November 12 Flu Update
Yesterday, we ran a story on the projection of a pandemic having severe health effects, but not a huge effect on the economies of large nation's. We had some comments on it, and now Effect Measure weighs in...saying we shouldn't base policy on something this uncertain.The Philippines has Tamiflu which will expire this month, so they are donating it to Indonesia and Cambodia.
We blogged this before, but it is worth another look. From Nature, a call for research during seasonal flu to find out how basic interventions like hygiene might effect flu transmission.
This is something others have done as well....the Macon County Department of Health is going to vaccinate large numbers of people (for seasonal flu) in one place and time, to see how it goes.
posted by Orange @ 10:51 PM
4 Comments:
Orange;
Oh boy, "model mania" drags on. I think Revere is as skeptical about the Australian McKibbon model as I was. Without examining the exact data and regression values that were input, closely, it is nearly impossible to determine the quality and accuracy of the conclusions. Revere is correct in saying that policy should not be made on something as inherently uncertain as this. The fact is though it will. U.S. policy for influenza planning is indeed being established and will be implemented, based on numerous U.S. agency parametric models of the pandemic. These models are kind-of similar to both the Australian and NIH types, in that they are based on independent variables whose properties and assigned data values, and are used to ultimately determine the characterics and predictive behavior of different pandemic scenarios. The government models are updated frequently as more information becomes available, so they are not static. Little if any of this information is made public - it's considered sensitive and/or classified, for obvious reasons based on the national security. The modeling information, as well as conclusions, are coalesced from agency experts, up to the highest cabinet levels, and planning decisions are formulated. The NIH/MIDAS model prediction that by closing international flights quickly, buys us a small amount of critical time to make preparations, is not weird - it's very credible and crucial.
Mathmatical modeling is how we routinely predict new generation automobile performance, it's how we plan for ultra-complex space missions, and it is also how the U.S. government is planning for the next ultra-severe influenza pandemic. Whether we like or not, our future may well be in the hands of the latest generation super computer (named Hal ?).
Wulfgang
Models are all about assumptions. And at some level, there is level of arbitrariness to them. Actually, I think the prevalent approach is to look at scenarios, which this study did. The policy question becomes which scenario do you plan for? As a society, we cannot plan for the worst case in every situation (who wants to live like that). What you need is a plan for a likely case--with a good idea what you would do from there if the worst case happens.
Signed Hal.
Orange;
Bulls eye on your over all comment. The federal government and our close allies are indeed planning and modeling multiple gamuts of senarios, from mild epidemics to super-type pandemics - all based on varying world wide and domestic assumptions. They do target the most likely and least disruptive case for our society and that's what is being worked publicly. However, within the government and the military, they must take it several steps further and develop viable strategies and options for a worst case, and be able to react accordingly. That's their job; the situation at that stage then clearly becomes a national security issue. They must be able to posture quickly to a developing worst case and maintain our fractured infrastructure, if it manifests.
What you are calling arbitrariness, we call "uncertainty" or "risk factors" - it requires excellent judgement.
Wulfgang
I have no doubt it requires excellent judgment. I think it requires some luck, too. Many things are too difficult to envision.
Post a Comment
<< Home