# Test results, part 2: how about negative results?

This is part 2 of this article:

To recap: there’s a pandemic going on, 1% of people have the virus. There’s a test that can detect the virus, and the test is 99% reliable (for both positive and negative results).

But this time, when you take the test, the result is negative. How much can you trust that result?

# Ideal case

If you know nothing else besides the test result, then it’s very reliable: 99.9898%, which is basically 100%.

I will not repeat the analysis, please refer to Part 1. But, again, this is the ideal case scenario. What happens in reality?

# In the real world

Turns out, the negative result remains very robustly trustworthy even when the infection rate grows very large (or you’re in a cohort with a large infection rate).

- at 10% infection rate, the test is still 99.9% reliable
- at 50% infection rate, it’s 99% reliable
- at 90% infection rate, it’s 91.7% reliable
- at 99% infection rate, it finally drops to 50% reliability (coin toss)

Here’s the graph:

If the general population is at 1% infection rate, you’d have to be in a very specific cohort to not trust the negative result.

For example, if this particular disease has **extremely specific symptoms** that do not occur with any other disease, and **you do have those symptoms**, and you get a negative test result, then **you should question the test**, because you’re in a cohort that’s pretty much guaranteed to have the virus, no matter what the test says.

Other than that, the negative result tends to be pretty robust.

Code:

healthy.chance <- function(inf.rate, test.rel = 0.99, tot.pop = 10000) {

sick.pop <- tot.pop * inf.rate

healthy.pop <- tot.pop - sick.pop

pos.sick <- sick.pop * test.rel

neg.sick <- sick.pop - pos.sick

neg.healthy <- healthy.pop * test.rel

neg.total <- neg.sick + neg.healthy

chance <- neg.healthy / neg.total

return(chance)

}infection.rate <- 1:100 / 100

chances <- lapply(infection.rate, healthy.chance)

plot(infection.rate, chances, col = 'blue'); grid()

That’s all, thanks for reading!