but typical of people who don't live here. The city's defining characteristic is its huge variety, such that if you come here to confirm your initial prejudices, you can do so easily and go away comfortable that you know what LA is like, and write about how it's an even more segregated and racist society than the deep south. And in some places, you're right, it is. You can find what you're looking for here, negative or positive.

he's experienced with situations like this and has demonstrated willingness to jump in and help out, he's more popular than Cheney and Bush combined, and he's in better health than Cheney.

Someone's already mentioned Chuck Hagel, although his voting machine company associations make me uncomfortable. Still, there he is. From the same state, Coach Osborne is being talked about for gov. Another repub senator who actually has made a presidential run and is truly a decent person, man of integrity, all that, is Dick Lugar - rock-ribbed conservative, etc, but an honorable man. And I think he might be younger than McCain.

For the dems, it's time for a west-coaster or three. LA Mayor Villaraigosa is a rising possibility, I like Barbara Boxer a lot more than Feinstein, and isn't there a very liberal physician from Washington state?

it seems cars these days are not what they used to be. In the old days you could build up a lethal concentration of carbon monoxide gas in a garage easily. But then along came emissions controls and the like, and as a consequence emergency rooms are seeing quite a few people who tried to go out the CO route and didn't make it - the gas levels don't build up, garages aren't all that air-tight, the car runs out of gas and CO levels drop, the patient loses consciousness but doesn't die before being discovered - with the result that there's been a significant rise in numbers of people with acute CO poisoning showing up, alive but damaged, in emergency rooms.

excellent bit of reporting. Although it was interesting, after reading the Salon coverage of Mardi Gras, to see the contrast with the mainline reporting.

A minor point, Sidney - it's "breech".

no relation to the scary bird flu, but it has been pretty widespread, ie schools have noticed it, and an alarming proportion of cases go into pneumonia, even in young patients, and that's after a week or so of a high fever. I've heard of some patients developing gasteroenteritis after the pneumonia, although that might be just coincidental. It's been a pretty nasty outbreak, but has received almost no press attention - although your state and local epidemiologists certainly know about it. Bird flu typically becomes much milder when it (if) it does migrate to people. There was a huge outbreak in Pennsylvania, mostly, that ruined a lot of poultry farmers, and that was in the late 70s or early 80s - but that never migrated into humans.

first, when you screen a mostly healthy population for a rare disease, even when your test is so good that it catches 99.5% of the true positives and 99.5% of the true negatives, you can be overwhelmed by the false positives, who by your screen will _look no different_ from the true positives, and who will now be very upset to think they might have a very scary disease. So what you then do is take that whole group, which can be overwhelming, and do further tests. How much do those further tests cost in physical pain, time, anguish, and money? This is a problem, and is not confined to medicine - just look in any airport and you'll see a screening test which picks up way less than 99.5% of the true positives while catching hordes and hordes of false positives.

The second problem has to do with something called lead-time bias. Suppose a patient develops a disease that runs a course of 15 years before death. Symptoms normally develop around year 10, let's say, and the patient gets in to be diagnosed around year 12. Three years later the patient dies. But, a new screening test is developed, that allows the diagnosis to happen in year 7, let's say. The patient still dies at year 15, but apparently has a much longer survival after diagnosis because we've pushed diagnosis back five years, from year 12 to year 7. This is a major problem in cancer screening and is often unrecognized even within the medical profession.

The final problem is even more subtle, and is called length-based sampling. Very briefly, it means that when people come in for periodic screening, and get a cancer diagnosed, they will be the people with the less aggressive cancers (with longer survival) because they manage to make it the year, or two or three, between screenings without becoming symptomatic. The symptomatic tumors, however, develop quickly and so are caught by symptoms rather than by screen - and those are the ones over whom we cluck our tongues regretfully and say, "if only she'd gotten her mammogram, we could've caught it." Probably not.

If you're interested in a clear, more detailed explanation, I highly recommend Leon Gordis' fine book, entitled "Epidemiology". (I am not Leon, and am only an acquaintance, not a friend or publishing agent). A PubMed search on screening, especially mammography, would be revealing - one very good paper addressing the methodologic problems that came out maybe 10 years ago has Polly Newcomb as lead author.

and a way to start creating some buzz. And to draft them if they're reluctant. I'll start: Dick Lugar. Perhaps Chuck Hagel.