Ann: Positive Preliminary Results from SMARTCOUGH-C-2 Study, page-301

The first US trial pneumonia results were presented showing accuracy statistics for “cough alone” and also for “cough, age, gender and symptoms”. In these trial results RAP neglects to inform readers of the components of the prediction.

 

15% of participants did not complete the study or provide analysable data. From the real-world performance perspective that the FDA is interested in this missing data needs to be modelled on a best-case / worse-case scenario.

 

Taken together these two factors suggest the accuracy estimates for pneumonia missed the 75% threshold (that the trial was designed to demonstrate) by a mile.

 

To raise problems with the WHO PEP standards and processes Ithink is silly. This is the gold standard. So what that primary readers disagreed and this was not resolved by secondary panels and required majority tertiary consensus. That is how the process works.

 

What RAP need to explain is why they added clinical diagnosis to those WHO PEP standards when deriving their reference category.  This would introduce the “noise” of clinical diagnosis – which is as RAP point out is context dependent – different healthcare / insurance systems; swamping whatever true diagnostic signal they were modelling.

 

The results from a FDA regulated - registered clinical trial conducted in three leading US hospitals with independent primary investigators with highly skilled adjudicators applying worlds best practice in diagnosis with the trial being implemented to very high standards are going to be pretty convincing.

 

And they trump the results from an unregistered Australian “study”conducted by an single investigator in a local Perth hospital with “clinical diagnosis”as the reference. The idea that the Australian “diagnosis” is better than the US diagnosis because it has better agreement with their app is just funny. The app has simply learned to measure the noise of clinical diagnosis.

 

But hope springs eternal. They neatly transferred trial risk to shareholders through the CR.  They have money in bank and a cult-like following on HC. They are pretty creative.Still plenty of money to be made here …. by someone.