There was a great response to last week’s Curbside Consult and my mention of the therapeutic powers of baking. Despite everything going on at the office, it ended up being a fairly low-key week, so the only things coming out of my kitchen were a pan of brownies and a batch of banana bread.
(I admit I played a little bit of the Mad Scientist game with the banana bread. Although it was good, it wasn’t significantly better than the original recipe, so maybe I’ll stop trying to mess with perfection.)
We made a fair amount of progress in our due diligence efforts around bringing the patient records from the practice we acquired onto our system. Although some people might find it boring, I actually enjoy rolling up my sleeves and digging in. It’s predictable work in some regards.
Our DBAs started looking at their system’s data structure to identify how many custom fields they are using compared to a vanilla version of the software. Some of our EHR analysts started looking at the actual user screens to identify custom fields from that perspective as well as to begin diagramming the workflow they’ve built in the EHR.
We’ll send people on site and work with their training team to determine whether the EHR workflow matches how they operate in the practices or if this is an opportunity to retire any custom elements that aren’t actually working in the field. I’ve seen plenty of instances where physicians have customized their systems to the point where efficiencies are lost. This tends to happen more when users don’t have adequate training or don’t agree with the design intent of the software.
Where there are customizations in the workflow, we’ll also do some statistical analysis to look at how many times custom fields are actually used. Just because they were built doesn’t mean anyone uses them regularly.
Our medical group has grown substantially over the last few years. Given the number of physicians who currently use an EHR, we’ve had to do a fair number of conversions. Some of them are simple, especially when the source EHR is fairly primitive or doesn’t have a robust data structure. In those situations, we might convert the patient notes to PDF files and bring them in as if they were scanned documents. It doesn’t give us a lot of discrete data, but in some regards it may be safer than trying to map imprecise data.
I’ve seen systems that don’t use any kind of formatting on data fields (such as restricting blood pressure entries to numbers only) that lead to garbage in the record. In those situations, I typically sit down with the physician and explain the choices: we can either bring the data as images (akin to scanning a paper chart as far as patient safety is concerned) or we can spend a lot of time and money trying to map it. In the latter scenario, they will need to sign off on any corrections.
Most physicians who hear about the time commitment for mapping data run shrieking out of my office and I never hear from them again until I see their signature on the checklist approving the test extract that’s been pulled into the imaging system. Those who aren’t scared off by the time commitment are usually scared off by the budget, which our medical group usually isn’t very keen on funding.
I’m surprised (at least at some level) but the number of physicians who realize they have dirty data but don’t do anything about it. They see the typo’d letters in the BP fields and authenticate their notes anyway rather than talking with their staff about data accuracy. Very few have thought to talk to their vendors about why the system even allows typing of letters into a BP field.
I guess I shouldn’t be that surprised, because I’ve seen even wackier things in the paper world, such as subspecialists who had their staff stamp consult letters with nonsense like, “Dictated but not read; signed by secretary to expedite.” Someone who is OK with that probably doesn’t care about potentially erroneous data in their notes.
So far, the potential conversion doesn’t look that bad from a technical perspective. Although there is a fair amount of customization, it’s not being used extensively. In fact, overall use of the EHR is pretty light. From a change management perspective, though, that’s pretty ominous, especially since our group requires significant commitment to documentation via discrete data. We’ll have our work cut out for us in helping them truly adopt EHR as well as in helping them adapt to our culture.
I almost wish the technical aspects of the conversion were more daunting because I could use that to buy more time with the powers that be. Our analysts still have a bit of digging to do and the workflow teams will find plenty of issues when they go on site, but I’m not sure we’ll have as much time to formulate an effective plan as I’d like. We’ll have to see how things unfold.
Regardless of what we find, I know we won’t have anywhere near as much budget as we need to do our best. We’re pretty good at delivering the impossible, though, so I’m hopeful. And when all hope is gone, there will always be pastry.
Email Dr. Jayne.