By Michael A. Goldman
June 15, 2003 | Larry Smarr, the outspoken astrophysicist who directed the National Center for Supercomputing Applications when it developed the first widely used Internet browser, NCSA Mosaic, turned his attention to digital medicine -- or the lack of it -- in his address to the International Biotech Summit in San Francisco last month.
“Why do people take better care of their cars than their own bodies? A new Mercedes has no fewer than 60 microprocessors to tell you how every critical subsystem is running. Why aren’t we doing this with people?” asked Smarr, now director of the California Institute for Telecommunications and Information Technology.
Smarr told the audience he envisions a day when sensors throughout the human body broadcast data over a personal area network (PAN) to a personal medical assistant (PMA), a kind of Palm device for the processing and coding of data. Subsequently, data are sent over a wide area network (WAN) to a medical monitoring station for remote analysis.
Sound far-fetched? Far from it, Smarr says: It’s practically here. Smarr is perhaps best know to the life sciences community for his 1999 report co-authored with David Botstein, department chairman and professor of genetics at Stanford University. Their report has since become the cornerstone of a National Institutes of Health (NIH) initiative in biomedical information.
Today, Smarr noted, he is connected to the Web 24/7 from his cell phone. Why not do the same for medical needs?
Wiring the Patient
A May 4 article in The New York Times suggested that anyone might one day mail in a blood sample with a check for $100, and get back a detailed readout of his or
her current phenotypic condition. Instead, Smarr suggests, why not put a chip right in the human body? For years, Finnish company Polar Electro Oy has been making wireless devices that transmit heart-rate data to a wristwatch.
“Just add Wi-Fi or wide area cell phone data transfer,” and you’ve got it, Smarr says. An Israeli company, Given Imaging, is marketing a video pill, the M2A Capsule Endoscope. It’s a camera the size of a vitamin pill that transmits pictures to a body-surface receiver as it photographs the intestines, perhaps someday rendering the colonoscopy an unhappy memory.
Don’t get too excited. The day of implantable monitors transmitting vital data to the Internet might be decades off. Astro Teller, CEO of BodyMedia, cites the difficulty of getting a long-lasting small device that can transmit data over even very short ranges (the M2A battery is dead by the time it reaches the large intestine), and the resistance people have to surgical implants. Teller’s company is focusing on wearable monitoring devices with onboard software, running sophisticated mathematical models to convert 32-times-a-second data points into useful feedback to the individual.
“Two-thirds of healthcare cost comes from the way people live their lives,” Teller says. BodyMedia hopes to provide users with constant feedback about exercise and diet. The company is working with Roche Diagnostics to produce body-monitoring devices and software for use in clinical weight, diabetes, and cardiovascular disease management. According to Teller, “BodyMedia software for Roche provides a summary report for how [the patient has] done, replacing a 5- to 10-minute conversation with clear, concise, objectively accurate information” about lifestyle.
To achieve these kinds of healthcare advances, “Biomedicine requires IT and telecom now,” Smarr says. “Historically, human brains could keep up with all the data. Now, only software can read all the data.”
“We’ve got 28 billion base pairs in databases, drawn from over 50,000 species, enabling cross-species pattern recognition from 4 billion years of evolutionary experiments,” Smarr says. With all of this information, “networks and the grid are emerging as the 21st century driver. It’s a scandal the NIH has been such a laggard,” he says, referring to the building of the information infrastructure for biological and medical research.
But progress is being made, Smarr accedes. The Biomedical Informatics Research Network (BIRN), led by University of California at San Diego and University of California at Irvine and funded by the NIH, is one example (see Mouse Hunt on page 46). According to the BIRN Web site, “The initial goal is to address the needs of biomedical investigators across the country to effectively share and mine data in a site-independent manner for both basic and clinical research.”
BIRN’s early focus is on neuroscience and imaging. BIRN’s brain imaging repository collects data on brain function from a host of collaborators, modeling, for instance, the neural circuitry involved in schizophrenia. Smarr says the NIH plans to expand this effort to other organ systems and other academic centers.
Smarr is proposing a National Science Foundation Large IT project that will allow tiled displays to view 150-million-pixel montage medical images of brain slides that already exist. He envisions artificial intelligence software trained on “millions of human image data sets, monitoring the life cycle of single individuals and providing automatic early warnings” about medical conditions.
Of security, privacy, and the proprietary nature of data, Smarr quipped: “I’ve seen this movie before in many fields. The power of collaborative sharing overwhelms the desire to keep data proprietary.” But the concerns about the confidentiality of private medical data have already resulted in the cumbersome new HIPPA medical privacy protection standards, and may thwart efforts to achieve a true digital medicine. Let’s just hope the exciting movie unfolding before us isn’t “The Twilight Zone.”
Michael A. Goldman is professor of biology at San Francisco State University.