Skip to main content

Part of SUS+ PbR reference manual

HRG grouping

Healthcare Resource Groups

Healthcare Resource Groups (HRGs) are the ‘currency’ of Payment by Results (PbR) for admitted patient care, outpatient procedures and A&E attendances. In the context of PbR, ‘currency’ refers to the units of healthcare for which a payment is made.

HRGs are clinically meaningful groupings of patient activity derived from NHS patient records, primarily using procedure and diagnosis codes. They support PbR by providing a means of determining fair and equitable reimbursement for healthcare services by providing consistent 'units of currency', based on expected resource use.


HRG design

HRG design is developed and maintained by the National Casemix Office, driven by policy and assured nationally through Expert Working Group consultation.

The design for each version of the classification is represented by a definitive set of rules and reference data. It is implemented by an algorithm, delivered via a software application, which follows design rules to interrogate reference tables to determine whether criteria for candidate HRGs are met by the incoming patient record data. Using a process of elimination, the most appropriate HRG is determined and assigned to the activity.


The Casemix Local Grouper

The Casemix Local Grouper is the software application that aggregates patient-level coding information into HRGs.

The local grouper performs validation checks before using a complex algorithm to assign HRGs to patient records and produces output files which contain the original input data along with the assigned HRGs. It also produces quality files that contain details of any errors or conflicts.


Grouper processing

Several validation and pre-processing stages take place before the actual grouping takes place whereby HRG codes are assigned to patient record data.

The stages are:

  1. Input data – patient activity (episode or spell data)
  2. Validation
  3. Radiology pre-processing
  4. Pre-processing (combinations)
  5. Unbundled activity
  6. Grouping – multiple trauma, procedure driven, global exception (such as PPNCD)

After grouping, Best Practice Tariff (BPT), Specialised Service Code (SSC) and Programme Budgeting Category (PBC) flags are assigned.

The data is then output as patient activity with assigned HRGs and flags.

Validation

Validation checks are built into the local payment grouper and are applied prior to the HRG code being assigned to a patient record. This checks the Office of Population Censuses and Surveys (OPCS) codes to ensure they are valid.

Reasons why a code may not be seen as valid include:

  • logical inconsistency (for example a paediatric procedure being given to someone over 18 years old)
  • an invalid or missing code being assigned (for example a treatment function code that does not exist or is clinically irrelevant for grouping, such as family history of diabetes)
  • if a deliberately ambiguous OPCS code is used (for example diagnosis/procedure ‘unspecified’)

Radiology pre-processing

Pre-processing occurs for radiology activity because it must be mapped to reference data before processing.

Pre-processing

Pre-processing involves creation of combination procedure codes, and logical deletes.

Unbundled activity

Unbundled procedures are processed separately to derive unbundled HRGs (See unbundled HRGs). The grouper then ignores these unbundled components when deriving the core HRG.

When all significant procedures in an admitted patient care episode or spell are unbundled, diagnosis is used to derive a core HRG for the episode. For outpatient care, if all procedures are unbundled the episode is allocated one of the eight relevant non-admitted care attendance HRGs as a core HRG.

There can be one or more ‘unbundled’ HRG codes assigned that can be used to identify the use of repeating resource use such as scans.

Grouping

Grouping is the main stage of the process in which one HRG code is assigned to the spell. This is referred to as the core HRG. Patient record data items, such as procedures, diagnoses, age and length of stay are used to determine the appropriate HRG code for the Spell.

Assign flags (BPT, SSC, PBC)

Best Practice Tariff (BPT)

Most Best Practice Tariff flags are generated by OPCS and ICD10 (International Classification of Diseases) codes. These may be required in combination and may also require qualification by other codes, for example site or approach codes.                       

Only certain HRGs can be flagged as best practice. In addition, age criteria and type of admission will determine whether activity can be flagged as best practice.

Specialised Service Codes (SSC)

SSCs are assigned based on the record meeting predefined reference criteria.

Programme Budgeting Categories (PBC)

The grouper maps the primary diagnosis of a patient record to a Programme Budgeting Category (PBC) which is then output by the grouper. There is no direct mapping of HRGs to PBCs. The programme budget category allows high level reporting on the amount of money being spent on specific treatments, such as cancer or heart disease.


Unbundled HRGs

A pathway of care typically consists of a number of different service elements such as diagnostic imaging, high cost drugs and rehabilitation. Unbundled HRGs account for these consumable elements, allowing them to be commissioned, priced and paid for on an individual basis. PbR data contains the first 12 unbundled HRG codes generated for the activity.

Unbundled HRGs are generally only assigned non-mandatory tariffs as set out by PbR policy. SUS only applies mandatory tariffs to the PbR data and tariff information does not appear in the Aggregate Unbundled Adjustment National data item. They do however allow for this type of activity to be easily identified for further analysis.


Differences between the SUS Grouper and Local Grouper

The grouper is implemented as a standalone entity within SUS. This means that the SUS grouper is essentially the same as the Casemix local grouper and uses the same algorithm and design logic. Therefore, if the SUS grouper and the local grouper receive identical inputs, both groupers will produce identical outputs.

It is important to note, however, that SUS applies cleaning rules and therefore it is possible that in some cases the results produced by the SUS grouper differ from those produced by the Casemix Local Grouper when using the same baseline data. Otherwise the results of grouping the same data in SUS or in the corresponding local payment grouper are always the same. There is no difference in the grouping engine. Any apparent differences between SUS grouper and local grouper output data are likely to be caused by the following.

Exclusions

Some episodes are excluded from spells according to PbR policy rules. Lack of such records in a spell can affect the spell-level grouping results. To get the same results it may be necessary to remove SUS excluded episodes from a local grouping run.

Critical Care Days calculations

The SUS derivation of critical care days is more comprehensive than in guidance provided to support local grouping. It also includes several validation checks that may result in a slightly different number of critical care days being allocated which may affect grouping results.

Spell creation

The local grouper uses the Provider Code and Hospital Provider Spell Number data items to group episodes into spells. SUS derives its own spell identifier and uses it in place of the Hospital Provider Spell Number. This can mean that different sets of episodes are treated as part of a spell. In normal use this type of disparity is expected to be extremely rare.

In outpatient care the SUS grouper can produce two HRGs. This can occur when the derived HRG is not tariffed. An attendance HRG is calculated by a means equivalent to grouping after removal of all procedures except those starting with ‘X62’.

In emergency medicine the SUS grouper produces a ‘dummy’ HRG DoA (Dead on Arrival) from which a tariff is derived if the A AND E PATIENT GROUP data item contains value 70 (Brought in dead).

In local grouping, users are advised to group appropriate sub-sets of critical care data separately and treat their results as unbundled HRGs as a part of the parent episode or spell. This is necessary because of the difficulties expected for many users when dealing with more complex relational input. SUS groups the combination of Admitted Patient Care and Critical Care simultaneously.


PrB excluded activity

PbR excluded activity is grouped but not priced. Excluded episodes or attendances do not have any tariff applied or financial adjustments made. (See also Exclusions)

Excluded OP (outpatient) activity

Exclusions of outpatient episodes are performed after HRG grouping has taken place. This means that SUS derives the appropriate valid Core HRGs for all outpatient attendances, regardless of whether that activity is later identified as being excluded from PbR (Note: prior to SUS R11 the core HRG value of excluded OP attendances was set to N/A).

All records (PbR included and excluded) are output in the OP extracts with a populated HRG code. Where the Grouper has derived a procedural HRG and SUS has then derived an attendance HRG (WF), both HRGs are output, even if the episode is then excluded.

Excluded EM (A&E) activity

Exclusions of Emergency Medicine (EM) episodes are performed after HRG grouping has taken place. This means that SUS derives the appropriate valid Core HRGs for all EM attendances, regardless of whether that activity is later identified as being excluded from PbR.

All records (PbR included and excluded) are output in the EM extracts with a populated HRG code. Where SUS identifies a ‘DOA’ (dead on arrival), this value is used to populate the HRG data item.

CDS 011 ECDS contains streamed activity which is handled as excluded – see ECDS

Excluded APC (inpatient) activity

Submissions up to and including 2016/17:

Admitted Patient Care (APC) episode-level exclusions resulted in the creation of Excluded Single Episode Spells. These spells receive a Spell in PbR/Not in PbR indicator of 7 and were assigned new, unique Spell IDs. The original Spell ID became the Parent Spell ID which was used to link the Excluded Single Episode Spell to the original spell.

A Parent Spell where all the constituent episodes were excluded (single or multi episode spells) received a Spell in PbR/Not in PbR indicator of 2 (All episodes in spell are excluded (prior to grouping) and a Spell Core HRG of N/A.

The APC Spells extract incorporated Excluded Single Episode Spells as standard, regardless of whether:

  • all episodes in the spell are excluded, or
  • one or more (but not all) episodes are excluded.

The APC Spells extract was sorted so that Excluded Single Episode Spells were grouped with the corresponding Parent Spell. The Parent Spell ID and the Spell ID (child) of the newly created excluded spells were reported on both the episode and spell-level extracts.

From 2017/18 onwards (SUS+)

The practice of creating single episode excluded spells was discontinued.

All episodes in a spell will have the same SUS Spell ID.

All episodes and spells are grouped and assigned an HRG. As before, the PbR spell HRG used for tariff will only use data from included episodes. All episodes are clearly marked as included or excluded from PbR. Each spell carries an indicator to show if all, some or none of the constituent episodes are included.

SUS+ also reprocessed data for years 2015/16 and 2016/17. Spell data extracted from SUS+ for these years will not contain the excluded single episode spells that may be found in SUS extracts for the same period.

Provider exclusions (equals sign)

The equals ‘=’ sign is used by providers to exclude activity from PbR when intentionally included in COMMISSIONING SERIAL NUMBER. This locally priced activity identification method does not affect the grouping process.


Short stay emergency admissions

In some cases the combination of the HRG and the Spell Length of Stay mean a record qualifies for the Short Stay Emergency tariff. The tariff is reduced as the patient has stayed for less time than expected for the allocated HRG. The number of days for each HRG where this applies is the short stay trim point. This is compared to the PbR adjusted length of stay.


Regular attenders

A regular attender is identified as an elective admission with 1 of the following:

  • Patient Classification 3 (regular day admission) with an Episode Length of Stay of less than 1
  • Patient Classification 4 (regular night admission) with an Episode Length of Stay of less than 2

All records that qualify as regular attender (REG) form single episode spells. All episodes with a Patient Classification of 3 or 4 that do not meet the above identification criteria as regular attenders are excluded from SUS.

Zero-priced regular attender tariff

Renal dialysis for chronic kidney disease uses the national renal dataset which is outside of SUS, and mandatory tariffs for unbundled chemotherapy delivery and external beam radiotherapy activity have been introduced in a staged way. Therefore, SUS does not apply a mandatory tariff to this activity. The only mandatory tariffs assigned for regular attender activity are where the core HRG is one of the five zero priced (£0) HRGs. These are priced in SUS and are shown below.

  • LA08E, chronic kidney disease with length of stay 1 day or less associated with renal dialysis
  • PB03Z, healthy baby
  • SB97Z, same day chemotherapy admission or attendance
  • SC97Z, same day external beam radiotherapy admission or attendance
  • UZ01Z, data invalid for grouping

All other core HRG activity tariffs for regular day and night attenders are negotiated locally.


Grouping Algorithm Version

The data item Grouping Algorithm Version identifies the version of the grouper used to group the submitted CDS data. The value refers to the HRG grouping logic identified in the reference data table REF_PBR_GROUPER_INFO.

A Grouping Algorithm Version of ‘1415’ refers to the 2014/15 Payment Grouper.


Prescribed Specialised Services (PSS)

SUS+ integrates Prescribed Specialised Services (PSS) processing as developed by the National Casemix Office in partnership with NHS England. The same PSS data items as produced in the PSS Operational Tool are output in the following extracts:

APC full online

FCE NPOC

National Programme of Care (NPoC) code output by the PSS grouper at episode level.

FCE Service Line

Highest ranking Prescribed Service Line code output by the PSS grouping process at episode level.

FCE Service Line List      

List of Prescribed Service Line codes produced as an output of the Prescribed Specialised Services grouping process.

Spell NPOC

National Programme of Care (NPoC) code output by the PSS grouper at spell level.

Spell Service Line

Highest ranking Prescribed Service Line code produced as an output of the PSS grouping process at spell level.

OP full online

Spell NPOC:

National Programme of Care (NPoC) code output by the Prescribed Specialised Services grouper at the spell level.

Spell Service Line:

Highest ranking Prescribed Service Line code produced as an output of the Prescribed Specialised Services grouping process at spell level.

Specialised Service Codes (SSC) codes data items are present in the extracts for the purpose of supporting established local processing routines. These data items have been identified as obsolete and will be removed in future. Prescribed Specialised Service (PSS) and NPoC codes will be output into new data items as detailed above.

For more information on the new codes, please refer to the Extract Specification on the SUS+ PrB guidance page.

For more information on the design of the PSS Operational Tool, please refer to the Casemix PSS page.


Last edited: 10 June 2022 4:33 pm