Technology

Really great article on the importance of UX to software development.

Digitization genuinely can be amazing. It means that (with proper practices) nothing critical is ever lost, key information can easily be accessed and communicated without risk of loss or duplication errors, and additional analysis can be quickly and easily done (minimal cost per extra set of data analyzed; it does take appropriate design and investment to design the algorithm), and a variety of other benefits.

But if you don't have a very clear picture of how the data will actually be used in the real world, none of those benefits matter because the whole thing is a pile of shit.

Physician and Chief Medical Officer here. I started my practice on the tail end of paper records and I have been through, and led, several different renditions of electronic medical records. Like any other software, the technology has evolved and matured over the last 20 years. Electronic Health Records are just about patient information anymore either. They incorporate patient info, clinician orders, billing, and also provide numerous cross checks and safety systems to improve patient care.

The article discusses the issue with digitizing handwritten notes. This is always problematic as they are not searchable, indexed, etc. and can be very difficult to work through. The system where I work has been on the same medical record system for over 10 years. This provides an efficiency that I didn't have when I was working in a paper system. The entire chart is searchable. Medical history sections are filled out and robust.

From a safety perspective, these systems give us so much more than we had. One prime example is bedside scanning. When a nurse administers a medication in a hospital environment the medical record plays a pivotal role. As a physician I enter and order for a medication. If there is an allergy, medication interaction, or a host of other things, the system will alert me to it before the order is signed. Once signed, the order then goes to a pharmacist to review and approve. The patient's nurse will then be notified of the order and be able to pull the medication from an integrated medication storage unit with multiple drawers and compartments to ensure the right medication is taken. Finally, the nurse must use the EHR to scan the patient's armband and the medication barcode. If there is any discrepancy, the medication is not given until the issue is resolved.

From a user perspective, I can now create notes and write orders faster than I could in a paper world, even with dictation. Most docs use a combined template/voice recognition approach which works well. There has been a great deal of work in the Epic world (one of the largest EHR systems) to increase efficiency and improve the user experience.

Are these systems perfect? Far from it. But things are better now, at least where I work, than this article makes it seem.

It's true, electronic medical records range from garbage to totally broken in terms of usability for healthcare workers. Then you realize the actual customer is the hospital system, and all the design decisions start to make sense. Because the real purpose of the software is efficient billing, not patient care.

Finally something I'm actually qualified to weigh in on! I'm the lead UI developer for an EHR software (not saying which one or getting into details--it'd be pretty easy to figure out my identity).

First, to be fair, it's possible that the software they're using is genuinely terrible. They don't say which EHR. I've heard this kind of thing from providers before, though, and it's usually that they don't know how to use the software. From the way the article describes the provider, it sounds like they're stuck in paper and don't want to learn a new way of doing things. On the one hand, fair enough--patient care should be their primary concern. On the other hand, patient care is so much easier, faster, and more accurate in an EHR.

In my EHR you select a patient and can get a full visit summary on any visit the patient has ever had with a couple of mouse clicks. Immunizations, clinical notes, radiology, goals, problems, vitals, education--everything that happened during the visit. There are built-in tools for reminders that automatically notify you of things that are important for the visit based on previous visits, contraindication checks for medications, tracking of fluid balance, integrated documentation for clinical reference and distributing to patients, etc, etc.

That's not even to mention things like compliance for clinical quality measure reporting, integrating with state immunization registries, easy export of data to external facilities (eg, CCDA), using digital signatures for non-repudiation of controlled substance prescriptions, automagically pinging requests and data around to the different departments, etc. So many things that used to rely on a human squinting at a paper now just happen, with a built-in audit trail.

As for billing: we (developers, testers, and project/product managers) HATE billing. It's a necessary evil, but we package it off as a separate plugin. It can pull procedure codes and the like from the database to do its job, but to suggest that billing is the only reason to use an electronic health record is astoundingly ignorant. Patient care is the primary concern of everyone who actually has hands on the application. Most of us are former providers who just happen to be alright at coding.

I see a lot of people talking about how this is an issue of capitalism through the eyes of who purchases (c-suite) the electronic health record or EHR. This isn't really applicable when it comes to healthcare delivery systems. Every system has c-suite representation from the clinical side. CMO, CNO, CMIO, CNIO, etc. In addition physicians have strong lobbying power within the orgs to ensure that they are listened to.

Ultimately trade-offs need to be made somewhere, and the real issue is that these pieces of software are incredibly complicated. Have you ever stopped to consider how much information might be in your medical chart, for a single doctors visit? Prior to the visit they need to have or collect a bunch of customer data on you - name, date of birth, insurance info, etc. They need to schedule an appointment for you at a location with a specific doctor which means they need a calendaring and scheduling system and all the data that comes along with that. They may need to collect and scan documents about you, or get information from other medical systems. Then when you show up, you interact with more people than just your doctor - you get checked in, they collect a form of payment or the actual payment itself (meaning they need to interface with insurance to understand what to bill), then a nurse or medical assistant takes you back. A bunch of vitals get recorded - height, weight, blood pressure, pulse oximetry. Some of these come from devices which are hooked up to their system. Then the doctor comes in and does any number of things to you- there's a lot of narrative that needs to be collected, pieces of information about why you are there, your history, and so on. They may collect some kind of material from you for which the system needs to at the very least record that it was collected and what a result is (realistically it's interfaced and sometimes the interface includes media such as images). I could keep going on, but I think you get the idea - the amount that needs to go into a system to make it useful to all the various staff at a place of service means that the product is very expensive and difficult to create.

The real issue with capitalism comes in here - it's an issue of very few companies providing good products. It's very difficult to create a competing product in the EHR world because the established giants have been developing for 30+ years. They've poured billions of dollars and man hours into creating software that can manage the extreme complexity of medical care. Even among these giants which do hire clinical professionals to help shape the front end so it's as user friendly as possible, medicine is huge and there are people of all walks practicing - some are great at tech and others not so much. Being able to appease everyone means you need a flexible UX which also means... more money and more man hours. This problem unfortunately can't really fix itself until it's possible to create a complicated system with less resources, which I don't foresee happening anytime soon.