Wednesday, October 26, 2016

The Push To Exploiting Health Information Seems To Be Building A Head Of Steam. It Is Not All Good News.

This appeared last week:

Top scientist wants to bring big data analysis to healthcare

A day after winning the Prime Minister's Innovation Prize, Mike Aitken, was busy pressuring federal and state governments, health funds, hospitals and doctors to begin a mammoth big data exercise in healthcare.
Aitken, who is chief executive of the Capital Markets Co-operative Research Centre, says there is an opportunity to save up to $20 billion a year from leveraging the power of personal health data.
His urgent call for action is timed to coincide with the publication of the first of three reports by CMCRC on the multiple silos of data in the health system.
The first report is called Flying Blind – Australian consumers and digital health.
Aitken says it "chronicles the opportunity loss to Australia from a failure to integrate more than 50 sources of health data in Australia".
The report was sent to federal and state health ministers this week.
It concludes that Australia has high-quality healthcare data in digital form. This could support real-time personalised healthcare.

Too fragmented

But the fragmentation of the data from multiple sources means it is almost impossible to prepare consistent healthcare consumer identifiers.
Instead we have multiple health information coding systems and multiple points in the chain where consumers hand over the same data given to someone else in the system.
A symbol of the dysfunction is the sight of people carrying hard copies of medical documents and radiology images.
The report says administrative data collections contain incomplete filtered and filleted data that limit their usefulness for planning and managing the health system. The same weaknesses work against the efficient monitoring of the quality of health services.
Consumers are the biggest losers from the failure to centralise and share personal data.
Lots more here:
I thought it would be useful to have a look at the report.
Here is the link:
Here is the Executive Summary:

EXECUTIVE SUMMARY

The Australian health system records sufficient high quality data in digital form to support consistent and targeted, real-time, personalised healthcare for each Australian. While, at a whole- of-population level, Australians enjoy high quality healthcare, our study shows that the acute level of data fragmentation creates an environment in which individual consumers, their next of kin and their service providers are flying partially or completely blind. The problem then extends to those charged with policy, resourcing, management and funding decisions in the public, private and non-government organisation (NGO) sectors, as well as health and medical researchers whose work is vital to the future health of consumers and the financial health of the country.
The fragmentation is the result of a number of key factors. Health services are delivered across a myriad of primary, secondary, hospital and allied healthcare settings, by a combination of private, public and  NGO providers,  many  of which  are  regarded as independent businesses. These are, in turn, funded and paid in many different ways and through different channels by consumers, state, territory and federal governments, and a wide range of insurers and related schemes. That each key player may use one  or more of a diversity of consumer identifiers and, in some cases, different health information coding schemes for their records, adds a lost in translation twist that further exacerbates effective use of data, even where this is brought together. Of equal significance are the multiplicity of state and federal laws and regulations that deal directly or indirectly with health data governance. While undoubtedly well intentioned, many present, or are interpreted as presenting,  significant,  and,  in some cases, insuperable  obstacles to rational combinations of health data. When viewing health as a market we have a clear case of market failure, a point   that is made directly and indirectly in recent submissions to the Productivity Commission’s inquiry into data availability and use.
The current situation does not result from a lack of goodwill or investment. Governments, universities and research bodies and service provider collectives (for example, doctors, hospitals, insurers) have made significant investments in improving health data collection and linkage over the past two decades.
However, these efforts have been largely focused around assembling fit-for-certain-purposes datasets that are generally de-identified, geographically bounded and do not contain complete health data even for the populations they are  seeking to study. Indeed, much cost and effort is often incurred in attempting to detour around the many barriers and data weaknesses that are described in more detail in our report.
In relation to consumer-centric health data initiatives, most effort has been directed towards what is now termed My Health Record. This initiative has been underway in various guises for almost two decades, but has yet to gain significant traction amongst consumers and health service providers.
Commentators have noted that tapping into complete, current and ‘fast flowing’ datasets of healthcare providers is an alternative model that delivers more comprehensive and ‘real-time’ benefits to consumers and their carers as well as providing the richest possible environment to support system planning and management, and  research.
We are effectively in the fifth decade of the digital era. Today, the power of data-and-evidence-driven product and service delivery   is taken for granted across many aspects of our lives. It is therefore ironic, that the most intrinsic and important aspect for each of us, namely our healthcare, remains such a significant outlier.
The perceived wisdom that continues to mitigate against joining up personal data to improve health is today significantly outof step with consumers increasing demand for instantaneous, personalised service delivery and their understanding of the social-contract necessary to achieve this. Generally, consumers have confidence that secure information processing regimes can deliver the benefits while mitigating the risks.
It is important to note that a range of other countries have tackled this situation boldly and have environments that leverage the power of health data for both preventive and curative healthcare.
This report is intended as an urgent call to action for the nation. While the notion of data may well be unexciting,    the power of complete, quality data to dramatically improve the fortunes of all health market stakeholders, particularly consumers, is assured. The knock-on benefits to burgeoning health budgets is an important by-product.
The time for piecemeal data initiatives is long since passed.   The nation must embrace an initiative that is truly transformative. This requires that we start from the perspective that data recorded at the point-of-service is the richest, most accurate and most current resource we have and that we should seek to capture once and use many times. It goes without saying that using state-of-the-art techniques to assure security and privacy are intrinsic to any solutions approach, but we must not let these issues deflect, hamper and ultimately defeat the initiative as has happened too often in the  past.
To succeed, such an initiative must be representative of key stakeholders, with consumers, clinicians and other health services providers determining the health outcomes required and policy makers, funders and researchers driving the what, when and how. Whilst acknowledging that government is a key stakeholder, this initiative should be driven as much by non-government stakeholders. The mission and funding of the initiative must be independent of electoral and government funding cycles. It must also be freed from the constant changes in stakeholder and management groups that are the characteristic of many national initiatives. We should adopt  an iterative, lightly engineered approach that will deliver immediate benefits and avoid costly mistakes and dead-ends.
At a detail level, we will need to adopt persistent identifiers  for consumers and providers, and consistent health information coding schemes as well as addressing legislative and  regulatory barriers that unnecessarily hamper consumer and community sanctioned collection and use of health data.
We are confident from our extensive interactions with stakeholders across the health market that there is broadly based support for such an initiative. As one stakeholder put it: “At the end of the day, our health is all we   have”.
----- End Executive Summary.
I have to say – having read this two or three times I am still wondering just what the group is really on about. The whole effort could really have benefited from some input from some e-Health experts who had some ‘on the ground’ experience. It is by no means clear to me just who is the intended target for the information etc. and just how issues of consent and privacy are to be addressed. I am also by no means convinced that the authors have a good feel for the complexity of health information and how deep the understanding of health data meaning is.
On a related topic I spotted this last week as well.

Can big data research lead to useful medical insights?

Antony Scholefield | 18 October, 2016 |
Many doctors have heard the phrase ‘big data’.
In fact, some might be tired of it already. There’s no perfect definition, but it usually describes huge datasets that reveal patterns of behaviour.
The Federal Department of Health has started to focus on using big data for medical research, and recently released a dataset based on anonymised Medicare data from 10% of patients.
It’s also working on a framework for secondary uses of MyHealth Record data, which will probably include using it for research.
But some UK and US researchers argue that big data will not lead to useful medical insights.
The rise of big data has come about because digital technology has made it easy to collect and store data.
One of the study authors, Dr Roger Highfield (PhD), director of external affairs for the Science Museum Group, UK, wrote an explanation of the paper for the website Wired.
He described big data research as being able to “take the ever expanding ocean of data [and] send a torrent of bits and bytes into a great hopper”.
“Then crank the handles of huge computers that run powerful statistical algorithms to discern patterns where science cannot.”
It’s the type of research that can tell doctors which demographics are more likely to be vaccinated, and who is more susceptible to developing kidney disease or catching a cold.
However, Dr Highfield and his co-authors warn that it’s the type of research that could also encourage people to incorrectly draw links between autism rates and vaccinations rates.
More here:
Good to see a healthy level of scepticism as to just what is possible and useful. Well worth a read:
Here is the full reference:
David.

13 comments:

Anonymous said...

The digital health honeypot has the bees buzzing. I'm guessing these folks want to make a case that they should be funded to fix this. Their complete lack of expertise in the area should not be a hinderance, right? With friends in high places, and enough pubic money, they ought to be able to pull something out of the hat eventually.

Anonymous said...

I would have to agree, the term'failure of the Market' is a clear pitch at the department who infact have failed to understand the market

Terry Hannan said...

David, excellent comments. They reminded me of the following points published in 2011 by Ashish Jha. I have inserted a comment about involving ehealth experts 'on the ground' because this lack of involvement is a major defect in eHealth developments in Australia.
Taken from one of my slides:
We have the lack of a robust measurement program. No nationally agreed-on methods for systematically identifying, tracking, and reporting adverse events. A shortage of good patient-safety metrics. Poor quality measures are plentiful. Current patient-safety indicators, which use billing data Poor sensitivity and specificity- their utility varies with hospitals’ billing practices.[Case-Mix, DRGs, ABF]
Data collected in a post hoc fashion-NOT at the time of care. Fail to engage clinicians (and as per your comments Health Informaticians) at the time of care delivery. Data unavailable for review until years after the care is delivered.
Ashish K. Jha, David C. Classen, M.DGetting Moving on Patient Safety — Harnessing Electronic Data for Safer Care..NEJM 365;19 NEJM.org 1756 November 10, 2011

Trevor3130 said...

From CMCRC's 'Flying Blind'
At a detail level, we will need to adopt persistent identifiers for consumers and providers, and consistent health information coding schemes as well as addressing legislative and regulatory barriers that unnecessarily hamper consumer and community sanctioned collection and use of health data.
A pretty good summary of the roadblocks to achieving the kind of data flows in health care that enable further steps to be taken.
Maybe, in other words, further capital investment in eHealth won't happen until there's broad commitment to removing the blocks.

Heather Leslie said...

Hi David,

The difficulty and complexity of health data is seriously underestimated in general. In the standards development environment we have massive amounts of effort put into collaboration on messages and documents with some limited interoperability occurring as a result. However historically there has been almost no work happening on defining the holy grail of health IT - the 'little data' - with the exception of openEHR archetypes and some other smaller initiatives.

It is one of the biggest elephants in the healthIT room as far as I'm concerned.

We deserve high quality data for the whole range of clinical use, but the lack of standardisation of data pattens, lack of real clinician engagement, and lack of understanding of good data design is holding us back. Traditionally each project, organisation or vendor has designed their data structures de novo - reinvention for every single eHealth activity. While that mindset persists, the variable quality, fragmented data silos will persist.

We need to acknowledge the problem and the explore it orthogonally if we seriously want to change what has been the usual approach for the past 30 years.

The openEHR approach to standardisation of the 'little data' as archetypes which are then aggregated into clinically relevant data sets using templates is now well tried and tested and underpinning a number of national eHealth programs.

In the international Clinical Knowledge Manager tool, there are more than 1500 clinicians and domain experts from 88 countries collaborating to develop agreed and verified clinical data patterns that can be freely available, shared and reused across many contexts - www.openehr.org/ckm

Norway, NHS England, a UK vendor group, Brazil, Slovenia, and most recently the province of Alberta in Canada are using this approach to standardise health data building blocks for use/re-use across EHRs, messages, decision support, research & analysis.

We have the tooling available here in Australia, under auspice of 'the Agency'. Most of the Australian usage, to date, has been focused on specific projects in the NT for indigenous health - http://dcm.nehta.org.au/ckm/.

I've blogged about some of this experience over the past few years. One post that describes how to manage the fractal complexity of recording examination findings here: https://omowizard.wordpress.com/2015/08/09/fractal-examination-findings-i/. There are a range of other posts about data quality, data collaboration, clinician engagement etc...

I fully support the reports call for the breakdown of silos. It has been my mantra for the past 10 years as we have developed this collaborative methodology to standardising health data. We are now seeing some significant momentum happening.

Regards

Heather

Dr Heather Leslie MBBS FRACGP FACHI MAICD
CCIO, Ocean Informatics
Clinical Programme Lead, openEHR Foundation

Anonymous said...

Attended the RACGP this week, so much talk about secure messaging. I came away scratching my head, they seem to have a solution but I am not sure they have a problem. Even the solution they have is at least twenty years behind the times. I know they mean well but I sense this is going to end badly.

Anonymous said...

The sharing of information is indeed an important goal, however I hope Tim is not doing this in isolation of the payload after all this is about supporting people to communicate with each other. Without semantics they are making some very concerning claims regarding interoperability. ADHA may well set themselves up for either a major disappointment or worse we end up again with everyone working in silos achieving little in value.

Anonymous said...


I agree there needs to be a lot of work to increase the value of clinical documents, flat PDF’s stored in MyHR cannot be relied on if aggregated and mining the coded information is not safe in isolation of the narrative so each document is a unique universe, so long as this is not a tactic of keeping certain voices quiet and then ostracise them when things go wrong. I support the drive to work with clinicians and industry to create specifications for critical clinical communications and embed terminology into clinical documents as required to support smarter working. But simply putting them in charge and asking them to design the solution seems unfair and poorly thought out (almost insulting IMHO. You don't ask your solicitor to design your house, get the architect to install the pluming and get the interior designer to pride legal advice. This coproduction is about the best application of the appropriate skills in the right area to represent those viewpoints, currently it looks like everyone is in charge but not of a well formed and focused team.

There is a seemingly clear appreciation that an increased emphasis on information flow electronically between (to and from) the patient’s GP and other healthcare providers and their organisations point to point using national standards for clinical documents such as a referral, prescription, pathology and diagnostic imaging order and result and a discharge summary is something that needs resolving, just make sure we use the best people in their respective disciplines playing the right role.

We might also want to listen (to understand) the legal, governance and privacy concerns (whether perceived or real).

I also wonder (and challenge) the professions in coming up with new models of care.

And don't forget the minister will need a new incisive scheme, clinicians are not on centre link but need paying all the same

Anonymous said...

Hoping someone can help me out here? I had a look around, there appears to be standards in place to support secure messaging delivery http://www.e-health.standards.org.au/Home/Publications.aspx. There seems plenty of products and providers on the market and in use. If it is about incomparable products using different standards or about providers not allowing messages to be exchanged then that should be down to the RACGP and its members to encourage their suppliers to work together or stop using the ones that don't play ball. I don't see it as the role of tax payer money to solve a commercial arrangements.

The payload is another question, after so long and so much money what has AMT actually provided back to us the investors and why NEHTA information models are so odd and hand written?

I am sure I am missing something? Must be surely?

Andrew McIntyre said...

A few of the issues are:

Many providers, especially the ones who would send like specialists don't have the certificates to allow it. No effort has been made to change this or make it easier.

The directory services government said they would provide are totally inadequate.

The endpoints do not reliably consume standards compliant messages and/or don't produce them

The renewal process for certificates is manual paper and post!!!

The government agencies often won't accept electronic communication and want fax.

The killer for safety is the lack of systems reliably handling compliant messages, so you can't be sure it will work without customizing messages for a particular version of a particular application. You may not know that unless you have a client installed at their end.

On top of this SMD is somewhat of a dog with very heavy inefficient choices of technology which were made because they were trendy. This is not really the barrier however. The lack of actual certificates and directories and reliability is.

It is a case of be careful what you wish for, having messages that don't work coming in a great rate of knots is a recipe for disaster. Reminds me of Census night....

Anonymous said...

Thank you Andrew, seems a straight forward set of complex tasks. I will wait for March 2017 for the plan to resolve this as promise at HIC.

Anonymous said...

The renewal process for certificates is manual paper and post!!!

The government agencies often won't accept electronic communication and want fax.

How very government, I would not be surprised it the end process involves scanning the fax to store electronically. Good luck changing the Civil service, they are like a Canberra procured torpedo, don't work and can't be fired.

Anonymous said...

Michael Moore makes the point that Trump is a human Molotov Cocktail that middle class Americans sick of the system of elite government hangers on plan on throwing at Washington and he thinks he may well get in. I know of a few people sick of the hopeless public "service" in Canberra who line the pockets of those same American elites in the form of consulting and "IT" companies who deliver little in return. Government has become the servant of crony capitalism and people are noticing. I think NEHTA and from all indicators ADHA are cut from the same cloth and deliver little while distracting Australian Health IT from useful work while virtually refusing to even explore local innovators and certainly imposing requirements on them that lead nowhere, but prevent real progress.

I think government needs to take notice of the Trump phenomenon, The Pauline Hanson vote indicates its alive an well in AUS. They need to hold the public service to account and stop endless waste or the Australian middle class will also revolt. Its just not good enough. NEHTA should have been axed. The ADHA has already made statements that indicate its just as lost as NEHTA, devoid of any real technical ability or insight. Its time to realize that the billions invested have been wasted and just stop doing that and stop employing expensive consulting companies that have no accountability or care factor. We need to support local companies and get out of their way rather than try and force a failed experiment to appear to succeed, because the PCEHR is pushing up daisies and trying to divert attention to other areas is a smokescreen.