IBM and the giant Mayo Clinic are teaming on a broad technology initiative aimed at speeding IT advances related to patient care and medical research that for the first time uses the Blue Gene supercomputer for one of its original stated aims -- large-scale mathematical modelling to better understand gene and protein structures and how they interact.

The wide-ranging initiative started with the integration of 4.4 million patient records that had been in various incompatible formats into a system that has security and privacy features built in and that complies with federal regulations, said IBM and Mayo. Although the two have worked together on medical care IT projects since 2001, the new initiative goes far beyond what they have done together to date.

IBM has a facility in Rochester, Minnesota, where Mayo Clinic is based, and is committing what it describes as substantial resources to the initiative. "Dozens of people are fully dedicated to this collaboration," said Mike Svinte, vice-president of information-based medicine at IBM.

"It's the real belief that we can reinvent patient care and health care by taking a whole new approach to the collaboration," Svinte said, regarding the new initiative "The collaboration between our organisations is with the desire to drive a whole new level of innovation in a number of areas."

Those areas include molecular modelling, more highly individualised patient treatment, data mining, a better system of storing records with the intent of improving diagnoses and patient care, and acceleration of medical research. Specific projects involved in the collaboration include:

• Mathematical modelling of diseases using IBM supercomputing, including Blue Gene technology. This project will use large-scale mathematical modelling of gene and protein structures and is meant to help researchers identify causes of diseases and work on disease prevention. Viruses also can be modeled using Blue Gene, which could help scientists develop vaccines and ways to treat viruses.
• Biomedical informatics related to integrate genomic and proteomic data with clinical records and public databases that physicians use. One important focus of this project is to better use information collected in 'unstructured text' of notes taken by doctors as they talk to and examine patients. This aspect of the initiative is the second phase of the initial collaboration between IBM and Mayo started in 2001.
• Individualised patient care, with Mayo physicians using pervasive devices and data-mining tools to create individualized information for each patient to the practicing physicians "on demand". The aim of this aspect of the collaboration is to enable doctors to use existing information about patients and to compare specific problems a patient has with medical research and experience of others who have similar conditions.

"Wouldn't it be marvellous if a doctor knew not just the exact location of the patient's cancer, but its gene characteristics and the outcomes of the last 500 patients with cancer in that identical location and with those identical characteristics?" said Mayo vice-president Dr Hugh Smith in a statement. "To do this, there needs to be a consistent way to link these kinds of data, not just in a single hospital, but regionally, nationally and globally."

As software and services are developed through the collaboration, IBM will release products commercially, Svinte said.

As an example of technology expected to come from the collaboration, Svinte cited last month's announcement of the IBM Healthcare and Life Sciences Clinical Genomic Solution, which helps research institutions and biopharmaceuticals companies integrate, store and analyse genotypic and phenotypic data for use in medical research and patient care.