What to you mean by EMR? Is it what the doctor uses, the nurse uses, the hospital or clinic uses?
The problem with standardization is that it ASSUMES that the goal and environment are stable or change very slowly. If you look at the history of the automobile, there was experiementation for over 100 years before "standard" features could be clearly determined. We are only 30 years into the desktop/display era of computing, and we are still evolving how a person works with inforomation.
The issues of standardization have so far approached the problem from the exactly opposite wrong end, mandating that people get EMRs to get experience with them, and then creating EMR's that function well for the goal they are intended. The first steps should have (like the Internet) focussed on how disparate systems can exchange information reliably, evolve in non-catastrophic ways, and change in order to improve the quality of the content and interaction.
Few, if any, of the government standards address this level of utility to advance the field. ICD-10 is a poorly constructed coding system built to accomodate increased granularity of data collection. It's a code with no mnemonic structure and a digital representation with no check digits to avoid common data entry and transcription errors.
The way the Internet came into existence was after years of experimentation by many different vendors (IBM, Digital Equipment, Intel, Xerox), we evolved to two working standards (IBM token ring and DEC-Intel-Xerox ethernet). Of these two, only DIX ethernet was a non-vendor specific format. When the National Science Foundation decided to connect the national supercomputer centers together, they elected to use DIX Stanford Research Institute's TCP/IP because it ran on both token ring and ethernet, and could be used across a wide range of connection technologies. From the ARPAnet base technologies, NSFnet was created and evolved into the Internet.
We should focus on interoperability not as a standardized product, but as a uniformly availble set of services which allow any system to find, query, transmit and receive, store, and share a constantly set of uniquely qualified and quantified clinical set of observations with each patient being uniquely and unambiguously identified, and services which negotiate the format and content of data they are able to exchange. We should elminate system conversions, standardized coding system updates (i.e. ICD-9 to ICD-10) conversions, and single point in time cutovers for every provider's system.
We need a better plan than brute force.