I agree in principle, but incredibly difficult in practice. Requires very complex analysis to separate out respiratory admissions that would've happened anyway from the ones that are potentially due to smoke/pollution
-
-
Replying to @GidMK @Asher_Wolf
I'm also not sure it's an easy thing to report publicly in this way, because there'll always be a fairly wide margin of error in that sort of estimate
2 replies 0 retweets 3 likes -
Replying to @GidMK @Asher_Wolf
Can't you just use aggregate numbers. i.e. this is how many respiratory cases and plot against air quality data to see the rise in number of cases. There's no need for medical examination to estimate whether the patient only presented due to the smoke.
2 replies 0 retweets 1 like -
Replying to @dave_w77 @Asher_Wolf
You can, but aggregates are often misleading
1 reply 0 retweets 3 likes -
Replying to @GidMK @Asher_Wolf
Yeah often. But here you would get mean and std dev for normal levels of pollution and have a good idea of how many extra cases are due to the rise in pollution. You can't say which specific cases but from a public health and policy perspective that's not important.
1 reply 0 retweets 1 like -
Replying to @dave_w77 @Asher_Wolf
Nah, you'd use an age-adjusted regression model, with some correction for seasonality. But more than that you'd need to take into account yearly trends in infectious disease, outbreaks, etc, it's a fairly large bit of work rather than a simple figure
2 replies 0 retweets 4 likes -
Replying to @GidMK @Asher_Wolf
I think this is what you'd do if you wanted perfect number. But I think you'd find that the number of respiratory cases would be so far above a baseline (in terms of number of std deviations) that we'd have a good idea of the difference in number of cases.
1 reply 0 retweets 1 like -
Or at the very least a decent estimate (without needing to go in to all the extra corrections that you suggest).
1 reply 0 retweets 0 likes -
Replying to @dave_w77 @Asher_Wolf
But without the correction all you've got is a meaningless number. No idea whether it's associated with the event in question or just a statistical anomaly
2 replies 0 retweets 1 like -
Replying to @GidMK @Asher_Wolf
Except that I'd expect from baseline number to be out by over 6 standard deviations. i.e. You can just totally reject that it is an anomaly, and it isn't meaningless at all.
1 reply 0 retweets 0 likes
I'd be very surprised if that was the case. Even if 10,000 people had respiratory hospitalizations directly related to the smoke, it'd probably be less than 2 SDs. There are tens of thousands of respiratory hospitalizations a month in both NSW and VIC
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.