Today's handhelds can range in cost from about USD$50 up to USD$750 (or more). The USD$50 price-point makes these devices more available to low-income population, which is important because not only high-income people get sick. (You could argue that low-income people may get sick more frequently, but that's beyond the scope of this discussion.)
The devices costing approximately USD$50, at the timing of writing, still have mediocre performance in three-dimensional (3D) applications. In the long run, I'd expect that restricted or prescription applications would be developed by using both two-dimensional (2D) and three-dimensional (3D) methods. If an application were to be released within twelve-months from the time of writing, it may be likely that some USD$50 devices may not be sufficient to meet the needs to allow immersion. I'd expect that there could be testing applications developed to detect if a handheld device was sufficient for running the actual health-related application.
As far as right-restrictions (digital rights management, known by its acronym DRM), the prescriber could provide terms similar to physical drugs, say 30 days or 90 days, at which time a renewal code would be required. It would be expected that the application itself could query the server (if an Internet connection were available) to detect if the prescriber electronically renewed the prescription. The first-time download of an app could easily be allowed by a QR code so long as the downloader has access to a prescription code to enter on the first application load.
I would hope that applications designed using these recommendations would cost less than USD$15, either one-time or on a monthly basis, putting it in line with a co-pay amount for a physical drug prescription. I would discourage a developer from expecting a USD$50 price tag on an application that would conceivably be playable on a device costing USD$50. There may be exceptions for apps limited to smaller audiences, but developers should realize there are will potentially be billions of users within the next decade.
In a final note, during alpha and beta testing phases, I would strongly encourage developers to liberally screen testers against the subgroups that could be harmed by the application design. It would be better to assume a tester not be eligible than be eligible if given information suspecting of a beta tester that may be disqualified.