by revere, cross-posted from Effect Measure

Monday morning, start of week three of the official flu season (which began October 4). CDC’s scientific spokeswoman on the flu, Dr. Anne Schuchat has said we are seeing “unprecedented” flu activity for this time of year, including an unusual toll in the pediatric age group. What does “unprecedented” mean? It’s not very specific on what precedents are included, but if we confine ourselves to the three years before this one, we can get a good idea of just how unusual this flu season is. This week CDC unveiled a new graphic for their Emerging Infections Program (EIP) (I liked the old one better; this one is not very legible), the component of the surveillance system that tracks laboratory-confirmed influenza-associated hospitalizations in children and adults. It collects data in 60 counties covering 12 metropolitan areas of 10 states (San Francisco CA, Denver CO, New Haven CT, Atlanta GA, Baltimore MD, Minneapolis/St. Paul MN, Albuquerque NM, Las Cruces, NM, Albany NY, Rochester NY, Portland OR, and Nashville TN). It appears CDC has added six new sites (in IA, ID, MI, ND, OK and SD), but the data here are for the original 10 so comparisons can be made. This is a passive surveillance system that reviews hospital laboratory, admissions and infection logs at sites chosen to allow calculation of rates per population in the hospital catchment areas. Here is what the graphic looks like as of last Friday:


CDC Source, full size here.

Weeks are on the horizontal axis. There are five age group panels (0 – 4 years at the top, seniors at the bottom) and at the right side of each panel you will see lines gradually ascending from week 40 (October) to week 20 (April), CDC’s traditional flu season. Everything to left of week 40 in each panel is prior to or just after the official flu season and is essentially zero for the three prior seasons, i.e., during the “off months” we don’t see lab confirmed hospitalizations for flu (the prior seasons are color coded but I can hardly read the legend and the fact that I’m color blind doesn’t help). The reason the levels are close to zero out of flu season might be because we haven’t been looking very hard. After all this system depends on someone doing a test for flu on hospitalized patients and if they don’t believe there is any flu during the summer, the test doesn’t get done. But it is almost certainly true that whatever flu is around at those times of year, it is much less than during the usual non-pandemic flu season.

As the season progresses you will see that the cumulative number of people who were infected with flu per the populations in each of the age groups served by these hospitals rises. Think of this rate as reflecting the chance that someone in that age group will get infected with flu and be sick enough to wind up in one of the EIP site hospitals with his/her case confirmed by a lab test. Some years are worse than others and you can see that the rates of infected persons per 10,000 population ends up at different points by the end of the flu season in the spring (week 20) during the three different years (population rates, left vertical axis; case counts on right axis). Flu is like that. Not every year is the same. It depends on what the predominant seasonal flu types and subtypes are that year and probably lots of other things we don’t know.

The “unprecendented” year we are having shows up in the left of each panel. The rates are a solid line running across the bars. the bars are case counts (note to CDC: lose the bars. They are obscuring the picture). To see how unusual this is, look at the top panel (0 – 4 years), where the solid line has risen to 1.4 cases per 10,000 infants and toddlers in week 40 (first week in October). In the three previous years that seasonal risk level isn’t reached until the 3rd week in January. We are 3 months in advance of the last three years by this measure, and the case count for that age group is still rising. For the 5 – 17 year old panel the difference is more dramatic. We’ve already reached the risk level we would normally see for the whole flu season, and we are just getting started. That’s the group being hit the hardest, and half the fatalities since September 1 have been in the 12 – 17 year old age group, confirming that.

For adults between 18 and 50, the picture is like that for the under 4 year olds: about 3 months in advance compared to the bad flu year of 2007 – 2008 and already exceeding that for the entire flu seasons of the other two years. This is the age group populating the ICU beds. Even though population rates may be smaller, there are many more of them.

When you look at the 50 – 64 year old age group you start to see a different kind of difference. The bad flu year of 2007 – 2008 is now clearly visible and the main targets were the over 50, and even more striking the 65+ age groups. That’s typical of a bad flu season during usual times. You can see that the 50 – 64 is still being hit, well in advance of usual, but the 65+ age group looks more like the usual seasonal year, although even here we are in advance by months. But so far this year the swine flu hammer has fallen elsewhere and it is this change in the epidemiology — the pattern of influenza in the population — that is signaling the presence of a pandemic strain.

It is true that seasonal influenza kills a lot of people every year. They are mainly seniors, people like me. People care about seniors, of course, but our deaths are considered part of the natural order of things. Old people die. If it’s not one thing, it’s another. And of course our mortality rate is very high, compared to all other age groups. Even if the number of infants, children, teens and healthy adults double or triple or quadruple, the number of deaths may not ever reach what happens normally to seniors during seasonal flu, but the psychological and social impact is considerably greater. That’s one of the fallacies in comparing the numbers of deaths from this flu with the usual seasonal flu.

Where this is going to go, we don’t know. But as CDC’s Dr. Schuchat remarked concerning the increased mortality in babies, children and teens, what we are seeing is “very sobering.” That’s not because of the numbers of deaths, but because of who is dying. And as a parent and grandparent, I have to agree. Those deaths are very sobering.