The problem is that high pressures can cause barotrauma. So you might end up increasing PEEP beyond what a patient's lungs can survive, and because you're using a CPAP machine instead of a ventilator there's not really any facilities to measure pressures that high. Or there are but they were never characterized because the machine doesn't measure pressures that high. And because they were never characterized you may not actually be delivering the pressure that the machine says you're delivering. So this is pretty crazy.
The other thing ventilators typically include is pressure alarms and high pressure limits. Under routine use the high pressure limits keep you from popping patient lungs in volume-controlled modes. A pressure-controlled mode like BiPAP may be okay in this case since it provides those limits. The downside of pressure-control is that the patient's lung compliance and alveolar recruitment may change, so you can start off delivering a certain tidal volume at a certain pressure and then because of compliance changes the patient's alveoli lose the ability to contain the same volume of air, so the patient is losing oxygenation unless you change the settings.
My point is that there are a lot of considerations beyond simply delivering a fixed pressure of gas into a patient's lungs and hacking a CPAP may not meet those requirements.
The idea is manufacturers don't want to loose money and do this stuff themselves, so somebody donates their time to bring this hack to the FDA, and then the hospitals can use it.
