Informatics 97nov2

Journal of Informatics in Primary Care 1997 (November):2-5


Meeting the requirements of specialists and generalists in Version 3 of the Read Codes: Two illustrative "Case Reports"

Fiona Sinclair MRCGP*, Erich B Schulz MB BS† , Colin Price MPhil FRCS†

*NHS CCC General Practice Specialty Working Group  †NHS Centre for Coding and Classification, Loughborough, LE11 2TN, UK


The Read Codes have been recognised as the standard for General Practice computing since 1988 and the original 4-byte set continues to be extensively used to record primary health care data. Read Version 3 (the Read Thesaurus) is an expanded clinical vocabulary with an enhanced file structure designed to meet the detailed requirements of specialist practitioners and to address some of the limitations of previous versions. A recent phase of integration of the still widely-used 4-byte set has highlighted the need to ensure that the new Thesaurus continues to support generalist requirements.


The Read Codes[1] were originally designed to allow general practitioners to collect summary patient data using personal computers and, in 1988, they were adopted as the standard for general practice computing[2]. In 1990, when the original 4-byte set was purchased by the Department of Health and became Crown Copyright, a new set of terms was introduced to enable hospitals to generate central returns using the Office of Population Censuses and Surveys Classification of Surgical Operations and Procedures, 4th Revision (OPCS4)[3] and the International Classification of Diseases, 9th Revision (ICD9)[4]. In this new 5-byte set, the procedure and disorder chapters were supersets of these classifications, firmly based on the classification structures but with the inclusion of additional concepts. At the same time, a more sophisticated file structure known as Version 2 was introduced. Amongst other features, it allowed the different terms associated with one concept to be identified by a term code. As most users of the 5-byte set also adopted the Version 2 file structure, Version 2 has generally come to mean the same as the 5-byte set.

Subsequently, it became apparent that a key requirement of the NHS's Information Management and Technology Strategy[5,6] was for a standardised clinical vocabulary to allow uniform and unambiguous information sharing throughout the service. The Terms Projects[7,8] (1992–95) involved all the healthcare professions in the enhancement of the Read Codes to meet the information needs of specialist practitioners. Forty-three medical specialties were involved and funding was provided for each to appoint a research worker.

The early phases of the Project involved a review of the existing 5-byte codes in an allocated domain of prime responsibility. New terms were included as required and those felt to be unsatisfactory were excluded. A process of consultation with other interested specialties was undertaken and, when agreement was reached, the restructured list of terms was submitted to the NHS Centre for Coding and Classification (NHS CCC) for integration into Read Version 3 (the Read Thesaurus)[9].

In the later phases of the Clinical Terms Project, at the NHS CCC, all 5-byte concepts were incorporated into the new version, either by matching to ones which had already been included as clinically useful (current), or by placement in the hierarchy as additional concepts (optional). These latter concepts were typically represented compositionally in the current set (as in Figure 1) or included unnatural classification language (e.g. NOS, NEC).

This 5-byte integration was the first stage in the development of the Read Code superset which aims to enable information recorded using early versions to be accessible from Version 3 and to allow the new hierarchy to support analysis of historical information. Since the Projects formally ended in April 1995, there has been ongoing refinement and an important facet of this work has been the integration of the original 4-byte set.

Version 3 File Structure

Two major structural enhancements, reflecting published work on the construction of controlled clinical vocabularies[10], and aiming to overcome recognised limitations of existing coding schemes[11], have been introduced in Read Version 3[12].

Firstly, there is a separate table to hold the hierarchical arrangement of concepts, independent of the alphanumeric code. A single concept, with the same code, can now be placed in several hierarchy locations, allowing different specialist views to be presented. Furthermore, whereas in previous versions of Read it was only possible to manipulate the hierarchy by adding to it, in Version 3 major rearrangement can be undertaken if required. This improved flexibility has enabled many of the documented deficiencies of earlier Read Code versions[13,14] to be addressed and has also allowed the amendments discussed later in this paper.

The second structural enhancement is the adoption of a mechanism to enable the attachment of additional detail through the use of qualifiers. This has allowed long lists of concepts that only varied in a single facet, to be replaced with much shorter lists in Version 3, whilst templates specify additional information that may be appended (Figure 1) to create complex concepts compositionally.

Figure 1: Core concepts and qualifiers

5 byte representation:
Total prosthetic replacement of hip using cement

Primary cemented total hip replacement

Revision cemented total hip replacement

Total prosthetic replacement hip not using cement

Primary uncemented total hip replacement

Revision uncemented total hip replacement

Version 3 representation:
Total hip replacement

Revision status: Primary/Revision

Type:                 Cemented/Uncemented

4-byte integration

One of the authors (FS), in liaison with the General Practice Specialty Working Group, reviewed each 4-byte term with a view to matching to those in the Version 3 hierarchy. As a result of this exercise, it became apparent that a number of useful 4-byte concepts had not been included in Version 3. Using two examples from the surgical procedures chapter, we illustrate how the flexible structure of the Read Thesaurus can, within the same clinical terminology, support the representation of both specialist and generalist views of clinical practice.

Two reports:

(1) Labyrinthectomy

Figure 2a shows a hierarchy from the Inner ear procedures section of the Thesaurus in which three specialised types of labyrinthectomy are presented. A simpler concept Labyrinthectomy, present in the 4-byte set, has not been included – possibly because the specialist authors did not foresee a requirement for recording such generalist data. A simple restructuring (Figure 2b) accommodates this 4-byte concept resulting in a satisfactory hierarchy for both specialists and generalists.

Figure 2: Clinical Terms Project hierarchy for labyrinthectomy:

(a) before and (b) after 4-byte integration



Operation on vestibular apparatus

Operation on vestibular apparatus

Vestibular neurectomy

Vestibular neurectomy



Ultrasonic labyrinthectomy


Membranous labyrinthectomy

Ultrasonic labyrinthectomy

Osseous labyrinthectomy

Membranous labyrinthectomy


Osseous labyrinthectomy

(2) Prostatectomy

In our second example, we examine the prostatectomy hierarchy through successive versions to show how, within the Read Thesaurus, all previous versions can be reconciled.

The 4-byte hierarchy for this procedure is shown in Figure 3 and extracts from the corresponding 5-byte hierarchy in Figure 4.

Figure 3: 4-byte hierarchy for prostatectomy

7B1. Prostatectomy

7B11       Transurethral prostatectomy

7B12       Suprapubic prostatectomy

7B13       Retropubic prostatectomy

7B14  Radical prostatectomy

7B15  Perineal prostatectomy

7B1Z       Prostatectomy NOS

The hierarchy of the 5-byte set illustrates the consequences of the inflexible code-dependent trees of the early versions. The five-level fixed-depth hierarchy does not allow placement of Transvesical two stage prostatectomy as a child of Transvesical prostatectomy (TVP). This makes both TVP concepts children of Open prostatectomy, but fails to represent the true relationship between the two, so that analysis on the code 7B362 (=TVP) will not retrieve any 7B364 (=2 stage) instances of this operation.

Figure 4: 5-byte hierarchy for prostatectomy

7B3.. Bladder outlet & prostate operations …

7B36. Open prostatectomy

7B360 Radical prostatectomy - unspec excis pelvic nodes

7B361 Retropubic prostatectomy

7B362 Transvesical prostatectomy

7B363 Perineal prostatectomy

7B364  Transvesical two stage prostatectomy

7B365  Radical prostatectomy without pelvic node excisn

7B366 Radical prostatectomy with pelvic node sampling

7B367 Radical prostatectomy with pelvic lymphadenectmy

7B383 Endoscopic laser ablation of prostate

7B38y Endoscopic resection of outlet of male bladder OS

7B38z Endoscopic resection of outlet of male bladder NOS

7B39. Endoscopic resection of outlet of male bladder or prostate

7B390 Transurethral prostatectomy

7B391 Punch resection of prostate

7B392 Other endoscopic resection of prostate

Italicised terms are optional.

In addition, the subsequent insertion of intermediate concepts within the hierarchy is not possible. The simple but useful concept prostatectomy is absent from the 5-byte set (in which the hierarchy was based on that of OPCS4 and forced the user to specify whether a prostatectomy was open or transurethral). If this concept were to be added to the 5-byte set, it would not be possible to correctly position the various types of prostatectomy as children.

Figure 5 illustrates the reworked Version 3 hierarchy developed during the Clinical Terms Project.

Figure 5: CTP prostatectomy hierarchy

Bladder outlet & prostate operations …

Open prostatectomy

Radical prostatectomy

Radical prostatectomy - unspec excis pelvic nodes

Radical prostatectomy without pelvic node excisn

Radical prostatectomy with pelvic node sampling

Radical prostatectomy with pelvic lymphadenectmy

Retropubic prostatectomy

Transvesical prostatectomy

Transvesical two stage prostatectomy

Perineal prostatectomy

Transurethral prostatectomy

Punch prostatectomy

Revision of transurethral prostatectomy

Transurethral vaporisation of prostate

The specialty group have utilised the unlimited hierarchy depth to enable true subtype structuring of concepts, that is, all the variants of radical prostatectomy are subordinate to that concept and 2 stage TVP is subordinate to TVP.

Comparison with Figure 3, however, reveals that the general concept prostatectomy is still absent. This has now been rectified, and Figure 6 shows the final hierarchy after integration of the 4-byte codes into Version 3 with appropriate rearrangement such that the terminology can now support all levels of detail.

Figure 6: Current prostatectomy hierarchy

Bladder outlet & prostate operations . . .


Open prostatectomy

Radical prostatectomy


Retropubic prostatectomy

Transvesical prostatectomy

Transvesical two stage prostatectomy

Perineal prostatectomy

Transurethral prostatectomy

Punch prostatectomy

Revision of transurethral prostatectomy

Transurethral vaporisation of prostate


Achieving the goal of a standard clinical vocabulary to be shared by many different disciplines throughout the Health Service requires a flexible structure in which the varying needs of diverse specialties can be accommodated whilst, at the same time, maintaining a useful relationship between concepts for the purposes of aggregation and analysis.

We have illustrated how the Read Code Version 3 hierarchy file structure enables these requirements to be met. Nevertheless, a key factor underpinning the creation of a shared clinical terminology is the need for each clinical specialty to understand and tolerate the information requirements of those outside the specialty. The NHS CCC continues to liaise with a wide range of specialties in the refinement of the Read Thesaurus to ensure that all viewpoints can be represented.


  1. Chisholm J. The Read clinical classification. Br Med J 1990; 300:1092
  2. General Medical Services Committee: RCGP. Joint Computing Group. The classification of general practice data: final report of the GMSC-RCGP. Joint Computing Group Technical Working Party. General Medical Services Committee, BMA, London, 1988
  3. OPCS. Classification of surgical operations and procedures (4th revision). HMSO, London, 1990
  4. World Health Organisation. International Classification of Diseases. 9th Revision. WHO, Geneva, 1975
  5. Information Management Group of NHS Management Executive. An information management and technology strategy for the NHS in England: IM & T strategy overview. HMSO, London, 1992
  6. Leaning MS. The new information management and technology strategy of the NHS. Br Med J 1993; 307:217
  7. Buckland R. The Language of Health. Br Med J 1993; 306:287–288
  8. Severs MP. The Clinical Terms Project. Bulletin of Royal College of Physicians (London) 1993; 27(2):9–10
  9. Calman KC. New national thesaurus. CMO's Update 1994; 4:1
  10. Cimino JJ, Hripcsak G, Johnson SB, Clayton PD. Designing an Introspective, Multipurpose, Controlled Medical Vocabulary. In: Kingsland LC (ed.). Proceedings of the Thirteenth Annual Symposium on Computer Applications in Medical Care. IEEE Computer Society Press, New York, 1989:513–518
  11. Coiera E. Medical informatics. Br Med J 1995; 310:381–387
  12. O'Neil MJ, Payne C, Read JD. Read Codes Version 3: A User-Led Terminology. Meth Inform Med 1995; 34:187–192
  13. Smith N, Wilson A, Weekes T. Use of Read codes in development of a standard data set. Br Med J 1995; 311:313–315
  14. Williams JG. Recording and retrieving data about depression and related problems in general practice using 4-byte Read codes. In Richards B (ed.). Conference Proceedings – Current Perspectives in Healthcare Computing 1996. BJHC Ltd., Weybridge, 1996:535–542


  • There are currently no refbacks.

This is an open access journal, which means that all content is freely available without charge to the user or their institution. Users are allowed to read, download, copy, distribute, print, search, or link to the full texts of the articles in this journal starting from Volume 21 without asking prior permission from the publisher or the author. This is in accordance with the BOAI definition of open accessFor permission regarding papers published in previous volumes, please contact us.

Privacy statement: The names and email addresses entered in this journal site will be used exclusively for the stated purposes of this journal and will not be made available for any other purpose or to any other party.

Online ISSN 2058-4563 - Print ISSN 2058-4555. Published by BCS, The Chartered Institute for IT