An informatics manifesto

edited August 2012 in General
I've created an Informatics Manifesto, of basic principles for NHS bodies to adhere to. It's written specifically with CCGs in mind, but could easily be adapted for other NHS institutions. Your comments would be appreciated.

Informatics is a tool, not an end.

Information handling within the CCG aims to be:
• minimally intrusive
• maximally cost-effective.

• WILL NOT ask for any data without a clear potential benefit to future clinical practice (or alternatively because of statutory requirements)
• WILL NOT ask practices to reveal confidential information about identifiable patients unless this falls within GMC guidelines.
• WILL NOT give people or organisations huge quantities of unnecessary data to analyse purely to look good, or to meet targets for ‘reflective practice’
• WILL NOT ask for year on year improvement once a sensible upper target has been reached. High quality will be acknowledged and further improvement will not be demanded. Beyond this level it will be enough to have plans ‘to continue at the present high standard’.
• WILL NOT judge individuals or organisations using metrics over which they have little or no control.

• WILL attempt to make every moment count which is spent on analysing data
• WILL squash any attempt to gather unnecessary, or unnecessarily detailed, data
• WILL look for honesty in the interpretation of data. It WILL NOT encourage ‘tick-box exercises’ over data analysis, nor ask for fanciful interpretations of data to disguise the fact that nothing practical can be done to improve future clinical outcomes. If improvement can’t be envisaged, users must say so, and will not be adversely marked nor criticised as a result.
• Targets and comparisons WILL ALWAYS be made on a truly like-for-like basis
• League tables WILL ONLY BE PUBLICLY PUBLISHED where like is truly being compared with like
• The CCG’s informatics’ handling of claims will operate a on a high-trust basis


  • On a practical point it would be helpful to number your points.

    I don’t agree with.

    “League tables WILL ONLY BE PUBLICLY PUBLISHED where like is truly being compared with like”

    And consider the proceeding linked point:

    “Targets and comparisons WILL ALWAYS be made on a truly like-for-like basis”

    Potentially too restrictive, although a valid principle.

    The route to improving quality of both data and the services they measure is through the publication of the data as-is warts-and-all followed by a meaningful debate about what the data tell us about both. I don’t underestimate how difficult it is to get such a debate with patients and the public or even many of the sadly functionally information illiterate who will be working in CCGs and the NHS CB ticking boxes.

    I go with the Government’s mantra “Publish not Polish”, but we have to educate the public and politicians, but recognise there are potentially vicious and virtuous circles here.

    BCS Health did some work aimed improving the chances of taking the virtuous route. Which I think is worth reading.

    I would also add to your manifesto

    1) CCG will ensure that their contracts with providers require them to provide the data necessary to properly support the ongoing process of care and evaluate the quality of the service provided in a timely manner with due regard to available standards.
    2) No person who believes the “Daily Mail” is a newspaper should be allowed to view or comment on any health related information.
  • Couldn’t agree more that “Informatics is a tool, not an end” but the rest of the “Informatics Manifesto” still worries me as the whole approach implies what seems to me to be the wrong "end"

    Especially the implication that it may be acceptable for the CCG to ask for any data which has “a clear potential benefit to future clinical practice (or alternatively because of statutory requirements)”

    Nearly 10 years ago a certain Ewan Davies wrote a paper entitled “Avoiding the ‘dataset mentality’” Informatics in Primary Care (2003) 11: 49-52 - (which I have only recently come across.) One of his main phrases is still essential.

    Whenever data for any purpose is being collected electronically one must always ask a) “Is collecting this piece of data essential to provide “Support at the ‘point of care’ i.e. helping clinicians to avoid human errors” (with data benefitting future clinical practice or for statutory requirements being collected invisibly) or b) is it that item of data being collected because someone somewhere believes that there is “a clear potential benefit to future clinical practice (or alternatively because of statutory requirements)”

    However worthy the cause (for CCGs or any other reason) our ability to provide good individual care is being increasingly compromised by our obsession with data-collection.

    And we will not make use of the potential for IT to reduce the chances of human error and to reduce the ever increasing data re-entry workload until we focus far more on whether each item of data collected will provide immediate individual “Patient Encounter Assistance” - with data for other worthy reasons being collected invisibly.

    I would particularly draw John’s attention to the workload/cost classification of electronic data collection as suggested in our Oct 2011 JRSM paper.

    Or put it another way I recently attended (with 60 other!) a 4 hour "Speed Awareness" course (which turned out to be much better than expected). The drop in road fatalities from 8,000/year in the 1930s to around 2000 last year was attributed to :

    This so well reflected my life-long mantra “Don’t just shout at the driver, try as well to see what can be done to improve the road.”

    So much of our NHS-IT effort is being devoted to Education (Shouting at Drivers) and Enforcement (Shouting at Drivers) and in hospital medicine far too little to prioritising Engineering (Improving the Road)

    In the light of this I would suggest adding to his bullet points:
    WILL NOT unthinkingly insist on data collection for secondary purposes until every attempt has not first been made too see if by standardising the data which is essential for individual patient care can be collected in a way that makes it suitable for secondary analysis as well.
    WILL NOT insist on a 100% sample for analysis when a random 1 in 100 sample would provide virtually identical information.
Sign In or Register to comment.