Bear Flag Republic Radio Weblog

"Major trouble" in U.S. computer science research. With everyone abandoning the field, future innovation will suffer. Blame goes to collapse of the IT sector, poor economic returns to those who pursue a Ph.D. degree, fewer foreign students (post 9/11) and of course, everyone's favorite scape goat, middle school algebra. He does acknowledge that "pure programming" jobs will all go offshore.

Interesting quote: "We are emphasizing a lot more with our students that they need to understand something besides computing--like business, biology, chemistry, mapping, geography, information retrieval or history."

My own observation is that it is probably better for many students to major in business, biology, chemistry, etc, and to minor or take courses in computer science. The demand for pure computer scientists is likely to be in decline as those "pure programming" jobs move offshore. But we do not need to train computer scientists who know something about business or chemistry. We actually need business and chemistry professionals who their way around computer models.
[Edward Mitchell: Common Sense Technology]
comments < 12:25:02 PM        >

U.S. Centers for Disease Control. It is that time of year when we must be whipped up into an influenza hysteria.

This widely distributed article summarizes presentations given at a recent medical conference, beginning with "Recent flu seasons dominated by more severe strains of flu and an aging U.S. population that's more vulnerable to disease have caused a jump in the number of hospitalizations and deaths from the respiratory illness". The article continues with "There has been a startling increase in the number of hospitalizations for flu" from 114,000 "a few decades ago" to 200,000. In the next sentence, the report states that mortality due to influenza has grown from 20,000 just 20 years ago to 36,000 last year. (Update: 9/25 - the linked Forbes article is now hidden behind a full page ad. It appears that the Forbes writer actually made up most of the news story, as I explain below. I did find the actual press release from health authorities and they did not say the things that the Forbes writer put in the article. Another news outlet misquotes the whole thing saying that flu causes 114,000 hospitalizations per year.)

Your junk science detection system has hopefully activated when seeing these numbers. For instance, this massive increase in influenza has occurred simultaneously while vaccinating vastly larger portions of the population? Hmmmm. Something is missing here. Does influenza vaccination cause influenza? Of course not, which suggests there are problems with the data or how it is being presented.

What happened and who forged the documents this time?

On January 7th, 2003, the Centers for Disease Control issued a press release noting that they had made made errors in their past estimates of influenza fatalities and that the estimate of 20,000 was much too low and they were revising upwards their estimates to 36,000 due to statistical issues.

The huge increase was based on a correction of erroneous past data estimates from 20,000 to 36,000. The CDC says so itself. Also look at this chart: Influenza Epidemics in the United States, France, and Australia, 1972[^]1997. This chart shows the number of people who have died, increasing a bit, over time, in the U.S., probably because our population is increasing every year (just over half due to in migration). Stranger, perhaps, is that the estimate of 36,000 is different than that issued in another CDC report.

Where does the estimate on 114,000 "a few decades ago" increasing to 200,000 hospitalizations come from? It appears to come from the study reported in this press release (JAMA, Sep 15th, 2004) which compared estimated hospitalizations due to influenza in 1979-1980 to estimated hospitalizations in 2000-2001. Unfortunately, typical of medical literature, most of it is not easily accessible to the general public so I am unable to directly access the study. An abstract is available here. From the press release, the authors did not report a figure of 200,000 people being hospitalized - instead, they estimated 200,000 more hospitalizations than previously estimated (but since the Forbes article is so messed up, I am not certain that the health care folks even said that.) However, as reported here, there was no increase in sick people; only the estimate was improved to more accurately measure who was hospitalized due to influenza.

I tried finding more data on hospitalizations due to influenza, and found this report from the American Lung Association, using data from the National Center for Health Statistics, Report of Final Morbidity, 1979 to 2001. According to Table 4, the mortality rate (per hundred thousand) went from 20 in 1979 to 21.8 in 2001 with a range from 20 to 34 (in 1998). Table 5 confuses me, however; it shows the age specific mortality rate in 10 year age groups declined in all age groups except for those over 85. Yet the total mortality rate for all age groups went up? It may be skewed by the 85+ category. The data in Table 5 implies that vaccination is working as do many studies. Since the influenza virus is different from year to year, as is the effectiveness (or ineffectiveness) of the vaccine, it makes little sense to compare one season to another single season 20 years later. We do know that influenza vaccine effectiveness falls off sharply the older we get - and that alone may account for the increase in mortality at age 85 and above, which in turn, resulted in a 9% mortality rate increase from 1979 to 2001 (which is well within the normal variability from year to year).

It looks like the influenza media hype machine is already in full swing and off to a great 2004-2005 season, replete with bogus scary numbers that exist only as a statistical correction. The hype helps sell newspapers, magazines, and eyeballs to TV advertisers. The story does not need to be based on facts, of course, as long as the goal of vaccinating more people is met! This one was even coordinated with Aventis, the world's largest manufacturer of influenza vacines.
[Edward Mitchell: Common Sense Technology]
comments < 12:24:22 PM        >

A Liquid That Turns Solid When Heated [Slashdot:]
comments < 12:21:32 PM        >