But couldn't you test the same person multiple times?

With all this talk of a zombie apocalypse I'm going to be really disappointed if there isn't one or two in my lifetime. And what am I supposed to do with all these guns I've been stockpiling for the occasion?

Reading this and trying to figure out the math makes me want to eat BRAINZ....

Wait a minute. So eliminating 90% of your uncertainty makes a difference? We must write more articles about this.

Only problem with that scenario is a "gas that only decomposes zombies" is impossible considering zombies are just sick humans.

MayoSlather: With all this talk of a zombie apocalypse I'm going to be really disappointed if there isn't one or two in my lifetime. And what am I supposed to do with all these guns I've been stockpiling for the occasion?

A whole bunch of guns isnt going to help you, you just need a couple...its ammo you need a lot of.

MayoSlather: With all this talk of a zombie apocalypse I'm going to be really disappointed if there isn't one or two in my lifetime. And what am I supposed to do with all these guns I've been stockpiling for the occasion?

Same thing people did with their cold war bunkers post-nuclear survival bunkers and stockpile.  Find something new and irrational to fear and prepare for.

I for one, recommend the coming robot uprising.  We are due, and it's going to be trendy in a few months / years.

The author is gaming the numbers by assuming an unusually low infection rate of 1 in 500. The positive predictive value is largely dependent on the prevalence of disease in the population. If you are assuming that this disease is reaching pandemic proportions then 1 in 500 is a very low estimate. If you instead assume that 25% of people will actually have the disease, then the positive predictive value of the 99% accurate test is 99%. Even if you assume a 10% infection rate, then the positive predictive value of the 99% accurate test is still 92%.

insano: The author is gaming the numbers by assuming an unusually low infection rate of 1 in 500. The positive predictive value is largely dependent on the prevalence of disease in the population. If you are assuming that this disease is reaching pandemic proportions then 1 in 500 is a very low estimate. If you instead assume that 25% of people will actually have the disease, then the positive predictive value of the 99% accurate test is 99 % 97%. Even if you assume a 10% infection rate, then the positive predictive value of the 99% accurate test is still 92%.

FTFM

couldn't you just ensure everyone was infected then administer the cure?

snuffydaman: couldn't you just ensure everyone was infected then administer the cure?

That was my thought.

If you know how to cure the airborne disease then you obviously know how to isolate said disease. Just intentionally infect someone and then administer the vaccine 15 minutes later.

Hell if the vaccine is simple enough to be administered by Rite-Aid and CVS just pump the disease through the air conditioning so everyone coming in for the shot is guaranteed to be infected.

insano: insano: The author is gaming the numbers by assuming an unusually low infection rate of 1 in 500. The positive predictive value is largely dependent on the prevalence of disease in the population. If you are assuming that this disease is reaching pandemic proportions then 1 in 500 is a very low estimate. If you instead assume that 25% of people will actually have the disease, then the positive predictive value of the 99% accurate test is 99 % 97%. Even if you assume a 10% infection rate, then the positive predictive value of the 99% accurate test is still 92%.

FTFM

Well, that was going to be my post, but yours is more succinct.

MindStalker: But couldn't you test the same person multiple times?

No.  It's generally not that the test is imprecise, it's that many times, even with perfect precision, the outcome has some amount of uncertainty because the probability distributions of test values overlap between the normal and diseased populations.

For example, say you have a test for a particular disease looking for elevated hormone concentrations in the blood, and of the people who had test values of 1.3 ng/dL, 75% had the disease and 25% did not.  You could repeat the test a million times, and doing so would make you very, very certain that the hormone concentration was exactly 1.3 ng/dL, but even knowing that, you still are no closer to knowing if this particular person has the disease or not.

In other words, the higher end of the normal spectrum overlaps with the lower end of the abnormal spectrum.  No amount of repetition can separate out those two conditions.