient Level Carstairs Index for 2011 Census;Patient Level Index of Multiple Deprivation;Patient Level Index of Multiple Deprivation Domains;Patient Level Townsend Score;Practice Level Carstairs Index for 2011 Census (Excluding Northern Ireland);Practice Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation Domains