EPAT General Discussion, page-77

The visual component may not be able to distinguish a baby in pain but combined with behaviour, audio and auxiliary data it may have a very good chance!

It may be a tough thing to develop though, even through in-house testing (excuse the pun) without the app in the first place how do you determine whether a child is in pain in the first instance? I suppose you could say the same for dementia patients as well though and they have succeeded here! I'll have to have a read through the synopsis available on data collection for the dementia app.