Doing some quick searches in response to our co-blogger’s co-blogger’s post about the 1918 Spanish flu epidemic, I came across the following chart detailing the ratio of reported cases to deaths in San Fransisco. Not only is it a priceless statistical representation of panic, it also captures the malleability of even professional opinion. To wit:
I’ve highlighted the number of cases in red because blood is the color of riot—and for legibility. With certainty, we can say the author of this study, W.H. Kellogg, captured something of cultural significance when he rocketed his data up and off the y-axis. But the convergence of the incidence and death rates between the 23rd and 30th of November may be even more interesting. How do we account for the fact that, for one short week, everyone who caught the disease died from it? Easy:
According to the 21 November article, because public health officials claimed that “the influenza epidemic had been stamped out,” at noon “[t]he shrieking of every siren in San Fransisco, blowing of whistles, clanging of gongs and the ringing of bells will . . . signal for throwing away the gauze face coverings” (9). Why were there no more new cases reported than there were deaths the next week?
Because someone said there wouldn’t be. So there weren’t. People caught colds and had the sniffles, but it wasn’t Spanish flu. Couldn’t be. The epidemic was over. Did you somehow sleep through the infernal cacophany last Tuesday? The city has no more need for mass-prophylaxis. Everyone who catches the bug now brought it with them on the boat, and everyone knows you can’t catch flu from boat-people. Wait—what do you mean, “How do I think it got here in the first place?” What? How come nobody told us—quick! Everyone! En masque en masse!
So said the San Fransisco Chronicle on 4 December, and back up the panic-axis we go . . .
5 comments
January 26, 2009 at 1:19 pm
Vance
It’s a solution to the charting problem Eric was facing a few weeks ago — suitable in this case because Kellogg doesn’t want to communicate more about the peak of the rate of cases reported than that it was high.
January 26, 2009 at 1:36 pm
Crazy Little Thing
From The Simpsons
Homer: Not a bear in sight. The “Bear Patrol” is working like a charm!
Lisa: That’s specious reasoning, Dad.
Homer: [uncomprehendingly] Thanks, honey.
Lisa: By your logic, I could claim that this rock keeps tigers away.
Homer: Hmm. How does it work?
Lisa: It doesn’t work; it’s just a stupid rock!
Homer: Uh-huh.
Lisa: But I don’t see any tigers around, do you?
Homer: (pause) Lisa, I want to buy your rock.
January 26, 2009 at 2:27 pm
michael holloway
I note that the profile, the pitch and duration of the first diagnosed line(Oct. 5) is identical to the re-diagnosis line(Nov. 23). A predictable function of a specific zeitgeist?
Is the graphical expression of communal fear similar to communal bliss?
Could one match a profile in a current negative hysteria graph to an approval rating graph of say, Obama-mania?
Or is it just a rock?
mh
January 27, 2009 at 4:35 am
hmprescott
Since I teach history of medicine, just thought I’d make some comments about this:
1. Since influenza is a virus, masks are useless in preventing the spread of the disease.
2. You’re assuming that the diagnoses and deaths at the end of November are the same people — this is not necessarily the case.
3. Keep in mind that public health officials in late November did not know that the number of new cases and deaths would spike a week later. From their perspective, the plummeting figures indicated that the epidemic was declining.
January 27, 2009 at 2:03 pm
Chris J
A couple of comments:
True — masks don’t block viruses. But they often do have indirect effects. They remind us that things are not normal, so we are more careful. For example, we may avoid dangerous contacts or we may wash our hands more (which does prevent viral spread). We see this phenomenon in standard hospital strict isolation protocols, which do include masks: the masks end up helping reduce spread, but not for entirely straightforward reasons.
Another possible explanation is that the case-fatality rate for the initial burst of infections was much higher than that of the subsequent spike in infections. This could be because the most susceptible persons died in the first burst. This would not be unusual for viral illness.