Future of Health IT: Trends and Scenarios

Coming soon via Harvard/MIT Health Sciences and Technology: Your vital signs, on camera

You can check a person’s vital signs — pulse, respiration and blood pressure — manually or by attaching sensors to the body. But a student in the Harvard-MIT Health Sciences and Technology program is working on a system that could measure these health indicators just by putting a person in front of a low-cost camera such as a laptop computer’s built-in webcam.

via web.mit.edu

I found this through ACM TechNews. The technology's in a pretty raw state at this point, but it points to one facet of where we're likely to be 3-5 years down the road. Here's a YouTube video in which the inventor describes it in a nutshell. There's a link in the lower right, only barely visible, which will open the video in YouTube itself. If you slide your mouse cursor down there, you'll get hover text ("Watch on YouTube") when you're over the link.

 

News Analysis - Medical Science and Practice in Conflict - NYTimes.com

The New York Times ran a story about a paradigmatic conflict of views coming to a head - the dichotomy between medical science and practice.

This week, the science of medicine bumped up against the foundations of American medical consumerism: that more is better, that saving a life is worth any sacrifice, that health care is a birthright.

via www.nytimes.com

As someone who has worked for years in the clinical research field and has done some research of my own, I tend to side with the scientists. Except, of course, when I am playing the role of patient - then cost is no object (I only have to deal with the co-pays), and my attitude becomes "leave no diagnostic test behind." This article came in the wake of the new recommendations of the US Preventive Services Task Force at the AHRQ.The recommendations were an about-face from the same group's 2003 advice, suggesting that much fewer screening tests could be done with the same net result in the public health, leading to lower overall costs at a time when the federal deficit is crushing and looks only to be getting larger as time goes by.

The challenge of persuading patients and doctors to accept such standards requires a transformational shift in thinking, particularly when the disease involved is as prevalent, as deadly, and as potentially curable as cancer. How do you convince them that it is in their best interest to play the odds when they have been conditioned for so long to not gamble on health? After all, for the one in 1,904 women in their 40s whose life would be saved by early detection of breast cancer, taking the risk would in retrospect seem a bad choice.

The problem is, in concise form, that self-examinations and even primary-care clinical exams lead to a high proportion of false positives, reflective of the "better safe than sorry" viewpoint taken by many if not most physicians and patients. Mammograms, the most common second-tier test, have risks of their own, given the radiation involved.

To further confuse matters, at about the same time the recommendations came out, GE's CEO Jeff Immelt announced the release of a handheld ultrasound screening device that would make breast exams (and many others) significantly safer and less expensive.

Which side are you on? I still have a foot on either side of the fence, and it's getting increasingly uncomfortable as time goes by.

Natural language processing: the key to the *omics revolution?

I wasn't the first (or the brightest) informatics geek to notice the similarity between the molecular structures of the various *omics diciplines and the concept of the textual corpus. A number of researchers here at Michigan are applying linquistic techniques in systems biology, so there's a natural curiosity about that similarity on my part, even if it's not where I spend my time.

Last week I came across a, dare I say, "textbook" example of the use of natural language processing (NLP) to advance genomics, in a news release from UC Berkeley entitled Improved method for comparing genomes as well as written text. Here's a fairly large taste:

Taking a hint from the text comparison methods used to detect plagiarism in books, college papers and computer programs, University of California, Berkeley, researchers have developed an improved method for comparing whole genome sequences.

With nearly a thousand genomes partly or fully sequenced, scientists are jumping on comparative genomics as a way to construct evolutionary trees, trace disease susceptibility in populations, and even track down people's ancestry.

To date, the most common techniques have relied on comparing a limited number of highly conserved genes - no more than a couple dozen - in organisms that have all these genes in common.

The new method can be used to compare even distantly related organisms or organisms with genomes of vastly different sizes and diversity, and can compare the entire genome, not just a selected small fraction of the gene-containing portion known to code for proteins, which in the human genome is only 1 percent of the DNA.

The technique produces groupings of organisms largely consistent with current groupings, but with some interesting discrepancies, according to Sung-Hou Kim, professor of chemistry at UC Berkeley and faculty researcher at Lawrence Berkeley National Laboratory. However, the relative positions of the groups in the family tree - that is, how recently these groups evolved - are quite different from those based on conventional gene alignment methods.

The computational results have surprised scientists in being able to classify some bacteria and viruses that until now were enigmatic.

The real deal has appeared online in the Proceedings of the National Academy of Sciences; here's the citation:

Gregory E. Sims, Se-Ran Jun, Guohong A. Wu, and Sung-Hou Kim, "Alignment-free genome comparison with feature frequency profiles (FFP) and optimal resolutions", PNAS: published online before print February 2, 2009, doi:10.1073/pnas.0813249106

I can't say I'm intimately familiar with the feature frequency profiles concept, but then I'm an NLP buff rather than an expert.  The link above takes you to a page where you can download the preprint as a PDF, which has a few typos, but gives a fairly clear grad-student-reading-level description of the algorithm and how they optimize it for particular genomes. Cool stuff!

I'm not sure how directly this will apply to other *omics disciplines, such as proteomics and metabolomics, which are where much of the real action is found in biomedical research. Still, it's an exciting development, and a harbinger of things to come, as the NLP community takes on the challenge of the greatest story ever told, the story of life as viewed through the lens of molecular biology.

Quickie: Doing quality control on computerized genome science

Ain't genomics (and *omics for that matter) wonderful?

Of course the *omics revolution is wonderful. But as with any new technology application, the glow imparted by visions of its potential often obscures realistic considerations. For example, consider this report of a promising new invention:

The Homework Machine, by Shel Silverstein

The Homework Machine, oh the Homework Machine,
Most perfect contraption that's ever been seen.
Just put in your homework, then drop in a dime,
Snap on the switch, and in ten seconds' time,
Your homework comes out, quick and clean as can be.
Here it is--"nine plus four?" and the answer is "three."
Three?
Oh me . . .
I guess it's not as perfect
As I thought it would be.

...which brings us to Ars Technica's news post entitled Doing quality control on computerized genome science.

Anybody that has followed our remarkable progress with genome sequencing is aware that the completion of each genome brings a number—usually in the low tens-of-thousands—of genes found in the DNA of the species involved. Some of these numbers gave human pride a body-blow, as we wouldn't have expected to have only a few thousand more genes than the worm, and half as many as a rice plant.

But we've never actually confirmed that the majority of these genes exist, and most were predicted by computer programs that scan the DNA sequence, searching for signs of a gene. Two recent papers have performed a quality check on these programs, and found that they do make an appreciable number of errors—just nowhere near enough to give us more genes than rice.

It's worth a read, just to get an idea of the kinds of errors the studies found, and their take on the potential risks. The post author's assessment doesn't completely jibe with the opinions of the study authors, from what I can see: we're talking about an error rate of 13%, according to one of the studies.

Obama and Biomedical Research Informatics: Reason for Hope?

Recently a friend at work sent an email to our workgroup in which he said:

In his weekly radio and video address describing what he called the American Recovery and
Reinvestment Plan, Obama spelled out five main goals. He said his plan proposes to:

-double renewable energy production and make public buildings more energy efficient;
-rebuild crumbling roads, bridges and schools;
-computerize the health care system;
-modernize classrooms, labs and libraries;
-and provide tax breaks to American workers.

So the computerization of our Health Care system
becomes a National Priority.

The implication is that Obama's presidency bodes well for our workgroup, which is devoted to biomedical research informatics. I wanted to share his thought and my reply, which follows:

While I share the enthusiasm for Obama's recovery plan, it may not
have much direct impact on research informatics units like ours, if
any at all. The effort to computerize healthcare was a key goal of the
Bush (W) administration, and while some progress has been made in
terms of determining evaluation criteria for EMR systems, from what I
have seen thus far, little or no interest has been exhibited with
respect to incorporating research informatics in the picture, in spite
of lobbying on several fronts.

The challenge of getting every American's health record computerized
is monumental, conceivably beyond the effort thus far put into space
exploration, and is fraught with contradictory economic and political
interests and motivations. This is not just an academic research
issue. Big Pharma has a stake in the integration of research
informatics into EMR systems (e.g. post-marketing surveillance), and
when Big Pharma's lobbyists can't get much traction, you know the
obstacles are formidable.

That said, another of Obama's campaign promises was to double the
budgets of the NIH and NSF. Let's cross our fingers and hope that
these promises are fulfilled as the massive economic stimulus and
recovery effort gets underway.  If they increased the government's
research agenda by even a fraction of what he promised, it will
benefit us significantly. Biomedical research informatics in
inextricably tied to biomedical research itself; progress in research
will be impossible without significant investments in biomedical
research informatics.

Moreover, all types of translation are becoming increasingly dependent
on informatics. Moving basic science discoveries into clinical
innovations and thence into evidence-based medicine and improvements
in public health can be greatly accelerated by computer and
information science.

All in all there is reason for hope, barring total economic meltdown
and the end of civilization as we know it.

In retrospect, I continue to hope Obama keeps his promise about doubling the NIH and NSF budgets. However, we have thus far in the past few months spent trillions of dollars on bailing out Wall Street, and a few billions on bailing out the auto industry. I wonder how much money will remain once we embark on our New New Deal.

Broken Is Fixed: Health informatics standards in light of Clay Shirky's "In Praise of Evolvable Systems"

I recently re-read an old essay by Clay Shirky (shirky.com) called In Praise of Evolvable Systems. It opens with a laughable description of designing a novelty protocol to slip past the IETF as a joke - however, the feature set described is an accurate description of the World Wide Web, as it existed when Shirky wrote the essay in the late 1990's. That lead-in ends with this:

HTTP and HTML are the Whoopee Cushion and Joy Buzzer of Internet protocols, only comprehensible as elaborate practical jokes. For anyone who has tried to accomplish anything serious on the Web, it's pretty obvious that of the various implementations of a worldwide hypertext protocol, we have the worst one possible.

Except, of course, for all the others...

It seems the WWW protocol family has turned out to be another example of "less is more", or more accurately, "broken is fixed".

Healthcare Scenario Planning: All systems go

Science is an evolutionary driving force; discoveries go on all the time, largely independent of other major influences like the economy or politics. Every so often, though, a revolutionary paradigm shift occurs that surfaces in multiple forms, each of them radical in its own right. Systems biology is a good example.

The Economist ran an article this week on systems biology that focuses mostly on - as one might expect - the economic implications.  Here's the link, with the caveat that you may need an Economist subscription to look at the full article: Biology | All systems go | Economist.com.

The rationale for systems biology is described in the article as follows:

Mindfulness meditation, anxiety, and mood symptoms

As a 33-year meditation practitioner, I am always interested in medical studies related to meditation. I enjoy reading the Positive Technology Journal blog because its author, Dr. Andrea Gaggioli, provides almost unbelievably extensive coverage of emerging technologies and their intersection with psychology and neuroscience. His latest post, Does mindfulness meditation improve anxiety and mood symptoms?, looks at recently-popular therapeutic applications of the Buddhist concept of mindfulness (also known as "bare attention" or "choiceless awareness"). Among these applications, mindfulness-based stress reduction or MBSR is one of the more popular, and one that is widely believed to be effective.

The Future of Leadership - The Enabler

CIO Insight recently provided some insight into where biomedical informatics is heading in an article entitled The Future of Leadership - The Enabler.  It tells the story of Sandy Aronson, Director of IT for the Harvard Medical School–Partners Healthcare Center for Genetics and Genomics.  He's wrapping up an advanced degree, not in computer science or business management, but in molecular biology. 

It's not often that an IT professional earns an advanced degree in an entirely unrelated field for the sake of his job. But to Aronson, it makes perfect sense. "As IT professionals, it's really important to understand the content area we work in with as much depth as possible," he says. "You need to understand the non-IT processes you are working to support, so you can play a positive, efficient role in the process." For Aronson, that process isn't just about cutting costs or improving productivity. Aronson's group supports a user base of more than 600 researchers, scientists, technicians and healthcare professionals, all working on genetic-based healthcare, a concept that could re-write the rules of modern medicine.

I understand totally where Sandy is coming from. My own Master's, which I'm inches from completing now, is in information science with a focus on Human-Computer Interface. I've learned a lot, but in the meantime I've been helping to maintain existing and design new clinical research informatics systems. I've had to learn continuously about many things there - HIPAA; the Common Rule for the protection of human subjects; 21 CFR Part 11 and other aspects of the FDA's Good Clinical Practice predicate rule as they apply to computer systems; information security; the culture of a major academic health center; clinic and laboratory practice; and even a smattering of biology and medicine.

I'm betting Sandy's experience is similar. No matter what you study in school, there is so much more to healthcare informatics that no single course of study is going to provide all the knowledge you will need.

It's an exciting time to be in biomedical informatics. I've written about the synergistic convergence of multiple disruptive information technologies in my Rogue Wave scenario paper, and this field exemplifies perfectly the characteristics of the convergence.

Disruptive technologies are uncommon but not rare. The twentieth century witnessed the birth of numerous such technologies. The major events of the past twenty-five years in the world of IT have been the result of the ascendance of two different disruptive technologies about a decade apart. The first disruptive technology was the introduction of the personal computer and in particular the IBM Personal Computer and Microsoft’s DOS operating system in the early 1980’s. The second was the emergence of the World-Wide Web in the early 1990’s. Each of these changed the face of life in the advanced economies of the world in ways most of us would certainly call significant. Yet the changes we have thus far experienced pale in comparison to what the next four years hold in store. The reason: as we stated earlier, what we are about to experience is not just another emergence of a single disruptive technology, but the synergistic convergence of multiple disruptive technologies.

What do we mean by “synergistic convergence”? Synergistic convergence means the technologies harmonize, their disruptive forces melding in additive fashion. Rather than mitigating the effects of each other, their energies join just as the energies of the sea combine to form the rogue wave.

Information technologies are by definition enablers for other technologies, which on the surface may not appear related to IT. I used to work in the automotive industry, and what I saw there was that IT was a pervasive influence, affecting the processes of design, manufacturing, and sales, and enabling better and faster diagnostics in the service bay.

This picture is very incomplete, of course.  A modern automobile typically has multiple local-area networks embedded in it, with almost every dynamic aspect of the vehicle's operation under computer control.  High-end vehicles are beginning to use information technologies like Doppler radar to implement adaptive cruise control, and can communicate by satellite with monitoring systems that respond instantaneously to events like extremely sudden stops and the deployment of the vehicle's air bags.

Modern automobiles are complex beasts, but any single organ in the human body is orders of magnitude more complex, operating under the dynamically adaptive control of genomic, proteomic, and metabolomic information systems (technologies?). Aronson's academic studies have no doubt provided him (directly or indirectly) with insights into the ways in which such systems can be represented and simulated.  Even if his work is in the mundane arena of data processing that is used to manage clinical or basic science research operations, this knowledge will pay off.

I'm glad I focused on HCI, having seen how abysmal are even the best user interfaces in research and clinical IT, but if my budget and my wife will allow it, I can foresee further education in my future. It may focus on the complex knowledge domains on which Sandy has focused in obtaining his Master's degree. On the other hand, there are equally complex challenges in better understanding the cultural and socioeconomic dynamics operating within and between healthcare research organizations. It's a difficult choice.  The Chinese curse applies: we have indeed been born into an interesting time.

[BIOMEDICAL RESEARCH] Silicon Valley futurists predict creation of synthetic life-form within a year

i'm catching up on a few weeks of not being able to pay much attention to blogging, so I'm going to be writing about some news items that are not totally up to date. This one, though, is about something that doesn't need to be from today's news to be riveting.

Silicon Valley futurist Steve Jurvetsen predicts the creation of the first synthetic life form within a year, according to a March entry in the HYPERtext blog entitled Silicon Valley looks beyond Silicon (and the Valley) for 2008.

The first synthetic life form will be created - another mind-blowing prediction from Steve Jurvetsen. Steve says that within a year we'll see the creation of a bacteria-sized synthetic organism designed for either a biofuel or medical application. Apparently, UC Berkeley professors are already working on a mutation of a bug that excretes an anti-malarial compound. Wow.