Urgent and Emergency Care Statistics Data Quality

Monthly A&E Attendances and Emergency Admissions Official Statistics

Specific guidance to users

As set out below, the completeness and data quality for this collection is fairly consistent over time which aids interpretation for users, however if data are missing for particular providers care should be taken in comparing trends in national estimates, especially for volumes such as A&E attendances although proportions such as A&E performance may still be comparable depending on the size of the missing provider.

An example would be during the period of May 2019 to May 2023 fourteen NHS Trusts undertook field testing of new UEC performance metrics and stopped reporting information on four hour performance. Specifically, the trial providers submitted their attendance data, but not their data relating to patients who spent more than 4 hours in department, so to get a national position, it is important to remove those attendances from the national totals. To make this easier for users, we have removed the attendances from those 14 providers in the performance tab, so activity volumes in that tab are not the same for those months.

The document below provides details on the impact of the field testing on the national A&E data:

Impact of CRS Field Testing on A&E Timeseries

Completeness

The published monthly A&E statistics are based on returns from all providers meeting the activity thresholds of and average of 200 attendances or 40 emergency admissions per month (calculated over a quarter), as set out in the official guidance (Guidance). This document sets out the definitions for the A&E Attendances and Emergency Admissions Monthly Return. Please note that this guidance covers current A&E standards; and defines attendances, those over 4 hours, booked appointments, admissions, waits for emergency admission, clinical streaming, exclusions and some frequently asked questions relating to this collection.

This is a well-established data collection and we consistently receive submissions from all providers for the key fields so there are rarely issues to report on the completeness of the data set. However any missing trust submissions in a month are highlighted in the Statistical Commentary for the month. This is accessible in each financial year’s page, which can be navigated to by using the navigation bar on the left of the screen or through the main landing page.

Quality Assurance

The flow of data from trusts to the publication, is explained in this document
MSITAE data flowchart V1.

First responsibility for the submission of accurate data is with providers.

As set out in the flowchart, NHS England then applies further checks on data and where figures fall outside certain thresholds that suggest potential inaccuracy, queries are raised with the submitting provider and if they confirm the figures is accurate, they are asked to provide supporting information. The submissions from each provider are compared with their previous submissions, as well as their submissions of the UEC Daily SitRep for the current month. Any data which looks to be out of line with historic submissions, is questioned with providers. Examples of this would be where a provider submits significantly more or fewer attendances than previous months submissions.

Finally, there is a twice-yearly revisions process to handle any material inaccuracies in what has been submitted by providers. This includes missing data that providers were previously unable to submit, or to account for post month validations.

Revisions are only made where the change is material at an aggregate level, and they are highlighted on the main publication page, as well as in the individual file names, and within the publication file.

Supplementary ECDS analysis

These data are sourced from the Emergency Care Dataset (ECDS) and are separate from the data published as part of the Monthly A&E Attendances and Emergency Admissions Official Statistics.

ECDS is a patient-level dataset and contains a high level of patient information, including clinical condition, discharge destination and patient demographics, enabling us to understand capacity and demand; and improve patient care.

Specific guidance to users

The completeness and data quality for this collection varies especially for non type 1 providers. These differences could impact on the consistency and comparability of A&E statistics from ECDS and it is important to bear this in mind when comparing statistics using these two different sources for providers, or looking at trend data over time.

Completeness

For information on ECDS coverage and comparisons to MSITAE, please see Data Coverage and Completeness statement linked below.

ECDS Supplementary Analysis Date Coverage and Completeness Statement February 2023

First responsibility for data completeness and accuracy is with data providers.

In the ECDS supplementary analysis publication spreadsheet, accessible via the links to monthly data from the main landing page (for February 2023 onwards). The last tab of each published supplementary ECDS analysis report sets out detailed information available on ECDS completeness and data quality within the publication at provider level for the month in question. This summary varies from month to month, so users should refer to the information for the month of interest.

In general, for type 1 and 2 providers there is good data quality for age, gender, ethnicity and 12 hour performance and more variability for chief complaint and emergency admissions via A&E. The data quality for UTC providers is more variable across all fields.

Quality Assurance

Where a provider fails a data quality check they are not included in the associated publication table. Where a provider fails all the data quality checks they are a missing provider. A list of missing providers is given at the end of the supplementary analysis publication spreadsheet and the statistical commentary file, accessible via the same route, each month. The data quality checks applied for each measure are explained in the monthly supplementary analysis publication spreadsheet and below. Each month, a provisional publication is made for the most recent month, and a finalised version of the previous month’s data. This allows for providers who are unable to submit in a more timely manner, to ensure their data are included, while still providing a timely comparison for the previous month.

Not all providers submit complete data in a timely enough manner to allow their data to be included across all analyses. Therefore, a number of different cohorts are used for each breakdown. These provider cohorts may differ in size across metrics and over time, but are still considered representative of England activity as a whole.

The cohorts used each month are:

1. Those with data for each day in the month (used for Age and Gender)
2. Those with data for each day in the month and at least 90% of records have a valid code for ethnicity
3. Those with data for each day in the month, at least 90% of records have a valid code for chief complaint, and more than one chief complaint code is used throughout the month
4. Those with data for each day in the month, at least 90% of records have a valid code for discharge destination (disposal), and more than one discharge destination code is used throughout the month
5. Those with data for each day in the month and at least 90% of records have a valid departure time (12hr performance)