Misdiagnoses happen all the time, and will continue to happen...

Misdiagnoses happen all the time, and will continue to happen regardless of if the app is used or not, that's just a reality. As for the clinician's confidence in the app, well they'll look to the past clinical trials and peer reviewed publications. They're either going to be confident in it or they're not going to use it. And your scenario has already been put into practice at Morayfield in which they said it helped them find cases that weren't detected in the physical examination. But sure there will be times when the app gets it wrong, but that's no different to what already happens. Remember the second US trial? The adjudication panel couldn't come to agreement on a number of cases, so they had to bring on another adjudicator just to get an agreement.

Also the scenario you present is just for one possible avenue. What about the other settings the company is trying to target? Such as screening/patient triage, Telehealth, etc. The alternatives there are either don't improve triaging patients, or more costly use some of their clinicians to personally triage patients. While for Telehealth it's either don't do diagnosis over the phone, or the clinician will try to diagnose from the sound alone, or send the patient additional hardware they can use on their end (which in this case how is it any different, the clinician still needs to be confident in that hardware and a misdiagnosis may still happen).

And while it's not very applicable here as RAP might not even be around by then, but eventually we will have to depend on a lot of this automated diagnostic tooling, as it's an inevitable shift that will eventually happen.