I first blogged about electronic medical records (EMR) almost three years ago, comparing the effort necessary to convert just the US medical industry to their use with the effort around Y2K, the effort to make computer systems survive the transition from 1999 to 2000.
The Y2K effort was huge. Most people today think it was a non-event, but the worldwide cost was estimated at more than US$300B. Oh, there were some amusing incidents like the inability of California to issue five-year driver’s licenses in 1995 and 1996 (the license showed up as expired when a police officer ran your license after a traffic stop). There was the hilarious story of the lady who bought a 10-year CD in January 1990, and the bank added a few million dollars of interest overnight (thinking that the CD had matured in 1900). Cars and planes and trains and ships, elevators and bank accounts and defense systems all continued to operate just fine. The boss of the company I worked for at the time brought in a big screen TV to the lounge, added a few cots, and scheduled 24 hour a day coverage by support and engineering teams starting December 30. We were all staring at the phones and emails when Guam hit midnight on 12/31. By the time midnight reached Hawaii, the boss sent back the TV and we all went home. Other than a couple of “hey, it worked!” emails, nothing happened. The many hours we had spent making it a non-event were effective.
The EMR issues dwarfs the Y2K issue. It will cost a lot more, and almost everybody will notice.
I went to an eye doctor two years ago. Behind the reception area was a large room with rows of tall file cabinets full of paper files, probably going back decades. The clerks had dozens of computer systems for appointments, billing and payments, interfacing with multiple insurance companies and governments, and lots of paper-based systems to keep records, doctor’s notes, prescriptions, and inventory. They were just starting to transition to an EMR system. They were ecstatic; soon everything would be in one system and they could get rid of the tons of paper. The doctors expected to be able to reduce the size of their clerical staff, get insurance payments quicker, and in general eliminate a bunch of the stress of just running a small medical practice. In preparation for my appointment, they had copied my paper records into the computer, and I spent 15 minutes with them reviewing it for accuracy.
I went back a couple of months ago for my annual (OK, bi-annual) checkup. There were some obvious changes. Oh, all the paper was still there, and all of the computer and paper systems still seemed to be in operation. What had changed was the addition of a couple more clerks, and an increased stress level on everyone. They were still in transition, and weren’t very happy about it. They had yet to get to any of the benefits, but the journey was “interesting.” I spent another 15 minutes with them reviewing my EMR for accuracy.
This is not a unique occurrence. I have yet to talk to any medical facility that has had a smooth transition. The best one was a transition that occurred in six months; most take one to two years, some are still going on after five years.
Scot Silverstein from Drexel University was quoted in a February 18, 2013, article in the Philadelphia Inquirer newspaper. Silverstein believes that we are rushing too fast to EMR and that the notion that they prevent more mistakes than they cause is not proven. He cites serious issues with some software components that printed orders for the wrong medicines, or the wrong dosage. “We are in the midst of a mania” to convert to EMR, largely spurred by government carrot and stick tactics: money if you convert, delays in payments if you don’t. “We know it causes harm, and we don’t even know the level of magnitude. That statement alone should be the basis for the greatest of caution and slowing down.” Silverstein does hold a minority view and he believes in the potential benefits of EMR over time, but “patients are being harmed and killed as a result of disruptions to care caused by bad medical IT.”
What happened? Why isn’t the transition to EMR as smooth as the Y2K effort? What happened is that IT folk treated EMR like they did Y2K – a huge project management problem, millions of separate things to be done in a specified time; 12/31/1999 was not going to slip. Yet it all involved computer systems. Fix the software, test it, and move to the next step. IT people know how to do these kinds of projects, no matter how large. The average person saw nothing, did nothing different other than, in a few cases, have to enter a four-digit year instead of a two-digit year.
But while Y2K was just an IT issue, EMR is an IT issue, a data conversion issue, and a people issue. The IT issue is complicated, but really has no unknowns. IT has done these kinds of software development projects before.
The data conversion issue is huge. The electronic data is relatively easy, although there is usually a significant code conversion issue: the old system used “measles” and the new system uses “783.2” and each insurance company has a different code. The hard part is paper. Even a small medical practice has tons of paper. For those skeptics, a box of paper (5,000 sheets) is 50 pounds. A typical four-drawer file cabinet even if not stuffed will have about 200 pounds of paper in it. It’s easier to count boxes or drawers – look in your doctor’s office file room and do a quick count. A ton is about 200,000 pieces of paper.
Much of the data on those sheets is hand written, often by people in a hurry and not known for good handwriting. A lot of it is second or third carbon copies, or faxed sheets. OCR (optical character recognition) technology does not work very well under these conditions. Each of those sheets has to be scanned, processed electronically, and then manually verified. A trained clerk can do much of that verification, but some will require a medical professional to figure out. This data conversion effort will probably be 99% correct, but that is hardly good enough since it probably means several errors per patient. Most are likely trivial, but some may be critical.
But the real problem is people. Every process changes: how you schedule appointments, admit patients, move patients between rooms, deliver medicines, conduct and review tests, bill, record doctor’s comments and directions, generate and fill prescriptions, …
The transition is a nightmare.
Some places try a slow phased transition – one system at a time or one ward at a time. In general, I recommend this approach because you learn something at each phase, but it has the problem that you have to keep both the old systems and the new system running for a long time, probably several years. When you have patients that move from an “old” environment to a “new” environment you have to scramble to get their information into the EMR system. Worse, if you have a patient who moves from the “new” environment to the “old” environment, you end up in a real mess that confuses everybody.
Some places try a cold-turkey approach. I have a good friend who is a senior doctor at a major hospital. They decided to switch everything at once. They picket 5PM Friday, since over the weekend it is primarily the ER that is really busy – everything else slows down significantly. They put on extra doctors, nurses, clerical people and representatives from the EMR vendor and concentrated them in ER for the weekend. I haven’t talked with her for five weeks. A mutual friend says, “She is very busy.” I suspect they all are.
This is what I think Scot Silverstein was really worried about: disruption. The addition of process stress on top of the normal stress caused by caring for people’s lives must lead to errors. A mistake can kill a patient, no matter where the mistake originates in the transition process.
The last word:
The value of EMR has not changed. When we get there, we will have a less expensive more efficient and safer care delivery environment. The journey is just longer and more difficult than anybody imagined. There is huge training effort required, which I believe is largely ignored or significantly short-changed. But it is a journey we all need to take. It needs to be carefully planned. Do not simply take the “migration plan” provided by the EMR vendor. If you do not have project planning and management people on staff, get some that work for you and have them create a workable plan. This process will take months, but is critical.
I repeat my recommendation from 2010: If you are a young software developer looking for a career, I have one phrase for you: “Electronic Medical Records.” But I now add that same recommendation to you as a business graduate. Medical organizations need even more help in the management of the transition to EMR.
Keep your sense of humor.