How smart do we want our electronic health record to be?

Somewhere between as dumb as a piece of paper and a pen, and too smart for our own good.

Many, many years ago, before we spent the majority of our office visit staring at a flatscreen LED and typing away, our charts were simple manila folders with those bendy metal bars that allowed you to insert new pages, separated into multiple sections with cool little colored divider tabs.

You’d open up the chart, find the most recent progress note, go to the end of that one, and just start writing.

Sure, there were red pages that listed allergies, little flags that said whether the patient was a smoker, yellow pages where a patient’s medical problem list were supposed to be kept updated, a separate page for medications, then a lot of laboratory and imaging and miscellaneous sections.

Paper and pen.

Not much room for natural language processing, population health management, data analysis, or artificial intelligence.

If you wanted to find out something, you had to go back and flip through the pages. Trying to remember when an event took place, whether this had happened before, trying to spot trends in their blood pressure or their renal function, all took place manually.

And remember all the chart reviews we had to do manually if you were interested in doing a research project?

Stacks of charts piled on your desk, incredibly labor-intensive work digging through all those pages trying to find what you’re looking for.

Today, we have something more, something smarter, and yet we are still going through growing pains trying to figure out how to have the electronic health record work for us instead of the other way around.

Those who designed it are convinced that it fills our needs, that it gives us everything we want.

Now, it’s certainly better than what we had before.

Remember the countless hours we spent tracking down lab reports, radiology reports, or consultant notes, and then manually having to put them into the chart?

Refilling medications has become a breeze, and despite all our complaints about the button clicking it takes, it’s clearly worlds better than in those earlier days, that exasperated feeling we used to have at the end of an appointment when a patient would say “Oh yes, and I need all my medicines.”

Handwriting out a dozen or more prescriptions on a paper prescription pad, or even worse a triplicate form for controlled substances, would leave us with cramps at the end of a long day of refilling meds.

And I can trend someone’s complete blood count to spot a gradual developing problem.

And I can reconcile medications from outside sources.

And for our patients who see specialists and subspecialists within our own institution, their consultation notes are instantly available for us to view, to know what happened at those appointments.

But some of the bells and whistles that have been built really don’t add much to care, and they were often created by people who think they are making our lives easier.

As part of our institution’s plan to do a huge conversion to bring every provider within the organization under the same electronic health record, we’ve been going through the functionality piece by piece, line by line, to see what’s there, to try and figure out what to put in for everyone to use, to try and standardize, minimize, streamline.

Each area of clinical care, emergency medicine, critical care units, surgery, ENT, OB/GYN, pediatrics, adult internal medicine, and so on, each has their core team of clinical advisors who are going through all of the tools and widgets that providers have available to document the care of our patients within the electronic health record.

The goal set for the team is to figure out what is useful, and what is not, to make a limited number of suggestions for things to add or change, and hopefully find a lot of things we can get rid of.

Just recently we started going through all of these in a weekly conference call, and all of the providers on the phone have been amazed at what’s there.

Most users of electronic health record barely scratch the surface of the functionality that hides within.

But are these worth it?

Do the macros, templates, forms, and smart sets, really make our lives and the lives of our patients better?

Take for instance those Smart Sets.

They were created to help streamline an episode of care around a single clinical situation that providers within that specific specialty frequently encounter.

They are an attempt to make it easier to collect the history, a review of systems, physical exam, and place any orders specific to the management of that particular problem.

There’s a Smart Set for just about everything.

Sinusitis. Routine well woman exam. Chest pain. Upper respiratory tract infection. Urinary tract infection. School health form.

The list just within outpatient adult primary care goes on for several pages, and each different specialty has their own list of shortcuts, macros, and templated texts.

Often these take the form of sections where you can move from question to question, toggling between yes or no, present or absent.

Take for instance gastroenteritis.

Here’s a little bit of the Smart Set that is built into the electronic health record for patients presenting to the office with symptoms suggestive of gastroenteritis:

Subjective: This <AGE> year old <GENDER> presents with these symptoms:

Abdominal pain:{YES/NO:***}

Nausea: {YES/NO:***}

Vomiting:{YES/NO:***}

Diarrhea:{YES/NO:***}

Fever: {YES/NO:***}

Duration: ***

Relevant dietary history: ***

Similar illness in contacts: {YES/NO:***}

Recent antibiotic use: {YES/NO:***}

Past Medical History/Diagnosis Date:

<PULLED FROM PAST MEDICAL HISTORY>

Objective: BP ***/*** | Pulse *** | Resp ** | SpO2 **% | Breastfeeding? No

Gen: well-nourished, well-developed ***male, NAD, healthy-appearing

Head: normocephalic, atraumatic

Eyes: EOMI, anicteric sclerae, pink conjunctivae

CV: RRR, no murmurs/rubs/gallops appreciated, PMI- 5ICS-MCL

Lungs: clear to auscultation b/l, good air exchange, no respiratory distress

Abdominal exam:

Bowel sounds: {ABDOMEN AUSCULTATION:***}

Tenderness: {ABDOMEN DEGREE OF TENDERNESS:***}

Rebound: {YES/NO:***}

Organomegaly:{YES/NO:***}

Masses: {YES/NO:***}

Ascites:{YES/NO:***}

***

Ext: no edema

Assessment: Gastroenteritis

Plan: Current Outpatient Prescriptions on File Prior to Visit: <PULLED FROM ACTIVE MEDICATION LIST>

Patient Education:

Pt counseled to drink frequent small amounts of clear fluids (especially Gatorade), then advance diet as tolerated. May try BRAT diet if desired, use meticulous hand washing, avoid dairy products for now, and rest as much as possible. Pt to call or return if high fever, worse abdominal pain, severe weakness or fainting, or bloody stools occur or if sx not fully resolved in 2-3 days.

These things are designed so that if there’s something I see frequently, using this Smart Set will make my life easier, as I can quickly flow through these fields to fill in the information to give me a fully fleshed out history, physical exam, and plan for care of the problem.

The problem is, for most of the things we do, there’s nuance in how patients present, how we interpret it, and how we proceed down the clinical pathway that helps us get to a decision.

Rarely is one presentation enough like all the others that these templates are really going to end up saving us much time.

True, there certainly are things that we do exactly the same way every time, and if we can find ways to save on keystrokes, clicks, and typing, then more power to those who provide them.

But for those of us who live and breathe in the words of the history, the subtleties of the physical examination findings, and a carefully crafted assessment and plan, this often feels like we’re trying to squeeze our patients into a rigid set of yes/no questions that they rarely comfortably fit into.

Just the other day, I saw an elderly patient with gastroenteritis, and it was only after I’d finished typing my history, documenting relevant findings on her physical examination, and clearly stating my plan for what we were going to try next, did it even occur to me to try typing the commands that brought up the gastroenteritis Smart Set.

In some ways, it looked a lot like my note, but it felt cold, sterile, and clinical — and not in a good way.

Maybe we should get better at using these, and maybe we should try to figure out ways to make them feel more natural, more like the way we really take a history and document what we find.

Our hope is that the electronic health record can evolve in the future, become more than just a glorified word processing system, and that together we can learn to harness its computational power and the new tools of natural language processing and artificial intelligence to help us identify gaps in care, recognize trends, start to become proactive and predictive in the care we provide rather than reactive.

I’m sure lots of folks out there using this functionality of the electronic health record really love using shortcuts, and I’m not saying we should get rid of them all.

But as we build a 21st-century tool to help us take care of our patients, we need to insist that it’s not just templates and macros we need, but a smart electronic system that stands by our side and augments the care we provide.

Fred N. Pelzman is an associate professor of medicine, New York Presbyterian Hospital and associate director, Weill Cornell Internal Medicine Associates, New York City, NY. He blogs at MedPage Today’s Building the Patient-Centered Medical Home.

Image credit: Shutterstock.com