Conundrums of healthcare software
As someone with programming background muddling my way through the medical field, I’ve come to realize there are monumental issues with the software that is used in the field today–different systems flat-out refuse to talk with each other, analytics software is non-existent, and interfaces are horrendously non-user-friendly. Why, oh why, in 2010, do we not have better? There are plenty of people qualified to code conversion programs, thousands of mathematicians trained in the most efficient algorithms, and colleges are pumping out design majors like never before. What gives??
1) Not enough doctors are involved in the creation process. I am pointing the finger at every physician, medical student, and pre-med who sees a computer as nothing more than a terminal for the web and some word processing, maybe a few powerpoints and PDFs here and there, maybe a few games. Who has always just taken broken computers into the shop and paid exhorbitant prices, or gotten a friend, or just “someone else” to deal with the problem because they can afford it, “don’t have the time”, or “don’t know anything about computers”.
This is a huge. fucking. problem. The grand majority of healthcare professionals are not really computer literate, and that doesn’t seem to be changing anytime soon. True story: I fixed a classmate’s computer the other day when a boot file got corrupted. It took ~5 min to run a disk check and solve it. I got the fix off Google. Does this make me some sort of computer genius? Hell no. Truth: I didn’t have access to internet in my home until senior year of high school. It was 56K dial-up. If I can troubleshoot a computer, so can you. All it takes it takes is a bit of “get-up-and-go(ogle)”.
How, if physicians aren’t comfortable with computers, if they can’t imagine the possibilities, can they help make the next generation of software that will serve their patients? Let’s face it, EMRs are here to stay, and the more physicians become tech-savvy, the better the quality of healthcare will become. Even if a physician doesn’t help create the software, if they can provide competent feedback about its pitfalls and strengths, that alone would bolster the status quo immensely.
2) Software tries too hard. Take a look at any EMR that exists today. I guarantee you every single one will present the physician with hundreds, thousands of data points. Every single one will try to list every single condition, symptom, test for every possible scenario. This is in(s)ane. All it does is create an overwhelming interface that the physician wastes time looking over, hoping they marked everything correctly. Or they mark a few things and don’t bother with the rest, which is arguably worse.
Simplify, simplify, simplify. We don’t go through 12 years of grade school, 4 years of college, 4 years of med school, 3+ years of residency, and then X number of years in fellowships to have all the stuff we memorized be spoonfed to us again. There are an enormous number of situations where leaving the physician a well-placed text input area to just type their observations, or having a button to record audio clips would be a hundred times more useful than another list of all the causes of [insert condition here]. Stop giving us more lists, and give us the freedom to use all that training.
3) Software is designed for billing. Every single EMR is designed around billing people. And indeed, why shouldn’t it be, you ask? It is, after all, how we make our bread and butter.
Issue: the fundamental cornerstone of software is that it will do what you designed it to do. Look at Google. It was designed to search. The revenue from ads came after they’d perfected their search product and people actually found it useful. Medical software is completely backwards–it’s made for money first and patients second. It’s being used more because hospitals buy the hype, and less for the reason that it’s actually well-made and a genuine improvement to the workflow. This needs to change. It’s the reason why there are no analytics, it’s the reason why user interfaces suck, and it’s the reason EMRs are slowing down healthcare with administrative burden rather than speeding it up.
4) Competing companies don’t make products that are compatible. They’re so caught up in gouging the medical field for incompetent software that, again, they don’t think of the patient. True story: at a very well-funded private hospital in California, a child presented to the ER one night. At a follow-up with his pediatrician a few days later, the pediatrician discovered that because the hospital’s ER and clinic utilized software from competing EMR companies, she couldn’t access her patient’s ER treatment record. She had to walk over to the ER, present her credentials, and have the record printed out for her because faxing patient records was an unacceptable security risk. Imagine if, instead of two different departments, this scenario had taken place with two different hospitals. In two different states. See where I’m going with this? It’s not hard to see that patient data often just falls through the cracks.
Companies need to focus on delivering systems that are capable of crosstalk–this would deliver a much more comprehensive and smooth patient outcome. They need to see past their profit-driven myopia and focus on the bigger goal of patient care. For that, (3) needs to change first. Whether that will happen sooner rather than later is something that remains to be seen, but I’m certainly not holding my breath.
~~~~~
Ranting aside, I’m really hoping that in the future I get the chance to help create an EMR that properly addresses these issues, but for now, I guess these will serve as my notes of what to avoid…
Am I wrong about something? Did I miss anything? Sound off below! 
about 1 month ago
While I think these are problems, I do not think they are the root problems.
I think rushing to market so the companies that make the software don’t get left out has a lot to do with it. Because of that rush, they forgot the UI designers. More than anything else they. need. UI/UX. experts. DESPERATELY.
Also, you seem to be against the lists. I don’t think the lists per-say are bad. You need to classify people in a consistent manner that text fields will not do. The analytics you want and EMRs need won’t be able to be accomplished with textfields. One doctor will abbreviate, one won’t, the next will spell it wrong etc. Textareas should be involved, but not for categorization. It’s the UI/UX of the lists that is the problem. I think color-schemed categorized tagging would be a hugely beneficial UX feature (but would probably have to be personal to the doctor and the doctor maps to the overall stored information one time). Card sorts are in order. Decision trees. Make the UX like the medical professional thinks (there should be an advanced role based permission system instead of what they seem to have which is access/no access or medical/billing/patient).
Which brings me to my next point: Another root problem is the engineering decisions. Why, oh why, oh why, are they client-based applications instead of internet applications? Don’t tell me it’s because of security concerns because both of the medical networks I’ve encountered in the last 2 years have given me online access to a specific level of my information. If Amazon can keep credit cards safe, hospitals should be able to keep medical records safe. AJAX/jQuery/etc. are a UX persons best friend. Compare that to programming in .NET… I’m just saying no wonder the UX is bad – the language doesn’t have the features. You don’t even have to worry about cross-browser compatibility because remember it’s being deployed in a managed, known environment that you can dictate if you want.