Information systems in medicine tend to be tools. Trials tend to be of the products.
I can see the possibility of a RCT between two decision-assisting algorithms for a single condition, but I think one of the larger elements to study will be how th eusers like working with them. Unfortunately at the moment that can very easily be subsumed by a discussion of how th epaymasters an get the managmeent information they want from the various system components, and how they can enforce particular behaviour on users. Given a few more generations of managers and programs this may tilt back, but RCTs are perhaps a subtler instruemnt than fits the software we have at present. -- From the Linux sofa or garden swing of Dr Adrian Midgley http://www.defoam.net/
