Sunday, February 25, 2007

Ethics of using non-regulated devices

Most administrators and physicians will be concerned about using a non-regulated device. It is not intended that such devices ever be used except when there is no alternative. The ventilator design I am proposing will of course be tested by the design team and possibly validated by some other party, but it does not seem feasible to get FDA approval. The FDA could not approve even a perfect design in which the prototypes tested flawlessly because the FDA also requires that there be very tight control of the manufacturing and assembly process.

If you get right down to it, a ventilator consists only of a few components; some valves, sensors, possibly a blower or regulator and a sophisticated control system. The valves, sensors etc would be of industrial instrumentation grade. This grade of component is extremely reliable and these components can run continuously for years in a harsh industrial environment with minimal failure. The project would identify suitable components, build and test prototypes ahead of time. All the tests for over-inflation, low rate, disconnection etc would be performed and documented by the development group.

Much of the value and the safety of a ventilator resides in the sophistication and reliability of the control and user interface software. This software would be developed in an open source format so that the source code is available for anyone to inspect and improve on. It has been shown that open source software can be as reliable or even more reliable than closed source software in this regard. I believe that such a device design approaching the reliability and safety of regulated devices could be produced, providing that the project team consisted of the right group of people.

Now if we looked at this reliability issue from a real world, worst case point of view, and a device was produced that killed say 10% of the patients that used it due to equipment failure. The FDA would view this as a horrible device and of course would never approve its use. Now suppose this (poorly designed) device was the only one available and you had to choose to use it or allow say a friend or relative to die without the ventilator, (100% chance of death as opposed to 10%) the choice is now quite clear.

No comments: