Have you ever wondered if the doctor who saw you in his office would recognize you if you met at the grocery? He or she might not, and it’s not just because you’re in street clothes. It might be that, with new health care regulations for computerized data gathering, he spent most of your annual checkup facing the computer screen, not looking at or interacting with – you.
Obamacare has done many good things: getting 20 million more people on health insurance; eliminating preexisting conditions as a reason to deny coverage; allowing young people to stay longer on their parents’ insurance. It also mandates conversion to electronic medical records (EMR), a great way to enable doctors and hospitals to exchange patient information, reduce duplicate tests, and improve patient care. At least, that’s the idea.
Things started well enough when hospitals had their own in-house computer gurus, but their ability to share records with other systems was limited. It still is today, and that’s not the only issue. For many providers, the process has become one of epic proportions, especially with Epic software, the company which controls a majority of the multi-billion-dollar conversion business. “On a really good day, you might be able to call the system mediocre, but most of the time, it’s lousy,” according to an emergency department chairman in San Francisco.
Clearly the system was not designed with the end-user in mind. As the author of Digital Doctor put it, Boeing engineers would never design a cockpit computer without input from the pilots.
Epic promised a comprehensive data collection system to maximize and speed up reimbursements. Docs say its software may be useful for a general hospital or clinic, but its protocol is not suited to facilities providing specialized care, such as for patients with cancer, burns, physical therapy or diabetes. Providers across the board are experiencing the pains of conversion. And patients should be prepared for scheduled appointments to be cancelled, prolonged waits, lengthy visits and even lost health records.
After billions of dollars spent converting to EMR, hospitals are still stymied in the exchange of patient records. Lack of interoperability has raised the ire of Congress. And other aspects of the software are decidedly not intuitive or user-friendly. If data aren’t properly entered, there will be coding problems that thwart reimbursement. And the decision trees do not necessarily conform with the ways doctors elicit information from and assess patients. Both doctors and patients are discouraged from asking open-ended questions and patients, from volunteering information, while doctors fill out rigid check lists. At other times, the system shuts itself off mid appointment, and doctors have to log in again.
Because the idea is to have the most comprehensive information on each patient visit, a doctor in a follow-up session can’t skip the far-reaching information gathered in an original visit, asking only the relevant questions and skipping others. In effect, the software architecture is controlling their thought processes. Some time-pressed doctors simply input old test data using the current date.
A doctor who is stymied and can’t “close an encounter” (that is, finish a report on a patient visit) is supposed to call an in-house “super user” or a help desk. Getting help can take two or three days.
Because Medicare reimbursements are structured to allow only so much time per patient visit, doctors can’t focus on the patients and input their notes immediately afterward because they are forced to move directly on to the next patient. The doctor can no longer dictate patient notes remotely but must do it sitting before the computer in order to access the 10-digit code of the patient visit. All this means that the doctor can see fewer patients. I’m told that Maine Medical Center, Lahey and others have actually lost money in the early stages of using this software.
Medicare cuts reimbursements to providers not complying with “meaningful” use of electronic medical records (EMR). The real meaning is that patients have to compete with the computer for the doctor’s attention. Neither doctors, nor nurses nor patients are happy about this. Apparently the only entities satisfied by the new software are the vendors themselves (including Cerner, Allscripts and Epic), whose lobbying enabled their participation in the shaping of federal IT regulations and standards for health care.
They were also lavish in their campaign contributions. Ironically, despite the government’s push for nationwide uniformity, the Pentagon last summer awarded a $4.3 billion contract to Cerner.
As one doctor wrote in a 2012 issue of the Journal of the American Medical Association, the federal government is spending in excess of $20 billion to incentivize converting to electronic medical records. We should all wonder how that money will be spent and what will be the human costs incurred along the road to electronic nirvana.
I’ve seen a few of the “old guard” doctors still jotting notes while facing their patients, not letting the software come between them and those seeking their help. But they seem to be the exception these days. Frustrated, many are being driven into retirement, just when there’s a shortage of doctors.
There is no going back, nor should we. But software vendors must be compelled to get end-users involved and be as actively involved in resolving problems as they were cashing in on the opportunities presented by Obamacare’s push for EMR. Only then will patients realize the quality care promised in the Affordable Care Act.
I welcome your comments in the section below. To be alerted when a new blog is posted, click on “Follow’ in the lower right portion of your screen.
8 thoughts on “Rocky road to electronic medical records”
Pingback: When public policy meets private patient by Marjorie Arons-Barron - RichardHowe.com
Pingback: When public policy meets private patient | Marjorie Arons-Barron
Interesting opinions; I second Ms. Arons-Barron’s reference to Wachter’s “Digital Doctor”, which is one of the most consumer-friendly treatment of my academic field (medical informatics) I’ve seen in the 14 years since I received my doctorate (I am a 1977 graduate of Winchester High School and saw this blog featured in the online Winchester *Star*).
I was a bit disappointed to see repeated the old and inaccurate trope connecting EMRs to Obamacare, however; the HITECH Act and the Patient Protection and Affordable Care Act were two distinct pieces of legislation; the PPACA does not mandate conversion to EMRs. HITECH does. See the federal government’s list of health IT legislation here for explicit descriptions of what each of these Acts does:
Or a Forbes magazine take on the relationship between the two here:
Thanks for the links.
Marjorie, This is excellent, I am now on the board at Bournewood Hospital, Nasir’s hospital in Brookline, and we have been grappling with the EMR situation for several years now, even before I became involved. It is a specialized psychiatric facility, even more difficult, not much out there in behavioral health. If you do a follow up, happy to put you in touch with our CFO.
That is arrant nonsense.
Medical care is part business — and therefore, structured data (coding, charges, days in the hospital, drugs used, and the like). That business part lends itself well to computerization, and in fact was computerized long ago with good results. No one argues that computers are good for the part of medicine that is like E-Bay, for scheduling and billing.
Medical care can also completely non-structured — patients come in to see the doctor because they “feel bad”. They often don’t have words for what they feel, and there is no objective sign of their distress. If they can describe what they are feeling, there may be cultural or linguistic mismatches that make accurate communication difficult. This is the situation that doctors are educated and trained to evaluate, and make decisions about. Computer software programs with all their dropdown boxes can not clarify and objectify medical evaluation. What it does do is force a patient into a specific decision tree branch — a sort of analog-to-digital conversion that might be accurate and if the patient came in with crushing substernal pain, diaphoresis, elevated troponins and ST elevation on EKG. But more often the problems are not so quantifiable– often it’s more like trying to extract and record the “meaning” of a Jackson Pollack painting. You can put a price tag on it, but you can’t really say what it is.
And of course, tere is the issue of “interoperability” — which is just a confusing way of saying that software vendors refuse to use a universal file structure that can be read from anywhere (one suspects an ulterior motive here — there are all sorts of cross-platform possibilities, and in any case, we don’t need to be able to WRITE to someone elses files, just READ from them). There is another set of government rules– HIPAA– that has virtually stopped medical communication in its tracks.
Electronic medical records could be designed with medical communication in mind, and they would work pretty well. AmazingCharts works well for a small system, and something similar could be scaled up — but it wouldn’t cost $billions, and so is clearly unsuitable to large hospitals.
In any case, you need some way for doctors to record what they do without becoming transcriptionists and file clerks filling out mind-numbing “bullet points” so the “coders” can do their work of objectifying the medical interaction so that an appropriate amount of money can be transferred (by a different computer) from one hard drive to another, and the CEO can justify his giant salary.
Thanks very much for your thoughtful reply. I can’t disagree with anything you have said, though it does seem clear that docs, nurses and other providers could have had more input into the design before it became activated, affecting one of life’s most sensitive and critical relationships: the nexus between patients and doctors, illness and remedy.
As someone who has built large systems, let me explain the fundamental problems here. While all of this may be true, here is why, and here is what is required to correct it.
First, it became obvious that someone was going to have to build a computer system for this. Modern medicine just couldn’t survive without it and, if you think erroneous medical billing is a problem then wait until you have seen it done on paper. So, this had to be done, like it or not, and whether anyone really understood the scope of the problem, or not.
Second, in terms of business requirements for the system — what the users actually need it to do — this thing is a monster that could kick Godzilla’s butt without half thinking about it. Without a doubt, this ranks as one of the most difficult engineering projects you have ever seen. I worked on one project that would be equal to only a small part of this one. The diagram of the table structure filled an 8 foot by 8 foot wall in 6-point type.
Third, in any project of this type, imperfection, and even non-function of critical parts are taken as a given. Because the project is so big, the team has to pick their battles on what they think they can make function first. Therefore, the number of new software projects that are loved by the users in Version 1 is tiny. The reason is that they are driving a Model T Ford when they were really hoping for, and needed, a Tesla. I would also say that, on new systems, complaints by users that the developers are not listening to them are common, and often true.
Fourth, any project improves over time. Whatever the initial failures were, somebody eventually figures out a way to work around. However, the fixes are invariably never fast enough. In this case, it is complicated by the fact that new requirements are probably being discovered every day. As soon as people get used to using one feature, they want another. See iPhones, for example, and assume that the process of installing a new app on each individual iPhone required a team of 20 people and took three months of planning.
Fifth, and most importantly, you are stuck with it. Like it or not, you are stuck with it. If you want to change to a new system, then you first have to develop a new system with the required functionality, figure out a way to transfer all the existing data, retrain everybody, issue new manuals, etc., install the new system at thousands of locations, etc., etc. You would spend less money and have an easier time building a tourist hotel on the moon – and probably wind up with more satisfied customers in the bargain..
Therefore, your only option is to try to fix the specific complaints you have. We are at the point where the cost of trying to replace it with anything different would hugely exceed the cost of making this one work. So, bottom line, your only real choice is to take that Model T Ford and upgrade it step by step until it becomes a Tesla.
Sorry, but that is the current reality.