Apr

13

 I am a physician who practices Anesthesiology at a community hospital. I recently attended an academic meeting in which a well known critical care physician gave a couple of lectures relating to glucose control in critically ill patients, and evidence based medicine. The physician in question, Dr Avery Tung, is a faculty member at the University of Chicago, and discussed his collaboration with some of the members of the University of Chicago Business School. I found his observations enlightening, and I thought that some of the key points might have some relevance for this forum.

The biochemistry of glucose metabolism and insulin's role in regulating it have been understood since the 1920s. Low levels of glucose are known to be dangerous, and high levels are known to lead to atherosclerosis, coronary artery disease, stroke and peripheral vascular disease. Very high levels are known to acutely lead to acidosis and coma. It is also well known that the closer one can get one's glucose into the normal range, the less likely that one suffers long term consequences. Short term, moderate fluctuations in glucose do not demonstrate any short term disability in healthy patients. Until recently, no one thought to ask what the affect of modest fluctuations in glucose would have on the critically ill. It is not unusual to see modest increases in glucose in the critically ill who are not diabetic as a result of the stress of illness. Van de Berghe and his colleagues published an article in the New England Journal of Medicine in 2001 that prospectively put Surgical Intensive Care patients into either a "standard " insulin regimen of intramuscular insulin for serum glucose of greater than 220, or an intensive regimen of intravenous insulin designed to keep the serum insulin less than 110. Only 13% of the patients had a history of diabetes. The results of the study were shocking. The group that had very tight glucose control had a 50% reduction in mortality compared to the "standard " therapy group. No other finding published in the intensive care literature has had the potential to improve outcomes as much as this one. The key breakthrough was that someone had the insight to question whether well understood data for non-acutely ill patients held for the acutely ill. Nobody asked this question for decades.

 The second lecture on evidence based medicine covered a number of topics. I wish to discuss one point in the lecture, the use of the pulmonary artery catheter. The pulmonary artery catheter is a flow -directed catheter that is inserted into a central vein and is advanced slowly into the pulmonary artery. It directly measures the pressures in the right side of the heart and the cardiac output (the pumping capacity), and indirectly measures the pressures in the left side of the heart. Since its introduction in the early 1970s, it has ben instrumental in training Intensive Care physicians and directing treatment algorithms. One of the most difficult clinical decisions can be whether low blood pressure is the result of sepsis (infection), or cardiac pump failure. The PA catheter had been at the forefront of the decision making and treatment algorithm for this issue more than two decades. Since the mid nineteen-nineties, multiple studies have been done trying to demonstrate a survival advantage in using this monitor as part of the care of Intensive care patients, and none has been able to do so. The first key point is that it took two decades to question whether this monitor made a difference. The information didn't appear to change outcomes. The other point is that while multiple studies have come out which have not demonstrated a survival benefit, the use of this monitor, while declining, persists. There was an economic incentive to place these monitors in the past, but managed care has removed much of the economic incentive. A generation of physicians used to making patient care decisions with certain treatment algorithms may be finding it difficult to make decisions with less information, even if the outcome is no different

The lesson for traders is never to assume that a question has been asked. In looking at conventional wisdom, investigate how a consensus has been reached. The question may not have been asked, or if it was, broad generalizations made that do not pertain to smaller subsets. The most spectacular results may come from investigating and challenging long held assumptions. Data may help you in you decision algorithms, but your decisions may be no better that guessing if you don't test whether the data affect outcomes. Understanding the mindset of those using data which you know to be useless may help you to decide to take the opposite side of a trade.

Yossi Ben-Dak remarks:

It is always the particular insight that comes from taking a fresh, uninhibited look at the system as a whole but also at the components that sometimes, or actually quite often, are considered by the majority, as of less than pivotal interest, that produces the struggle with the true optimum. This is especially true in engineering an effective technology solution that benefits from other fast changing sub technologies for a constantly redefined purpose range, as compared with technologies that are more purpose-limited and are from known manufacturers.


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search