Solving the Last Mile Problem in Healthcare IT (Part 2)

In my last post, I described the Last Mile Problem in healthcare information technology (IT). To summarize: The Last Mile Problem originated with supply chain and delivery. It is relatively easy to get a manufactured good from the factory to close to the purchaser (note the word relative!). That last mile – from the post office or local storage depot – is by far the most difficult. Are the roads big enough for the truck? Will the driveway support heavy equipment? Will the customer be there when you deliver? Will the object fit through the door? You get the picture.

In the world of healthcare IT and the electronic health record (EHR), it might seem like we’ve solved most of the issues that doctors and nurses used to have with paper. Old charts are easily viewed. Lab results can be graphed and trended. Orders and documentation are neatly typed so no more guessing due to physicians’ abysmal penmanship. We’ve come so far, but precious few clinicians are happy. In fact, the EHR in the United States is blamed (fairly or not) for a significant role in physician burnout. Although our technology has gotten us far, we need to get past the last mile.

Newsflash: we won’t solve our problems with healthcare IT with more healthcare IT. At least we won’t do so with the tech that is commonly available today. Sure, it makes sense to work toward the day when the AI scours all available medical history and presents a clinically-accurate patient summary along with an ordered differential diagnosis and solid suggestions for discrete next steps. That’s great; I want that. But until that comes, I’m going to continue to argue that we need a combination of human touch and other ancillary technology to get past the last mile.

There are non-tech tools that can help. Let’s start with the concept of team-based care. If we really had wide adoption of teams in inpatient and ambulatory milieus, doctors could concentrate on things that only doctors can do. Imagine if a nurse or medical assistant (MA) looked ahead a few weeks on a doctor’s schedule to see what kind of patients were coming up. A nurse could make sure that routine labs were ordered and resulted. An MA could review specialists notes and letters that were generated since the last visit and queue that information up for the physician. (By the way, there are software vendors that can automate a good chunk of this prep for the nurse – say hello to Healthfinch for one.) None of the work that I just described inherently needs a physician to routinely perform it, so why do we ask doctors to do it, and often to do it in front of the patient at a visit?

While some see the very concept of scribes as an admission of EHR failure, I do not concur. Lawyers have long used paralegals to supplement what they do. For example, a lawyer might instruct her paralegal to prepare the standard contract but modify the intellectual property section per the client instructions. The lawyer is certainly capable of doing the detailed work but realizes that it makes more sense for someone else to do the actual typing, cutting, and pasting while she does work that only an attorney can do. Why do we eschew this in medicine? Scribes can help doctors produce high-quality, readable, actionable progress notes in real time. Since we don’t seem to be able to control the crazy documentation requirements in order to practice in the United States, we might as well seek help from others as opposed to what we generally do: just make the doctor do it.

The A/P is where the “good stuff” lives.

Another under-utilized ancillary tech tool we should use more of is voice recognition (i.e. you talk, it types.) Is this as nice as having a human help you out? No, it’s not. But it’s a heck of a lot cheaper, and it allows the physician to generate high-quality, concise documentation where it’s most needed: the assessment and plan sections of the progress note. Most of us in the medical field will freely admit that the A/P is where the “good stuff” lives. All those other data are important, of course, but the meat of the note should be in the physician’s assessment and plan. Modern voice recognition software can also do cool things like control the EHR. Today, doctors can say, “Order the routine 3-month diabetes labs” and “Have the patient return in 6 months” and magically, the lab orders are queued up and waiting for review and signature and the return-to-clinic box has the appropriate discrete data. This functionality is available today in enterprise-grade EHRs and voice recognition software. Today, people. Why aren’t more doctors using it?

That last question was rhetorical. I know why more doctors aren’t using advanced voice recognition tools. It’s because they lack training. Oh yeah, I said it. But you won’t find me blaming the training team and your local hospital or clinic. Oh, no. Typically, the CIO and CMIO are begging doctors to spend more time initially and ongoing to learn how to really use the electronic tools of the trade. But so few physicians take advantage of these offerings. I get it. Doctors are busy, and even when they get home, they often spend more hours finishing up work from earlier in the day. Who’s got time for more training? Dear reader, do you understand what is happening here? Doctors can’t (or won’t) make time for training that will help them to more efficiently use the tools that they have. Spend two hours with me today, and I might be able to save you a half-hour or an hour every day for . . . well, forever. That’s a good deal, but so many doctors won’t make that deal.

We’re tantalizingly close to making tech more doctor-friendly. I really believe that. But we need to push past the last mile and get the delivery done. And for that, we need all parties to come to the table and keep on pushing.