Thursday, May 14, 2009

60 days till midnight...

Swine flu is the Y2K of our time, the beast that roared but never bit, the River Phoenix of the disease-disaster constellation....right? Wrong...

The first few days of swine flu seemed like the aporkalypse. 1800-odd cases, 80 deaths and plane-loads of coughers and sneezers on their way to threaten our 'border security'. Antivirals were in short supply, pork was suspicious (if not unsafe) and we needed to close the airports and stick thermometers in tourists in order to keep ourselves safe. The ABC was doing its bit for the nation with some hard hitting reportage on scurrillously unhindered passengers at Sydney Airport. The end was nigh...

But then it wasn't.... no one that mattered was getting sick, there wasn't any Australian cases and masks were not the newest fashion accessory in Mosman. Besides; Madonna was getting remarried, the government was inventing words and abusing idiom (see "nation-building", "forged in the fire") and NRL players were objectifying women again. All was right with the world, the danger had passed.

At this point, most right-thinking media consumers probably saw a furphy...the eggheads and politicans had sold us a pup!! The cynical saw an availability heuristic in action; "suspicious thinkers" saw a consipracy at every-turn. Pockets had been lined, Roche had sold a bucket-load of Tamiflu and curious troop movements were seen in the Arctic...we'd been hoodwinked and there was actually nothing to fear. Just another beat-up by rent-seekers and panic-button pushers.

The problem with this perfectly reasonable view is its distance from pesky reality. The latest scientific information, based upon initial case and fatality data from Mexico, turns up some interesting information. Firstly, the speed with which the virus spreads seems to be relatively slow...for a pandemic (in germ geek-speak; R0=1.6). The number of people who will likely die once they have been infected also seems to be quite low (CFR <0.6%). style="font-weight: bold;">is
a potential Lilliputian.

Of course, that doesn't mean it's not going to affect you, just that it might take some time before it does. The impression that the H1N1 flu was a fizzer is tempting, but misleading; especially when the models of how disease spreads through a population seem to tell us that pandemics can have relatively long periods of stasis before reaching a tipping point beyond which the disease spreads rapidly and effectively.

Therefore, to put the provisional R0 of 1.4-1.6 in context, we need to go down in the weeds a little bit. For those of you allergic to graphs...shield your eyes now.


(A) Given our value for R from the Science paper. The delay before an epidemic-inducing 20 concurrent infectious cases in Australia (or any country outside of North Am) has a 95% probability of being between 40 and 70 days from initiation of the epidemic in the source region.
(B) Number of travellers leaving source region and arriving in at-risk country per day has a negligible effect on the median delay once number of travellers >100 (i.e. remains at ~60 days for R0 of 1.5)
(C) Screening incoming travellers for symptoms has a negligible effect on delay
(D) “In general, the additional delay achieved by introducing non-pharmaceutical border control measures is generally small in comparison with the natural delay”

The (highly) general conclusion from this is that we may have to wait up to two months for the natural introduction of the pandemic into Australia....and that border screening will do very, very little to alter this. Sixty days is a long time for the media-land goldfish (and its ADHD brother, the blogosphere), and hence the tendency is towards swiftly shifting attentions and the phenomenon of rapidly appearing and dissappearing threats.

The idea that disease could take a long time to get to Australia also works counter-intuitively in a world were the bananas on your kitchen table were in the Phillipines two days ago. But the confusion is lessened when you realise that the probabilities of any one passenger on a flight to Australia being infectious with a disease that has only infected a few thousand people in a country of 100 million, while rising towards 1, are relatively small in the initial stages of an epidemic.

Hence the idea that we are 60 days from midnight...or atleast 60 days from when I'll be proven wrong...

3 comments:

Adam Man Tium said...

Question for you, captain. Can you post some form of update to this page, as to current levels of infection/mortality etc and ho the pre-pandemic predictions are now looking?

aerobic dvds said...

I am afraid that this statement is not correct "The idea that disease could take a long time to get to Australia also works counter-intuitively in a world were the bananas on your kitchen table were in the Phillipines two days ago.
".
aerobic dvds.

aerobic dvds said...

I am afraid that this statement is not correct "The idea that disease could take a long time to get to Australia also works counter-intuitively in a world were the bananas on your kitchen table were in the Phillipines two days ago.
".
aerobic dvds.