I'm not talking about reassortment but about co-infection (which might in turn lead to reassortment but doesn't necessarily do so). If a patient is infected with two strains simultaneously, will they both be picked up by testing, or will the test only pick up one of the strains?

It's easy to construct “just so” stories which would allow two virus strains to assist each other to spread. For example, one virus could irritate the airways, causing coughing, while the other increases fluid secretion, meaning that the coughs are more productive, spread more droplets, etc. In this case, the two viruses together would transmit more efficiently than either virus alone. This is a possible reason for the unusual out-of-season spike in the seasonal H1N1 and H3N2 subtypes – if swine-H1N1 was somehow facilitating their spread (and/or vice versa).

However, if this is true, it suggests that some proportion of the patients typed as having seasonal flu are in fact infected with (seasonal flu AND swine flu), and conversely that some proportion of the patients typed as having swine flu are in fact infected with (swine flu AND seasonal flu). I'm asking how likely it is that the serotyping procedure could be missing / mis-scoring double infections as being only one strain or the other.

If increased testing were the only factor, I would (naively perhaps) expect seasonal flus and influenza B to increase proportionally to each other. The fact that we see a further increase in the seasonal flus over and above the increase in influenza B is what leads me to suspect there may be something else going on here – some kind of cooperation between H1N1 and the other seasonal influenza A strains, where they facilitate each other's transmission.

To be crystal clear what I mean by a mixed infection I mean a single person infected with H1N1 and H3N2, who then coughs on someone else, they catch both H1N1 and H3N2, and so on. Depending on how well the two viruses replicate within the host, one patient could have “mostly” H3N2, while the next could have “mostly” H1N1, and some could have a balanced mixture. What would the serotyping tests make of that? Would they be able to pick up a mixed infection? Or would it show up as either one or the other, or a noisy “unsubtypable” category?

Coinfection might of course lead to reassortment if a single cell is infected by both viruses, but that's not necessarily the case. The seasonal flus could be infecting the upper respiratory tract and H1N1 the lower respiratory tract. Or reassortments could simply be unstable / selected against and thus not seen.

with 100000 estimated people having Mexflu in USA and another 100000
having seasonal flu, the chance of coinfection is only 1e-7 or 33 cases in USA

This extends the “paraspecies” concept you covered a few posts ago to a wider level of grouping, with a population of different flus transmitting together.

The increase in positive rates for season flu was in week 17. If seasonal flu positive rates continue to drop in week 19 and beyond as they did in week 18 you're looking at the expected pattern other than an across the board increase in positive rates in week 17.

If looking for differences at week 17 it would be good to also look at the weeks just prior. If submission practices at a time of high concern overstate flu rates, then the tail end of the flu season might be a time where submission practices understate flu rates. I wonder if there are any papers on a late season under reporting effect?

Maybe find the timing and location of the bottlenecks and fight flu there !
Either by killing it or by introducing milder strains to replace virulent ones

Three influenza-associated pediatric deaths were reported to CDC during week 18 (Arizona, California, and Texas); one was due to influenza A virus (rapid test positive) and two were due to influenza B virus infections. The deaths reported this week occurred between April 5 and May 2, 2009. Since September 28, 2008, CDC has received 59 reports of influenza-associated pediatric deaths that occurred during the current season.

This is very disturbing to me… The media are completely missing this tragedy because they aren't deaths caused by the new doomsday H1N1. Sorry for the cynicism but I think there is a loss of perspective here.

How else can you explain the upsurge in influenza @ week 17 other than patients and physicians both being on high surveillance mode looking for novel H1N1 due to saturation media reporting? Could there be another explanation?

Which begs the question of the accuracy of the normal baseline numbers. And it would be interesting to see if this reporting spike had any other particular characteristics relative to baseline: geography, race, urban/rural, income, education.