Contents:
- Introduction
- The Dunedin Meeting
- Review of the General Domain Model
- Country Identifier Standards
- The Requirements Document
- The Access Work Item
- Discussion
- Conclusion
Introduction
Nineteen delegates from five nations met in Dunedin to discuss the 'work items' which this group has been preparing for the International Standards Organisations Health Records Committee (ISO/TC215). WG1 (modelling) is one of four work groups which include WG2 'messaging', WG3 'content', WG4 'security', and more recently WG5 'Healthcards'. They are intended to develop their work areas separately, but finally to work together to achieve a standard for electronic health records. This process was planned to take about 2 years from the inaugural meeting in Orlando in August 1998.
The lead-up to this project is fully documented in the reports of other meetings on this site. After the Tokyo meeting (attended by David Menkes, representing New Zealand), we thought our task was in accordance with the resolutions of the WG1 meeting reported in the minutes. This was to model the Access process, in a way that would inform the deliberations of WG4 and hopefully lead to an implementable standard. It was agreed that this was to be a collaborative report with them, a modelling exercise to produce a product that they could then implement. There had been debate about implementation-orientated modelling versus policy-orientated modelling. The resolutions and minutes from WG1 and WG4 in Tokyo, plus David Menkes' verbal reports, suggested that we in New Zealand should follow a middle way.
We interpreted our role to be that we should not attempt to design an implementation of security and privacy processes, nor should we produce just general statements about privacy policies. Our 'middle way' should actually point to something implementable, and should provide a means through which diverse policies could be expressed. We thought we should seek international collaboration to achieve this.
We had thought that securing such an important assignment for New Zealand was an honour and an opportunity for this country. We had considered that we could act as a focus for international views in this area. Despite wide dissemination of invitations to participate from us both here and overseas, the silence that greeted our efforts was disconcerting. We decided to press ahead regardless. We decided to continue the project 'as if' we were preparing for a New Zealand implementation.
Our draft report on Access to the Electronic Health Record actually outlined an implementable global model which integrated the work of the Security Work Group Four on digital certificates and attribute certificates with a mechanism that although real, was culturally sensitive and endlessly flexible to different record architectures. It was well received by all the delegates except by one United States representative who was dismissive not only toward that item, but to the whole ISO process. The implications of this are discussed below.
The Dunedin Meeting
Of all the eight work items that there had originally been set up by Work Group One in Sydney (January 1999), only three remained 'on the table' at the Dunedin meeting. The others simply had not received attention from their proposers at earlier meetings, or had been withdrawn in Tokyo (Peter Schloeffel's 'requirements for the EHR' work item).
These three were:
- The General Domain Model
- The Country Identifier Standards
- The Access Work Item.
Peter also re-introduced his work item in Dunedin. Apart from this, our own Access work item was the only one which had been at all extensively worked up for this meeting, but the others are briefly reviewed below.
Review of the General Domain Model
This model had been the brainchild of Peter White, who had represented Australia at the WG 1 meetings before, but who had retired from the group. It had previously had extensive support from the US, particularly from Robert Mayes of the US healthcare funding agency (HCFA). It consisted of a set of universal categories which different countries could map their own national data collecting activities into. It was re-presented by Elizabeth Moss of Australia, who asked that the group endorse it, and identify 'candidate models' from their own jurisdictions.
In an extensive commentary, Brian Love from the UK put the General Domain Model in the wider context of attempts to find a global model, and discussed the US 'RIM' (Reference Information Model). This was 'no more or less than an attempt to comprehensively map components of information that is needed to populate information exchange'. It was 'an attempt at mapping the ground of information that needs to flow and be maintained'. He noted that the UK view of it was that the clinical concepts in it were inadequate, and noted the parallel developments in the USA of the PRA (Personal Record Architecture) and the USAM (Unified Service Action Model), both from the HL7 organisation from Duke University in the USA. He discussed the relation between these, and the relation again to European work. He suggested that there was a choice between trying for a model that would satisfy everybody (a kind of amalgamating model), or instead going for a 'preferred approach' which we might all try to join. He indicated that there were people in the UK taking a serious look at the USAM model in this light. He suggested that the separation of the Work Groups might have outlived its usefulness, and emphasised that the UK experience had been that a model for 'messaging' should not be separate from 'modelling' per se.
Gary Dickinson from the US also made some wide-ranging comments. He also suggested that the USAM model was 'close to the kernel of the problem'. He suggested that any model should be 'looking at the process, not just the information' He pointed out that the present RIM was based on HL7 v2 messages. He suggested that the process of medicine may not be reducible to conceptually simple terms, and agreed that the RIM has difficult problems, in particular because of heterogeneity in systems, policies, procedures, and national approaches.
Peter Williams from Australia (the Chairman) mentioned the concept of a 'model of models', and implied that the 'General Domain Model' tried to be this. However, Peter Waegemann, head of the US delegation, was dismissive of the enterprise, pointing out that its chief supporters, Peter White of Australia, and Robert Mayes of the US healthcare funding agency were not at the WG1 meeting, and were 'not coming back'. He suggested that the concept should be dropped by default. He made negative comments about the utility of WG1 altogether, and suggested that ISO were thinking of disbanding it. These remarks set a very negative tone for the meeting.
Country Identifier Standards
This was the next item to be discussed. The Canadian team, who were not actually represented at the meeting, submitted a detailed paper. Some members doubted the utility of having this work item at all, since it appeared to duplicate pre-existing standards in this area, including existing ISO standards and telecommunication standards. The paper submitted by the Canadians was referred back to them for further work. Originally, this work item had been seen as a way of starting an ISO standard, since it would be non-controversial, and some real progress might be made with it. However, the whole point of having a separate ISO/TC215 Country Identifier standard remained obscure to many delegates, and with its rejection by the meeting in its present form, not even this non-controversial area achieved agreement.
The Requirements Document
Peter Schloefel from Australia again introduced this work item, which had been withdrawn in Tokyo for logistical reasons. Real efforts had been made to enlarge the concept to represent all views on this matter. However, there were negative responses from the US contingent, and on the final day of the meeting, Peter gave this work item as an example of one which could be worked up in time for Vancouver. The Americans comment was that it would still look like GEHR (the Good Electronic Health Record Model). It appeared that the US HOD would oppose anything that was associated with this approach. However, Gary Dickinson of the US was more supportive of it, and indeed Dr Schloefel has subsequently presented papers in the US, and I understand there are ongoing contacts and developments.
The Access Work Item
We had half of one day of the meeting to present this. Our presentation was introduced by David Menkes and incorporated input from myself (Michael Mair) on our access model, Neil Price on the review of existing National Standards, and a presentation by a Maori group from Otago who addressed the Maori perspective on Access to Health Records. Our report is to be found on http://www.health.nsw.gov.au/iasd/imcs/iso-215 and the user name 'wg1' and the password 'berlin' will gain access to the 'areas under development section' of this WG1 site where the paper is located. A Power Point version is available here.
Each part was introduced by its author, and the Maori delegates emphasised the reality of their cultural concepts and even sang to the group. Their presentation was warmly appreciated by all, and confirmed (if such confirmation was needed) that there are genuine differences in cultural attitudes which will affect any global Access regulating standard to the EHR. This was amply confirmed subsequently by discussions with the Japanese and Korean delegates, who were very supportive of our presentation, and Peter Williams and Elizabeth Moss from Australia also emphasised the relevance of this concept for indigenous minorities in Australia. The UK delegate Brian Love also spoke positively of our presentation, and suggested that although it did not suggest a record architecture as such, it had 'architectural features'.
Yasu Hirose from Japan made a brief presentation of his work with the Access Control Matrix concept at the end of our presentation, and his wonderful diagrams well complemented our ideas. There was some subsequent discussion that his paper, of which the ACM diagram was part, would be posted on the WG1 site, but unfortunately this has not happened. The two approaches appeared to complement each other well.
The US HOD was as negative about the Access presentation as he had been about the General Domain Model, suggesting that we had misunderstood our task, which had been to develop a model of 'policies', not a model which broached on implementable models. Documented evidence was produced at the meeting to counter this claim. Again he suggested that WG1 as such might well be disbanded by the ISO secretariat, and was derogatory about 'coming all this way across thousand of mile of ocean' for the meeting.
The Access work item was accepted by acclamation by delegates after the US delegation left, and supportive resolutions were drafted. It was suggested that our model be developed further, and be presented at the forthcoming meeting in Vancouver, at a joint session of WG1 and WG4.
Discussion
The full plenary session of the committee is shortly to be held in Vancouver (16-23 June 2000). It is necessary to be sensitive to the immediate conditions of the argument as it has evolved over the last 2 years of the ISO/TC215 process. In Dunedin we were subject to a form of extemporisation on behalf of corporate interests that it is hard to confront. Rather than seek confrontation with such interests, we should adopt a dual strategy of trying to work with and around them.
Functionaries from the European Standards organisation (CEN), the US HL7 organisation, and paid employees of government agencies such as the UK NHS or the Australian Standards organisation are the main attendees at these meetings. Most delegates therefore have positions to defend or causes to espouse, and hardly any have the practice of medicine as their principle occupation. Many of the organisations may have a weak or even negative commitment to the ISO process. Doubt has even been expressed in some quarters about the sincerity of some participants.
The minutes of the meeting, which are available at http://www.health.nsw.gov.au/iasd/imcs/iso-215/meetings/modelling-nz.html, omit the crucial statement, which was that the group 'accepted the reality of cultural differences in practices and behaviours in healthcare'. This form of words, which was developed with some care in discussion between the delegations, summarises what may be the enduring legacy of the Dunedin meeting. It signs the introduction of cultural relativity as a significant determinant of any global access standard. By the end of the meeting, a feeling of identity of interest and trust between members was established at least among the Asia/Pacific group, and this may prove important as the committee continues its work.
Conclusion
The Dunedin meeting had some very positive and some very negative aspects. On the positive side was the consensus demonstrated between Asia/Pacific members, both on the Access work item and a feeling that a community of trust and common interest has been established in this group. There was strong support for the notion that New Zealand's Access work item should be presented again in Vancouver. On the 'down' side was the attitude of one member of the US delegation, who appeared to be keen to close down WG1 of ISO/TC215 and was dismissive of the work item we developed. Also disheartening is the paucity of serious contributions to the WG1 effort, and a growing realisation that by relying on volunteers in this complex and difficult area, ISO may not be in a position to facilitate the standard that it is committed to developing. The New Zealand delegation will now attend the Vancouver meeting, and a further report will be forthcoming after that.
Mike Mair
(for New Zealand delegation)