Shared Electronic Health Records are Not Easy

A day or so an interesting report appeared in the web magazine ehiprimarycare.com revealing that the plans for the Shared Electronic Health Record (SEHR) for the NHS spine had been significantly scaled back. To quote the relevant section:

"Dr Gillian Braunold, one of CfH’s national clinical leads for general practice, said major diagnoses and major procedures will no longer be included in the initial upload to the spine, a change in policy from proposals issued by CfH in the autumn.

She told EHI Primary Care: “We feel it is sensible to limit the initial record to just prescriptions and allergies because we want to make sure each set of records is fully correct. The profession felt strongly that although globally a practice might feel its records were good it was important to look at each set of records before they went up.”

This outcome can only serve to remind us just how difficult it can be to come up with a useful SEHR and to ensure the accuracy and utility of the data held as well as address all the other issues that arise (privacy, security etc).

Australia’s effort in this regard, termed HealthConnect, presently has a rather ambiguous status. While it has not officially been killed off, it now seems highly likely that it will not be built as originally conceived. This conception was of a large centralised repository which held an initial patient record (with standardised content) and then a series of “event summaries” which were created each time an individual had an encounter with the health system. Over time the event summaries were to form a longitudinal health record which lasted the patient’s life.

The key barriers to success of this project would seem to have been:

1. The lack of clarity about just what the information contents of a summary records should be to meet the benefits objectives of the project.

2. The inability to clearly recognise the importance of, and the need to standardise, the client systems which were to be the sources of the information contents for the SEHR.

3. The lack of provider and client identifier approaches as well as the lack of an agreed clinical terminology. (the need for these is now recognised six years on – and they may be available in 2-3 years)

4. The lack of properly conducted, transparent, well funded trails of the various components required to reach the 2001/2 vision.

5. The difficulty in working through the privacy and legal issues, as well as the consent issues, associated with the whole concept.

6. The lack of a properly thought through and compelling business case for the overall project – leading to a lack of budgetary commitment for the whole enterprise.

7. The lack of recognition that the quality and sophistication of the client systems was crucial to the overall justification of the project.

8. An inability on the part of the sponsors of the project to develop political and public support for the implementation of HealthConnect.

It seems clear to me the UK has been sensible in scaling back what they are attempting and to approach the whole endeavour incrementally – hopefully learning as they go on. I am firmly of the belief “big bangs” are not the way to success in Health IT.

David