I'm coming to the discussion from the perspective of a physician who became a technology entrepreneur, starting a company that provides a patient engagement solution on smartphones, tablets and computers.
While I'm very much a supporter of evidence-based medicine, I can now appreciate the complexity involved when it comes to combining research & technology development in a fruitful manner.
One challenge I've observed is that we often jump to doing a study evaluation of a web/mobile health intervention that's barely been used - with the perspective that the study is the opportunity to test. And this means the software cannot be changed/improved during the study.
However, unlike drugs/medical devices, software development by nature is an iterative process, and the expectation for software is to use continuous feedback to constantly improve the product. In other words, I believe studies evaluating effectiveness should be saved for later on in the application's life.