Sign In

Communications of the ACM

BLOG@CACM

Encouraging IT Usage in Future Healthcare; Quality in CS Education


View as: Print Mobile App ACM Digital Library Full Text (PDF) In the Digital Edition Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
BLOG@CACM logo

http://cacm.acm.org/blogs/blog-cacm/140646-a-futuristic-health-it-scenario/fulltext
Nov. 8, 2011

The information technology trends supporting "cyber as a fifth dimension" are clear: Big Data, cell + cloud, wisdom of the crowds, co-robots, cyber-physical systems, Internet of Things, brain-machine interfaces, bio-molecular machines, nanocomputing on the one hand and exascale on the other, and quantum is still a teaser. Let's project these trends onto a point of convergence in the future and consider the following scenario relevant to health and well being.

Imagine the day when an elderly woman in India feels ill. At birth, her genetic code had been entered into her medical record. Since birth, she has been able to record a complete history of time- and location-based measurements of her physiological features (for example, temperature, blood pressure, height, and weight) and of her environment (for example, air and water quality, interactions with people). Ubiquitous sensor networks would collect this information. Today she might record this information using her cellphone and store it in the cloud. Today she might be illiterate but still be able to manage this information with speech input. These recordings are part of her personal medical record, which also includes past interactions with health and wellness professionals, such as diagnoses, interventions, treatments, and medical test results.

She contacts her doctor. They meet in cyberspace. Today this real-time communication could be in a virtual world through avatars or it could be through a wall-sized touch display in her home, projecting an image of the doctor. They share information visually. For example, she can demonstrate the pain she gets in moving her body in certain ways. She can show the location and pattern of her rash. Her doctor can explain the meaning of a test result by zooming in on a medical image or by replaying a videograph. Tomorrow the doctor might be able to palpate the sore area in investigating the problem.

The doctor consults the world to help diagnose and treat her. Based on populations of people with similar genetic makeup and similar histories (including physiological paired with environment) who had similar symptoms and reported the effectiveness of their treatments, the doctor can determine the most appropriate treatment for her. A treatment for an elderly Indian woman will be based on populations more similar to her rather than on, say, middle-aged male Caucasians who grew up in the U.S. Today some of the data is already here to mine; tomorrow there will be more data, more ways to determine relevance, more ways to spot trends, and hopefully more ways to more quickly prescribe more effective treatments for the individual. We won't need "special populations" for clinical trials anymorewe can take any subset of the world population as needed.

The doctor's knowledge base also includes a model of the human, a multiresolution, multiscale, many-dimensional, highly parameterized computational model of a human body. It relies on exascale (or beyond) data and processing capability to simulate all systems of the human and their interactions, from the molecular level to the systems level. Any "What if?" is possible.

The doctor decides surgery is needed. The surgeon directs a robot with nanoscale or molecular-scale precision to aid in the operation, implanting an embedded programmable device for continuous monitoring or periodic drug activation. As her condition improves or as new discoveries in medical science are made, the device can be reprogrammed through unintrusive software updates.

She returns home from surgery. The doctor is able to monitor her recovery remotely and continuously. Her medical device communicates wirelessly to the cloud, adding entries to her personal medical record. Device readings are accessible by only the doctor except that alerts are also sent to emergency specialists. Her home co-robot makes sure she takes her medication. It helps fix her meals and cleans her house as she recuperates.

I probably missed the boat on some of these points, not just the computer science, but certainly the medical science. But for sure, our technology is the change agent for healthcare for the future.

I will close with two caveats. First, privacy. The medical profession upholds a principle of privacy that poses difficult technical challenges for computer scientists to tackle. How can we give the doctor access to a population's data and still preserve the privacy of the individuals in the population? How can we protect her personal medical record stored in the cloud? Privacy in healthcare is an emerging area of research in computer science. Second, ethics. As with any technology, just because we can does not mean we should. As computer scientists, we are responsible for explaining the benefits and limitations of information technology and for participating in open debate on its ethical consequences. Technical solutions will not suffice; we will likely need new regulations and changes to social norms.

Back to Top

Mark Guzdial Enrollment and Quality: Does it Matter to Measure?

http://cacm.acm.org/blogs/blog-cacm/143715-enrollment-and-quality-does-it-matter-to-measure/fulltext
November 29, 2011

Lord Kelvin has been quoted as saying, "If you cannot measure it, you cannot improve it." (But he also said, "There is nothing new to be discovered in physics now," so what did he know?) In contrast, quality guru W.E. Deming wrote, "the most important figures that one needs for management are unknown or unknowable." What can we measure in computing education, what can't we measure, and does it matter whether or not we can? I've been thinking about enrollment and qualitywhat can we measure, and what does it matter?


For sure, our technology is the change agent for healthcare for the future.


Enrollment: Network World declared computer science to be "the hottest major in campus" recently. Enrollment has risen dramatically at the top CS departments. But has it really risen nationally? Internationally? I recently visited Swinburne University in Melbourne, Australia, for their Melbourne Computing Education Conventicle. CS enrollments are down in the state of Victoria, and applications for next year are down 10%.

Most of what we know about CS enrollment in the U.S. we know from the Computing Research Association's Taulbee Report, which gathers data from Ph.D.-granting research institutions. There have been efforts to gather data more widely in the U.S. (called Taurus for "Taulbee for the Rest of Us"), but those have been small and not adequately funded. The U.S. Department of Education tracks undergraduate enrollment in their IPEDS database, but only for first-time and full-time students. Part-time students, and adults returning for more education, are not counted. In reality, we don't know how "hot" CS is as a major. Nobody has the broad view.

Is that a problem? Many were concerned about a lack of enrollment in computer science. Some are now concerned about the rise in enrollment. We don't really know what the enrollment is, up or down, and maybe it doesn't really matter. We simply respond, and mostly invisible market forces will drive the students in ebbs and flows. If it is important to us (for example, to the IT industry, to those concerned about the economy), then we need to figure out a way to measure it.

Quality: I have argued in the past that we have only a few good instruments for measuring knowledge about computer science, and these aren't used often. We need these measures in order to figure out what works in computing education. I recently finished reading Richard DeMillo's new book, From Abelard to Apple. He talks about the challenges facing Universities today, from issues of cost, to issues of accessibility. The for-profit institutions threaten today's non-profit higher education institutions because they offer lower-cost and more flexible alternatives.

The argument is posed that the for-profits offer lower-quality offerings, that the non-profit colleges and universities offer better quality. Do they? How do we know? Rankings of collleges and universities are based on prestige and reputation, not on measures of learning outcomes. If a student wanted to choose an institution based on the one that could provide the most learning opportunities, how would she find that institution?

If learning quality matters, then we should try to measure it. But it might not matter. DeMillo recently pointed out (in a response to a blog post) that land lines offer higher quality phone calls, but cellphones won out because of the importance of flexibility and accessibility. The quality is good enough on cellphones. Does the added quality (if any, if measurable) of colleges and universities make the increased cost worthwhile? Or are all higher-education alternatives equally good enough, so choice is based on cost and accessibility? If quality matters, we should figure out how to measure it and demonstrate the value.

Back to Top

Authors

Jeannette M. Wing is a vice president of Microsoft and head of Microsoft Research International.

Mark Guzdial is a professor at the Georgia Institute of Technology.


©2013 ACM  0001-0782/13/05

Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from permissions@acm.org or fax (212) 869-0481.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2013 ACM, Inc.


 

No entries found