Submission to the PAC by John Mason (7 Nov 2006)

From Nhs It Info



This document is triggered by the Richard Bacon and John Pugh suggestions and is a comment on the present state of IMT in the NHS, with suggested ways forward.


While the NHS Information Authority may have been thought to be a cumbersome organisation, a considerable baby was lost with bath water when it was dismantled. This baby included:

  • continuing work on the NHS number to accurately identify patients and clean data on existing local systems
  • confidentiality
  • information standards and the need to share information between healthcare
  • messaging
  • the need to include the patient in the equation.

There was professional advice available from BMA/Royal Colleges both Medical and Nursing and Allied Healthcare Professionals, together with an Information Group chaired by Sir Kenneth Calman which placed an emphasis on the patient pathway, and included social services input.

It was recognised that the information needs and the shape of the record would be different for several groups, and that there is a particular difference between so-called primary and secondary care.

Historically, the hospital record has a high content of lab and specialist test results together with input from speciality groups e.g. anaesthetics, cardiology, oncology, genetics etc. It is often typed (legible), contains correspondence, and in addition the activity is coded by professional coders using the International Classification of disease ICD10, and our UK Office of Population Census and Surveys (OPCS) codes for procedures. There is a National summary front sheet HMR1 which must have coded and dated diagnostic and procedure entries for each patient discharge. The diagnostic inpatient information is required by the World Health Organisation from all countries in ICD coded form. Latterly the data quality of the hospital returns has been subject to audit.

Historically, primary care or more accurately General Practice used a National format, the Lloyd-George envelope, a brown A5 sized container with an external summary of uncertain quality, often containing hospital discharge letters and lab results. General Practitioners led the way in the use of computers in practice, originally, it has to be said, to keep track of remunerable items of care e.g. cervical smears particularly the sending of recall letters etc. Systems improved, and there was a need to record clinical data. James Read recognised that the use of a computerised hierarchical system allowed speedy recording of clinical information to the level of detail felt appropriate e.g. heart disease which embraced valve disease, coronary heart disease, congenital heart disease and so on. These Read Codes exposed the first problem of leaving untrained clinicians to code, which was the tendency just to use the highest level rather than being precise e.g. to record heart disease rather than mitral valve disease. ICD and OPCS are of little use for general practice coding as there are few social and disability codes in ICD9 although this has improved somewhat in ICD10. Equally the attractive speed of entry of data led hospital doctors using local systems to look for something which covered hospital medicine and National work produced expanded the Read codes to ReadV3. Disappointingly General Practice suppliers were unwilling to alter systems which had embedded the earlier codes; hospitals, still intent on the existing coding methods (now essential for payment with the arrival of resource management) put no pressure on suppliers to incorporate the codes and nothing came of it all.

Newer GP systems recognise the need for input from nurses and other staff involved in the care of a patient, whereas hospitals tended to develop separate systems for nursing, physiotherapy, renal clinics and so on. The central hospital Patient Administration Systems (PAS) are exactly that, recording clinic appointments, waiting lists and inpatient occupancies, only latterly taking on the HMR1 role with coded admission and outpatient summary information. It should be added that many of these systems although ageing are very effective and robust administrative tools, even with the bolt-ons now needed to provide waiting-time data etc. The arrival of Choose and Book has made upgrading of these systems now essential. Finally many hospitals already recognised the need for electronic image storing of X-rays and had started to attempt to fund PACS as individual trusts before the advent of NPfIT.

Historical problems

During this time there have been repeated changes in the management structure of the NHS. These changes have, or will even yet, seriously damage attempts to build information systems which will meet clinical needs and which will only be supported (i.e. funded) by managers certain of obtaining the data demanded by the management and financial structure of the day.

Although input from the professions was funded and sought, it has to be admitted that the Royal Colleges failed to grasp the chance to have major input. The BMA, very much GP oriented felt that existing systems for general practice filled a need. The organisation perhaps rightly became much more concerned about confidentiality of patient information and made valuable progress in that area.

Attempts to get agreement on the structure of the record were an uphill struggle, and when agreement was reached between the healthcare professionals, the advice from the growing speciality of medical informatics was that such a structure was impossible in an integrated medical record with messaging. This could be an area where input from the 23 academics could be of great value.

In spite of agreed advice¹ on training needs it would seem that training has fallen behind in the competition for funds. In particular there is shortage of funding needed to backfill staff absent for training in IM&T, and no provision for this has been built into any implementation plans.

When the time came to transfer electronically the patient record GP to GP all the issues of lack of structure and data quality emerged and are taking years to correct.

The decision to make choose and book a priority has been an error. It was not high on the professional agenda and was badly scoped. At one stage it had taken no note of the quite complex hospital processing of referrals using human intelligence not easily replicated electronically. The lack of thought about adding clinical information particularly drug information is a fatal flaw, as was the lack of thought about the actual processing of appointments in the primary care setting.

The National Library for Health (NLH) is a successful venture; NHS direct has good points and bad points.

From here

Although it may not be a politically correct question to ask, what do healthcare professionals really want?

  • GP to GP record transfer which would eliminate the delay in transmission of the paper record. That is without asking for a justification of the inordinate delays which occurred in the transfer of the paper record.
  • GP systems which as well as recording clinical information from many sources allow the recording and retrieval of information required as part of remuneration, and allow reception of results from laboratories and hospital discharge information to the patient record.
  • Hospital staff need specialised local systems for branches of surgery, anaesthesiology dermatology etc as well as a central hospital record. The local system often has to also meet the needs of the clinical secretary responsible for the correspondence from the unit.
  • All of these systems would benefit from an ability to transfer limited data without endless re-transcription. That core information will be demographic, but past evidence shows that huge amounts of time could be saved and patient safety improved by including a current drug list and allergies, and a list of problems agreed by patient and prior medical attendants. The full record is often of little value.
  • To allow immediate care of known problems e.g. out of hours GP cross cover of the ill patient, then some means of recording immediate information is needed applying only to the group of carers and social services. This has already been achieved in ERDIP pilot studies. Again, structure of that cover may change in scope, and perhaps this was the thinking behind the grand plan of a National record.
  • Social service input

Much of this will be most readily achieved by building systems which allow the transfer of information rather than creating a single huge repository. The transfer of information needs a vehicle and such exists and is in use. Health Language v7 (HL7) exists, having started out in laboratory use. It should now become the required messaging standard for the NHS.

Earlier versions are used to deliver Lab results in a safe and reproducible way to GP systems. Within the vehicle, in the seats as it were, needs to be appropriate and reproducible clinical information. Mention was made earlier of Read V3. Work done with the American College of Pathologist’s Systematic Nomenclature of Medicine (Snomed) has allowed the introduction of all of the Read V3 clinical terms work and the classifications (ICD10 and OPCS4) into Snomed to produce SnomedCT, and the NHS has a licence to use this tool, and well tried systems for updating the content. It will generate ICD10 and OPCS codes for HMR1 returns. It should be mandated for use with HL7. There is a related drug product, previously UKCPRS now known as dm+d (dictionary of medicines and devices), which links with Snomed and would allow an accurate reproducible and transferable drug and dosage list to be created. This must become the required standard for all clinical systems.

Local systems for Hospital, private sector and other use should be encouraged. There should be a proviso that any such system must be able to produce summary information for each episode of care in a format appropriate for messaging in HL7, to the main hospital record, and onwards to primary care systems. The present large annual financial turnover requirement for companies tendering to supply local clinical systems eliminates all but the largest suppliers. A sensible compromise should be reached to allow the small companies to thrive.

An opportunity to make data quality a clinical responsibility was lost in the new NHS contract negotiations. Quality data makes quality control easier.


If clinical information is to be safe, accurate and transferable it has to be structured, and staff training is needed. Speed of access and of data entry is critical. It should be unnecessary to state that the NHS number must be used by all providers of healthcare to identify patients.

Having core information available in a reliable form for electronic transfer into local records can save large amounts of clinical time.

The means of transfer could be by smart card or on-line messaging.

Such information will then be in a form which allows meaningful analysis for audit and epidemiological purposes.

Links to NLH could be built using this structure to allow quick decision support and to allow patients to find reliable information about their specific problem.


1 Learning to manage Health Information – a theme for Clinical Education BL ISBNo. 0 953 27190 8

Further reading

Audit Commission For your Information. A study of Information Management and Systems in the Acute Hospital London HMSO Publications 1996

Information for Health. An Information Strategy for the modern NHS 1998-2005 NHSE publication

The Reduction of Uncertainty J R Mason British Journal of Surgery 1998 85 115-116 Kaiser Permanente’s experience of implementing an electronic record: a qualitative study BMJ 2005 331 1313-1315


About the Author
John Mason is a retired General and Vascular surgeon with an interest in audit who implemented hospital wide local clinical systems from 1985. These were based on an existing GP system. As medical director in a Trust he was involved in the commissioning of a Cerner Pathology system, and is very aware of the problems of translating an American system to UK use. He was a member of the Royal College of Surgeons Audit Group, chairman of the Academy of Medical Royal Colleges Information Group, and member of the Chief Medical Officer’s working group on clinical information until 2000. He served on the American College of Pathologists Snomed editorial board until 2003. He is a member of the Worshipful Company of Information Technologists Medicine and Health Panel.

Return to Unpublished Concerns Regarding NPfIT

Your Ad Here
Personal tools