design challenge update
This is an entry from 27/4/2005 which blog program didn't
seem to want to publish so trying again.
A few hours ago I emailed my project code and writeup to the
Circuit Cellar contest coordinator with mere hours to go before the
deadline. It was probably one of the most satisfying things I've done
in a while as when the project was over, I had a very useful system
which I'm just starting to use on a regular basis to further
characterize physiologic variables during sleep.
Now that my project is in the hands of the judges, I can
freely discuss it. In my previous blog posting I described what I was
doing so I don't have to repost the material. RSN I will be putting the
full source code and schematics for the units I have built on my web
site, but this entry is less about the details of the project than it
is about how neat it is to have actually pulled this off.
The above material was written last week, and then I went off
on a long tangent that really had nothing to do with the project.
Hitting delete solved the problem.
Now that I have had a week to think about what I've done,
I've concluded that doing this project was a really good thing as it
has gotten me back into quantitative medicine. Medicine, as it is
practiced for the most part today, is very qualitative. It was one of
the things that really bothered me about medicine when I first became a
medical student, but then I began to realize that an experienced
physician was a very good diagnostic instrument that utilized unique
data acquisition methods, performed calculations in non-deterministic
algorithms buried in wetware and produced surprisingly accurate
diagnoses. I was suspicious at first, but eventually these diagnostic
algorithms were programmed into my wetware through the process of
internship and residency.
I've practiced qualitative medicine but it bothers me to do
so. My scientific background is too strong and I have a need to measure
things as I don't believe you know anything about an object or process
unless you have quantitative data. Medical expertise is very nice, but
what has finally convinced me to get back into quantifying things is
patient responses to my diagnoses. What I've found is that patients
like quantitative data, the more the better.
Where qualitative medicine still reigns is in psychiatry. I
really enjoy psychiatry, but what turned me off this path was the
perverse satisfaction that many psychiatrists had with their specialty
being non-quantitative. During my psychiatry rotation as an intern, I
commented about the lack of measurable variables in psychiatry and the
psychiatrist who I was working with at the time tried to stress the
advantages of psychiatry over other specialties; no equipment aside
from the psychiatrists brain, a pad of paper and a pen. He didn't
convince me although I did learn a lot of psychiatry from him.
The most difficult task I've had as a physician is trying to
convince patients that they have a psychiatric diagnosis. To me, the
diagnosis is obvious and when I am able to get patients in to see a
psychiatrist, they almost always concur with my diagnosis. The main
criticism I hear from patients is: "what evidence do you have that I
have this diagnosis?". Actually I have plenty, but patients have a
rather low opinion of the psychiatric evidence I have carefully
collected from them using a structured psychiatric interview. A
gastroenterologist would have the diagnosis of a gastric ulcer based on
history alone, but they will then perform an endoscopy and provide
patients with high resolution color photographs of their ulcer.
Patients love photographs of their GI tract, readily share them with
family and friends and post them on personal web sites. Showing a
patient a picture of their ulcer, or other dramatic GI lesion, makes
them much more ready to accept a physicians diagnosis and suggested
There are quantitative tests in psychiatry such as HeartLink
monitoring which has high sensitivity for diagnosing major depression
based on the results of 24 hour heart rate patterns. A HeartLink report
consists of a graph of minute averaged heart rate data over 24 hours as
well as movement data. When I've performed HeartLink monitoring on a
patient, I can show them their heart rate graph and representative
graphs of various psychiatric conditions. This approach has led to far
greater acceptance of antidepressant therapy in patients and much
better compliance than if I'd simply told them they were depressed.
Also, if the HeartLink monitor shows a normal pattern, I reconsider my
Now will try to bring this back to my Zigbee project which
involved wireless monitoring of patient movement and respiration during
sleep. This was a modest start to my goal of having a person totally
instrumented so that multiple physiologic variables can be measured in
real time and on an ambulatory basis. The idea is to walk around
engaged in normal daily activities while multiple wireless monitors
quantify every step, every heartbeat, time breaths to msec precision,
measure eye movements, EEG, peripheral oxygen saturation, skin
conductance as well as lots of physiologic variables that don't come
immediately to mind. This is all possible now, and in the near future
one can go internally with realtime monitoring of blood glucose every
second, pCO2, pH, electrolytes, testosterone, cortisol and a myriad of
blood constituants which are now measured at a single time point every
few weeks with considerable effort.
The ultimate monitoring system would be a system of nanobots
which would be ingested as a capsule which would distribute themselves
throughout the body measuring any physiologic variable one could think
of for a day or so and then downloading their data to a remote
When the first nanobot monitoring system is available, I
predict huge lineups of people waiting to be monitored as people seem
to have an immense need to find out what is going on inside them. Some
patients of mine think nothing of forking out $1800 for a total body CT
scan which seems to be used primarily to marvel at the neat 3D computer
reconstructions of their own body produced by software. I have no
worries about my place in this scheme since they have to come somewhere
to find out what it all means.
Based on the crude measurements I have of movement on
respiration during sleep, I've discovered a new world to explore. Sleep
is a key factor in many psychiatric conditions but something that very
few physicians even ask patients about. I've found that asking a
patient questions about sleep as well as questions relating to other
basic physiologic processes are far more useful than asking about mood.
by Boris Gimbarzevsky Comments