Minister for Health and Ageing

MEDIA RELEASE

15 August 2011

ACCENTURE TO BUILD AUSTRALIA’S PERSONAL EHEALTH SYSTEM

Accenture, a world leader in IT, has been selected to lead the building of the IT infrastructure for Australia’s national personally controlled electronic health record (PCEHR) system in another major milestone for national health reform.

“A consortium led by Accenture has been selected as the National Infrastructure Partner for the development of the PCEHR system,” Minister for Health and Ageing Nicola Roxon said.

“Accenture will be responsible for designing and building the physical PCEHR system, which will be used by people to register for and view their eHealth record.

“People will also use this system to allow their GP and other health professionals to view their record, helping to make sure their records are available whenever and wherever they are needed.

“Health professionals will use the system to view their patients’ eHealth record, which over time will include information such as health summaries and discharge summaries. Health professionals will also be able to update this record with the most critical and up to date medical information to deliver safe and efficient healthcare.

“Accenture’s track record in the successful delivery of Singapore’s national electronic health record earlier this year was an important factor in their selection during the highly competitive independent tender process with applicants from around the world.”

Minister Roxon said Accenture will provide leadership to a consortium of companies, which also includes Orion Health and Oracle and will play a critical role in developing the PCEHR, including:

·       developing internet portals so people can view their personally controlled eHealth record, control access and enter their own medical information

·       developing a portal to allow health care providers, when given access by the patient, to view and update a patient’s record

·       providing for an audit trail that will show when and by whom a person’s record was accessed

·       providing a reporting capability for critical information about the PCEHR system itself, including performance and usage.
“A key responsibility of the National Infrastructure Partner will be to ensure that the PCEHR system has high availability, including in a medical emergency, and that it works efficiently and securely for patients and providers alike,” Minister Roxon said.

“Personal, secure electronic health records are a critical part of the Gillard Government’s national health reform. They will drive improvements over time to the way people engage with our health and hospitals system – making it better coordinated and more efficient and effective.”

The National Infrastructure Partner, under the instruction and oversight of the Commonwealth Government, will ensure stringent protections are in place to safeguard the privacy of patient records. The infrastructure built will be the property of the Commonwealth.

see also:

• The Australian news article – “Accenture would receive $47.8m to develop the personally controlled e-health record system.The department will also pay $17.8m to Oracle in licence fees for access to e-health records stored within all PCEHR repositories, and $11m in fees to Orion for operating a portal.”
• IT News states “The New South Wales Government has previously contracted Accenture to consult and advise the development of an Integrated Clinical Information Program (ICIP) for the state’s healthcare systems. The State Government is also named as an Australian customer of Orion Health.”

Google news on Accenture:

Accenture is a global management consulting, technology services and outsourcing company. It has over 223,000 employees including 4,800 senior executives, services in 53 countries and is generating net revenues of $US21.6 billion as at 2010. It’s revenue apparently primarily comes from consultancy and outsourcing.

It will be fascinating to see how they manage to deal with the risk of hacking given other major players have fell foul of the hackers recently, and whether they can create a system for $77m for over 20 million population when Singapore’s solution allegedly cost $140m for only 5 million population.

Accenture also has just won the tender to develop an Australian national electronic system for property exchange and conveyancing called PEXA- see their media release.

Accenture was also mentioned in the Jan 2011 settlement by Oracle to pay $46m in fines to US govt to resolve alleged kickbacks paid by Sun to companies such as Accenture, when these consultancy firms managed to convince a federal agency to purchase a Sun product – seems such kickbacks are rife in the US government contracts, but surely not here in Australia.

In 2009, then Labor minister, Peter Garrett opened IT services company Accenture’s new Sydney office – see here. It seems the Labor Govt has been courting Accenture for some time, so hopefully their decision to appoint it as lead in this project has been well considered.

Accenture’s troubled ATO IT system:

Accenture was also apparently awarded a $435m contract for an IT system for the Australian Tax Office (ATO) called Change Program in 2004 but seems that project may not have been smooth sailing with cost blowouts of nearly $400m – see here when apparently the ATO terminated Accenture’s contract in 2010 for the troubled system, and engaged a separate consultant, CPT Global to advise on how to make it work during the coming tax season.

In April 2010, 140,000 tax refunds were issued without cheques, contributing to a backlog of some 210,000 tax returns. A further 16,000 taxpayers were affected by erroneous, automatically generated decisions.

This does not seem so promising.

The ATO project got off to a controversial start when the tax office dumped the tender process to find a project partner after it judged all the bids non-compliant, and thus the early design and systems integration work was awarded to Accenture – see here.

The Inspector-General published concerns in May 2011 related to the ATO Change Program and the risks ATO took on including Accenture’s ability to deliver – see here.

Hopefully the lessons have been learned and history does not repeat, although past behaviour is one of the better indicators of future behaviour.

The Federal Labour goverment can ill afford more egg on its face from failed projects.

• increasing staff dependence upon computers to document their clinical work and access timely patient information and clinical decision support.
• lack of physical space for additional desktop computers or computers on wheels to address the above point
• current laptops and Windows-based tablets are too heavy for carrying around by staff.
• this means we need wireless networks throughout hospitals and staff will increasingly use their Apple iPads, perhaps their iPhones and Android phones and other devices to help them manage the increasing pressures to find an available computer when and where they want it.
• the resultant need for high performing, rapid data entry, database intense, cross-platform software with 3D modeling multimedia capabilities (eg for documentation of injuries using 3D models of humans, potentially for 3D views of CT scan images, for rapid analysis of data streams on small devices) which does not degrade when zooming in or out on smaller hand held devices (this requires vector graphics based graphical user interfaces instead of current bitmap pixel based interfaces).
• the need for better use of computer chips given that we have largely stopped making substantially faster chips, and additional speed in new chips is largely enabled by adding more chips which has implications for software design.
• the ability to rapidly modify database applications in response to rapid changes in business needs of hospitals such as introduction of new business rules regarding KPI’s such as the 4 hour rule, etc.

Microsoft appear to have abandoned their .NET and Silverlight technologies as a cross-platform technology and, along with Apple and others, appear to have adopted CSS and HTML5 as the cross-platform technology.

CSS and HTML5 are great technologies and have brought us web-based applications such as Facebook, etc and can be leveraged to create many clinical applications BUT they are dependent upon the browser and device used and they are no where near as rich a user interface nor as responsive as a native application can be.

Enter a new candidate from left field which may be game changing for rapid development of database intense, native cross-platform business applications built upon a vector graphics engine which offloads the 3D and much of the other graphics processing to the device’s graphics processing unit (GPU), thereby taking the pressure off the CPU.

I had the honor of attending an all day presentation yesterday of this exciting new technology from Embarcadero which they have called FireMonkey.

It is a brilliant, revolutionary extension of their existing RAD Studio product and allows one to program in either C++ or Object Pascal in Delphi, and the same code can then be compiled to create NATIVE applications for 32bit Windows (eg. most current hospital desktop computers), 64 bit Windows, Mac OS X, Apple iOS (iPhones and iPads – and the applications can be uploaded to Apple iTunes store for easy deployment), and in 2012, they will add Linux, Android and Blackberry support (and C++ iOS compile support).

The code can even be debugged and edited WITHIN Apple’s xCode programming environment, in the original Object Pascal code – now that is very exciting indeed!

Completely cross-platform compliant code needs to ignore certain platform specific functionality such as Window’s COM interfaces into Microsoft Office integration, and certain features of the iOS such as camera technology, but the programmer can user compiler directives to include any of these technologies for a specific target platform as needed.

It fully supports the creation of their DataSnap multitier technology which means it becomes quite simple to create rich, visually stunning business database intense clients on any of these devices using 3D vector graphics if needed with no direct database connections, but a common database connection to a DataSnap server (which can use JSON or XML for communication, across https if needed for internet access) – a Windows-based middle tier application created using the same technology which runs on a server and is responsible for all the database connections. This means one does not need drivers and database connection settings to be deployed to all these devices making maintenance much easier and furthermore, adds an extra layer of security.

There is one important issue at present, remote desktop or Citrix environments cannot currently utilise any GPU functionality (I believe Citrix is working on how individuals can access a shared GPU on a server but don’t hold your breath). This means 3D modelling for rapid visual analysis of complex data streams or for ease of documenting injuries, or GPU optimised manipulation of 3D CT scan reconstructions will NOT work within Citrix at present and possibility into the future.

FireMonkey will be available as part of the RAD Studio XE2 software which will be released in the next couple of months. Clinical developers of middle tier servers will need the Enterprise or higher version so they have access to the full range of database connectivity which allows connection to the main database servers such MS SQL Server, Oracle, Sybase, etc as well as access to new Cloud Computing database stores such as Microsoft Azure and Amazon.

No longer will a programmer need to know and be an expert in multiple programming languages (currently, SQL, C#, html, javascript, java, PHP and Objective C) and multiple programming environments (MS Visual Studio, java Eclipse, Apple’s xCode) to achieve high quality cross-platform deployment of applications with different code sets for each platform making maintenance a nightmare.

Now all they really need is to know SQL and Object Pascal (a very readable programming language similar to C#), and a single programming environment (although in the initial release, one may need to use xCode and Eclipse but only for final compilation).

This is indeed an exciting development which at last may solve the many issues currently preventing staff satisfaction with clinical systems in hospitals.

Of course, the technology is one thing, clever implementation of it by careful analysis of end user needs is another thing!

There are over 2 million Delphi developers worldwide making it the 2nd largest Windows developer community after Microsoft, and Delphi sales have been growing 15% pa for the past several years, and with this new cross-platform technology, sales are likely to be boosted subtantially. It is clear Embacadero believe they are on a winner!

More technical details here.

CT scanning overall had a sensitivity of 93% for detecting SAH, but if done within 6 hours of onset of headache and interpreted by an experienced radiologist, the sensitivity rose to 100% picking up ALL 121 patients with SAH of the 953 patients scanned within 6 hours.

Looks like its time to modify our practices as long as we have access to experienced radiologists – overnight may be an issue when such access may not be readily available.

Their related study (see pdf here)  trying to develop a Canadian SAH rule of who to do a CT scan upon needs further validation but suggests that consideration for possible SAH and thus possible CT scan on all patients with sudden onset headache (reaching peak intensity within 1 hour of onset), which has not occurred more than twice over the past 6 months (thus excluding chronic recurrent headache patients), who have any of the following features:

• age >= 40 years
• witnessed LOC
• complaint of neck pain or neck stiffness
• onset with exertion
• arrival by ambulance
• vomiting
• diastolic BP >= 100mmHg
• systolic BP >= 160mmHg

Of the patients included in their study, 54% had benign headache, 27% migraine, 6.5% SAH, 4.3% viral, 1.8% TIA, and 1.5% had post-coital headache.

They used 3 rules using variable items above, and each of the rules had 100% sensitivity for SAH, while specificity ranged from 28-39%

It will be interesting to see how the validation studies pan out.

It seems that the consensus was that for EP’s to be productive with such software they each need to have a personal scribe to ensure adequate timely documentation while allowing them to have adequate time at the bedside with the patients.

Furthermore, it seems that the majority of prescibing mistakes are now because of software design issues and how the software interfaces with end users.

Perhaps Australian administrators should be taking heed – just because it “seems” to be a good idea and it is used in the USA does not mean it should necessarily be adopted here.

See details on the wiki

Microsoft itself avoided using .NET for its main Windows applications such as Office.

Today, Microsoft unveiled early version of Windows 8 – and is promoting that it will be designed for two main types of apps – current Windows apps and yes, you guessed it, HTML5/javascript cross-platform apps.

Seems all that hype over the last decade of the supposed benefits of .NET is coming to an end.

I am not sure where that will leave the major medical IT players who have designed their whole systems around .NET, although I would think Micrpsoft will continue to support these “legacy” apps for the next decade at least.

I am really glad that I stayed with native Windows developent with its ultra rich client experience, and now its time to branch into HTML5/javascript so the plethora of Android slate devices (and iPad) can access health information within intranets and also in cloud computing technologies.

Now we just need NEHTA to get on with completing its preferred datasets for Australia and we can all start making exciting new products now that the air has cleared somewhat and there is more clarity in technology directions.

“Sally McCarthy, president of the Australasian College for Emergency Medicine, said the review by Professor Jon Patrick confirmed that the system, loathed by doctors and nurses, is unsuitable for its purpose”

But then, NSW ED staff have been complaining about it ever since it was introduced, and still it seems Cerner has not fixed its problems.

Nevertheless, the increased risk seems consistently increased as radiation dose increases and thus the results may be plausible.

Given the average age of these patients being ~61 years, some 14% were diagnosed with new cancers in the 2-5 years following their AMI, thus a relative increased risk of 3% per 10mSv is something to stress us!

See here for the paper.

This will inevitably put pressure on us NOT to refer patients for stress MIBI scans, nor diagnostic coronary artery CT scans.

In an ideal world, we would have timely access to a sensitive and specific radiation free investigation.

Unfortunately, stress ECG testing has a relatively low sensitivity and in younger women a higher false positive rate.

That leaves us with stress echocardiography, but availability and operator dependence means this also has issues at present.

Furthermore, the added morbidity associated with cancer risk may also impact who we should actually be referring for angiography and/or angioplasty.

Time will tell, but in the interim, perhaps the least we can do is provide adequate informed consent.

See wiki for more details of stress testing.

The midwives are pushing to rapidly expand planned home births through Community Midwifery programmes, hoping to emulate the success of the long existing UK practices.

A report recently published by the Western Australian government on analysis of perinatal mortality for the 3 years 2005-2007 appears on face value to be quite damning of planned home births in that state, with mortality rates which have not improved on the previous 3 years and which are ~20x higher perinatal mortality than that for ALL hospital births over 37 weeks gestation when looking at perinatal hypoxia causes.

Put this another way, if a midwife in WA selects a patient as being low risk and suitable for home birth, that child has a 1 in 120 chance of dying from perinatal hypoxia compared with a 1 in 3000 chance of dying if it were delivered in a hospital (perhaps even lower risk as hospital data includes moderate and high risk patients).

Now that appears to be a very damning statistic for planned home birthing as currently implemented.

BUT, there is one important piece of the puzzle that is not available – morbidity data from perinatal hypoxia – and without this, once cannot make good sense of the data.

For instance, it could be that the difference in mortality rates could be entirely accounted for by increased severe hypoxic brain damage rates in the hospital group.

Of course, it could also be true that there is actually an excess of hypoxic brain damage in the planned home birth group.

The report advocates collection of morbidity data for the next report which may shed light, but only if they collect it for BOTH groups, and preferably compare apples with apples such that moderate or high risk pregnancies are excluded from the hospital data.

See the report here (pdf).

Check out his blog here.