EMR's in the ED, Who Pays? by Rick Bukata, MD

Physicians are supposed to be men and women who embrace science and the scientific method.  So I would like to take a look at electronic medical record (EMR) use in the ED setting and try to bring together some data focusing more or less specifically on the economics.  Some of the data are easily challenged, some are not.  But data concerning EMRs (or EHRs, your call) in the ED setting are not particularly common so we may have to take what we can find.

And given that information about EMRs in the ED setting is not particularly commonplace, one good question to begin with is “why not?”  Since it is estimated that the annual number of ED visits is about 130 million, you would think that a lot of folks would have focused on the economics of the use of EMRs in the unique setting of the ED.  

People make a number of assertions about EMRs – one is that EMRs will improve quality.  Try and prove that.  Sounds good, but where’s the data?  You would think there would be some compelling data to support this fundamental assertion.  Where is it?  Where is the convincing evidence that quality is improved by EMRs?  We may even be willing to pay more for EMRs if, in fact, it could be convincingly demonstrated that they improve quality in the ED, but I’ve not seen convincing data.  

From what I’ve seen of ED quality measures, I prefer throughput measures – they are easy to determine and fundamentally speak to the functional aspects of an ED.  I like measures establishing variability among physicians when there are solid data that indicate there shouldn’t be much – for example, the work-up of suspected PE patients.  The data are straightforward – in the several good studies that have looked at variability in this setting, findings have been embarrassingly bad – variability is huge.  And I’d be willing to bet that all of these reports were from hospitals with EMRs.

The only way that EMRs will improve care is if they incorporate well-developed physician decision support – which most don’t.  Software that advises physicians of the cost of tests being ordered, software that advises physicians of the radiation associated with imaging studies being ordered (and perhaps converting that to cancer risk), software that provides physicians with proven diagnostic algorithms for PE, low-risk chest pain, dizziness, headache, otitis media, pediatric UTIs, etc, etc.  

But currently it seems that the only physician “support” relates to horrible programs regarding potential drug interactions of medications that are being ordered or prescribed for patients.  Generally each specific EMR has a company-developed drug interaction program that drives physicians nuts because it requires that they click through and acknowledge all of the potential reactions.  And do these basic systems improve patient care in the ED – who knows?  We probably use the same 30 drugs over and over again, so you would think we would know most of the main interactions by now.

So enough about quality – my assertion is that you can’t prove that EMRs do anything of consequence in the unique setting of the ED.  Perhaps the only situation in which they might be helpful would be when systems are linked together between hospitals and, as a consequence, an emergency physician could ensure that a CT scan done three days ago at another facility (which was likely unneeded in the first place) does not get duplicated at another ED seeing the patient for the same problem.  But how many hospitals have systems that allow physicians to see data generated at other facilities?  Few and far between.

Since it seems that it hasn’t been definitively established that these systems ensure quality improvement and provide a definite benefit in the ED, we can now focus on what it costs to buy them.   Costs have to be looked at broadly.  The literature has focused on three general areas – charting, throughput and, most recently, test ordering.  

From the charting standpoint, the vendors claim that hospitals will make money from improved charge capture.  Hospitals are into microbilling.  Every 4x4, bandaid, bottle of lidocaine, ibuprofen tablet, bag of IV fluid, the tubing for the IV fluid and the extension tubing, and the intracath with the clear plastic dressing all need to be charged for – item by item.  So, when hospitals insist on billing this way, it is easy for nurses to miss a few charges for stuff that costs virtually nothing in the first place.  But, the hospital wants those charges.  That $1 bag of IV fluids with all of its accoutrements will bill out at well over $200.  It is customary, here in California, to be discharged from the ED with at least a $1,000 bill – minimum.  I’ve written lots of essays about hospital charges in the past.  

And here’s a very recent example of what would be, most charitably, called fraudulent charging.   A woman goes to an ED after slicing off the very tip of her finger.  You know the problem – there is nothing to sew back and you have to get the bleeding stopped.  She was discharged with a $120 charge for a pulse oximetry.  Yes, it is true.  A totally unindicated test that takes one minute at most to do – and which is likely being charged to every soul who enters that ED.  Where’s the whistleblower?  So, yes, these systems can probably improve charge capture if hospitals think this modus operandi is reasonable – which it isn’t.  What about some global charges?

So charge capture gets better.  Does it get too much better?  According to a number of prominent articles, charges to Medicare have gone up considerably from EDs with EMRs.  Specifically, the numbers of level 4 and 5 visits at some hospitals have gone through the roof.  Are the patients sicker now that EMRs are in place?  Hardly.  When administrators are asked about the sudden increase in charges the response is predictable – patients are sicker and we are doing better charge capture.  Or are we submitting fraudulent charges?  Have these new systems facilitated up-coding?  If you look at the data in some of these newspaper accounts, the answer seems pretty obvious.

In fact, the Department of Justice has said they are going to look at the use of macros and cut-and-paste charting that is creating records that are larger than life.  And, obviously, the concern is on both the hospital and physician side of the ledger.  Given that every contracted ED group uses a professional billing company that is usually charging a percentage of collections, these groups and billing companies are highly motivated to maximize charges – and they are very, very good at it.  So which is the more likely scenario – patients are sicker, patients were undercharged due to poor charting and billing practices, or it is so easy to create a level 5 with computerized charting?  Here’s a snippet of information from a NY Times piece in the September 25, 2012 edition.  It refers to a NY Times investigation reported on September 21:

• “The Obama administration has issued a strong and much-needed warning to hospitals and doctors about the fraudulent use of electronic medical records to illegally inflate their billings to Medicare”
• “Attorney General Eric Holder Jr. and the health and human services secretary, Kathleen Sebelius, cited “troubling indications” that some providers are billing for services never provided and vowed to prosecute.” 
• “A Times analysis of Medicare data compiled by the American Hospital Directory found that hospitals received $1 billion more in Medicare reimbursements in 2010 than they had in 2006, at least in part by changing the billing codes they assign to patients in emergency rooms.” 
• “The findings involved two kinds of potential abuses. One is ‘cloning,’ in which a doctor cuts and pastes information from a patient’s electronic record to suggest that the services were performed again at the later date, or possibly uses the same documentation for other patients as well. The other is ‘upcoding,’ in which hospitals may exaggerate the intensity of care provided or the severity of a patient’s condition to justify higher billings.” 

Here are a few quotes from the original investigation (September 21):

• “The most aggressive billing — by just 1,700 of the more than 440,000 doctors in the country — cost Medicare as much as $100 million in 2010 alone, federal regulators said in a recent report, noting that the largest share of those doctors specialized in family practice, internal medicine and emergency care.”

• “For instance, the portion of patients that the emergency department at Faxton St. Luke’s Healthcare in Utica, N.Y., claimed required the highest levels of treatment — and thus higher reimbursements — rose 43 percent in 2009. That was the same year the hospital began using electronic health records.  The share of highest-paying claims at Baptist Hospital in Nashville climbed 82 percent in 2010, the year after it began using a software system for its emergency room records.”

In a January 8, 2014 NY Times story entitled “Report Finds More Flaws in Digitizing Patient Files” it was noted that

• “Although the federal government is spending more than $22 billion to encourage hospitals and doctors to adopt electronic health records, it has failed to put safeguards in place to prevent the technology from being used for inflating costs and overbilling, according to a new report by a federal oversight agency.” 

• “The report was especially critical of the lack of guidelines around the widely used copy-and-paste function, also known as cloning, available in many of the largest electronic health record systems. The technique, which allows information to be quickly copied from one document to another, can reduce the time a doctor spends inputting patient data. But it can also be used to indicate more extensive — and expensive — patient exams or treatment than actually occurred. The result, some critics say, is that hospitals and doctors are overcharging Medicare for the care they are providing.” 

• “The report being referred to was entitled ‘CMS and Its Contractors Have Adopted Few Program Integrity Practices to Address Vulnerabilities in EHRs’ and was published in January of 2014 by the Office of Inspector General, Department of Health and Human Services.” 

So charting may be a problem.  In addition to the $22 billion being spent by the Feds to help introduce EMRs into the medical system, CMS may be paying twice given the concern regarding upcoding.  

What about test ordering?  I had been begging for someone to do the most simple of studies – compare test ordering before and after the introduction of CPOE – computerized physician (technically “provider”) order entry.  My intuition was very simple – a mega-increase in testing due to the use of “order sets.”  Oddly enough, the use of order sets seems to be the only part of CPOE that emergency physicians like.  And, if I have intuited correctly, hyper-ordering can help to increase overbilling since the amount of data that you are considering is part of the criteria used to set EP levels of service. 

So here is the story, short and sweet.  In a single hospital ED (34,000 visits) 24 weeks after the implementation of an EMR with CPOE, lab testing increased from 225 tests per 100 patients before implementation to 374 after (a 66% increase), and EKGs went from 23.7 per 100 pre- to 35.7 post- implementation (a 51% increase).  Imaging remained stable – perhaps it was impossible to order more imaging than was already being done.  See “TRANSIENT AND SUSTAINED CHANGES IN OPERATIONAL PERFORMANCE, PATIENT EVALUATION, AND MEDICATION ADMINISTRATION DURING ELECTRONIC HEALTH RECORD IMPLEMENTATION IN THE EMERGENCY DEPARTMENT,” Ward, M.J., et al, Ann Emerg Med 63(3):320, March 2014 (EMA 8/14-#10) 

If the findings of the prior study are approximated in other EDs, it’s easy to see that payors will have significantly higher bills for the treatment of patients who are no sicker than in the past.  I very much admire the authors for publishing this first look at lab and imaging utilization post-EMR implementation – it is a landmark paper in this field and truly important.  This study should be easy to duplicate by others and it is surprising that it has not been done up until now (at least according to my knowledge).

Finally, what about throughput?  The fundamental issue – does charting with an EMR take significantly longer than other methods?  We’ve already noted that EMR charting may result in higher charges (hard to conceive that ED charges could be any higher), but do these higher charges offset the fact that charting with an EMR may be slower than the use of other methods?  Slower charting would result in an “opportunity cost” – could we see more patients if charting were faster? 

Emergency medicine is truly unique when it comes to charting.  Each patient is considered a “new” patient and each requires some degree of a new history, physical, interval notes regarding test results and the progress of the patient and discharge note charting.  Physicians in their offices are usually seeing “established” patients and only interval progress notes are needed. 

So I would contend that EPs do more charting than any other specialty.  The more time that is spent on charting the less that can be spent on direct patient care – seems pretty straightforward – either caring for the patients you already have or new patients.  And here’s an approximation of the economics – physicians may collect $120 or so per patient and hospitals about three or four times that amount.  So every new patient is worth about $500.  See one-third more patients per hour and in 24 hours this represents collected revenue of about $3,200.  See one-third fewer patients and you lose $3,200 a day ($1,165,000 a year).  And this number is very conservative.

Let’s start out simply.  In the WhiteCoat’s Call Room blog dated March 3, 2012 in Emergency Physicians Monthly the author looked at the allocation of his time during a 12-hour shift.  Time spent with patients was 219 minutes (about 30% of the shift) and time spent on the computer was 365 minutes (yes, about half the shift).  This is certainly not a scientific study, but a place to begin.  Many physicians have told me of the extra hours they spend after their shifts completing their EMR records.

A friend who is with a large, multicontract ED group (60 contracts) says his experience is that after implementation of an EMR, productivity drops sharply and gradually returns toward baseline but never returns to baseline.  What about the study by Ward, et al, noted previously?  At 24 weeks post-implementation the median length of stay was 179 minutes – it was 185 minutes pre-implementation.  But median patient volume was higher pre-implementation and, not surprisingly, the median interval from clinician to disposition increased from a very rapid 95 minutes pre- to 135 minutes post-implementation – a marked change.

The paper, “4000 CLICKS: A PRODUCTIVITY ANALYSIS OF ELECTRONIC MEDICAL RECORDS IN A COMMUNITY HOSPITAL ED” by Hill, R.G., et al, Am J Emerg Med 31(11):1591, November 2013 (EMA 8/14-#9), was also a look at a single hospital’s experience.  The authors tracked 16 attending physicians, residents and advanced practice clinicians for 30 hours and recorded time spent on various activities -- 44% of their time was spent on data entry, 28% on direct patient care (remember, it was 30% in the Ward study), 13% in consultation with staff and consultants and 12% in reviewing test results.  Mouse clicks approached 4,000 per busy 10-hour shift!

But not all studies have demonstrated a decrease in throughput.  A retrospective, before-and-after study involving 23 community EDs from the Schumacher group (Ann Emerg Med, June, 2014) with the majority of EMRs optimized specifically for ED use assessed four throughput measures and four operational characteristics.  The lead author was Michael Ward, who was also the lead author in the single-center study cited previously.  The long and short of this study – baseline and post-implementation steady state results were remarkably similar.  But the Achilles heel of this study – most of the EDs were using an EMR designed specifically for the ED (T System).  It is really very unfair to compare these best-of-breed systems with the enterprise-wide systems that are gradually replacing these ED-specific systems.  In addition, only one hospital had an ED admission rate over 20% (24%).  The large majority had admission rates that were low – 9 of the 23 EDs had admission rates of 10% or lower (EDs likely to have substantially less complex charting)

So here are multiple considerations suggesting that EMRs (most likely enterprise-wide systems) are very costly – not only with regard to the cost of purchase and maintenance (not even addressed in this essay), but as the result of potential upcoding, overtesting and throughput issues.  Nobody is going to remove their EMR (except perhaps to swap out of best-of-breed ED systems for clunkier enterprise-wide systems), but hospital administrators and physician groups owe it to their patients and those who pay the bills to assess the economic consequences of their EMRs.

W. Richard Bukata, MD
Medical Editor