That's why you'd need to have the rendering engine blur things that should not be in focus. From what I understand on how the medical systems work, you'd shine an IR light into the user's eye. The return from the light would allow you to measure the focal distance of the eye. You'd then adjust the lenses of the headset so that the display is exactly that focal distance away. The focal distance of the eyes would get passed to the renderer. The renderer would do proper focal distance bluring to objects that are not at the correct focal distance. The main limitations would likely be the measuring of eye focal distance in realtime, having lenses respond to it in real time, and the possibility that the always in focus screen door effect would cause issues.