The concept of the trial is as follows:
Healthcare providers (hospitals, GPs, diagnostic providers and community services etc) in two specific areas treating two specific categories of patients (the 65 + in the Hunter and the young children in Westmead) have an encounter record created each time they attend a participating facility.
The idea is that over time a longitudinal health record on each individual will be built up.’
On each encounter both demographic and clinical information will be captured.
The encounter from the GP, for example, will capture the following information automatically at the end of a consultation. Quoting directly from an Healthelink information sheet.
“The following fields will be extracted from GP electronic health record systems following completion of each consultation:
a Patient demographic data
- Surname, Given Names
- Date of Birth
- Medicare Number
- Contact details (e.g. address, phone numbers)
b Consultation data
- Provider/clinician – name, role and location
- Diagnosis/reason for visit
- Pathology/radiology test results (when a GP associates a report with a patient record)
- Medications - Prescription Order Information (Script number, GP name and ID), Prescribed Medication (Name, Route and Instructions) and Medication Administration Information
GPs will be encouraged to enter the data into the coded fields (e.g. reason for visit, medications) in their practice clinical systems as only these data will be sent to Healthelink. Free text information entered into the clinical notes for example will not be sent to Healthelink. The coded field only approach means that GPs can be confident that a sensitive free text comment entered into the clinical notes will not be sent to Healthelink and will not be available for other providers or the patient to view.”
With this background a few comments and discoveries made while researching the progress of the project are worth sharing.
First, the exclusion of the text notes and the expectation of using the coded information in the diagnostic fields is quite problematic. Experience from HealthConnect trials in the NT showed coded data was often not provided and this often led to the situation where it was impossible to understand just what a particular encounter was about. Healthelink looks likely to have similar data-quality issues.
Second, a review of the patient searching capabilities provided to users of the system confirm the worst fears of those concerned with privacy protection. A search on a partial name and approximate age (+/- 5 years) generates a weighted (for match) listing which contains a pretty full demographic record with full name, address etc. Just the sort of thing abused wives etc want to keep to themselves! The system should insist on a much higher level of initial knowledge – not that even this would help in the case of an abusive spouse.
Third, the system has no apparent concept of location etc – so if you are on the data-base and have not opted out you are findable. Extension of such a system state wide is clearly very problematic.
The record displayed by the system once the patient is located is also not segmented into more or less confidential information which means the patient has to make a choice to opt totally in or out on the basis of the most sensitive piece of information in the record as far as they are concerned.
All these issues are in addition to the lack of initial consent to having information stored as would be the case had ‘opt in’ been selected. For reasons best known to itself the NSW Department has decided to ignore not only its own earlier reports on the matter but also its own explicit Health Privacy Principle.
“HPP15 Linkage of Health Records
“Personal health information must not be included in a system that links health records in one service with health records in another health service unless the individual it relates to has expressly consented”.
A specific regulatory exemption from HPP15 has been provided for the Healthelink pilot. Whether this exemption should be retained after the end of the pilot will form part of the pilot evaluation.
The regulation provides for the use of an “opt out” consent model. This ensures health consumers will continue to be able to exercise the choice as to whether or not they participate in Healthelink and to exercise this choice at any time.”
I am aware that at least some physicians are by no means convinced this NSW regulation does not violate the Commonwealth Privacy Act.
Leaving the whole privacy issue aside there are a few other points to be made and question to be asked.
1. Just why has all this taken so long?
Taken from the March 2006 Project Outline:
"In March 2000, the NSW Health Council set an objective for the NSW health system to implement the Electronic Health Record (to be known as Healthelink) by 2010. Health Ministers also endorsed the proposal for a National Health Information Network, known as HealthConnect. The NSW Healthelink pilots were endorsed as one of the trials for the National HealthConnect program.
In May 2004 the Health Minister announced that the Care Junction Consortium consisting of LogicaCMG, Orion International and Healthlink had been selected to work with NSW Health to implement NSW electronic health record pilots – expected to commence operation in the first quarter of 2006.”
We now know what was then EHR*NET was the original proposal and the rename came a good deal later. We also must wonder why the whole thing should take six years from concept to pilot. At this pace 2010 does not look all that good for a total state roll out. (Note the 2002 implementation plan of EHR*NET indicated the project would be complete by June 2007 – time to really pick up pace!)
2. In a very recent (25/09/2006) GP information document we find the following:
“How would the Practice apply for the GP Testing Grant?
The Healthelink system is complex and although the components of the system will be thoroughly tested centrally, assistance will be required with testing the system once installed at practices and as used by GPs. Healthelink is therefore requesting that practices offer to assist the testing process and become early adopters.
A grant of $10,000 will be available to these practices subject to signing up to participate, meeting the eligibility criteria and being in a position to install the Healthelink software by 30 November 2006 or a date otherwise agreed with Healthelink. More details are available in the GP Testing Grant Agreement. “
I wonder how the announcement of this largesse was missed by the press and when the decision was made to be so generous!?
It seems to me this is a sign of desperation on the part of the project to have practitioners concerned about liability and indemnity to get moving.
3. It seems to me there will some real issues involved in the scalability of this system in a technical and user sense. Having a search operate across million people data-bases may turn out to be a real challenge.
4. There is also a real issue about just what impact the system as presently conceived will have on practice work flow and what value the practitioner will receive. The lack of proper integration between Healthelink and the practice management system may prove to be very difficult to use in practice - switching between one system and another. To have the system used properly it needs to be quick to access and consistently provide useful and relatively complete information.
5. The project's approach of making records unavailable for 30 days clearly makes very little sense and needs to be reviewed in the context of an overall privacy approach review.
I am afraid this project has all the signs of a project that is on the ropes. I hope not since it has taken so long to get this far. In many ways I am not surprised. As those who have read this blog for a while will know I have always been convinced that the complexities and difficulties associated with shared summary electronic health records have been very underestimated.
Maybe a strategic re-think and revision of the privacy approach can grab victory from the jaws of defeat. I hope so.