Sometimes it’s hard to explain to people outside of healthcare the complexity of our workflows and the level of information density that providers have to deal with on day-to-day basis.
To help illustrate, I pulled this video from content creator, Dr. Conan Liu, a hospitalist at UC Davis, that demonstrates what should be rather straightforward task: how to admit a patient (to the hospital). In this case, our protagonist will be using Epic, the EMR industry leader by marketshare in the acute care hospital space (~44% of hospitals, and ~90%+ of AMCs).
As this video rolls, I challenge you (assuming you’re Epic naive) to keep up with what Dr. Liu is doing as he clicks through menus, tabs through fields, and fires up macros. Watch as he goes through his chart review exercise (starting about 5:00) looking at prior encounters, labs, and imaging and take a gander at HOW MUCH INFORMATION he has to analyze. Watch the brain scrambling process of putting in orders starting at 9:45.
This video demonstrates in the clearest possible fashion why we need artificial intelligence and automation to support providers of all types on the front lines. The video helps us undertand that the software interfaces are overwhelming (note that Epic is probably the best inpatient EMR re: UX/UI), the amount of data that we have to evaluate is overwhelming, and the operational challenges of patient care are overwhelming.
Interestingly, this video was uploaded on September 27, 2022, only 2 months before the launch of ChatGPT on November 30, 2022, which ushered in the new era of AI. Since then, tremendous progress has been made using LLMs for ambient scribing, intake automation, chart summarization, voice agents, revenue cycle management, etc. etc. Working at a company building these technologies, it’s amazing to see how quickly the field is evolving.
Hopefully one day soon, all of these manual inputs into a computer to document a patient encounter will be a thing of the past. Until then, know that there’s a lot of human time, energy, and potential wasting away on data entry.
