Project Summary
BG Medicine, Clinical Data Appliance
An elegant solution for organizing complex data.
BG Medicine
Life Sciences
Practice Area
Analytics & Data Management

BGMedicine develops diagnostic tests for certain heart conditions. To support clinical trials for a new cardiovascular test, Quoin designed and implemented a network-based application to evaluate and score clinical patient data in order to predict adverse patient outcomes. Our solution was essentially a real-time Business Rules Management System for classifying laboratory data.

The Challenge

BG Medicine is a life sciences company focused on the development of cardiovascular diagnostic tests. To support the clinical testing of Galectin-3, their first commercial product, BG Medicine company wanted to build an easily-deployed software and hardware solution to help them access patient data, apply rules-based calculations, and generate the resulting index value as a new result for a patient's medical records. Since BGMedicine would deploy the appliance to a number of clinical environments, the system needed to be simple to deploy and compatible with the local hardware and networking. BG Medicine engaged Quoin to work on this Indexer Appliance application.

Our Solution

Despite the complexity of the problem, Quoin rapidly defined and implemented an effective technical solution within a three-month timeframe. We assembled a project team consisting of a technical lead and two software engineers, who were responsible for technology selection, architecture, design, implementation, and testing. 

Our approach was to construct a standalone software and hardware appliance – a Java-based application and open source rules framework (JESS) deployed to a dedicated Linux server. The BG Medicine AMIPredict application integrated with a specific LIMS middleware system (Data Innovations Instrument Manager) to enable access to clinical data. The appliance essentially emulated a laboratory system, using the appropriate Instrument Manager API to read clinical data and write the results after applying the rules-based scoring. Quoin’s technologists designed the architecture to be lightweight and to require limited processing capabilities, so that the 'server' could be any small-footprint Linux device. We also ensured that the application supported easy updates to the rules, as the scoring algorithm for the Galactin-3 Test evolved.

Throughout the project, our technical lead worked closely with senior management at BG Medicine to ensure that our technical approach supported life science and FDA compliance requirements. We collaborated with client stakeholders to ensure that AMIPredict could be used as part of the clinical trials. This required a series of internal reviews and modifications to the original application design, which were carried out in addition to the core technology design and build-out.