Melbourne Health is a public hospital with an operating budget of around $700 million. It serves a population of nearly one million people, featuring more than 7000 staff and more than 1000 beds.
Melbourne Health recently contacted BioGrid Australia, in order to allow researchers access to a collection of archived MRI images. BioGrid provides infrastructure to integrate research and clinical data from multiple institutions and disease groups.
Specifically, the researchers wanted BioGrid to consolidate 12 years worth of magnetic resonance imaging (MRI) scans from 1000 DAT tapes - representing two terabytes of data - into one single database, and to convert the images from proprietary MRI to DICOM files.
After initially looking at using the Oracle Database 10g, Naomi Rafael, senior database administrator at BioGrid, was encouraged to beta test version 11g, as its more advanced ability to manage DICOM images had the potential to better handle the store of images. Rafael’s team has since been testing DICOM image storage, metadata extraction and DICOM image retrieval with satisfactory results.
“The DICOM datatype as part of the offerings in Oracle 11g were extremely advantageous to the advancement of this project,” Rafael says. “The datatype has built-in functions such as extraction of the object metadata, anonymise function and ability to create views on the DICOM object itself.”
In particular, the project team - while facing huge data growth in the coming months and years - has welcomed the image compression aspects of the new release. The database currently represents 45,000 examinations, comprising over seven million individual MRI images.
“The compression functionality has yielded 42% compression of the data content,” Rafael says.
Furthermore, the team is happy with the security features of the solution. As Rafael points out, they allow the grouping and sharing of multiple patient MRI images between systems, without divulging confidential details such as a patient’s name or private details.
The security and compliance elements of Oracle 11g improve on the strict role-based regime from the previous version. It ensures only the right personnel can access sensitive data.
Additionally, options exist to enable the hospital to audit access history to ensure privacy standards are maintained and that breaches are easily identifiable.