By John Russell
April 28, 2009 | BOSTON--Microsoft introduced Amalga Life Sciences this week at the annual Bio-IT World Conference & Exposition. Amalga LS is the next step in Microsoft’s ambitious plan to create a single information highway and data aggregation/modeling platform – Amalga – that spans basic research, clinical trials, health care delivery, and consumer health information needs.
“We think of this as an information supply chain where information can flow in the right way, at the right time, and in the right context. Think of five broad categories of activities: collect, validate, store, analyze, and share. Amalga LS can navigate through this data and connect even dissimilar forms of data, such as biochemical data, EMR, pathway data, gene ontologies, expression data and so on,” says Jim Karkanias, senior director of applied research, Microsoft Health Solutions Group.
Amalga’s roots stretch back to 2006 when Microsoft purchased Azyxxi, an enterprise platform developed for hospitals by a group of Washington D.C.-area physicians. The problem they tackled was that very few of the disparate systems in a hospital network shared data in a way that allowed physicians or administrators see everything at once or to include different data in common models. Microsoft renamed and repositioned the product, which was first offered in two versions, the Amalga Unified Information System (UIS) for the U.S. market and Amalga Hospital Information System (HIS) for developing countries. To some extent, the release of Health Vault, Microsoft’s Personal Health Record product, is the tip of Microsoft’s Amalga consumer vision of the supply chain.
A guiding principle in Amalga (Azyxxi) development from its beginning has been that memory (hardware) is inexpensive and the most efficient way to deal with the myriad data types generated in health care settings is simply to capture all data and metadata at the point of generation (labs, human resources, et. al); store it in a consistent, computable form; and provide easy interfaces and some modeling capabilities. Many early, centralized strategies failed to capture critical metadata, limiting their usefulness.
Now comes Amalga LS, pushing the platform into basic research and clinical trials. The official description is: “Amalga Life Sciences is designed to manage and analyze heterogeneous experimental and operational research data. The Amalga Life Sciences “hybrid store” manages both simple and complex data, including experimental protocol descriptions and higher-level knowledge representations, giving organizations the opportunity to manage and redesign processes to increase productivity, improve decision making, and reduce errors. In addition, Amalga Life Sciences interfaces with existing and future systems as an open and active software system. This unique system allows life science organizations to provide its investigators with a research environment that assists the scientific investigation and validation of clinical and biological systems.”
Coinciding with the introduction, Microsoft announced several early prominent users – the Fred Hutchinson, Cancer Research Center is one – highlighting the platform’s diverse uses.
“Our researchers face an overwhelming challenge to collect, analyze, interpret, and share complex data from a wide range of diseases and experiments,” said Lee Hartwell, Ph.D., President and Director, Hutchinson Cancer Research Center. “We look forward to exploring the potential of Amalga Life Sciences to help us understand this data in a rich and efficient way and ultimately help us meet our vision of enabling personalized medicine.”
At the other end of the supply chain, Microsoft announced specialty diagnostics supplier LineaGen will use the platform and also described an early LineaGen collaboration with Microsoft and Navigenics to eventually collect and evaluate the genetic and health information from millions of individuals, and perform interventional trials to improve the effectiveness of preventive health strategies.
“Ultimately we want to reach to the consumer as well so that they use some form of this environment so they understand the relevance of their particular state as it shown to them in Health Vault. I want to be careful to point out that today Health Vault doesn’t connect to this system in that way yet and it is our intent to create a connect framework like this. We’ll be announcing pieces of it over time,” says Karkanias.
Michael Paul, LineaGen president and CEO, says “We are very excited to collaborate with Microsoft and Navigenics to deliver a large-scale, innovative platform for understanding the linkages between genetics and disease, and to delivering this understanding as novel products and services in the personalized health care market. Combining our expertise and technology assets, we will uncover patterns in health and genetic data whose outcomes will make an impact on the ability of individuals to manage their personal health and wellness more effectively.”
Perhaps not a surprise, another early Amalga program participant is the University of Trento (Italy) Centre for Computational and Systems Biology (CoSBi), which is 50 percent owned by Microsoft. CoSBi’s primary goal is to perform research activities and develop specific languages and mechanisms of modeling, analyses, and simulations in the fields of medical science, biology, and complex systems in general. One can imagine CoSBi developed technology finding its way into Amalga.
Says Karkanias, “The current nature of multidisciplinary research and operating models has led to a research IT landscape that lacks seamless integration and extensibility of data, expresses clinical knowledge inconsistently, and which operates on incomplete biological knowledge described by different research contexts. Amalga Life Sciences is designed to overcome these issues and enable research organizations to implement new models of research and development for personalized medicine.”
Amalga LS features a “Concept Browser” for the UMLS (unified medical language system) ontology, which enables the user to input and connect multiple concepts via their ontological relationships. The user may choose to connect via all possible connections, to select the depth of connections, or to display only the shortest path. The output is displayed to the user as a visual network of concepts and relationships. Once the data is accessible and manageable, subject matter experts can then apply models and theories to data, identify the best solutions to complex problems and predict new relationships.”
“Consider a concrete example. There might be a pharma that has very advanced modeling tools based on genomic data they’ve gathered in their clinical trials; those modeling tools only can look at the data in the very narrow sense of what they were designed to look at and can’t really correlate that information with the EMR data that might be available for these patients. That’s a connection we would be able to make with our platform,” says Karkanias.
Early customer targets include pharma and biotech and academic medical centers – places where there’s a strong desire and opportunity to link research data and clinical data. Pricing wasn’t available at the time of this writing. Microsoft insists Amalga adopters needn’t use a “forklift” strategy in which the new system basically replaces the existing one; instead Amalga can be incorporated into existing systems.
Microsoft also insists it won’t compete with application providers – “We’re not going to build an EDC system, for example” – yet some functionality (e.g. modeling) embedded in platform certainly intrudes into the application space. Amalga LS will include an API to allow other vendors to build plugins for the platform.
Amalga is hardly Microsoft Office for the health sciences. Still, Microsoft is putting together a very interesting strategy that seems flexible enough to accommodate ever-advancing point tools (instruments and software) as on- and off-ramps to its health information highway and repository, while at the same time building in fundamental application-like capabilities (e.g. modeling) into the roadway and database architecture.