IBM Moves into Information-Based Medicine Without Software

By M.L. Baker  |  Posted 08-08-2005
IBM hopes to be the backbone of information-based medicine—without developing specialized software applications.

The company recently announced two deals following this strategy. On Friday, genomics and proteomics company Protana Inc. announced that it had picked IBM partner Turboworx Inc. to automate certain kinds of data processing and integrate analyses of multiple forms of data. Also Friday, GeneOS Inc., a clinical informatics company, announced that it had selected IBM partner Waban Software Inc. to manage and analyze clinical and genomic data.

IBM's strategy is to let smaller, more focused companies develop specialized software applications and then tie into the infrastructure and integration capabilities of IBM. Then IBM and its partner will share in the marketing and the profits, says Kareem Saad, business executive of IBM's Information-Based Medicine division.

Saad said the integration requires much more in-depth collaboration than hooking together a few servers. The process of melding the software's capabilities with storage, integration and security provided by IBM is something that takes months, according to Saad. It also requires that IBM staff appreciate the forms and standards used to store and communicate the data. Saad himself studied for a doctorate in biochemistry before changing his focus, and nearly everyone in his group has expertise in both basic science of how cells or genetics work and more-applied science of specific diseases, plus information sciences.

Based in Toronto, Protana contracts with pharmaceutical companies to look for clues about what drugs will work for what patients. It runs patient tissue samples through test after test, looking to see what genes are on and off, and how levels of proteins differ between patients who subsequently get a disease or who recover after a particular treatment.

Right now, doctors treating cancer, depression, or even heart disease tend to pick one way to treat a patient, then wait a few days or weeks and see what happens. The hope is that information-based medicine could identify the best option or, better yet, detect diseases before they become dangerous. That means hunting for and evaluating a variety of biomarkers—measurements that can be used to make predictions.

Complicated devices such as microarray readers, mass spectrometers and other sensors report copious data straight to a computer interface, but looking at results from separate laboratory procedures has been difficult or impossible, requiring scientists to manually process and transfer huge loads of data. For Protana, implementing the TurboWorx solution doubled the analyses that could be done in a given time, said Mark Pearson, Protana's CEO. More than that, he said, analysts are able to ask new kinds of questions and probe for previously unrecognizable relationships in the data.

Computer models calm fears of flu pandemic. Click here to read more.

Based in Helsinki, Finland, GeneOS has access to years' of Finns' medical records. The company is examining treatment outcomes along with genetic information to figure out what treatments are most effective under given circumstances.

For this approach to work, data from reams of paper medical records must be correlated with information from blood samples—all in a way that both follows patients longitudinally and protects individuals' privacy. (Although the company works only with data that has been stripped of identifiers, patients have the right to request that any of their samples be destroyed and removed from the study. That means there must be a way to track samples to individuals.)

GeneOS picked Waban Software to create a centralized database that can store and integrate information from various registries. Steven Leherer, CEO of GeneOS, said his company looked at several technology vendors before settling on Waban. Because of the different kinds of data involved, he said, the company had very specific needs for querying the database and reporting and analyzing the data.

Waban's platform was sufficiently rigorous and flexible to provide that. Further, IBM could provide reliable security and networking between GeneOS and hospitals.

IBM's Saad said that these collaborations give his company much more access to the growing field of personalized medicine than if it tried to dabble in producing software. The company would just compete with itself if it went that route, he said. Partnering widely means IBM can benefit from wide success.

Check out eWEEK.com's for the latest news, views and analysis of technology's impact on health care.