3 ms·
That feels too low to me to be practically useful. I would guess declaring all your patients to be healthy already gets you there for many, if not most, conditi
by Someone 7y ago
That feels too low to me to be practically useful. I would guess declaring all your patients to be healthy already gets you there for many, if not most, conditions you mention in https://news.ycombinator.com/item?id=20752782 https://news.ycombinator.com/item?id=20752782.
It also seems to imply that you will have many false positives, where people will unnecessarily get stressed while waiting for follow-up tests or even receive treatments they don’t need.
- SergeyDavydovGA 7y agoit must be taken into account that there will be no blood sampling, and a preliminary diagnosis will be made on one exhalation. And the price of the analysis will be quite low. Do you really consider it a bad result that today, 9 out of 10 people will be able to know exactly about the presence of cancer by taking 5 minutes of their time?
- Someone 7y agoAs I said, one likely can do better in one second, by telling everyone they’re healthy. If five out of a hundred people have a disease, you even can get 90% accuracy by declaring those five healthy and five others unhealthy. Accuracy is fairly meaningless for this kind of thing. What sensitivity and specificity (https://en.wikipedia.org/wiki/Sensitivity_and_specificity https://en.wikipedia.org/wiki/Sensitivity_and_specificity) will these tests have?