Garvan Institute rolls out new information architecture
- 17 July, 2009 12:10
The Garvan Institute has deployed a new database platform to consolidate and centralise more than 30 separate databases as the medical research organisation moves toward a new information architecture.
The platform, Sybase’s Adaptive Server Enterprise (ASE) 15, has allowed Garvan to merge patient information, such as symptoms, demographics, surgery results and post-surgery results with information collected from lab research.
It has also increased its ability to compile and analyse research data increasing the accuracy of Garvan’s research results and findings, and enabled the generation of detailed reports that merge lab research with information on patients suffering from a variety of diseases such as cancer, diabetes and osteoporosis.
Researchers now have the power to independently run queries on data via a new query tool without the need for the IT department to code up a query, compile the findings and present researchers with a spreadsheet if findings.
According to Jim McBride, information technology director, at Garvan, the move to the new database platform and information architecture was driven by the impending move to a new $100 million Garvan-St. Vincent’s Cancer Care Centre.
The centre will more closely combine Garvan’s cancer research work with St Vincent’s Hospital patient care service in an effort to find a cure for cancer and provide better care for cancer patients.
McBride says that in comparison to the organisation’s information challenge, building a $100 million facility is relatively straight forward.
“We have to make sense of a vast collection of clinical information about patients, their experience, disease progression, and the laboratory information across about 30 distinct databases,” he says.
“Each of the disease areas comes in from an historic solution. It may be a scientist who has an Access or FileMaker database that has gotten out of control, or sometimes even a collection of spread sheets that a researcher is trying to handle and keep current. Then, there are the errors that come with application versions problems and cross-correlation issues when you are trying to do information management with the wrong tools.”
With more sophisticated medical research tools coming emerging, such as gene chip analysis, higher quality images from microscopes, and sequencing tools, Garvan also had to get a grip on a growing data management problem.
“There is just terabytes of data,” McBride says. “We had to put together an information architecture which lets researchers make sense of that without having a PhD in computer science at the same time.”
With the implementation of Sybase’s ASE 15, Garvan has been able to use it as a foundation to tie together disparate applications, image management systems and clinical information systems in one information architecture, he says.
He says ASE, as the relational database management system, sits at the bottom of Garvan’s architecture stack with a layer of meta data on top of that, followed by the application layer.
“That way, when the data model changes, we can reflect those changes in the meta data layer and our applications can continue to function,” McBride says. “Researchers can ask questions of that data without an IT person having to go in and re-release the query a tool or code the query on behalf of the researcher.”
McBride says that the data migration process was pretty arduous and required getting researchers and other staff on side for the project. The key to this was educating non IT staff on the highly iterative process of data migration.
“Users gave us their data once, then thought they were finished,” he says. “But we had to go back to them and ask, ‘well, you have an ‘m’ and an ‘f’ in your gender column, but you also have 100 people who have a question mark, and three who have an asterisk – what does this mean?’”
The process of refining data is a major component in the migration process, McBride says, sometimes seeing more than 50 percent of project time being spent on data manipulation rather than interface design and application deployment.
“It’s only because we can demonstrate that once they have taken that journey they have a much more powerful environment in which to pose questions, and that their data will be there in six months’ or six years’ time, and will be securely backed up, that they come on the journey,” he says. “They’re now lined up at the door to do more.”
McBride adds that he is pretty confident of the flexibility of ASE to underpin Garvan’s information architecture and change with the business and its requirements.
“I’ve been around applications long enough to know that you can have a good application one day and a few days later people need it to be changed,” he says. “If you don’t have the ability to keep the application current – particularly in the research business – then you look like a turkey in six months and you’ll need a new job in 12.”