EHR Certification – A Delicate Balance

ARRA requires that EHR systems be certified. ONC published an IFR outlining certification criteria supporting Meaningful Use as defined by CMS. Later on ONC proposed rules for the creation of EHR certification authorities, and NIST is currently defining brand new testing procedures. In the past CCHIT certified EHRs and large portions of CCHIT certification criteria came from HL7’s work on various functional profiles.
All this means is that hundreds of functional points were required to be present, in very specific form, for an EHR system to achieve certification.

There are functional points mandating what data to capture, how it is recorded, how it is stored, how it is displayed and when it should display. Incorporating all these certification requirements into a software application created very comprehensive products with the ability to do many things, or as they came to be known “bloated EHRs”. But there is another, probably unintended, consequence when trying to enforce such incredible detail; the certification process was in effect dictating particular work flows and distinct user interfaces. That was the point where the grumblings about lack of EHR usability and complaints of EHRs not being built with physicians and patients in mind, started gaining steam.

I cannot count the times when my only response to a complaining physician was that the offending piece is required by CCHIT. Many vendors searched for places where the system need only have “the ability” to do something and created customizations allowing physicians to turn off said ability. Nothing more so than the endless, and mostly useless, alerts. At one point you could tell at first glance if an EHR was or was not CCHIT certified. All the certified ones had this strange little dropdown box next to each diagnosis in the problem list, where you could select if the problem was acute, chronic, resolved, bi-lateral, controlled, etc. I have never seen it used in real life. Does anybody think that unless we mark it chronic, there is a danger that a doctor will assume that DM is acute or bi-lateral?

Well, if it’s all so counterproductive, why are we certifying EHR technology? Honestly, I don't know that we should. The original thought was that certification will assure prospective buyers that the system is able and willing to do the job, thus separating the wheat from the chaff. Unfortunately, certification acted in a more simplistic way and separated the heavy from the lite, regardless of usability, value and stability.
The new thought at ONC is that certification should provide a guarantee that one can achieve Meaningful Use by employing a certified system. So instead of trying to dictate every detail, the ONC certification is all about the specific criteria of Meaningful Use; 25 criteria for Stage 1. As a result, some small uncertified vendors of very lite EHRs are expressing interest in the ONC certification, if the price is right, of course. CCHIT gracefully turned on a dime and created a much slimmer ARRA certification process. But is it possible that the slimming and trimming process is going too far?

Is it possible to certify too little?

It is probably too soon to tell if the new ARRA certification will be able to strike an acceptable balance between allowing vendors the freedom to innovate and ensuring that Meaningful Use can be achieved with a certified product. However, the proposed NIST test procedures indicate a possibility of certifying too little.
For example, certifying that a piece of software is capable of receiving electronic lab results in discrete format, while allowing the vendor to choose the structured format in which lab results are received could lead to the inability of future users of the software to meet Meaningful Use requirements, if the vendor selects anything other than HL7 for their preferred structured format. Laboratories do not send results out on a regular basis in either CCR or CCD, which are specifically allowed by NIST, or any format other than HL7. An EHR, or EHR module, capable of receiving lab results in a CCD format exclusively, may become ARRA certified, but its proud owner will be unable to become a Meaningful User.

A different and more troubling issue arises from looking at a requirement in a very narrow way and neglecting to account for ramifications not directly related to the particular Meaningful Use criterion. For example, testing for the ability to modify a medication list without concern for historical integrity of the list creates a strange and dangerous situation where adjustments made to the SIG (strength, dosage,etc.) of a particular prescription are allowed to overwrite the entire history of said prescription (think Warfarin). The narrow scope of testing that modifications are possible fails to assert the need for those modifications to be made in accordance to general principles enforcing the integrity of the medical record. Presumably this larger concern is left to vendor discretion. Existing vendors that certified their products under the old regime will probably provide a reliable product anyway, but new innovative vendors, writing software to published specifications and test procedures could easily create a certifiable, yet clinically unsafe product.

We are all aware that an over prescriptive certification process has not produced massive improvements in EHR offerings. The temptation to err on the side of oversimplification is large, particularly since the time is short and we must come up with a temporary certification process sooner rather than later. However, if we look at certification as a process designed to help vendors build better products and assist buyers in selecting tools to achieve Meaningful Use, then we must make sure that, while not overbearing, ARRA certification is indeed meaningful.