news Queensland Health’s beleaguered IT operation has turned its focus to a sizable IT replacement project slated to cost the state up to $438 million and see a 22-year-old patient administration program replaced, as the fallout from its billion-dollar botched payroll system upgrade continues to be felt in the state’s public sector.
The system, known as HBCIS for the Hospital-Based Corporation Information System, has been in use for several decades in the state’s many hospitals and is used to administer patient information. It is a critical tool used by medical specialists in their work; without it or an equivalent tool being in place, clinical outcomes would be severely impacted.
However, Queensland Health believes the system is in urgent need of being replaced. In September 2011, for example, the state’s then-LNP Opposition tabled a leaked extract from Queensland Health’s ICT portfolio plan, which included a strategy to replace the system, at a projected cost of $438.8 million.
The work to replace the system needed to start from July 2012; if HBCIS wasn’t replaced, then its obsolescence could result in “the inability to register, admit, transfer and discharge patients as well as effectively manage clinical and financial patient information”.
As first reported by iTNews this morning, Queensland Health chief information officer, who has been on deck at the department since September 2009, was interviewed in early August by the Queensland Parliament’s Health and Community Services Committee, which was seeking answers from the public servant in relation to Queensland Health’s e-health program following an audit of the issue in 2012. Brown’s testimony to the committee is available online in PDF format.
In his opening statement to the committee, Brown noted that significant progress had been made within Queensland Health on a range of IT initiatives over the succeeding years since the department first initiated its e-health program in 2007.
For example, he noted that new systems relating to radiology, discharge, viewing of clinical data repositories, mental health, breast cancer screening, anesthetics record-keeping, intensive care management and emergency department management had been initiated.
Brown said: “In 2007 the Queensland government provided funding of $396 million excluding depreciation to implement an eHealth program. As a result of this, the eHealth program was established within what was then Queensland Health to support clinicians across the Queensland public health system and to improve communications with private providers. This move towards electronic healthcare solutions for Queensland’s public hospitals is being undertaken in a considered, staged approach to ensure a successful transition from paper based medical records to electronic patient records for clinicians supporting staff and, in particular, patients.”
“The first phase of the eHealth agenda saw not only the delivery of speciality clinical systems to meet an immediate need and growing demand but also the laying of foundations for the delivery of an integrated electronic medical record to support a more sustainable public health system for Queenslanders into the future.”
However, the HBCIS replacement appears to be a trouble spot for the department, with most of the department’s successfully delivered IT platforms over the past several years having been ancillary systems supporting specialist areas, rather than addressing the centralised HBCIS platform. The July 2012 date for replacement of the platform to kick off has already been passed.
Said Brown: “The department has commenced a patient administration system investment planning project to deliver an implementation approach architectural framework and business case to guide the selection and implementation of a new patient administration system for Queensland. The department is focused on implementing a solution that will provide a foundation for future innovation in health service delivery and enable greater collaboration between healthcare providers across the state whilst addressing the risk associated with the current system. The development of the implementation approach architecture framework and business case will be completed by January of 2014. The work will be structured to enable review, input and approval at multiple points.”
“The HBCIS replacement activity is particularly complex, as HBCIS is more than just a patient
administration system. It also includes a range of additional functionality associated with managing
the registration and flow of patients. There are 165 separate HBCIS instances with local
configuration to support the business processes of 202 business units across the state. There are
over 80 clinical and business systems that share or integrate with HBCIS. The department will be
recommending the appropriate approach to address the replacement of HBCIS following the
completion of the investment planning project in early 2014.”
Brown noted that the replacement of HBCIS would incorporate recommendations from recent reviews into his division, as well as other audits such as the Commission of Inquiry into Queensland Health’s massively botched payroll systems implementation.
Queensland Health’s initial plan to replace HBCIS has also been pushed aside by more recent work.
“We did have a draft business case prepared for this,” the CIO told the comittee. “Again, that approach has been totally reviewed in light of the recent reviews, reports, audits and documents that have been provided. All of those documents are being reviewed and a new approach is being documented as part of having those documents ready by January 2014. The intent is not to have this as an in-house application as it currently is. The intent will be to have it as a form of a managed service.
The difficulty around HBCIS is that there are many layers to the onion, and we have to start
to unwind some of the functionality that has been incorporated into that system over many year — it has been in place for 22 years—and then really get things back to what that core patient
administration system is. There is nothing that just replaces HBCIS or a service that does exactly
what HBCIS does today. The planning activity around that is going to be significant, but the intent in terms of the ICT as a service model will be moving to that approach.
News of the imminent HBCIS replacement comes as a number of recent audit and ombudsman’s reports published into state government IT projects right around Australia have shown a marked inability by state governments to deliver IT services or major IT projects on a competent and sustainable basis.
The Queensland’s Government’s first comprehensive ICT audit published in June, for example, found that ninety percent of the state government’s ICT systems were outdated and would require replacement within five years at a total cost of $7.4 billion, as Queensland continues to grapple with the catastrophic outcome of years of “chronic underfunding” into its dilapidated ICT infrastructure.
opinion/analysis
Queensland Health. A major ICT project replacing a central system which has been in place for two decades. A budget in the hundreds of millions of dollars. What could possibly go wrong? This should be fine, right? Right?
Are you sure HBCIS is only 22 years old? It was considered ancient when I worked there 10 years ago, and I’m sure I remember one of the old timers telling me it dated back to the early 80’s.
I also remember a project being in place to replace HBCIS way back then. Why am I not surprised it hasn’t gone anywhere in a decade?
Well that’s what Ray Brown said — that’s the only info I have on it ;)
I wouldn’t put much faith in Brown. He wasn’t that great in Police.
sounds to me like Greg W knows what he is talking about…
I don’t know him personally, if that’s what you’re implying.
Even better. Have a look at the Keelty Review.
What gets me is the layers upon layers upon layers (the onion reference).
Replacing a system like that with essentially another more modern version is just begging for trouble in another 20yrs.
They should be looking at first starting with the core service of the system, making it modular, then, instead of layers on top of that, modules which take over the functionality.
You might end up with a system that has 100 modules instead of 100 layers, but if you need to add a new module, simple, if you need to upgrade a module, that too should be simple.
One does have to wonder though when it comes to putting a price tag on stuff like this, why not just your levels a tad higher, look at it and say, 400million? lets triple it, call it 1.2billion, that’s probably a more reasonable figure.
Product selected in 1988 so older than you think, HBCIS came from the HCIS project that was used to base user requirements on and was used by the Toowoomba and Royal Woman’s Hospitals until HBCIS replaced it.
But as Renai said, the only information he has to go on, is what Ray Brown has said.
Before anyone goes anywhere near this the need to read this: http://www.joelonsoftware.com/articles/fog0000000069.html
I have no doubt that the system needs to be replaced, but there is simply no way that you can replace a 22 year old system successfully.
The only way to approach this is to follow the “Grandfather’s Axe” philosophy.
ie. It’s had 3 new handles, and 2 new axe-heads, but it is still my Grandfather’s old axe.
A classic article.
Nice article.
I get to go back to some of my old code from a few years back, generally just to fix up the design of the webpage in question so that it looks like the rest of the system (i’ve got some pages that are 4yrs old and they are 6 design versions out of date).
I also get to go back and fix the webpages of the point of sale which i develop for the company i work for where the page loads slow and by slow i mean takes longer than 5secs to display, hell, even 2secs is too slow for all our sales staff.
In that instance, when i go back and re-visit it, i turn a page that takes 2 secs to load into a page that takes 0.300ms to load.
I’ve never actually completely ripped out an entire part of the system and started from scratch though, not yet anyway.
Although in saying that, i sort of have i guess, since 4yrs ago we started making the move from the old console based point of sale to the more modern web-based point of sale, some of the code is from the old version just reworked to actually function in a web environment, other things in the new one never had counterparts in the old one.
Great article – loved it when he wrote it, and it hasn’t aged. And though it’s a timely warning, they are not proposing to rewrite from scratch – they are proposing to peel off functions and replace with existing off-the-shelf solutions. Neat trick if you can, but eventually you are left with a great big ball of wool that stands or falls as a whole, with no more layers to neatly unroll. But this can be done – has been done many times. It’s not like every project fails. Time will tell whether this one does or not.
No, I think it stands alongside Brooks essays as a discussion of irreducible, fundamental, issues with these sort of projects. Issues that are a result of he fact that people are involved, independent of specific technology, or domain.
The critical risk here isn’t technical. I had the good fortune to sit in on a conversation between two of the world’s best software engineers (including an IEEE fellow). They explained in detail how there is a size threshold beyond which we cannot extract a specification of what existing software system does. While the threshold is fuzzy, I seem to recall it hovers around 250kloc or ~12man-years worth of code. HBCIS far far exceeds this. So any attempt to ‘replace’ it cannot succeed, no matter how much money you spend on it.
The problem with the “peel off functions and replace” is two fold. The first is, you can’t rely on COTS or you buy yourself a world of integration headaches and massive cost blowouts. If you try to rely on bespoke you have to navigate a morass of political quagmires. This could be managed with a sufficiently engaged minister, however the QLD government just sacked the technical staff essential to the programs success.
The only solution I can see is to spend whatever is required to keep the current system running for the next 15 years. Then to setup an architecture unit to work with a technical feasibility team to figure out how the replacement system will be integrated (to itself, and the existing system). Then to fund the maintenance sufficiently generously that after 3-5 years, there are sufficient engineers and support staff that some can be seconded to various migration spike projects (along the “peel off” model) to get a feel for how the architecture works. Finally to fully fund two adversarial technical teams: one, to with the task of finding the odd, bizarre and quirky system requirements, and to demonstrate why the proposed architecture won’t solve them; and the other, to act as a privacy/security tiger team to determine possible issues in the proposed architecture.
Then to fully fund the above for 10 years, rotating staff between maintenance and replacement projects, and slowly but surely replacing the HBCIS system piecemeal.
Of course the above doesn’t deliver the minister ribbon-cutting-ceremonies and press-releases. Worse, the major announcements won’t happen for 10-15 years, when who knows which party will be in power.
So instead we will waste $1-2bil on a project that is guaranteed to fail.
Why wait? I’ve worked a little on the sidelines of HBCIS – the answer to “can we do X to make things better?” is nearly always no. So there’s an opportunity cost to not doing anything, one measured in healthcare outcomes. All the risks you mention are real and important, but they’re not going to get better. Except perhaps governance concerns, but they could get worse too. In fact, now, in the wake of the payroll problems, isn’t that the best time?
Actually that was my intended point when I said “The only solution I can see is to spend whatever is required to keep the current system running for the next 15 years.”
We will know the Dept’s of Health and Public Works intend to avoid another massive screwup when they announce they are committing to the ugly, messy, unglamorous, and unpleasant task of keeping the current system working, and slowing evolving to meet the needs of clinicians and administrators.
What is decidedly unserious, but what I expect to see happen, is the “Clean slate” approach, exemplified most recently by the payroll debacle.
QLD Health did go as far as a trial for another product (ironically of Australian origin – abnormal for QLD health) but the project died with litigation both ways.
http://www.zdnet.com/qld-health-buries-trakhealth-suit-1339295927/
From memory QLD health paid some form of compensation in the end.
The problem boiled down to the onion layers. Onion layers were missing, not being developed fast enough, not as extensive as required, or not presented in a manner the staff were used to receiving.
IMHO, what matters most is the information in the system. So, here are some questions:
Do they fully understand the information and business processes that manage that information in the current system?
Do they have complete and accurate Entity Relationship and Data Flow Diagrams?
Are they changing the underlying information models?
If they are not, then the grandfather’s axe approach is very sensible, but the new system will not be much better than the old one, apart from how it’s built and maintained
If they are changing the underlying information models then they might get a better business system, however the question then becomes:, do they have target Entity Relationship and Data Flow Diagrams?
If they do not have all that lot, then they will fail – because they won’t know what the technology is supposed to be doing.
If they do have all that lot, then there is a chance they might succeed. ERDs and DFDs are necessary (critical, in fact) but not sufficient for a successful project.
We should find out by looking at the RFT.
The whole saga of the HBCIS upgrade is another debacle. QH has known since 2006 (and before) that the system was being retired by the supplier (let’s face it, it was written in Pick) and were offered FREE upgrade to new software. QH have continually declined the upgrade since. QH have been happily spending milions and millions on Cerner and yet keeping delaying the decision to upgrade. I note with interest that in about Nov 2012 (just before the State election) that Anna Bligh announced in Parliament in November (in response to question from McArdle) that the process of replacing the PAS was underway and QH were getting the “best of the best”.
oops – I mean 2011
The track record of the Cerner product isn’t good either. NSW started rolling it out and then switched to another vendor for the remaining areas. Victoria has been struggling with it for 5 years.
Comments are closed.