There are hundreds of various EMR products, which typically cost $30,000 per physician to buy, and $5,000/year/MD for maintenance costs. After paying all that, the EMR products are incompatible with each other, so that records cannot be sent from MD to MD, except to print out the information and scan the pages into the other MD's system. If the digitized information is lost, then the value of the EMR to track information is also lost.
A better solution would be to encourage ro require the use of a single backend product, which would be open sourced, such as the VA's system. This system should be capable of handling all patient encounters: hospital, clinic, OT, PT, labs, etc. Separate front-ends could be created for different situations. The variety of these various situations (hospital front ends, ENT clinics, family practice clinics, pediatrician's office, and so on would be a great environment for private investment. A generic open source front end could be maintained for minimal cost (such as the current VA front end. This would lower the cost of projects, while increasing diversity in front ends (all the healthcare workers care about) and create an opportunity to share digitized information between providers who have permission to share this information from the patient.