Skip to main content
Uncurated Low Latency Hospital Data Sets for Admitted Patient Care, Outpatient and Critical Care

The Secondary Uses Service (SUS) is used to create additional uncurated data sets such as this which offer access to more current data, though less complete and assured. 

Overview of uncurated hospital data

NHS Digital routinely collects data from hospital providers regarding a patient's time at hospital as part of the Commissioning Data Set (CDS). This is then processed and is returned to healthcare providers as the Secondary Uses Service (SUS) data set and is used by the NHS for operational purposes.

Most NHS hospital trusts submit data on a monthly basis by deadline following a two-phase reconciliation process to arrive at a final agreed position for each month's activity defined in the NHS Standard Contract for payment purposes. This data is consolidated, validated and cleaned and then used to create the Hospital Episode Statistics (HES) data set which is released on a monthly basis as official statistics.

A number of trusts now submit data as part of the CDS more frequently such as weekly or even sometimes daily. This subset of more frequently reported management information data reported in SUS provides opportunity for additional insights from that which might be immediately available in HES data. To help unlock the opportunities this SUS data can provide, a number of new products are being developed to support an emerging need amongst researchers to have access to more timely though less complete and assured data.

This includes what is being termed un-curated Low Latency Hospital CDS data. These new products are available in NHS Digital’s Trusted Research Environment (TRE) and access to them is managed via the Data Access Request Service (DARS).

Initially the data contained within these products will be updated on a monthly basis but will be taken at different times to the pre-announced monthly deadlines that drive the primary commissioning use of the CDS data itself. This will influence the quality and completeness of the data when compared to products like HES.  

The data offered are un-curated. This means that a number of the data cleaning and assurance processes associated with products such as those articulated on the HES processing cycle page are not undertaken in this product. 

Available products

The following 3 products are available: 

  • Uncurated Low Latency Hospital Data Set for Admitted Patient Care
  • Uncurated Low Latency Hospital Data Set for Outpatients
  • Uncurated Low Latency Hospital Data Set for Critical Care

Customers can choose from 2 packages for each of the data sets. Each package will include a list of pre-determined pseudonymised fields:

  • package 1 will include all derived and non-derived pseudonymised fields
  • package 2 will include non-derived fields pseudonymised fields only

Use the Uncurated Hospital Data External Specification document for full information on the fields available in each package.  

Uncurated Hospital Data External Specification

The data in these products is the data exactly as it has been submitted by hospital trusts.

The established HES product provides a curated view of this data which have been quality assured and refined before being released.

It is important to note that the products detailed have not been through the HES service. 

Advisory note when using this data

Customers wishing to access these new products need to be aware of the potential issues with the data and will be asked to sign an agreement that they have understood the limitations and the inherent risks. 

All data within these products should be considered provisional in nature and may be incomplete or contain errors for which no adjustments have yet been made.

Counts produced from provisional data are likely to be lower than those generated for the same period in other datasets. This shortfall will be most pronounced in the final month of the latest period and will be more pronounced in this product than is the case with HES due to differences around when the monthly snapshot of this data is taken compared to the pre- announced commissioning submission deadlines.

It is also probable that clinical data are not complete, which may in particular affect the last 2 months of any given period. There may also be errors due to coding inconsistencies that have not yet been investigated and corrected. 

Data quality review of frequency and diagnosis coding completeness for activity within SUS APC data set by organisation and week

An indication of potential impacts this may have on more frequent data snapshots from the CDS.

Accessibility of this tool

This tool is in Microsoft PowerBI which does not fully support all accessibility needs. If you need further assistance, please contact us for help.

All data available have been de-personalised and no person identifiable data are made available. National data opt outs do, therefore, not apply to these products and have not been applied.

Derived fields

Two versions of each product will be made available:

  • package 1 version is the raw SUS data with none of the derived fields included
  • package 2 version will include both the raw CDS data and a number of derived fields” or 'derivations'

The derivations have been generated by SUS following published derivation rules which are used to derive additional data items based on:

  • data held on other national systems (Patient Demographic System, Organisation Data Service or Post Code data)
  • nationally defined policy (Payment by Results, HRG Grouping and Commissioner Assignment Method)

No fields that are submitted to SUS by the hospital provider will be amended nor will any records be added or removed (for instance suspected duplicate records will not be removed). 

Exceptions to this are legally restricted data whereby patient identifiers are stripped out if a provider sends them in error.  

Derivations within this product that are produced utilising reference data are derived at the point of time of processing of that record and will not be subsequently updated within the financial year.

This means that there is a risk that derivations for records relating to identical activities on the same date may differ when queried due solely to when updated reference data available in the back end becomes available throughout the year.

The lag time of availability of such reference data may vary. This differs from HES data where data is rerun each month to ensure consistency across all records within the derived data against a consistent version of that reference data.

Unlike the published HES products, the derived fields presented in these new products have not been fully assured by NHS Digital and so should be treated with caution. 

Support model

This early version of more timely uncurated data products is still in an early developmental and assurance phase with data being made available to facilitate an increasing number of research needs whist longer term strategic initiatives are being progressed to provide more assured and rapid hospital activity data. 

It is envisaged upon their completion that these initiatives will replace the requirement for the new products offered here in due course. 

As such it is not possible to offer a support facility to aid with queries relating to the data contained within these products.    

General requests for support

More general requests for support for these products is strictly offered on a 'best endeavours' basis and there are no guarantees that support can be made available.  

Support requests should be made by email:

Last edited: 4 November 2021 11:43 am