PainChek General Discussion, page-13059

Accuracy is a hard thing to quantify, it's complex and FDA knocked back RAP because the possible harm outweighed the possible good.

RAPpers have been blaming the system and a perceived bias from the FDA towards them when it's almost certain that it is the product, the application, the trials and documentation that is lacking.

Consider this simple scenario sample of people, of whom half are sick and half are not. If you have a device that diagnoses everyone as sick you can claim that your device successfully identified every sick person. But clearly that is not a useful device. You need to consider false positives and negatives and weigh up the possible harm vs possible good.

What about one that diagnoses every sick person, but also 10% of healthy people as sick (false positive).

Or one that only diagnoses 80% of sick people as sick (false negatives)? Assessing harm versus good becomes more complex.

The definition of a 'working' device is complex, even for something as simple as separating two binary options let alone a scale or multiple options. RAP found this out the hard way. PCK have been far smarter and patient. Starting with clinical trials and now with probably 100000 real world cumulative assessments they have a powerful case for proving a net good with their FDA submission