Aaron Berdofe, MHI – Health Informatics

The ease of which any healthcare organization will meet “Meaningful Use” requirements will depend on what data they have captured and how easily they can report on it. As long as you have a “certified” system in place that captures the required data and have a way to analyze that data and report on it to the government, obtaining stimulus funds should be a fairly low-intensity undertaking.  The committee tackling this really should be applauded as it apparent that they truly understand the value of an EMR…it captures data and makes it transparent.  This is what I outlined in my first post.

It is amusing that almost every single EMR vendor has shouted loudly that there is no doubt that you will obtain stimulus funds by using their product before the actual requirements are set in stone.  GE and athenahealth, amongst others, have slapped “Guaranteed” stickers on their products as a marketing ploy to make you really question those non-guaranteed products.  In reality, all EMRs that pass the certification criteria (when that’s official) will help you obtain stimulus funds with about the same ease.  Unless one of these vendors actually creates all of the reports to present the data for you, they’re really no different than their competitors.  This would be a great short-term business idea though if no vendor takes the lead.  In reality, it will be most likely that your vendor has given you the tools to report this data by yourself while stating that their software has the capabilities of capturing the required information.  Because the majority of my technical expertise comes from the MEDITECH world, I’m focusing this post on their reporting options, but please note that the overall strategy can be applied to any system.

MEDITECH systems store data in a hierarchical fashion meaning that it is much easier to continue down a “branch” of specific patient data than it is to look across the entire patient population.  This structure is by no means unique to MEDITECH because that is the nature of how clinical data is at input.  The data is input patient-by-patient, not blood pressure-by-blood pressure.  While reporting on the data of a single patient or off of an existing index, MEDITECH Report Writers will note that the return is pretty fast.  They will also be very quick to tell you that reporting across the population can be excruciatingly slow.  Relational databases are much more efficient at querying across populations.  John Halamka, MD, had a very informative post last year about this concept.

Good IT systems management practices dictates that any system be comprised of 3-tiers:  One to input the data, one to store the data and a one to report on the data.

EMRs and Practice Management (PM) systems are really only good at the input part given their nature even though a lot of them attempt or make claims to be all three.    Therefore, you should be dumping all data from the EMR and PM into a relational database as frequently as possible.  MEDITECH’s Data Repository will aid with this transfer, but make you sure you have contacted LSS Data Systems to turn it on for the Medical and Practice Management Suite as well.  Choosing a relational database is a lot less of a gamble than choosing an EMR and I won’t claim any favorites. As long as they meet an organization’s security and data integrity needs, they’ll work.  Choosing how to report on the data is not a big gamble either.  You could most likely hire an intern versed in SQL statements or buy a database analytical tool to find all of the ARRA requirements more cheaply and quickly than having to write MEDITECH NPR (Nonprocedural Representation) reports.   Although there are a good handful of NPR report writers out there (myself included), we can tell you that the going gets tougher when the information requested isn’t stored closely together.

If you do choose to tackle the “Meaningful Use” requirements using NPR reports out of MEDITECH and LSS, please start now and begin referencing the existing matrix.  Reading through the 2011 measures, there is a large amount of these requirements that are items that can be tracked by MPM’s Health Maintenance/Disease Management functionality.  Therefore, if you aren’t using this functionality and subsequently not saving data to these structures it will be more difficult to determine say, the % of females over 50 receiving annual mammograms, than if you are utilizing that functionality and having the information stored on those structures.  The idea is to get the related data stored together so it can be accessed more efficiently.  Otherwise, the report writer will be sifting through numerous modules to find the data for a single report.  The sifting across modules and DPMs causes the long compile times.  The more data that can be captured under the PBR.MPI.VIS (MAGIC) or EAR.MPI.VIS (C/S) “roof”, the better.  This DPM is where the vast majority of patient data is stored in MPM and where most reports will be written.

Please read Adam Bosworth’s post on Learning from data.