Thursday, September 24, 2015

Physician's choice of intervention - DNR

This doctor has a tattoo that reads, "No CPR" as a living will.

Unfortunately many of us will, one day, have to make choices about aggressive interventions aimed at prolonging our lives. The default is, of course, do everything possible since life is precious and our family and friends want to keep us around. It has been known for quite sometime however that doctors themselves are less likely to opt for aggressive interventions near the end of their natural lifespan. I first heard of this on an old radiolab podcast I listened to recently, The Bitter End.

Somewhat related is a freakonomics podcast I also recently listened to about the costs associated with end-of-life care and whether we should be given the option to take the money instead of the intervention.
I don't know what I find more disturbing: that some doctors give interventions to their patients that they would not have done to themselves, that some doctors are so traumatized by what happens frequently during end-of-life interventions, or that some of those with the most experience think that modern medicine has the wrong emphasis for end-of-life care. Maybe I'll ask for the money.


Friday, July 03, 2015


If I didn't break my finger mountain biking a couple of weeks ago, I'd like to be on my bike right now. So I've been watching lots of videos about bikes. Most are of the !Xtreme MTB! variety, however there are also some great Bayblab worthy videos that I would like to share.
Bikes are a pretty cool intersection of physics and muscle memory. This first video is about the physics of how bikes are so stable.

This second video is about the backwards bicycle and some insights into neuroplasticity
And then, of course, bicycles and robots.


Monday, April 27, 2015

Dr. Siri is hiding on your wrist

The Apple Watch harbours undisclosed hardware capable of measuring blood oxygen content as revealed by iFixit's teardown. The hardware is not activated yet, but I don't understand how securely it has been made inaccessible by third party software. There is speculation that Apple is waiting upon FDA approval to enable the device for medical indications.
The currently inactive hardware is a pulse oximeter which, as previously mentioned, enables the noninvasive measurement of blood oxygen content. You may recognize a pulse oximeter as the red light that is clamped onto a patients finger in a hospital.
If the FDA approves the device for any medical application the implications would be interesting. The amount of health related data that could be collected would be enormous, and would likely reveal some interesting and unexpected correlations.
I'm somewhat unclear as to the range of potential applications for the individual user. The pulse oximeter seems to have applications in an acute medical setting, but I'm not clear on the usefulness in an everyday setting. It may have applications for analysis of sleeping disorders, chronic disease, or for serious athletes doing high altitude training, but otherwise I'm not sure of its utility. Could it be useful for alerting care givers to out-patient emergencies? Perhaps the use of the device will become apparent once deployed on a large scale and everyday activity data is collected.


Monday, April 13, 2015

Time your next heart attack to conincide with national cardiology meetings

Surprisingly, high-risk patients with heart failure and cardiac arrest admitted to US teaching hospitals during dates of national cardiology meetings had lower 30-day mortality rates. The surprise is that outcomes improved despite the absence of the cardiologists who attended the meetings. While the cause of this correlation and the generalizability of the methodology are unclear, the finding is very significant. A Freakonomics podcast covering this study expressed the magnitude of this effect in a powerful comparison. While the combination of common interventions (beta-blockers, statins, aspirin, and blood thinners) reduce mortality risk by 2-3% in these patients, this effect reduced mortality risk by up to 10%. The Freakonomics podcast also entertainingly asks some cardiologists attending a cardiology meeting about the findings. The most compelling reason presented in the podcast to explain this effect is that the health professionals not attending national cardiology meetings use more conservative interventions during this time.


Wednesday, April 08, 2015

The rising cost of cancer drugs

A recent research letter in JAMA oncology presented a quick analysis of the costs of cancer drugs. The article cites evidence that cancer drug prices are rising faster than prices of drugs in other therapeutic areas. The authors found no significant price difference between next-in-class drugs and novel drugs. Drugs that were granted US FDA approval based upon disease response rate were priced significantly higher than drugs approved based upon overall survival or progression- or disease free survival, however no significant relationship between cost and the percentage improvement in end point was found. The authors concluded that current pricing is "not rational but simply reflect what the market will bear." The price which the market will bear is indeed rational from the perspective of a business though I would think. There was also no consideration given to the costs of drug development but perhaps it is insignificant, I don't know. Nonetheless, an interesting analysis that makes me wonder about drugs in other therapeutic areas.


Friday, February 13, 2015

Dendroclimatology: The divergence problem

I was reminded about dendroclimatology when reading a book about the geological evidence of climate change by E. Kirsten Peters. Dendroclimatology is the study of inferring past climactic conditions based upon tree ring width and/or density. The resulting data is high resolution since a tree ring is formed every year. Tree ring width and/or density correlates well with various climate parameters like sun, water, and temperature. Using techniques from dendrochronology (tree-ring dating), long climate records of thousands of years can be reconstructed using this technique. For example, summer temperature anomalies for the past 7000 years in Siberia were constructed using tree ring proxies in the figure below.

R.M.Hantemirov - Institute of Plant and Animal Ecology. Summer temperature anomalies of the Yamal Pennisula.

Dendroclimatologists have developed methods to ensure that the samples examined contain tree ring properties that best reflect only the climactic parameter of interest. Despite this there are confounding factors, as outlined in the dendroclimatology wikipedia entry, including nonlinear responses and environmental conditions and events that can otherwise affect tree ring width and density. The most interesting confounding effect has only been evident since the 1950s and is known as the divergence problem.

The divergence problem was first identified in Alaska by Taubes (1995)[1] and Jacoby & d'Arrigo (1995)[2]. The recognition that this problem was widespread in high northern latitudes was published in 1998 by Keith Briffa[3]. A study by Cook in 2004[4] demonstrated that the problem is unique in the past 1000 years, suggesting the possibility of an anthropogenic cause. The problem is that, in northern latitudes, tree ring proxy measurements have diverged from instrument-based temperature data since the 1950s (see figure below). Growth of trees at these latitudes is declining despite instrument-based temperature data that would normally correlate with increased tree-ring width. The cause is unknown but it is likely to be a combination of local and global factors such as global warming-induced drought and global dimming.[5]

Twenty-year smoothed plots of tree-ring width (dashed line) and tree-ring density (thick solid line), averaged across a network of mid-northern latitude boreal forest sites and compared with equivalent-area averages of mean April to September temperature anomalies (thin solid line). (Briffa 1998)[3] taken from wikipedia

Dendroclimatology seems like a fascinating field that, given some reasonably inexpensive equipment, could be done as an amateur. It would be a great excuse for a hike or backcountry ski while collecting data and learning about botany, local climate, local geography, statistical analysis, and sampling methods.

[1] Taubes, G. (17 March 1995), "Is a Warmer Climate Wilting the Forests of the North?", Science 267 (5204): 1595–1526.
[2] Jacoby, G. C.; d'Arrigo, R. D. (June 1995), "Tree ring width and density evidence of climatic and potential forest change in Alaska", Global Biogeochemical Cycles 9 (2): 227.
[3] Briffa, Keith R.; Schweingruber, F. H.; Jones, Phil D.; Osborn, Tim J.; Shiyatov, S. G.; Vaganov, E. A. (12 February 1998), "Reduced sensitivity of recent tree-growth to temperature at high northern latitudes", Nature 391 (6668): 678.
[4] Cook 2004
[5] d'Arrigo, R.; Wilson, R.; Liepert, B.; Cherubini, P. (February 2008), "On the 'Divergence Problem' in Northern Forests: A review of the tree-ring evidence and possible causes", Global and Planetary Change 60 (3–4): 289. 


Wednesday, February 04, 2015

Skiing Robots!

While I welcome our new robot overlords, I'm unsure that I want them to be better at skiing than me. Hats off to a team from Slovenia and another team from the University of Manitoba for making robots that are cool enough to ski.