The dirty secret of all electronic medical records (EMRs) is this: they are not primarily designed to help doctors record, review, or share information about their patients. No, they are primarily designed to capture the data necessary to submit bills to insurance companies and the benefit to doctors and patients is merely incidental. If you’ve ever seen a printout of a medical note, you have noticed how long, bizarre, redundant, and confusing it seems. This is because it is designed to be most legible for the purposes of billing. But it’s not just incomprehensible printouts — this purpose is reflected in many different aspects of EMRs’ design.
By contrast, here’s a fascinating story about the 40-year history of the Veterans Administration’s ugly-yet-functional electronic medical record (EMR). What distinguishes the Veterans Information Systems and Technology Architecture (VistA) from every other EMR is the fact that it was built to capture and use information for clinicians, not for insurance companies. Of course, maintaining a legacy system like this is expensive, and so some the VA is moving away from VistA to Cerner, the EMR that Mayo Clinic decided to ditch after years of working with it. (I have never worked with Epic, the system that Mayo switched to — but most people I’ve talked to say that it’s slightly better than all the others). Ostensibly, this is so that the VA and DoD can share the same system — though the two different patient populations differ greatly and given the massive cost and inconvenience of changing systems, it’s unclear what benefit the VA will see in the end.
The ideal EMR doesn’t exist and may never exist. I’ll grant that it’s incredibly difficult to create a records system that collects, organizes, and displays information that’s useful to a doctor in a way that doesn’t force him or her to spend the entire visit staring at a screen AND THEN allow that information to be shared between all the hospitals and offices you might use while keeping your personal details incredibly secure. But we’ve made things even worse with by subsidizing bad EMRs that can’t even talk to one another: if you usually go to Dr. Smith’s office but you had a test done at St. Mary’s hospital 6 months ago and you walked into the ER at St. Miriam’s hospital because you started feeling chest pain, you have to call Dr. Smith and St. Mary’s hospital in order for both of them to fax you a printout.
Unfortunately, there isn’t an easy solution at this point. A more “market-based” solution probably won’t emerge because once you’ve bought an EMR, it’s incredibly expensive to change it over. As long as the pain of the EMR is mostly felt by frontline staff and patients struggling to communicate around the screen, there’s not a particularly strong incentive to change. Even solving the interoperability problem is full of landmines. But the whole thing is a cautionary tale about the rush to adopt new technology that doesn’t really solve the problem it was created to solve while making everything else worse.
Matthew,
Great article. I took a job as an EMR analyst shortly after the ACA spurred on a glut of Health IT jobs, and I’ve been working in that space ever since. I think you’re spot on here; EMRs were indeed designed primarily to collect the data necessary to submit billing to payers. But perhaps the big missing piece in your article is that the massive migration to EMR systems was also intended to usher in an entirely new payment model; moving from a low-tech fee-for-service model to a data-driven model wherein physicians and health systems are paid for health *outcomes* and adherence to quality standards.
Moving to a system that pays providers for outcomes – rather than simply the services provided – certainly makes a lot of sense. Obstensibly it will help align the incentives of doctors and insurers with their patients. But it’s also indicative of our cultural moment, which puts considerable faith in data to solve our problems. And certainly data is not without its benefits; research potential, syndromic surveillance, etc. But we also shouldn’t be naive to it’s costs.
In my experience, a data-driven outcomes-based model can cause a significant diminishment to the patient-provider relationship. Clinicians now spend up to 2/3 of their time documenting in EMRs, and even in the exam room only slightly more than 50% of time is spent talking to/examining a patient. But the problem is not just that clinicians’ jobs are becoming increasingly clerical; the focus on macro-scale population health outcomes inevitably changes the very lens through which clinicians view their patient interactions. Rather than simply developing a clinical plan based on the individual assessment, physician are being implicitly trained to plan interventions in accordance with population-level quality metrics.
“I know it’s March but… would you like a flu shot?”
“Lower back pain? Let’s go ahead and order some imaging.”
“You smoke? Please take this smoking cessation form…. again.”
“Now I’d like to ask you these ten standard questions to rate your possible level of depression.”
Of course there are benefits to population health management strategies. But we need to be careful and aware of how a data-driven health system can de-personalize care and abstract the individual to a population metric. Big data may very well improve overall outcomes, but also threatens to diminish our sense of “care”.
Thanks, Sam! You are exactly right about the goals (and unintended consequences) of data-driven population health practice. And on the one hand, I’m in favor of such practices since so many doctors are so bad at it, but I also hate the way that such a mindset infects our clinical practice and care (as you note).
I’ve argued before that a lot of that work could be shifted from doctors to community health workers and/or midlevel providers — who have better training in and more responsibility for thinking on a population level and following checklists. They’re also probably better at integrating data entry into their interactions which could then be used on a population level (assuming that they would have a more geographically sensible population!) But doctors are a cartel, so I doubt it’ll happen soon.
[…] things that might have been useful or beneficial to the patient. I’ve written before about how EMRs aren’t designed primarily for doctors or patients, and while I am still deeply concerned about the high cost and low benefit we’re getting out […]
[…] doesn’t mention that EMRs are designed for billing, not for helping doctors care for patients. But other than that, it’s a great […]