Random Thoughts from a Restless Mind

The late, great Larry Weed, M.D., Professor of Medicine at the University of Vermont predicted both the age of EMR as well as the advent of IBM’s Watson, “Big Data”, and machine-learning in the practice of medicine. With the problem-oriented medical record in the form  of the SOAP note (Subjective -> Objective -> Assessment -> Plan) he codified a universal approach to essentially any medical problem evaluated in any patient. What was then called the Medical Center of Vermont implemented a data warehouse which allowed instant viewing of test data by computer throughout the institution in the early 1980’s (the first “EMR” if you will), and sister institution the Maine Medical Center solved the problem of the handwritten orders by adding computer order entry (CPOE) in 1984 or so. Despite all of the hoopla surrounding the Accountable Care Act’s carrot and stick drive to digitize the medical record, the horse was already out of the barn and slowly walking in that direction in the 1980’s.

Why, then, is the EMR landscape such a mess in 2018?

Our American healthcare landscape is blessed with a number of very large, prestigious institutions. They are self-professed and incessantly self-promoted as leaders in both thought and action when it comes to the advancement of medical care in all ways in the United States. It is right here in the laps of the leaders of those famed institutions that blame rests for the debacle that is the modern EMR. As early as 1990 and as recently as 2008 the opportunity to lead presented itself to our most august institutions. When given this opportunity to develop a new, better type of medical record that would aid in every aspect of caring for patients, our most important medical institutions punted.

When you think of the best medical care in the country, who do you think of? Pretty easy to answer that, I bet. The Cleveland Clinic, The Mayo Clinic, Yale, Stanford, the hospitals that made up what has become Harvard Pilgrim Health like Mass General, Brigham and Women’s and Beth Israel, Johns Hopkins, Baylor. Household names, all. Every single one of these institutions seeks to portray itself as the ultimate example of excellence in medical care, devoted above all else to the development and provision of care better than any and all competitors. Not only that, each wishes to project the most pious of images, one that espouses their monk-like devotion to doing what is best for their patients before all other considerations. With a building consensus that record keeping the old pen and papyrus way was hindering both present and future care, and indeed might be contributing to harmful care, the era was ripe for any or all of these presumably noble, altruistic non-profit institutions to answer the call.

When American healthcare was ready to look to any of these institutions to lead us into the digital information age, each and every one of them abdicated. The leaders of these and other great institutions had the chance to develop a true medical record in digital form that was first and foremost a tool to be used to improve the care that was provided in their institutions. They had the resources. Any one of them could have taken a leadership role in its development, not unlike the kind of leadership many of them have taken as the first institution in on cutting edge medical care such as organ transplantation or new generation cancer care.

Instead, both early and late, the leaders of each one of these major institutions chose a path with an eye not toward how the EMR would engage in the care of a patient, but in how it would engage with accounts receivable. Each institution opted to prioritize the growth of revenue over improved care. Everything is about maximizing the income of the institution, while at the same time minimizing the risk associated with billing.

J’accuse.

Think about that second part for a moment. EMR’s are not designed to promote the safety of an individual patient as she goes through her care experience (despite what the marketing brochures may tell you); for safety they are designed to limit the likelihood that a payer audit will find a lack of documentation that supports the charges. The bigger the company making the program, the greater is this emphasis. In the early 00’s any one of the above institutions (and Texas, and Ohio State, and Dartmouth, and…) could have launched a program that met all of the MEDICAL criteria for a good record. If they wanted to make a profit they could have sold the rights to use it.

Why don’t EMR’s communicate with one another? Were you aware that even institutions that run software from the same vendor do not have the ability to simply put notes from one another into a universal chart? Crazy, huh? Frankly I’m not really all that sure who is to blame for that particular bit of nonsense, but the obvious answer as to why your Epic chart can’t communicate with, say, Nextgen lies with that abdication of responsibility I spoke of above.

By not taking control of the process of EMR development at the outset all of our major medical institutions learned that 1) they never really bought an EMR, they just rent it which means that 2) they no longer really own their own information. What better way to remain in control if you are Epic than to prevent The Cleveland Clinic from banding together with The University of Pennsylvania as a bargaining unit than to prevent them from sharing patient information ON THE SAME DAMN PLATFORM?

J’accuse.

To their collective shame our most prestigious medical institutions and their leaders sold their souls by prioritizing their role as commercial entities rather than as leaders in medical care on behalf of patients. In the process they allowed themselves to be enslaved by the commercial interests that now control the medical record. Worse than that they created an additional barrier between a patient and his own medical record.

There has to be a bright spot, right? Some shining beacon, a last bastion, someone willing to stand against toute le monde and defend the honor of academia, to not become the next rhinoceros?  Certainly some institution was willing to stand up and do the right thing by saying “screw it”, we’re gonna make a killer EMR that does everything that Larry Weed said it should do first, and then figure out the billing crap later, right? Perhaps the medium sized Intermountain Health in Utah is on the right track, but all of the really big institutions turned belly up to submit to the demands of payers, hoping for a treat and a  belly rub. Surely UVM, the home of Larry Weed didn’t cave, right? The University of Vermont must surely have been driven by its early entry into the world of digital information management and created its own EMR that both houses information in a clinically relevant way, as well as allowing for computer-guided decision making, right? RIGHT?

Nope. Sorry. The University of Vermont runs on Epic.

“The essence of Medicine is story—finding the right story….Healthcare, on the other hand, deconstructs story into thousands of tiny pieces…for which no one is responsible.” –Victoria Sweet, M.D.

Being forced out of your comfort zone in any endeavor is always painful. In my experience it is also conducive to learning something new, and at least in my case it is a catalyst for creative thought. What, then, have I learned from our forced-march, point-of-a-bayonet transition from one EMR system to a new one? Are there any lessons to be learned on a broader scale, beyond the walls of SkyVision? Can I take this bowl of lemons and create lemonade that can be passed around the much larger table that encompasses the broad landscape of American medicine?

First off, our collective experience with our transition reinforced my long-held contention that you simply can’t effect change in a system of any type without either being a functional unit in that system, or shadowing those who work in the system you wish to improve. Imagine designing the cockpit of the next generation fighter jet without ever actually either flying one or sitting next to someone while they fly it. Take a look back at my essay “EMR and Underpants”; our information ecosystem was designed by engineers far, far away from the point of care delivery. It’s roughly the same as giving someone the job of choosing what underpants to deliver for your daily wear without ever having seen what you look like or talking with you about how you wear your clothes.

After all of our struggles there does appear to be one, huge 30,000 foot lesson in all of this that should, by rights, become the foundation of the next wave of innovation in EMRs: the spoken word is the goal. What made our traditional scribe process so successful in both efficiency and accuracy was the development of charting based on a spoken narrative. The doctor would dictate exam findings. The scribe would then intuit the various diagnoses from the conversation occurring between the doctor and the patient. While the doctor then went on to outline the plan of action this, too, was transcribed into the medical record. It was a natural and familiar way for all of the players in the room to communicate.

Why can’t I do that with any of the EMRs available on the market? Why is it that I can’t talk to an EMR and have my verbal encounter become what we would all recognize as a progress note? Heck, I’d be thrilled if there was an interim step in which all of the BS clicking we are doing to check all of those boxes could turn into something that looked more like spoken English (although our new EMR is OK and getting a bit better on this). With all of the hundreds of millions of dollars being raked in by EMR behemoths like Epic you mean to tell me they can’t find the resources to make this happen? Please.

You see, the essence of every healthcare interaction is the spoken word. When you have to stop talking or listening you have devalued time. Think for a minute from the patient’s point of view: it doesn’t matter whether it is a doctor of some other kind of worker in the room, once attention is shifted from the patient to the screen quality plummets. Make me a poor man’s AI interface that I can cue verbally to let it know what I’m doing and put it in the right box so that Uncle Sam won’t ding me for being a poor data entry clerk. I’d even be willing to talk to Mrs. Pistolacklioni about her smoking at every 3 month follow-up for her severe glaucoma (a disease that has no increased risk if you smoked, by the way).

While I’m at it, and as long as we are talking about communicating (cue Paul Newman in Cool Hand Luke), may we please find a way for the real medical record to be freely available on every platform? Seriously, how did this one escape the cloistered engineers and double-blinded underwear salespeople? Your Samsung cell phone can call your buddies iPhone and vice versa. An airman flying a MIG 22 can communicate with an inverted Tom Cruise in a 3g dive because there is a single standard for radio transmission and reception. Come on. This is basic stuff, the equivalent of declaring the gage of railroad tracks. You mean to tell me that the same people who think they know so much about how things must be that they have an opinion on the shape of operating room hats somehow missed this? Again. please.

I’m not kidding about the OR hats by the way; some DA administrators simply declared that bouffant hats were safer because they think so and won’t come off that even in the face of randomized control studies to the contrary.

Seriously, go all the way back to Dr. Larry Weed at UVM in the 1980’s and return to his beloved premises. There is too much information to be contained in any one doctor’s head, and doctors cannot avoid their biases and frame of reference when making medical decisions. Having true interoperability across all platforms would allow the free movement of information at the direction of the patient, the person who should be in control of that information after all. (Note: Carbon Health is on to something)

As a society we’ve allowed ourselves to remain captives of the trial bar’s defense of the status quo when it comes to malpractice lawsuits. This, in turn, has prevented us from examining repeating errors to determine if there might be a common thread that could be altered and thereby reduce their frequency. Interoperability would allow just the sort of root cause analysis that is needed, and because it would be done using anonymous information no actionable disclosure would be necessary from the doctors involved. As a bonus this would probably allow us to create true, vetted care protocols for the majority of patients with the majority of problems, and this evidence based care would then have to be admissible in court. All that would be necessary would be for doctors to explain in their chart why they decided to deviate in an individual case if that came up. Bingo, a data-driven solution to defensive medicine, all from better communication.

My new vendor is unaware that I am writing this, but interestingly has invited me to consider joining their advisory board and to speak at their annual convention. Who knows if those invitations will continue to be extended once they read this, but if they are I will have two very simple, very basic messages. This whole medical record thing should be about communication, just like it’s always been from the days of Hippocrates. That, and that Larry Weed was right. Before we go any further forward go back and read Larry Weed.

All we need is a little electronic SOAP to clean up this mess.

Being an eye doctor in 2018 means that you will take care of patients whose care is covered by a government program of some sort. In order to be able to get paid for your labors you need to record your work in an electronic medical or health record (EMR), and that EMR must be able to comply with  certain diagnosis and quality reporting standards. Failure to comply with these requirements does not mean you can’t take care of these patients, nor does it mean that you won’t get paid for doing so. It just means you will eventually get paid roughly 22% less for that work than someone who has an EMR that does comply.

15 months of effort to get our legacy system into compliance led to 3 months of research culminating in the purchase of a new EMR with a very sophisticated, dedicated ophthalmology/eye care format. With our purchase came on site training (with overtime pay for staff) and literally hundreds of man-hours of preparation work (on the clock) performed by both staff and doctors before we went “live”. The entire adventure was nothing less than a series of “OMG, you have GOT to be kidding” surprises for each one of us, starting with this killer: I would have to pay to retain access to the information SkyVision had gathered on our patients over 13 years. Yup. You heard that right. Even though we would never enter another electron of information into our old system, in one way or another I was going to have to ransom my own medical records.

As embarrassing as it is to admit it, I probably own that particular surprise. Really shoulda seen that coming.

What I also didn’t see coming, indeed what none of us saw coming, was just how different it is to practice medicine in the age of EMR. From Hippocrates through Osler and on to Marcus Welby and whatever the name of the doc played by George Clooney in “ER” was, medical care proceeded in the same orderly fashion. Once again we have Dr. Larry Weed to thank for codifying this process in the form of the SOAP note. Subjective -> Objective -> Assessment -> Plan. You listen to your patient’s story, cataloguing her symptoms and their salient characteristics (onset, severity, duration, etc.). Next comes the collection of data including your exam findings and any test results you may have. From this accumulated knowledge you make a diagnosis, or at least assemble a differential diagnosis, either of which launches a plan of action. The flow is so obvious that it’s somewhat astonishing that it took Dr. Weed to publish this as a process breakthrough.

From the minute we sat down with our laptops and tablets in front of us to learn how to use our new EMR, every single SkyVision staff member fell through the looking glass into a world gone, at best, sideways. Charting to billing, documenting everything that goes into taking care of a patient from the primary point of view of the payers, renders the SOAP model moot. Everything begins and ends with the diagnosis, the Assessment in SOAP-speak. What you plan to do comes next, and you now have to justify what that will be by demonstrating that the diagnosis can be found in the data. Your patient’s complaints have to be explained by your findings. Our tidy little straight line progression handed down from Hippocrates has been scrabbled. SOAP has become APOS.

How perfect is that?

Everyone is aware of how time consuming it is to enter data into a compliant EMR. There is just an endless number of boxes to click, even if you ignore the nonsensical sections that apply to worthless quality measures (childhood vaccine history review at the dermatologist? Smoking cessation at every eye doctor visit?). Even with the pre-loading and on-the-fly development of protocols that “pre-fill” all of the boxes for very common evaluations (e.g. cataract surgery in my world), it just takes a boatload of time to enter all of the information that is demanded. I hear those clicks in my sleep.

Remember, I already used scribes to enter information; if they are slowed down patient flow slows down, too. If I stay and enter information myself my schedule backs up downstream. If the scribe stays with the patient in the room after I’ve gone on to another patient there is no place to put the next patient in line. Leaving the charts “open” so that they can be “finalized” later is an option, of course, but one with three penalties. The practice gets socked with overtime expenses, the staff is overworked and can’t be home, and believe it or not that open chart is “timed” as a quality measure as if the patient was there waiting all that time. Doing a better job ends up dinging your quality score. Merde.

So what did we do and how did it go? We started 5 months ago with 3 charts in the new system per doctor per 1/2 day session. Sounds pretty reasonable, huh? Ease your way into it. Try not to upset the whole apple cart. Maybe just bruise an apple or two. The plan was to slowly increase the number of charts filled in the new system each week by slowly expanding the type of visits we recorded. You know, post-ops before massive, complex pre-op evaluations. New patients who didn’t have any data in the old system. It sounded pretty good when our trainer suggested it. Naturally, as soon as we expanded our universe of new EMR patients we crashed the entire office flow. What had been a finely tuned machine that seldom ran even five minutes behind on a single patient became a battlefield filled with folks waiting 30, 40, even 60 minutes for their exams within an hour of the opening bell.

It was like jogging in quicksand.

I’d really love to tell you that 5+ months in it’s all unicorns and rainbows. That we are now up and humming along, seeing the same number of patients we always have and running on time like we used to.  I’ll admit to occasionally coming across a random footprint that might have been left by a unicorn, and every now and again we catch flashes of color, a rainbow seemingly just out view. We had to hire a part-time tech to assume the task of “pre-populating” the new EMR charts with information from the old system. Every staff member has had to drop parts of their duties to take on the tasks of entering patient information on the front side or finalizing the chart entry so that it is consistent with our billing on the back. I will have to buy access to my old records in the old format, at least temporarily, so that we don’t get slowed down learning a new way to look at old data.

The best way to describe where we are after 5+ months is that we are now running rather than jogging in that quicksand. Exams that once kept a patient in our office for a maximum of 67 minutes now take closer to 90 (we really do track that kind of stuff). Where we rarely had a single patient more than 15 minutes behind schedule we now routinely have  5 or 6 who run an hour late every single day. A couple of week ago I was worried that this one change was going to drive us out of business because of the increased costs, and what I assumed would be mounting ill will from patients who were disappointed in their wait times and stopped coming to see us. Not gonna lie, it didn’t look very good.

A funny thing happened on that road to ruin paved in quicksand: my staff and my patients collectively said “no way.” Crazy as it sounds, two groups of folks who were suffering alongside me looked at the alternative and said “no”. Oh sure, there were certainly patients who trashed us on rating sites because we ran late on a single visit, including some who’d given us straight 5 out of 5 stars for years. But most of them read our “Under Construction, Pardon Our Dust” signs, gritted their teeth, and basically said that we’d earned their patience. Staff is coming in early and staying late. They are huddling and brainstorming ways to restore our flow. Our charting is no better than before but we do send out better letters. Some day we may even be able to do some of those things that Larry Weed talked about when it comes to managing large amounts of information and making complex decisions.

But for now it’s still nothing but pain. It’s hard and the hardship is slow to abate. We all feel the sense of unfairness, that we were forced into this position, and that what we have now does not make our patients any better off than they were before. I would not have chosen this path, not for any reason, had I not been forced to do so. I have no idea, and I will never know if it would have been easier had I picked the other option. Beware all ye who travel here. You are about to embark on a journey where each step is taken in quicksand. It will be a long, long time before you are cleansed of the residue.

Remember, your SOAP has been replace by APOS.

Unsportsmanlike conduct, piling on should have been the call, but alas, no flags were thrown. After roughly 15 months of crossing our fingers and hoping that our original EMR vendor would be able to guarantee our compliance with the twin pitchforks of ICD-10 and quality attesting it became clear to us that we could no longer afford our “spend and pray” strategy. We felt forced to initiate divorce proceedings with our legacy software and begin the search for a new program that would ensure our compliance, and in so doing allow SkyVision Centers to survive as an independent entity.

Well, that’s what we thought we were doing anyway.

Our original search for an EMR program in 2004 was undertaken from a position of strength in the skinniest of markets. There were only a few vendors who made a product suitable for eye care, and we were making a “want” buy rather than a “need” buy. In this position we had the luxury of working with the ultimate game plan: we could play to win. By this I mean we could clearly state what our objectives were and lay out in clear terms how we expected our new EMR to enhance our business. I mean enhance in every single manner, most definitely including the bottom line. Our search was meant to bring in technology that would not defend against loss but to help us bring in more revenue, to help us win.

SkyVision Centers entered the eye care market with a single, borderline maniacal focus on enhancing a patient’s experience in the office. Face it, there is no way that any but the most sophisticated patients are going to be able to differentiate between doctors and practices based on quality measures having to do with outcomes and safety. Our medical world is quite opaque in ways both inadvertent and purposeful. Discussing “bad doctors” or “bad hospitals” just isn’t done. By the same token, touting better outcomes or safer care is considered borderline slander by other doctors and institutions. Hence we simply considered all of this–quality, safety, etc.–to be the “table stake”, an assumption that each patient made, and we decided to set ourselves apart by our focus on how each patient FELT during their care.

If you think about this, we should also  be able to make all of these assumptions about something as basic as an EMR, right? That it should enhance both the doctor’s and the patient’s experience during care, or at least not have the opposite effect. Our original EMR was quick and efficient, was adaptable to our existing care and process protocols rather than the other way around, and allowed us to maintain eye contact with our patients for >90% of any care experience in the office. This, more than anything else, explains why I hung on for so long after any objective outsider would have already jumped ship to a new, regulatory compliant program.

It took precisely 90 seconds to realize that our patient-centered ideal was going to take a hit by changing our EMR. That’s how long I had to listen to the consensus best patient/doctor interface among the “modern”, compliant programs. Every single program arrives with a pre-set protocol, an indelible and unalterable set of mandatory processes that you WILL implement into your practice. Another 90 seconds spent reading the front page marketing pitch of this new batch of EMR candidates makes it clear that you are not looking at a program designed from the doctor/patient interface out to the billing office on its way to the payer. Every single product now sold that will comply with the various and sundry “quality” and reporting requirements is built from the billing interface back to your exam.

That’s an awfully tough pill to swallow.

Let me take just a minute to address the subject of scribes, staff members whose job it is to transfer the data that a doctor obtains and put it into the medical record. The solution to all of the problems created by EMRs is supposed to be solved by using scribes. Many (most?) doctors who are new to EMR are also new to the concept of using a scribe. Not so, though, in ophthalmology, at least at the level that I have practiced since I left residency training. For the most part I have had a scribe in the exam room with me from my very first day of practice. Our scribes were not just Carol Burnett Show era secretaries but highly trained technicians who simply rotated through their turn writing exam findings, diagnoses and plans with a pen on paper. I’ve always had this, and we utilized scribes in EMR v1.0 as well. Part of the unfairness was that I anticipated the need to hire MORE scribes simply to tend to the software’s protocols, so we were losing before the game even started.

What then does it mean to be playing “not to lose” when choosing the next EMR? The very first premise is little more than trying not to lose money you’ve earned by receiving a penalty for inadequate fidelity to all of those reporting rules. After that it becomes something more like trying not to lose your soul. Which program would allow us to maintain as much of our substance and our style while allowing us to continue to take care of the same number of patients? I was willing to forgo growth (loss #2 before game time). How long would it take for us to transition between platforms? Was there a program that would let us go fast enough that our patients would forgive us the obvious change in what it felt like to be in the office, even if that change was a 25-50% longer experience? I mean AFTER the transition. 25-50% longer in the office AFTER we are good at the new program. That was the reality I encountered in my search.

I’m not gonna lie. My biggest fear was that I would choose the wrong program. Well, check that; my biggest fear was that I would choose a program that would hurt us more than another one I might have chosen. I did my homework. I reached out to colleagues who do what I do and had chosen an EMR in the last 3 years. I went into offices and watched staff members and doctors use the programs I looked at. It was unbelievably depressing in all honesty. The lack of eye contact with patients and the slavish attention demanded by the computers in the room was appalling, not to mention the drudgery. Death by a thousand clicks.

Check that. A million clicks.

A consensus arose among eye doctors, one that I agreed with, that there are two EMR products on the market that are better than all of the other options, and that it is essentially a toss-up between them. Every colleague I chatted with felt this way, including those who were content with their present programs; they would choose one of the two (and not their present program) if they were starting from scratch. After narrowing down my choice to two it was almost a coin toss to be truthful.

After pouring lots of money into that car I already had in the hope of returning it to functionality I was ready to buy a new car. To spend yet more money on one of the two choices before me. No matter which one I chose, I was choosing something that would mean an existential change in how we practiced medicine at SkyVision Centers. Because there were only two, no matter which one I chose I would forever be haunted by the question of whether life might have been just a bit less unpleasant if I’d chosen the other one.

Changing your EMR because you have to is like the War of the Roses: there is no winning or losing, there are only degrees of losing.

Next Chapter 3: Jogging in Quicksand (where only the “A” counts in SOAP)