EPM executive editor Mark Plaster sat down with emergency medicine opinion leaders and informaticists for a candid discussion about the future of electronic medical records (EMR).
Featuring Rick Bukata, MD; Nicholas Genes, MD, PhD & Bruce Janiak, MD
Mark Plaster: How are the meaningful use rules involving CPOE going to impact the use of scribes?
Rick Bukata: In the process of doing our courses every year we probably talk to 1500 doctors and we have a session about what have you done to make your ER better. And so, every year I get a large cross-section of opinion about what people think. I think CPOE is going to be a kind of nuisance but I don't think it's going to be any kind of deal-breaker. I think the use of scribes is growing and I think some of these commercial companies who use scribes are getting more and more customers. But apparently the scribes aren't allowed to be entering into the computer the lab and X-ray and drugs that the doctors want to put in and my take is that there is a substantial subset of physicians who like computer physician order entry. When you ask them why, it's about order sets. As soon as I hear that it makes me nervous because they push that abdominal pain button and blow about a thousand dollars worth of tests. They all claim that they can remove tests, but the question is how often do they do that. I would really love to see a before and after in terms of what kind of bills are generated by CPOE versus the old system. I think this whole topic is absolutely chock full of unintended consequences. CPOE is kind of the most benign of the bunch. The idea that it’s supposed to be safer for the patient to have no middle man involved. In our database I have one article that states there are 21 ways to make errors with CPOE. The literature tends to be some positive, but there's definitely lots of negatives in there. Some pediatric hospitals, when they initiated CPOE, actually had an increase in deaths. Whether that was due strictly to the CPOE, who knows, but the whole point is it’s supposed to be safer than what we were doing before.
Nicholas Genes: That’s a great introduction to the topic, and Rick, you bring up a lot of good points. I’m just coming from a meeting at Mount Sinai where we’re going from a best-of-breed system to an enterprise-wide system and I just spent the last couple of hours with our vendor going over our order sets. Its an extraordinary process. We already had order sets and we were happy with them and we though they were good and evidence-based in a lot of cases, and efficient. The new vendor is offering us even more choices. What would you like to have displayed? What would you like to have expanded? You don’t have to just order a CT but you have the option for CT abdomen pelvis, with or without contrast, do you want that expanded or collapsed? What do you want pre-checked, unchecked. You have a lot of options. To say that just having the presence of CPOE leads to more orders . . . I don't think that's necessarily so. It depends, as always, on how it's implemented and how it's displayed to the user and what's pre-checked and what’s unchecked and what’s available. In terms of what gets ordered versus what gets paid for, I think the current state in a lot of emergency departments is that things are done by nurses and doctors, but they are not billed for because they're not officially ordered. There are repeat vitals, there are O2 Sat interpretations, there’s EKGs, things that just happen in the course of care but because they're not explicitly ordered, the insurance company can't legitimately bill for those things and the ED loses out on those revenue. Having CPOE and having these order sets, these simple little things that always seem to get done but never seem to get explicitly ordered, that is one of the many ways that you can start to recoup your investment in electronic records.
Plaster: What I hear people saying about CPOE generally is that it’s supposed to make the ED more efficient. Does it? It’s also supposed to streamline care and make it less expensive. What I hear you say is that it might actually make it more expensive.
Genes: You can tailor your order sets such that the medications that are available in your ED are highlighted, are marked, and so the users will gravitate towards those medications and not towards medications that have to come from pharmacy that takes thirty minutes and have to be tubed down. And so medications might be administered faster with an order set. You can imagine doing this the same way with a paper system but often these things change, they need to be updated on the fly and you don't want to print out a thousand new sheets of order sets. You can simply go in and change it in the emergency department. The order sets allow for evidence-based orders and I think this is why the government wants there to be no scribes involved when labs are being interpreted and when orders are being placed. If you're clever about this, if you take the time, you can have links out there to ACEP clinical policies, you can have links out there to new literature, new guidelines, and you can make sure that the medications that are being ordered are based on current understanding and best practices. That is how people think that order sets are going to be more efficient and improve care.
Plaster: Will we eventually get to the point where order sets have to be justified? Take ordering an antibiotic. Will we have to meet a criteria for that to go through?
Genes: We already have that. We can’t order Linezolid without filling out an additional form. That’s all on paper though.
Bruce Janiak: When you talk about evidence-based order sets, my concern is that those are approached by saying, what would be the right bunch of labs you’d want to get for someone with colesystitis, rather than an abdominal pain order set. An abdominal pain order set is all-encompassing and it then becomes pretty easy for someone in the department to check a box and get a bunch of tests that don’t need to be done. So someone would have to prove to me that it would be cost effective.
If you read many of the emails coming through the Benchmarking Alliance recently, you will notice a lot of concern about implementation of CPOE and increasing errors at least in the short term and in trying to figure out workarounds. Workarounds are one of the reasons that we’re going to be able to live with anything. In doing multiple visits lately to various hospitals and chart reviews, I have noticed that emergency physicians tend to increasingly rely on blanketly ordering everything for everybody and secondly, directors of various departments tell me that once you implement, you can expect if you're in a teaching institution to have markedly decreased face time with your residents. And if you’re any institution, markedly decreased face time with patients. So there is a cost to all this in terms of human interaction. You’re taking the highest-priced person and having them screw around with a computer. When I go to my physician, he doesn’t even make eye contact anymore. He just sits there and types.
Genes: If you’re at a facility with residents, they are the ones that are generally going to be putting in the orders. They're not the highest price person. As for error rates, who knows what the error rates are with the paper. It is much, much harder to really track. If on your order sets one particular medication keeps getting ordered the wrong way then that's a red flag that maybe you've implemented it wrong. You can go back and change the default dosing. You can clarify it, add something, remove something. If something is going wrong it lends itself to being fixed. If something is going wrong on paper, I don't think you would even notice for months.
Bukata: But most hospitals don’t have residents, and when you look at the number of medications being given in the ED, it’s really quite small, and most of them are quite benign and the nurses have been giving them over and over and over again. They’re not going to give magnesium for a broken leg when it should have been morphine sulphate. Some of these examples that people use are really absurd. I think one of the most important things here is that we’re talking about some major, major expenses, not only in terms of the computers that are going to be put in place to do this work, but in terms of who's being the data entry person and on tons of other assumptions. This is ready, fire, aim. Show me the data that says this is better, faster and more cost-effective. I don't think you can do that. I follow this stuff relatively carefully. There are the zealots who really like it, but there are the people on the other side of the equation. And since there are both, it would seem that as men and women of science, we should say, “Lets look at this and do a comparison.” One of my concerns, frankly, is that fairly dysfunctional emergency departments that have really long throughput times will say, “Well you know it really hasn’t affected our throughput time,” because they’re so off the bell shaped curve already that a CPOE will not make it any worse. But they don’t know any better. When you talk about throughput times of two hours for a discharged patient, then see if CPOE makes a difference or, more importantly, an electronic medical record. I think that hospitals and departments have gotten used to embarrassingly long throughput times and after a while it becomes the norm. I may be out of line, but I think it's particularly true at places that have residencies, where throughput is not one of the drivers. At least out here, when I go to USC, I don’t see any sense of urgency. It is absolutely packed, it’s always packed, it's been packed for 20 years and it will be packed for the next 20 years. If you come to my community hospital, if they see the waiting room full, they basically pick up the pace.
Genes: Perhaps Bruce can speak to the community vs. Academic divide, since both realms are represented in the Benchmarking Alliance.
Janiak: What I see in my own shop is the same partly legitimate sense of calm when it gets really busy because, ‘you can’t do it any faster because of the system that we have with residents.’ On the other hand, when I worked in a private hospital, you pump it up. In general, the data show that although the throughput times are not three times more than in the academic setting, they are longer, and a lot of it has to do with not just the tier of residents or the ED itself but the second tier of residents who you transfer patients to, who are never happy until you have a repeat CBC hemoglobin on every patient so that their attending doesn’t give them a hard time. That kind of clinical behavior in education, which I believe is abhorrent, is endemic and it is extremely difficult to fight. I fight it every shift. It does increase throughput times for patients and I think the benchmarking data will support that. Fortunately, the Benchmarking Alliance is now adding several hundred hospitals and so this next round of data for 2011 is going to reflect not only a much larger database but also a much larger database of new CPOE implementations. So we’re really looking forward to this fall.
Plaster: What is the goal of the M.U. legislation?
Bukata: One of the theoretical benefits of CPOE is supposed to be decision support. These things are supposed to be clever systems, not just push the button and then blow out the test. Theoretically the machines should ascertain whether the patient is a low risk for, or intermediate risk for PE, and if it's a higher risk you shouldn’t be ordering the D-Dimer. So there are some kind of theoretical reasons why decision support is a good idea. But the devil’s in the details. Whether, in fact, it can be done quickly and effectively is a whole other matter and, frankly, I’m kind of doubtful of its ability to do that in any meaningful way. Genes:
Rick, you have spoken so eloquently about practice variation and how it's unconscionable that in the same department you could have the same patient getting different tests from different practitioners. There is no way to really capture that variation, I think, without reporting.
Bukata: I agree that variation is huge, and the result is a bit embarrassing. However, I do think there are other ways to capture data that are very meaningful, but not captured by this process.
Genes: I agree. I think that's where the ED Benchmarking Alliance is so important.
Bukata: There is literature that says that if you are a hospital that orders tons of CTs for belly pain, then all of the doctors tend to get into line because they believe that that is what should be done. But there are other hospitals where collectively the physicians order many fewer tests because its monkey-see-monkey-do. Your environment helps set the threshold for ordering tests.
Genes: That’s what Hospital Compare is all about, to compare different environments. I think the ED Benchmarking Alliance does a better job because it stratifies the EDs based on geography, volume, etc..., but I think the government wants to do the same kind of thing.
Janiak: Rick makes a great point about culture. I have seen that as well. In doing consults, when you get to a department that has long throughput times, they invariably will have lots and lots of tests ordered. Also, the story you will hear is that, “Our patients are sicker then all the other patients in the United States and that’s why we have to do more tests.” I've heard that so many times and every single time it’s a culture of overtesting. That’s why, in the long run, I hope that CPOE will benefit us if we carefully mine the data appropriately to be able to change behaviors. Right now, everybody is so frustrated with just the implementation process that nobody's looking at output yet.
Plaster: Is it not the goal that the government will actually start to become more involved in clinical decisions?
Janiak: The Benchmarking Alliance now has the biggest database out there, so we should be able to at least be interesting to the Feds, and be able to provide them with information that can change meaningful use meaningfully.
Genes: Exactly. So When the government proposes a new metric to capture we can point to the ED Benchmarking Alliance data and say, “Actually what you’re proposing would be disastrous and here's a better way to do it and we've been collecting the data.”
Bukata: Although honestly, being not involved in this at all, its hard to conceive that anybody would be stupid enough to say, “Well if you admit X percent of patients, you are out of line.”
Genes: It’s hard to conceive, but you have to arm yourself and be proactive and know the facts and that's why the ED Benchmarking Alliance is so helpful for us. And having our own computer system and comparing notes with other places that are already collecting this kind of data, which is very helpful question.
Plaster: What about the transmission of data. In my practice, knowing what the patient has received in the past, what their past tests have said, would be very valuable. That seems to be the direction we are heading, but I’m not so sure it goes far enough, and if this is actually the kind of data they’re going to generate. Is it just the diagnosis, or will patients and physicians have access to their complete health records?
Bukata: This transitions us from CPOE to the larger quagmire of doing an H&P by some kind of electronic method. I don't know whether your hospital or you are aware of it or not, but in most hospitals, anything that’s been typed, you can see by some kind of system. Most of the hospitals have these older systems where you can bring up those kinds of reports (radiology reports for the cat scan six months ago) you can see those without much difficulty. So, great, your hospital can come up with an even better system so that you can see what's going on, but this failure of communication between the doctors office and some other hospital, that’s really going to be where the issues are. Even in our very uncomputerized hospital we have access to a lot of those reports. No one is going to argue that occasionally that stuff is valuable, but how often is it valuable? Yes, I think it's just good and it may cut down on the ordering of necessary tests because if they just had a cat scan a week ago you don’t need to reorder it. But, it seems that the challenge will be connecting these disparate, proprietary, user-unfriendly databases so that when you go to hospital X, hospital Y can see the information. In New York, they might already be doing that, but how is that able to be done in Los Angeles?
Genes: It's a technical question, but it turns out if you’re already in an electronic ED, there’s some standard communication protocols that can be shared. It’s not perfect. I can’t trend the patient's sodium level because one hospital might describe it as “sodium” while the next describe it as NA and the third hospital calls it some fancy number. But you can at least see in a place like New York, where there are so many hospitals and so many nursing homes, and one of these hospitals is always on diversion, so the patient can’t always get to their medical home, it's very nice to be able to log into a health information exchange and see recent lab values. And you know, we’re just trying to make the system better one data element at a time. We don't have EKGs available from every hospital yet, or EKG images, but we have cath reports and CT images. We just keep adding what we can as each hospital gets ready to standardize their data and broadcast it to the network.
Bukata: They’re all using different computer systems?
Genes: First of all, New york has almost a hundred hospitals and our network is only eight at this point, although more are interested in joining. And not all the data is displayed the same way, but if you log into the system and find your patients you can see their past values.
Bukata: That will be the holy grail, to see the integration of information from disparate sources.
Genes: In the health information exchange, we talk about conditions that are amenable to health information exchange. If you have an ankle sprain, you don't need to know what the patient’s sodium was two weeks ago. But if you have an altered mental status, a transplant patient, some sick older people, health information exchange is invaluable. We’re hopeful and where trying to show – as is Memphis and Indianapolis – that this is a system that saves money and increases good outcomes.
Bukata: I don't think you’re going to get a lot of people arguing that the ability to exchange information isn’t valuable. But that is a kind of smokescreen for the more nitty gritty issue about doctors generating H&Ps in a computer system. In emergency departments we are disproportionately documenting. You have 2.5 transactions in an hour, so we’re cranking out histories and physicals and progress notes much more than any other department in the hospital because of the way we do business. And so we disproportionately spend time on medical records. The concern is whether or not there is a return on that investment that is anywhere near the consumption of resources by having the doctor do that stuff. Because honestly, most of the stuff is unnecessary. The family doctor wants to know what the diagnoses was and what tests were done. They are not not interested in whether you heard an S-3 something or other. It’s another case where we've made some really fundamental assumptions which may be wrong. My passion is that we’ve got the wrong person entering the data, particularly in emergency departments that are busy. You could have a monk doing H&Ps with a quill pen if you see one patient an hour. Our whole issue will be throughput, throughput, throughput.
Genes: We’ve published our data about return on investment for electronic records and people say ‘you're an academic center’ or ‘you have slow throughput’ or ‘you have residents’. There are good experiences and bad experiences; it really depends on how the implementation is conducted. You can still employ scribes for dictation, for putting the H&P down, you can have your vendor make macros and standardized H&Ps. You can record that information almost instantaneously. You can tailor it to the patient. You can go in and change little features to personalize it. It doesn't have to be that you’re stuck on the computer for an inordinate amount of time. And what we’re finding is that even if there is a hit in productivity, it’s more than made up for in terms of additional charge capture. So your hospital in the end will appreciate the bottom line and your department will appreciate the bottom line, that more money is generated per patient.
Plaster: We know there is a trend towards enterprise systems. Some of these are incredibly unhelpful because they are one-size-fits-all. If hospitals move toward enterprise-level systems, will the ED be benefited or hurt?
Genes: I am in the midst of helping to change our ED’s best-of-breed system into an enterprise solution and believe me, there are a lot of pros and cons. The right solution depends on your environment. For us, in a tertiary care center with EM residents, but also with medicine residents and psych and surgery residents rotating through our department, we have a lot of complex patients with complex med regimens. The ability to seamlessly pull patient data from their inpatient stay from their prior visits has big advantages for us. And the residents will appreciate the familiar interfaces thats just the same as it is in clinic and just the same as it is upstairs. We think that will mitigate some of the potential performance hit. Our attendings and PAs may miss the old system 's efficiency, but we’re working hard with the new vendor to preserve and carry over what we can from order sets to the familiar layout. We’re confident that this is going to be a good transition.
Bukata: You know the Pennsylvania ACEP has had every year, for years and years, an informatics meeting. Probably the only one of the country. This is where all the niche vendors would show up and groups would go to this meeting with the nurses to look at the various systems. Last year that meeting was canceled and this year it had something like fifty attendees. I talked to them about what they thought transpired, and basically they thought that the Cerners of the world had won. Some of these niche systems – the lucky ones – have been bought by the bigger guys. The other ones will die.
Genes: With meaningful use now including the emergency department, the ED’s role in the hospital is important. So no hospital is going to neglect the emergency department. No smart hospital at least. They're not going to foist something upon the emergency department that is disastrous. At least that’s the hope, because the ED is so important to the meaningful use denominator.
Janiak: That depends on the definition of disastrous. It’s the administrator who makes up the definition.
Genes: If you invested in a best-of-breed system and you want to keep it, I think you have a good case to make to the hospital. You have a fighting chance to keep it. And I think these systems are working hard to integrate with, be interoperable with, these enterprise solutions. They have to be to survive.
Janiak: I agree with that.
Plaster: How is telemedicine going to impact care in the United States
Janiak: I still have great hope for those sites in which there is relatively little medical supervision and current system is, whether they come from the psychiatric hospital or the nursing home, “send ‘em to the ED.” We all live with that everyday. A primary care physician gets a phone call from an LPN at 3 AM that Mrs. Smith, who the EP has never seen before, has a low pulse-ox. The LPN doesn't really know how to put the sensor on right and so patient comes over, you put the sensors on right, it’s 98%, they go back to the nursing home, and there goes $2,000 in transport costs. Plus it ties up our bed. If we had utilization of telemedicine, we might be able to free up some time and be more cost effective. I always say this when I talk about these things, but we’re not actually saving money, but shifting money, from maybe the EMS transport service to some other area of health care. But at least we’re spending our money in a more sensible manner. So I think telemedicine can impact healthcare in the United States. In the future I also think we will be using technology like cell phones to triage patients. I think we can make these things happen over time.
Genes: Has there been any progress in getting reimbursed for those teleconsults? Seems like you’re only liable. You can’t earn money, you can only lose it.
Janiak: In Georgia, we get paid for telemedicine consultations at the same rate as a face to face. And the Feds have just changed their laws with regards to the definition of an urban area. It used to be, for instance, that you could not consult with a nursing home right down the street if you're in a densely populated metro area. And they’ve now rescinded that and you can get those consultations for local nursing homes as you can with rural nursing homes. So I think things are actually looking up, although I have to tell you I have not spent as much time as I should with the American Telemedicine Association, so I've rejoined, and looking at this next meeting, so I can get some sense of what the legal issues and the reimbursement issues are. In a state of Georgia, though, we’re making money.
Genes: That’s terrific. The other critique I hear is that the telemedicine industry – Cisco and the like – they want you to spend a fortune on equipment that ends up making you see less patients.
Janiak: That's going away too because the T1 lines are being unplugged all over the country and we’re switching to internet-encrypted transmissions, so the financial requirements are now a PC with the camera and that’s it. There is an issue with retrofitting old nursing homes with WiFi, but in terms of the monthly $400 to $700 rental fee for T1 line, that's disappearing, as are the $20,000 per unit special telemedicine screens and cameras. Capital expenses are dropping like a rock.
Plaster: In your opinion, will the EMR initiative that is currently underway make the U.S. spend less money on medicine?
Janiak: I would say no. We’ll continue to spend more and more money no matter what.
Bukata: I would agree. There may be a time when these systems get really, really cool and you’re, in fact, spending the same amount of time as you are now. Or maybe it even gets better and faster but it's going to take a long time, I think, to get there.
Genes: I'm not sure it's going to be less money, but at least maybe the rate of increase will slow and maybe the money will be spent more efficiently. Less duplication of testing, more of the right testing at the right time.
Janiak: That’s what I’m going for, more efficiency, not necessarily less expenditure.
Plaster: So you feel that this initiative is likely to result in better quality care?
Bukata: Honestly I wouldn’t go there. I think it's an opinion and how many billion are they going to spend on this?
Genes: 18 billion. 35 billion over five years. This is a drop in the healthcare bucket. Don't think this is even going to fund our military for a few hours. It’s not a huge expenditure.
Bukata: The horse may be out of the barn, but I really believe that we need to be much, much more critical of what we're doing. I mean, do you remember like ten years ago when every hospital was buying every physician’s private practice? Some hospitals saw the other hospitals doing it and it became monkey see, monkey do. Well, ultimately that was a huge initiative throughout the country that turned out to be, in most cases, a disaster. The doctors stopped working, the doctors’ productivity went down, they became employees of the hospital, referrals really didn't change. And yet, it was just intellectually embraced as a good way to do business. I’m suggesting that we need to be much more careful here because I don't think physician time is being adequately considered and throughput and productivity.
Genes: The pace of change is being accelerated because we spent ten or fifteen years dithering about the evidence and the best way to proceed. And now these changes are being foisted upon us because not enough of us were proactive. I don't know any other industry that has been as resistant to technology and asked for as much supporting data.
Bukata: I’m reflecting the sense of the physicians in our audience who, if you were to poll them, the vast majority are not positive about this. We’re saying to them, “you just need to get with the program.” It’s uncharitable to suggest that they’re just slow on the pick-up here. I think many of them have experienced the technology and have not seen the efficiencies, particularly in the community hospital setting where your income is derived by the patients that you see. To be candid, I don’t know what employed emergency physicians think, because they get paid whether they see one patient or four patients in an hour. But independent contractors live on their volume. I just see audiences that generally, the majority of the people unequivocally say we’re moving in the wrong direction.
Genes: It’s an incentive program that will change the way Medicare gets reimbursed by a few percentage points. It’s eighteen billion over five years which, spread out over the whole country, adds up in many systems to less than a few hundred dollars per employee. So it will change the industry in the long term, but if there is a reasonable objection in a hospital, if the numbers don't make sense and it doesn't make sense to go forward, then they don't have to implement.
Bukata: The other thing is that ERs are going to be disproportionately expected to generate a proportion of the orders through CPOE because those doctors are controlled by the hospitals, they work for the hospital, they want to make the administration happy. But I think about our hospital, a community hospital, that has about five hundred doctors on the medical staff, and when I think of telling them that they need to start generating electronic histories and physicals, all 500 of those doctors will go absolutely ape-shit. Genes:
This is a process that's been going on over the country, slowly and now faster. There are good ways to do it. You take time with your vendor, you take time talking to the doctors, you generate reasonable order sets, fast ways of charting, you minimize the amount of clicks and you can implement it with a modicum of pain.