A roadmap toward a European healthgrid

This is an exciting work towards building an environment of sharing of resources across heterogeneous and dispersed health data:

• molecular data (e.g. genomics, proteomics)
• cellular data (e.g. pathways)
• tissue data (e.g. cancer types, wound healing)
• personal data (e.g. PHR, EHR)
• population data (e.g. epidemiology)

Also, an application which can be accessed by all users as a tailored information system according to their level of authorization and without loss of information. Collaboration is the heart of this especially across multiple disciplines. In addition, many standards have not yet been realized (e.g., HL7 (Health Level Seven), the US based Standards Development Organization that currently offers asynchronous messaging), this makes a grid approach far more powerful where organizations join the grid and make their information available for querying, processing, analysis, etc.

Few challenges remain, such as:

• How do we secure and maintain high performance of such distributed structure of data integration and computing?
• How do we close the gap between grid standards and health-related standards [some nice work's been done here by Power, et al]?
• How do we go about next-generation open source ontologies for medical informatics?
• How do we close the gap between hospital policies, public health policies, etc. and the grid approach?
• How do we go about consumerism and patient ownership of her or his data?

If anything, I hope this raises more awareness of the grid application in health and public health - much still remains to be answered.

Successes are already underway in the health community, for example:

• in the European health community (HealthGrid©)
• US CDC [presentation on the Public Health Grid by Ken Hall (BearingPoint, Inc) and Dr. Tom Savel (US CDC) at the recent HimSS February 2008 meeting]

But also there are lots of lessons which we can learn from the innovative thinking of efforts like the Google Cloud (click here to learn more about cloud computing).

I'd be very interested to hear your thoughts on this...