& btw if you want to happify me this holidays π
Here is a hypothetical scenario:
The NHS develops a new piece of software which can analyse the public and– using a combination of image recognition, medical history, geographic data, and machine learning– determine if a person has undiagnosed appendiceal cancer.
Via randomised control trialing they determine that the software is 99% accurate, and so run it on the entire British population and offer treatment to all positive cases before the cancers develop.
What percentage of those diagnosed will have actually had cancer?
More than 99%
Exactly 99%
Between 99% and 95%
Between 95% and 90%
Between 90% and 75%
Between 75% and 50%
Between 50% and 25%
Between 25% and 10%
Between 10% and 5%
Between 5% and 1%
Exactly 1%
Less than 1%
For reference, the British population is ≈70 million, and appendiceal cancer is found in ≈20 people per million.