This is the second issue of a slimmed-down Digest. In it, we focus on advances in computing and communications, technologies likely particularly to affect health futures given that their existing (never mind their potential future) capabilities have been largely untapped. The one computing and communications device that just about every doctor and patient happily uses – the cell phone – may be about to turn on a veritable gusher of a tap.


* * *


For five consecutive quarters, Japanese consumers have been buying fewer desktop and laptop PCs and more cell (mobile) phones, TVs, and other devices. Given that millions of people already download music, do email, watch TV and movies, surf the Web, play games, socialize, take pictures and video, and shop on their cell phones, the implication is that the cell phone will become as powerful as, and a replacement for, personal computers – a prediction we’ve plugged before.


If you’ve ever wondered why cell phones and other electronic devices shrink in size every other month while offering ever more power and functionality (and therefore obliging you to buy a new one long before the battery runs out on your old one), it’s mainly because of the acceleration in the power of the computer chips inside them. And it’s not slowing down: In 2009, IBM server computers will have chips that run 35 percent faster and consume 15 percent less energy thanks to a nanoengineered insulator that prevents electrical energy leaking between copper wires. Remember, this increase will come on top of any other speed increases resulting from other advances in IBM’s chip technology.


But Intel would seem to have beaten IBM to the punch with this month’s launch of a revolutionary new generation of microprocessors with 45 nanometer circuits and using hafnium to overcome silicon’s problems in leaking electrons at that tiny scale. In any event, ten years from now these advances may look paltry. A Semiconductor Industry Association (SIA) panel recently listed some of the exotic science and technologies underlying processors that are in the pipeline and about ten years from production.


For example, spintronics technology (harnessing the “spin” of electrons) will be used in microprocessors to compute, as it is already being used in memory chips to store data. Other approaches the SIA panel thought promising include chips with logic gates comprising a single metal atom suspended between two carbon-based molecules, and chips made of carbon nanotubes. However, the panel did not think that the Holy Grail – quantum computing – was “doable in our lifetime,” nor that photonics was viable for in-chip processing though it was likely to become a dominant way to transport data between chips.


Meanwhile, back in January HP Labs demonstrated a “hybrid” chip architecture with standard-sized transistors but nanoscale crossbar switches, resulting in chip densities with up to eight times as many transistors compared to standard architecture) while requiring less power, generating less heat, and providing greater reliability and greater tolerance of defects in the chip. The new architecture could result in less expensive FPGA (field-programmable gate array) chips, whose circuitry can be reconfigured on-the-fly via software commands and which are commonly (but expensively) used to design the customized chips used in many electronic devices.


Even with all these advances in chip technologies, at least until quantum computers arrive there may never be enough power in a palm-sized, or even a human brain-sized, device for some needs. Let’s say you need a human brain-sized artificial human brain. Today, it takes a large room-sized supercomputer, IBM’s Blue Gene/L, to simulate just half a mouse brain. That’s roughly eight million neurons, each with up to 8,000 synapses. A ten-second simulation on Blue Gene/L ran at a speed ten times slower than a real mouse brain, but work is under way to speed up the simulation and make it more “neurobiologically faithful” by simulating structures in the brain.


In June, IBM unveiled Blue Gene/L’s almost certain successor as world champion supercomputer, Blue Gene/P, whose potential three petaflops (3,000 trillion calculations per second) performance is an order of magnitude more powerful than Blue Gene/L’s 360 teraflops. The first Blue Gene/P production unit is being installed at the Argonne National Laboratory in the US, with two more on order for US laboratories and a fourth for a UK laboratory. They will be used for complex simulations, including biological simulations such as brains that will advance our understanding and medical capability no end.


But in the meantime, existing computing power available to ordinary mortals and healthcare institutions is not even breaking a sweat in processing health information. We apply more computing power to some games and toys than we do to health information technologies (HIT) on which life sometimes depends; in particular, electronic medical record (EMR) technology.


It’s not for want of trying, and Siemens is trying again with an EMR offensive in the form of US$8 health cards that can carry the equivalent of 30 pages of medical records. Over the next 18 to 24 months, up to 100,000 patients of some 45 hospitals in New York and New Jersey will receive the cards, which are encrypted and protected by a personal identification number. At the hospital or doctor’s office, the card is inserted into a $12 smart-card reader attached to a personal computer. The patient must enter a PIN to unlock the data, though that requirement can be overridden in an emergency.


A player new to the Great EMR Game is the US banking industry, whose Medical Banking Project plans to launch a computer-based platform called “BoardTrust” that would let banks share information, including medical records, and provide standards to govern that process.


(A corporate member of the Medical Banking Project asked: “If you trust your bank with your money, why wouldn’t you trust it with your health records?” Most of us do trust the banks with our money, but many of us do not trust the banks with our privacy. Your money may be safe with the bank, but try to get a credit card without agreeing to let the bank share your data with whomever it is in the bank’s interest to share it.)


Nevertheless, HIT is gaining ground in US hospitals, and as a result more and more “nurse informaticists” are materializing to bridge the gap between proliferating EMRs and clinical practice. Many learn the new role on the job, though a handful of universities offer doctoral degrees in nursing informatics.


EMRs are just one form of HIT. Another form involves analyzing health data and making predictions from it. A significant recent advance in that sphere is IBM’s release of open-source (i.e., free and modifiable) software called STEM (Spatiotemporal Epidemiological Modeler) that graphs how an outbreak of infectious disease is likely to spread over time and geography. One analyst said it would have helped better manage the 2003 SARS epidemic, the recent outbreaks of “mad cow” disease in the UK, and the case of the globe-trotting tubercular American who earlier this year traveled extensively on commercial airliners.


Perhaps the sensor-studded prototype (no release date or price has been published) “Wellness” mobile phone from Mitsubishi and NTT DoCoMo, unveiled in October, would have advised the American to stay home and call the doctor. It is designed to serve as a personal trainer, taking one’s pulse, counting one’s calorie intake and expenditure, checking one’s body fat, timing one’s jogs, telling one if one has bad breath, asking one questions to assess stress level and even delivering a pep talk or relaxing music.


The cell phone on chip-driven steroids is undoubtedly where the action, in HIT as in entertainment and other spheres, is at.


Leave a Reply

Your email address will not be published. Required fields are marked *