Monday, February 16, 2009

Structure, process, outcome

An academic group I belong to has been having a pretty lively discussion on the initial difficulties with starting an EMR. There is consensus in this group (in which everyone is using different EMR systems) that, for the first little while, there is a lot of loss of efficiency. You are quite a bit slower on the computer than on paper. The gains don't come until later, and many people still find that recording a consultation in the EMR (while the patient is in the room) is slower than on paper.

Why is that? Paper is simple. You write things down. It doesn't matter how you write things down. Paper doesn't crash.

The benefit of the EMR is that you can have point of care reminders and you can mine your data for useful information at the practice level. This doesn't happen until (and unless) you enter information consistently, and in the right areas--in other words, you have a well-tempered chart. It is harder to do that than to just jot something down on a piece of paper, and the EMR can make it difficult to enter things in the right area. I went to a conference where the keynote speaker, a GP from England, showed a short clip of how he struggles to write a prescription--after more than a decade of using EMR.

The initial difficulties are magnified by the start-up problems inherent in this technology. Let me explain what I mean by that.

The EMR involves both software and hardware. The hardware means lots of different machines that all have to work properly together. The number of hardware permutations and combinations (servers, routers, firewalls, connectivity, networks, printers, scanners, labelers and all their assorted drivers and software applications) is very large. Failures due to issues involving hardware are very common since there are so many possible points of failure, and can be very difficult to diagnose. There is no dedicated funding for hardware maintenance, nor local expertise in most small medical offices. The EMR companies know more about their own software than about our hardware; they cannot possibly be aware of all the different hardware pieces present in many different offices (unlike corporate branch offices, each medical office is an independant operation, with an individual IT setup). Sometimes the EMR company can't help with hardware problems; we have machines from different vendors, and it can be very difficult to know who to call for help. We also lack the knowledge to plan for hardware failure and redundancy.

"Structure" failing means your printers won't print, your computer doesn't work, you can't connect to your server, your speed is slow. If you don't have a sound Structure, you can't even get to your Processes, and you certainly can't change your Outcomes. Structure is where many EMR projects fail to launch, and the root causes of this are systemic (see above). Paper has a very simple basic Structure (filing cabinets, file folders, pieces of paper), and can't fail at this stage. EMR can, and does.

"Process" is the way in which we do things. If you can't access your records because of a Structure failure, this is moot. If you have solved the Structure problems, then you have to tackle Process; not always easy or intuitive--takes time and thinking. Those that have not bought into the EMR, but are just there along for the ride because their group got one, are much less likely to invest in this. They may benefit from herd immunity, because the group is now functioning better, and the front staff is more efficient--but will be slower with their own patients and in most aspects of their practice.

"Outcomes" is where the big payback happens. This means that the quality of care actually changes: more of your diabetics now have their BP at goal. This requires the ability to not only enter data properly (good Structure and Processes), but also the ability to audit your data and then decide what processes to change, and then re-audit. This is present in potential form in the EMR, but cannot and will not happen unless and until we solve the earlier steps--and if this doesn't happen, you will continue to see studies showing that the EMR is not making much of a difference.

We had a FHN meeting recently; we regularly have EMR booster education sessions at our FHN meetings now. I demonstrated changes since the recent upgrade, such as improvements in our drug module and consultation letters; I also handed out notes. I think it is important to have on-going EMR education. Interestingly, the mood was quite a bit different than in some of our previous meetings: we had less complaining (but not zero), and more interest in what the EMR can do for us and our patients. One of my FHN colleagues talked about how he uses his diabetic reminders to recall people who haven't shown for their regular 3 month visit: he can see the list of overdue visits. This represents a change in Process that is likely to produce a change in Outcomes. It took us three years to get to this point.

Structure - Process - Outcome is a very common way to assess the quality of health care. I think it is equally applicable to EMR implemention.


No comments: