Tuesday, January 22, 2013

EMR Wannabe - Part 1


Enough bitching like a Spice Girl, it's time to structure my wish list and share it with you, dear vendors. Let's start with my favorite topic if you don't mind: Clinical Documentation and Data Input.

So here's my story from A to Z...if you wanna get with me, you gotta listen carefully.

When it comes to clinical documentation, physicians enter what I call the “choose your poison” zone. We are pretty much forced into the input method the system can handle, unless it’s the one our budget can afford. Sometimes it’s the good old dictation-transcription route, sometimes we get front-end voice rec, but most of the time we get such a piece of crap we end up sticking to paper like a piece of old bubble gum.

Thing is: even if we were offered THE one revolutionary method against another, it would NOT square the crazy-making circle. Why? Because I don’t see doctors using one or two input methods. I see them using a whole bunch of them, depending on who they are, where they are and what they are doing. So if you asked me to sum up clinical documentation 101 in a sentence, I would say it’s a generous mix and match of various input methods supported by hardware flexibility at the point of care. Period.

By supporting all (and I mean ALL) input methods, you factor in the most important ingredient out there: the human one. It’s taking into account the fact that physicians have various comfort levels towards technology, the fact that physicians and nurses work differently, the fact that expectations and workflows vary from a department to another, you name it.

So if you wanna be my EMR, you gotta get with my friends. And since you can’t teach physicians new tricks because they don’t have time – I know I’m repeating myself here but your neurotransmitters need to get the point eventually: WE DON’T HAVE TIME), you have no choice but to get it right. We’re paying you, vendors, to take it on your own time to understand the way we work and simplify clinical documentation to the point where, yeah, finally, we’re able to spend more time with patients. That’s our Holy Grail. You can slam your body down and wind it all around for as long as you like, if you don’t get that right, you’re zigazig out.

Then for those who get it right, they still have to price it right. But I’ll keep that scary spice for a later post. For now, let’s concentrate on the clinical documentation cocktail I’d like you to serve on a silver platter – on the rocks. Please make it last forever and our friendship shall never end.


If you wanna hear my nurse friends sing “zigazig ha” while filling in their Braden and Glasgow scales, give them what they crave for: smart electronic forms that are fun to fill in. Yes, I said it: “FUN”. Give them drop downs, radio buttons, check boxes, sliders, colorful YES-NO buttons and thou shalt be rewarded. But don’t forget to upload their input in the patient file. Sounds obvious, but hey, we never know…

Even for physicians in certain specialties, forms remain a powerful tool to collect structured information in a familiar way. It’s context-sensitive, it’s simple and mandatory fields force everyone to fully document patient care. And while the sounds we hear while filling in one of these is “Check, check, check”, what the system hears is the corresponding medical music, which then shows up in the final report. That’s synoptic reporting at its very best. But it won’t work without a profound, authentic desire to structure data and one heck of a cool, intuitive user interface (think Apple and you’ll stop serving us oranges).


Of course we do. Thou shalt therefore support computers, smartphones and stupid phones all at once. OS include, of course, Windows Mobile, BlackBerry and Apple. Connections span Wi-Fi, VoIP, WiMax, 4G, LTE and email. Ah, I almost forgot: thou shalt also support offline mode while ensuring patient data never sits on the device itself. Last but not least, we want to be able to start a dictation from one device - let’s say a smartphone - and finish it on a computer since we’ve been interrupted 5 times in the course of the dictation and we ended up in 5 different places. Yep, that’s how we work.


Don’t force us to correct our own reports if we don’t have time for that. The question is not whether back-end is better or cheaper than front-end, the question is: why aren’t we systematically offered both options at any point in time? Speech recognition without flexibility makes no sense whatsoever, given the crazy environment we evolve in, not to mention the needs and workflows inherent to each department. While shadow gazers will love wrapping up their own reports at laser speed, my road-runner friends from the pit will want to dictate notes on the fly and have them routed to secretaries in the background for a later sign off. On my end, I’ll feel like front-end on Mondays and back-end on Fridays. So what? It’s gonna cost the equivalent of a Posh Spice wardrobe? Too bad. I won’t change my mood (I’m a doctor for Christ’s sake). So don't go wasting my precious time, get your price list together and we will be just fine.


Fuzzy grammar extracts structured data from sections of unstructured text while using pattern matching or more general grammar-based techniques. Example: I say “60 over 79” and the system goes straight to the diastolic pressure field and fills it in. Cool hey?

Even cooler: feedback on errors and low confidence results, ability to navigate the entire system with voice commands, ability to toggle languages and specialty-contexts anytime. Food for code…


What I mean by that is canned text that can be inserted into a report via voice control or a dedicated panel. For instance, I dictate a single word and the system expands that into a term, line, paragraph or even completed report.  Demographics or data from the patient dashboard can even automatically be filled in this way. That’s another way to eliminate repetitive data entry and related (stupid) errors.


On tablets and touchscreens. If handwritten notes still account for a good chunk of documentation methods, it’s not only because EMRs drive physicians nuts. It’s because it makes sense sometimes. So we want an engine that translates handwriting into structured data. Come on, don’t be hasty, give it a try.


Because that’s perfect for describing a wound, for instance. In this picture-perfect world, we want to pull up those images directly from the system and do our little marks and scribbling on them. But if I, Doctor Picasso, feel in the mood to draw it all by myself, I should be able to upload my own artwork as well.


What I mean by that is that a report is, in fact, a collection of sections (i.e.: allergies, past medical history, physical exam, medication, etc.). Physicians can enter information into each section of the document separately. Why is this important? Because when data is structured in sections as opposed to blurbs of text, it becomes searchable...and reusable. Sections, whether created by me, Fred the Bonehead or Cath the Jockey are considered can be inherited from one report to another, saving us time since we don’t have to dictate the same thing twice.

The system assembles each section in the background to form a completed report based on pre-designed layout, formatting, demographics logo, headers and footers as initially and centrally set up by the Medical Records Department in the system.


Once a patient is discharged, you gotta, you gotta, you gotta…create a pending report that reminds the physician or the department that a discharge summary must be created and released.  Easy as 1-2-free.

That’s it for today…apologies for the long post. Please feel free to comment with suggestions and ideas from your own wish list! Next week: I’ll tell you what I want, what I really really want when it comes to Training. And I’ll try to keep it zigazig shorter this time.

Cheers from the pit,

Doc Nado

No comments:

Post a Comment