It seems like the investigation didn't harp on it much, but honestly I think the biggest issue is the repeated warnings the patient might have been suicidal. I

It seems like the investigation didn't harp on it much, but honestly I think the biggest issue is the repeated warnings the patient might have been suicidal. I was helping out with data analysis for a psychiatry study that didn't even really involve an intervention (looking at how well vocal properties from daily audio clips predict clinical scales), and it was made very clear how imperative it was to act if we had any reason to believe the patient was at imminent risk for self harm. So it is crazy to me that they didn't unenroll the patient upon receiving that warning. If a patient threatened suicide in one of their recordings we would have to get in contact with their main doctor ASAP, and this is actually one of the things the IRB was concerned about - if we didn't review the recordings in a timely fashion but patients thought we did, it could mean a suicide threat gets missed that might not have otherwise. Refusing to change the treatment plan of a known suicidal patient because of a research study really is awful.