Linking CPRD Primary Care Data to HES Diagnostic Imaging Data; A Data Resource Profile

Problem

When using electronic health data to study patient care pathways it is often useful to see beyond that which is administered and recorded in primary care. The Clinical Practice Research Datalink (CPRD) links primary care data to secondary care data sources including the Hospital Episode Statistics Diagnostic Imaging Dataset (HES DID). HES DID is a collection of detailed information about diagnostic imaging tests carried out on NHS patients in England since April 2012. HES DID contains information about referral source and patient type, details of the test, plus items about waiting times for each diagnostic imaging event, from time of test request through to time of reporting. HES DID does not contain the images produced or the results of the test.Whilst linkage between these two data sets has been in place for a few years to date nobody has attempted to describe what it offers. The aim of this study is to profile the linkage between HES DID and CPRD data in terms of content, coverage, completeness and concordance in order to define information that can be gained from the linkage.

Approach

The HES DID linked data will be described in terms of coverage of the data source over time and by geography. We will also investigate how complete individual scan records are. Analysis will look at the type of scans recorded in HES DID, which are the most frequently recorded imaging scans and how does that compare with records of scans in primary care data.We will look at the concordance between scans being recorded in CPRD data (both as referral records and scan outcomes) and the record of a scan taking place in HES DID. We will explore whether the linkage between these two data sets and other HES datasets linked to CPRD allows researchers to follow a patient journey from referral for an imaging test through to diagnosis and treatment. We will also carry out a descriptive analysis of the average waiting times between each stage of the patient’s care.

Findings

The latest release of HES DID data linked to CPRD covers the period April 2007 – June 2019. Initial exploratory findings have shown that there were ~15 million CPRD patients who had at least one scan recorded in HES DID during this period with the number of scans recorded in the data increasing slightly year on year. The most frequently recorded scan type was chest x-ray with over 5 million records.

Consequences

The linkage between CPRD primary care data and HES DID has the potential to provide a valuable resource for researchers wishing to study treatment pathways. This descriptive analysis will allow researchers to assess the value that linked HES DID data could add to their study.