Ann: Trading Halt, page-199

Just with regard to additional clinical studies. Yes, there will definitely need to be one more and possibly two.

The pilot study was to gather data for training purposes. This requires capturing the data that will be used for diagnosing the condition (cough sounds and any symptoms being used) along with an independent diagnosis. Using some of the data for training and some for testing will allow the team to determine:
a) whether a predictive model can be built, and
b) if it can, how accurate it is.

But the pilot isn't a double-blind study. You can't use blind data for training because you need the independent diagnosis.

So at the very least a double-blinded clinical trial will be required to prove the algorithm' accuracy with cases not seen in training.

In addition, if the training is promising but not sufficiently accurate to go straight to a clinical trial, the team might decide that additional training data could improve accuracy enough to then progress to a clinical trial.