Thursday, October 02, 2014

Article Draft: The Standards Setting Processes In Australian E-Health Are Broken and No One Seems To Care.

Before discussing the present issues in Australian E-Health setting a little background in two areas seems justified. We can consider these issues by responding to two questions.
1. Why do Standards matter?
Standards are important in all industries to ensure the quality, safety and predictability of all goods and services. We have food Standards to help prevent food poisoning and building Standards to help ensure buildings are safe, fit for purpose and proof to reasonable environmental impacts (earthquake, bushfire etc.). Another vital function of standards is to permit safe interchangeability and interoperability, think of things like screw threads etc.
E-Health has, as its primary concern, the safe and secure transmission of clinical information between healthcare providers and between providers and their patients. To this end Standards have been developed covering such things as electronic health records, secure health communication, health data and information Standards and so on.
You can see the scope of the various Standards that have been developed for Australia (in association with the IT-14 Committee of Standards Australia) on this web-page:
 Interestingly the Australian Institute of Health and Welfare (AIHW - www.aihw.gov.au) also has a role in developing a range of data definitions and the National E-Health Transition Authority (NEHTA) has also had a role in clinical terminologies as well as attempting to manage / interfere with (depending on your perspective) the Standards development process over the last few years.
Before discussion processes it is also important to point out that Standards adoption by commercial and government entities can minimise risk and reduce the costs of bringing new services and products to market while also assisting in broadening the capabilities that may be offered. It is quite usual for companies to find that there are distinct advantages in Standards adoption and use - not the least in facilitating future-proofing development efforts. Of course, for some Standards adoption is compulsory and enforced by laws (e.g. food Standards)
2. What is the present process in Australia for their setting?
In ordinary times - and the last few years in e-Health can hardly be seen as ordinary - there was a stable and well thought out, consultative and inclusive process for e-Health Standards development. Work began seriously on the Health Informatics domain around 20 years ago with Standards Australia creating a steering committee (to become termed IT-14) to develop and implement a work plan. Key to this plan was to leverage all available international work (especially EU and USA) and to only develop new Standards if international approaches could not be reasonably adopted. The intent was to maximise the level of international engagement of the local development committees.
IT-14 has a range of sub-committees and a wide range of members. This page provides a useful collection of links:
Once a work plan is developed for a period the chairs of the various sub-committees - facilitated by Standards Australia staff convene meetings of volunteer experts representing a full range of stakeholders to develop a draft Standards document. The draft is then workshopped as needed and a balanced consensus document is developed. Only once this is fully agreed is an Australian Standard, Technical Report or Technical Standard published.
The keys to the process are seeking conscious involvement of as many interested parties and technical experts as possible and a consensual open way of working and document development.
3. What has changed and what has gone wrong?
By my reckoning the system of e-Health Standards setting was working pretty well until around 2010. Most needed Standards were being developed with reasonable speed and quality with the help of a large and expert volunteer workforce and reasonable support from Standards Australia. The Commonwealth Department of Health was funding access to the relevant Standards for no charge and providing funds to ensure Australian experts were able to attend relevant major international meetings of bodies like HL7 and the ISO-TC-215 Committee so Australia was well linked into international trends and contributing where appropriate.
In 2010 the then Government announced it was going to implement a National Personally Controlled Electronic Record (PCEHR) and that it was going to be operational 1 July, 2012. In an attempt to have all the needed Standards in place well before the deadline (so they could be actually implemented)  many of the usual processes were abandoned with NEHTA led ‘Tiger Teams’ developing specifications which were then to be signed off by the volunteers. Not surprisingly there was considerable dissension and friction with many of the volunteers simply walking away and being replaced by NEHTA Staff. Additionally there was quite substantial funding provided by Government in this period to shape what activities were ramped up and what were deferred.
Also there were a range of quite fundamental technical directional choices where the was considerable disagreement between NEHTA staff and many of the volunteers and many of the volunteers felt they were being ignored and so became rather disillusioned.
To make matters worse NEHTA continued to pressure the acceptance of untested and unimplemented specifications which rather went against the usual standardisation approach of incremental development, testing and proof of concept and then implementation. Only when successful would further development take place.
In more recent times we have seen an accelerating loss of volunteers, failure to develop, agree and fund the 2014-2016 Work  Program for IT-14. As of the time of writing (Oct 2014) there has been no update of the Standards Australia e-Health site since May 2014 and most funding for meeting travel has been curtailed or cancelled.
Also as of the time of writing the fate and plans for the PCEHR are still in limbo with no outcome after a consultation period which closed late August 2014.
It would be fair to say, I believe, that for the present the Australian e-Health setting system is essentially ‘down for the count’. Whether it will ever recover in any useful form would have to be a completely open question. Essentially a working an valuable system has been damaged out of recognition by Government / NEHTA blundering, impatience and incompetence. Sad that!

-----
Comments welcome!

David.

10 comments:

Anonymous said...

I have been commenting on this since 2008- am volunteered out now but all subs are on APF web pages or in verbal submissions in Hansard. I'm a bit tired of repeating myself these days, its dull.
Juanita

Eric Browne said...

David,

I agree with the title of your article - the "Standards Setting Processes in Australian E-Health Are Broken and No One Seems To Care".
Except perhaps a very tiny few, including you, care.
I don't agree with your proposition that the system was working pretty well until around 2010. Perhaps it is "down for the count", but I don't think it has ever worked particularly well, especially for those standards related to exchange of health and healthcare information. For those of your readers who would like to understand some of the major factors in the world context, Thomas Beale ( who obviously also cares ) has contributed superbly on his blog with his series of postings "the crisis in e-health standards.

To add to Thomas' commentary but specifically in the context of the e-health standards produced through the Standards Australia process and similar models such as the current Pathology Information, Terminology and Units Standardisation (PITUS) Project, I have observed the following generalisations:-

The "standards" are not sufficient to meet the expectations engendered or the levels of interoperability claimed if these standards were to be adopted.

There has been very little gathering of requirements or documentation thereof prior to the committees or working groups meeting and drafting - the standards bodies aren't equipped to gather requirements.

The processes aren't open, but rely on the limited knowledge and experiences of the invited "experts" on the working groups.

There has been virtually no use made of collaborative development tools such as subversion or github, let alone asynchronous social networking or wikis.

There has been almost no engineering input into the standards artefacts or understanding of the deployment contexts, reflected in the lack of direct implementability of most of the standards and the lack of support for change.

There has been no ability and little interest to ascertain how widely the standards have been used, nor to what level of quality.

The standards bodies have acted more as clubs operating in isolation, rather than part of a larger e-health governance process. This may not be the fault of the clubs themselves, but I don't think they have done much to help. In other words, the clubs have established themselves and produced things, but what they have produced is not necessarily what is wanted, nor needed, by the community more broadly.


But the largest problem I see with the "standards" produced by these clubs to date has been the lack of implementation testing. Anecdotal reports by members of the clubs does not constitute testing. Nor are one day artificial connectathons. It is in the area of real implementation testing that a standard is to be judged. Even then it is of little value until there is widespread adoption and benefit.

On that score, it wouldn't take much for standards to be knocked to the canvas.

Perhaps, in conclusion, these are some of the reasons no one seems to care.

Dr David More MB PhD FACHI said...

Many thanks Eric a great and thoughtful set of comments. My perspective aligns with yours and Tom's among others.

I guess my central point is that things have got worse with Government / NEHTA interference, money and pressure - and so my suggestion was rather a relative one.

Will have a close look to see what I can bake in to the final.

Thanks again for your effort!

David.

Grahame Grieve said...

I just wanted to respond to some of Eric's comments

"There has been very little gathering of requirements or documentation thereof prior to the committees or working groups meeting and drafting - the standards bodies aren't equipped to gather requirements."

no. but who is? Who will respond? The usual pattern is to refuse to engage, then abuse the provider for not gathering requirements. This is as true for a vendor doing an implementation project as it is for standards. And the solution to this problem in all the Projects-for-Dummies books is to find a champion on the user side. So where's the champion? I don't think there can be one, and as long as there can't, then the standards process is broken, but there's nothing better.

"The processes aren't open, but rely on the limited knowledge and experiences of the invited "experts" on the working groups. There has been virtually no use made of collaborative development tools such as subversion or github, let alone asynchronous social networking or wikis."

The committees aren't closed - the technical committees are always open in my experience - but it's true that the open collaborative tools haven't generally been exercised. Maybe I'll do a blog about the perils of these collaborative tools in standards (I have probably as much experience as anyone in health standards now), but it's definitely true that Standard's Australia's extremely conservative approach to committee processes is a net harm to the outcome.

"There has been no ability and little interest to ascertain how widely the standards have been used, nor to what level of quality."

That's wholly untrue. What's true is that there's no interest on the part of the implementers in providing that information. I'm trying really hard to get this information right now, and even providers are actively prohibited from supplying it to us (the people who engage with standards are not authorised to say anything that might be construed in a negative fashion). And people wonder why they get standards that don't match their requirements... (right now, we - FHIR - are trying to set up a trusted private reporting system so providers and vendors can report to the standards process anonymously, but we don't know if this will work)

"The standards bodies have acted more as clubs operating in isolation, rather than part of a larger e-health governance process"

what e-health governance process?

"It is in the area of real implementation testing that a standard is to be judged. Even then it is of little value until there is widespread adoption and benefit."

if Eric has any suggestions, then I'm all ears. It's too easy to criticise, it's very hard to do something better.

"Thomas Beale ( who obviously also cares ) has contributed superbly on his blog with his series of postings "the crisis in e-health standards."

That blog post - and my discussions with Tom around that - were part of the blueprint for FHIR. We haven't done everything Tom says, but we've tried to do most of it. Time will tell whether FHIR will really be different.

Eric Browne said...

Grahame wrote "if Eric has any suggestions, then I'm all ears. It's too easy to criticise, it's very hard to do something better."

Well I think you've shown with FHIR just how vastly better the processes can be, in almost every respect. My criticisms certainly weren't fired at FHIR!

I think there is reasonable consensus that Australian e-health standards processes should focus on trying to improve the sharing and use of medication and diagnostic test result information in particular. If we had any decent e-health governance at a national level, then these would be given priority and appropriate resourcing.

For medications, the Australian Medicines Terminology goes some of the way to help. It is rather heavyweight for many systems to implement and doesn't provide adequate links to other knowledge bases, such as real product identifiers, substance knowledge for allergy checking, approved Product Information, Consumer Medication Information and whatever key knowledge is needed by clinicians to prescribe safely and effectively, such as that of the Australian Medicines Handbook. But at least the AMT is something upon which to build.

What we don't have is a strategy for how it can best be maintained, deployed, improved, quality assessed, funded. Neither do we have a plan for how it can best be maintained, deployed, improved, quality assessed, funded. The AMT standard is of little use until the fundamental strategy, policy, planning framework is in place. So rather than suffer NEHTA continually touting that it has ticked off clinical terminologies and the AMT in particular on its work program, we need an organisation honest and willing to admit it is operating in a governance-free zone and struggling to make headway with standardising the names and identifiers of medicines in use throughout clinical systems in Australia. My suggestion is to establish a national e-health governance framework, and replace NEHTA by an organisation willing and capable to embrace the sort of open processes that you have introduced with FHIR.

Now for pathology and diagnostic imaging. We have a standard that has been developed by Standards Australia for messaging of test orders and results between healthcare providers and diagnostic laboratories. It is one of the few working examples of "standards-based" sharing of clinical information in Australia. For some uses, particularly results reporting from a single lab to hospitals, it seems to work pretty well once the parties have invested sufficient time and resources. However, the standard and its inadequate deployment struggle to meet all the needs of clinicians and don't even address the needs of patients' electronic access to their results. There are a complex set of interdependencies at work which would take too long to discuss here.

For NEHTA, pathology has been a strategy and policy-free zone for many, many years. In that vacuum, the RCPA have initiated a new standardisation process centred around terminologies for test names, units, and organisms as well as information structures for synoptic reporting of anatomical pathology - for cancer staging etc. I applaud them for their initiative. I have struggled for the last 15 years to understand why these things can't be standardised across reporting. The RCPA's PITUS process is closed in that all decisions are taken by a club whose meetings are not open, whose rationale for decisions are not aired, let alone discussed in public, and whose published artefacts, consequently lack the quality needed for implementation. The club claims endorsement from NEHTA, yet NEHTA makes no such claims itself about the role of the PITUS developed terminology.
[ to be continued, hopefully ]

Eric Browne said...

[.. continued]
Some of my suggestions for diagnostic test standardisation would be :-

1. Establish an e-health governance framework and replace NEHTA by an organisation willing and capable to develop and implement strategy for improving diagnostic ordering and result reporting.
2. Have that organisation develop a strategy and plan that has a range of goals to meet the priority needs of clinicians, patients, researchers
3. Open the PITUS processes using online infrastructure - they don't need to be as sophisticated as FHIR, but at least open and accessible.
4. Establish similar infrastructure for the diagnostic imaging community,
5. Review the PITUS decisions around the use of SNOMED and LOINC terminologies, documenting the pros and cons leading to each decision.
6. Reconcile the ordering and reporting test names with those used in both the RCPA manual and Lab Tests Online.
7. Establish environments for efficiently testing the standards
8. Establish and document processes for efficient distribution of quality, implementable standards and reference sets - this is not the same as throwing a bunch of Excel spreadsheets over the fence and hoping for the best - we need to move beyond a cottage industry approach.
9. Establish and document how the standards or terminology sets will evolve over time.

Grahame Grieve said...

hi Eric

Building on my FHIR experience....

"Establish an e-health governance framework and replace NEHTA by an organisation willing and capable to develop and implement strategy for improving diagnostic ordering and result reporting. "

Forget about it. No organisation can do this. It has to be a palace revolution. Some single person with enough credibility and determination has to stick their neck out and do it - they have to take the risk themselves. And then they have to be able to bring the pathology services along with them. As soon as you put an organization in place, then all the need for "strategy" and "risk management" came into play, and it's all over.

but who's willing and able to take the risk...?

which brings me to:

"Open the PITUS processes using online infrastructure - they don't need to be as sophisticated as FHIR, but at least open and accessible."

There you go. *Michael* can, and he has my utmost respect. As for PITUS bring not open enough, well, you work with what you can, and I think Michael's doing a great job of bringing the people who count with him. I don't think that the FHIR approach could bring that community along (it's at a lower level, and a more diffuse and varied community, where as this one is tighter and has a different task)


Eric, you're a very capable guy. But what are you going to do move things along? We're at an impasse - the existing structures are moribund, and they're moving deck chairs around the titanic while a tsunami of icebergs loom (to mix my metaphors badly). At the same time, the harbinger of change - the social media support for open projects - is now here. But it needs individuals to take charge - and it involves huge personal investment, and real risks.

Andrew McIntyre said...

I think standards need to be developed by the people with real world experience at implementation who actually care that the resulting standard actually works in a practical sense.

While the quality remains patchy you would have to point to the Diagnostics committee (IT-14-6-5) as a resounding success story as we have widespread electronic delivery of pathology and radiology results. That committee was filled with people from path labs, especially in its early days, and workable standards that were implemented resulted.

What has been lacking has been any governance to ensure vendors adhere to those standards and interoperability has suffered as a result, but its still working.

Many of the other committees had a frustrating lack of implementation experience and the standards have not been as good as a result. I firmly believe that you can only have decent standards when the implementers are in the room developing it. Much of the standards issues are related to having government funded participants who push their own standards without the implementation experience to know that they are far off the track in the weeds. With respect I think Eric, in his early days at HealthConnect/NEHTA V1 lacked the health IT experience and fell into this category. This is not the case now.

Governments need to encourage vendors to send technical people to standards meetings and contribute and if they actually enforced compliance this would happen. Vendors have learned that they can ignore standards with impunity as the powers that be are clueless when it comes to understanding technical details and they can bluff them that they are compliant when they are not.

FHIR is still in evolution and may well succeed but in the meantime we need to take the standards that are working ie V2 standards and actually try and improve safety and interoperability by insisting on compliance. What we don't need is another NEHTA to "do it". We also can't jump to the latest "next big thing" but should consolidate what we have as its actually very capable when done well. To be done well vendors need to invest and compliance testing is the only thing that will make that happen.

Eric Browne said...

Andrew wrote:
"I think standards need to be developed by the people with real world experience at implementation who actually care that the resulting standard actually works in a practical sense."

I couldn't agree more that such experience is vital, but even more so, is a necessity to thoroughly test the standard through the use of real world data, either in real systems or a reference implementation.

The US has spent more than $20billion on meaningful use based on the HL7 Consolidated CCD. Thousands of conformance statements. The official sample probably passes validation through the NIST validator, yet they can't even get blood pressure correct. Probably hundreds of people, if not thousands with "real world implementation experience" have stared at the XML, yet no one seems to notice that diastolic BP is missing from the coded entries. Clinical safety issues abound.

Andrew McIntyre said...

Medical-Objects started as an offshoot of a day surgery so everything is developed in a clinical setting and put to use as soon as its ready. It takes a lot longer to be ready as clinicians have to tolerate it, clinicians other than me! We have 98+% electronic delivery of results now and have for some years. Incremental development with real world testing is more demanding and slower because the feedback is not always positive. However you don't spend a Billion dollars and get nothing, which is what can happen when you don't get feedback and ignore criticism.