A quarter of a century in the NHS

This week marks the completion of my 25th year working for the NHS, so I am being indulgent and sharing a few thoughts on the journey I’ve been on. If my memory has failed, let me know as a comment to this article.

1995 – fresh into a hospital and faced with the implementation of ICD-10 as a big-bang change to our core diagnostic classification. I had no idea what ICD-10 was, never mind an understanding of its value (and to be honest I don’t think many people did realise how valuable it is). After very brief consultation I decided we would implement by “coding” in Read codes and deriving the ICD-10 at the point of classification from the terms. By 1998 I took the bold decision to use Clinical Terms V3, the “strategic” terminology. 

National progress today towards classifying from clinical terms and records – at best graded a D- verging on an E. 

Also 1995 – recovering a failed “Casemix” system implementation. This gave the foundation for a single computer record of the patient data we held with a common identifier, and each type of event in a common model. The prototype for a longitudinal care record.

Hospital network on 10Mb ethernet with an X25 gateway and Racal network for pathology using Kermit mailboxes.

1998 – planning for the millennium bug. Our Oracle CFIS Casemix system, written by Oracle isn’t Millennium proof. (now called Technical Debt.) I lost faith in Oracle as an application provider and started developing a replacement in SQL Server 7, initially using the beta version. We got agreement to make the system cover the three trusts in our locality so extended the longitudinal care record to include interfaces to community, mental health, diabetes, maternity, radiology and A&E across the patch. Looking back, the information governance was weak, but it enabled analytics off the back of operational data from the whole locality. I also took the opportunity of the establishment of the NHS Wide Clearing Service, precursor to SUS, to introduce fully automated overnight submission of Contract Minimum Data Sets (precursor to the CDS). National progress towards that – mixed at best. Around this time the Internet gateway to NHSnet came in (a glorious 64Kbs for the whole organisation).

2000 – Millennium day – sober and boring – no incidents. Success isn’t always something that’s shouted about.

2001 – take the health record system into the private sector, returning to the NHS in 2002 working with the Modernisation Agency to help clinicians reduce their waiting times by understanding waiting list data. I suspect we will be back here next year.

2005 – closure of the NHS MA so transferred to NHS Connecting for Health, picking up the closure of the NHS Information Authority Birmingham office and rebuilding a Data Dictionary service having lost the experts. Recruited a wonderful set of people to succeed them, most of whom are still with NHS Digital. First big project – 18 weeks and moving from the old clinically unsafe waiting time measurements to “referral to treatment”. Introduced the NHS to “Patient Pathways” drawing on the experiences in NHS MA.

2008 – The logical Care Record, developing the thinking for a true Longitudinal Care Record. It soon became clear that we would only get semantic interoperability across care providers, care professionals and care settings if we use SNOMED CT as a reference terminology. Hopefully the key national infrastructure to enable that – the terminology server – will go into private beta later this month. Sometimes strategic need takes time to come to implementation.

2009 – Introduction to the world of ISO standards in Edinburgh. Came away with a belief that the People’s Front for Judea in Life of Brian had more grip on reality. Now I’m the ISO convenor of the joint working group for safe, effective and secure health software/IT. Building international consensus is hard but so important. Also 2009 engage in the development of ISO 80001-1, led by Maureen Baker CBE, setting out the safety framework for IT in healthcare providers. The successor standard is now out for final ballot for approval.

2010 – Directing the Data Standards and Products team as the health and social care act came in. Brought quality management to the team leading to ISO 9001 certification. Without it our cost base and defect rates would be higher and our customer satisfaction lower. SNOMED CD and dm+d become fundamental standards for the English Health and Care system (though implementation is taking some time to complete in all settings). I also agreed to lead the development of ISO 13940 – the business concept model for healthcare (see https://contsys.org)

2011 – The birth of FHIR, Graham Grieve sets out his vision to a packed bedroom on the edge of an HL7 working group. Standing room only with people in the corridor. A radical new look at how healthcare IT is delivered. It’s not perfect and has naïve handling of SNOMED CT but the focus on making it work rather than covering all possible uses was a substantial leap forward.

2013 – Health and Social Care Act comes into play; NHS Connecting for Health becomes a part of the new Health and Social Care Information Centre (aka NHS Digital). I take on a portfolio of roles in Information Governance, Quality Management and continue international standards development.

2014 – Develop the statutory Code of Practice on Confidential Information, agreed by NHS England board, NHS Digital Board and Secretary of State. Its not perfect, but the advice is solid today and covers data handling across its (often very long) lifetime.

2015 – ISO 13940 is published – a remarkable foundation, providing an internationally agreed view of how healthcare is delivered.

2018 – the move to a formal Data Architecture role and able to tackle many of the strategic data architecture questions discussed leading to the start of the terminology server framework contract, the acceptance of the need for a corporate metadata catalogue and the re-platforming of the NHS Data Dictionary to provide the springboard for the broadening of its scope to describe all data used in health and care systems.

2020 – still much to do, but many of the things I achieved 20 years ago locally are now in train to be achieved nationally. Longitudinal Care Records are emerging, in some cases for specific conditions and in other cases joining across settings. The value has come front and centre with the excellent work to establish the shielding / vulnerable patient list. The use of SNOMED CT is now widespread and its use as the reference terminology for the Nation is now something we can deliver. My home 350Mbs internet connection puts the original 64Kbs hospital link to shame (and yes its more than I need).

Reflections

Over the last 20 years society has changed, and the data representations of society too. New problems have emerged with systems now confusing once certain concepts such as male / female. The replacement of ICD-9 threw out the now incredible classification of homosexuality as a mental disorder, yet our GP clinical systems retained that model for many years after 1995. We also we are much more accepting of diversity in society (I don’t think I would be criticised for promoting someone who was pregnant now as I was in the ‘90s) though we still have a journey to travel.

I said at the beginning this is indulgent. I am optimistic that we are now able to make possible nationally the things I aimed to do locally at the start of my career. The strategy in the 1990s was to have a “three legged stool” – common language, common identifiers and common network. We have these now in SNOMED CT, “New” NHS Number and all the Organisational Data Services products backed up by cloud-hosted infrastructure / services accessible by secure connections over the internet. Its easy to reflect on today’s challenges but looking back I’m proud of the progress our health informatics industry has made and my modest contribution to that.

Thank you to all of you who have joined me on that journey – let’s keep moving forward. We are making a difference!