Sorry Rappers, page-178

good results from a second trial in the primary outcome could do this, with a prodigal child effect

that outcome measure is to me made less certain by the small numbers we've seen

it ideally would be tested in an area with lots and lots of pneumonia, enough cases to be very confident in under 5s and over 5s.

if you look at case numbers in SMARTCOUGH-C vs Joondalup/Princess Margaret you'll see much greater numbers of upper respiratory cases in the USA centres and much greater lower respiratory cases in WA. These aren't small differences. They seem almost inverse in relationship, from recall.

nobody wants good results to be a roll of the dice. Results should reflect the capability of the software. An alternative would be "sound analysis" by a resapp health technician on call, much as the existing situation with tissue histology or radiology in some centres