EPIC is a medical software company handling the electronic medical records (EMR) of 78% of people in the US. Many healthcare providers complain about EPIC because of it’s clunky user interface. But, there are also some –very few– tech savvy people who like what EPIC is doing.
Personally, I can’t say much since I haven’t started on rotations yet and our hospitals use Cerner. But, I’ve done a ton of chart reviews in the past using EPIC and I can say I did not like what I saw. I also like to think that I am pretty good with technology. To be fair, I think the user interface has been getting much better over the years, but still nothing close to what it should be given the amount they charge hospitals annually –tens of millions– especially when you compare the experience to using social media apps or websites.
Judith Faulkner is the billionaire who started the company 45 years ago. She got her masters in computer science then proceeded to build an EMR empire in her basement with an investment of 70K coming in from family and friends. The company is still private today (i.e., they didn’t receive any outside investment). So pretty much, her family and friends made a very good investment. For a company reporting an annual income of 2.9 billion in 2018, they also only have 13K employees (as reported in 2023) which is not much when you compare that to Cerner which has 29K employees globally but a smaller market share than EPIC. Also, pretty random but the EPIC campus in Wisconsin is massive and is Harry Potter Themed. It’s actually kinda cool. Here’s is a picture of one of the dining halls:
The company is constantly growing and have already expanded their health system to several European countries. But, what’s a bit surprising to me is that there seems to be such few competitors in the EMR market. The main ones –all of which are well behind EPIC– are Allscripts, athenahealth, Cerner, units of IBM, McKesson, Siemens, and MEDITECH.
The main reason I am writing about EPIC is that they recently shared that they are integrating LLM’s into their software. This is great news because I think it has the potential to address a lot the problems they face in the past. LLM’s like GPT 4 have the potential to make the user experience so much more pleasant. The challenge will be making sure on the back end that patient information can be retrieved properly based on certain prompts (e.g., if you write a prompt asking if a patient’s lipid panel on a specific date was normal, the system would have to first find that info then give a response based on GPT’s interpretation).
The whole idea of improving EMRs by building on top of LLMs is exciting and has the potential to significantly increase efficiency. Whether it’s EPIC or some other software company that someone is starting in their basement somewhere, I look forward to the improved EMRs in the coming years.
References:
