Electronic Medical Records

Death By 1,000 Mouse Clicks: Electronic Medical Records And Physician Burn Out

Physicians are being sucked into the vortex of their electronic medical records.

I’m on our hospital’s electronic medical record physician advisory committee and so I have access to data about how physicians use the EMR that is not available to the average physician. I was astounded by the amount of time I spend in the EMR and my data is not unique.

I’ve been using our EMR for 9 years and I’m pretty good with it – I close my encounters promptly and at the end of the day, I’ve managed all of my messages, phone calls, and test results. I see patients three half-days a week in the clinic (total 12 hours) and cover our inpatient pulmonary/critical care service 1 weekend a month. Nevertheless, I spend a huge percentage of my time in our EMR.

My average encounter length is 7,369 characters – although most of those characters are automatically imported from templates and the after visit summary, it still results in a lot of typing. For progress notes alone (fewer characters than the entire encounter), as a percentage, 20% of the content of my notes is copy/paste, 15% is manually typed in, and 65% is imported via templates/SmartTools.

I spend 32% of my time generating notes & letters, 24% doing clinical review, 16% managing my inbasket, 11% entering orders, 9% in the visit navigator, 4% in the schedule & patient list, and 4% doing other EMR windows. Granted, the EMR is open a lot when I am doing other activities (in fact, it is open while I am writing this post) so these percentages can be a bit misleading. However, our EMR automatically logs us out after 15 minutes of no use.

I use the EMR a lot of the day. Currently, we are not able to track the total hours per week we are in the EMR (but we are trying to get that data and if so, it will result in a new blog post!), but we can see our EMR activity as a percentage of use and my results are pretty alarming. The graph below shows my own use activity for a 3-week period in January 2017 and this is pretty representative of my use activity in other months.

My EMR use starts to ramp up about 7:00 AM and then tapers off around 7:00 PM with a second peak of inbasket management after dinner around 9:00 PM. To examine use in a different way, we can look at the percentage of time I was in the EMR while seeing patients in the office (scheduled time) versus times when I was supposed to be doing other things.

If you analyze this data, what you see is that I spend about 2/3 of my total daily EMR time when I am seeing patients and 1/3 of my total daily EMR time when I am not seeing patients. Another way of looking at this is that I spend 2 extra hours doing EMR work for every 4 hours spent in the clinic. And compared to my peers, I’m very efficient. For me, I’m pretty much done with the EMR by 10:00 PM but many of my peers are still in the EMR at 1:00 or 2:00 AM. I’m particularly fortunate because I work with a great group of nurses in the office who really streamline my EMR use.

The literature on physician use of EMRs is surprisingly sparse. A 2013 study in The American Journal of Emergency Medicine showed that during a 10-hour shift, the average ER physician spends 44%of their time doing data entry, 28% of time in direct patient care, 12% of time reviewing test results and records, 13% of time in discussion with colleagues, and 3% of time doing other things. The average ER physician had 4,000 mouse clicks per shift.

An often-referenced 2016 study in the Annals of Internal Medicine looked at 57 physicians during ambulatory care (family medicine, internal medicine, cardiology, and orthopedics) and found that they spent 49% of their time in EMR & desk work, 33% of time in direct patient care, 1% doing administrative tasks, and 19% doing other activities (including personal breaks, travel time, etc.).

A 2013 study by the RAND corporation commissioned by the American Medical Association identified 9 reasons that physicians disliked their EMR:

  1. Time-consuming data entry
  2. User interfaces that did not match clinical workflow
  3. Interference with face-to-face care
  4. Insufficient health information exchange
  5. Information overload
  6. Mismatch between meaningful-use criteria and clinical practice
  7. High cost of acquiring and maintaining the EMR
  8. Requirement of physicians to perform lower skilled work
  9. Template-based notes degraded the quality of clinical documentation

A 2015 study in the Mayo Clinic Proceedings found that of the 6,375 physicians surveyed, those that used EMRs had a higher burnout rate (57%) than those that did not use EMRs (44%). Physicians who used EMRs also reported lower satisfaction with the amount of time spent on clerical tasks.

In a previous post, I satirically outlined the strategy hospitals can use to ensure that their physicians get burned out. Electronic medical records are clearly one of the factors contributing to burnout and it is incumbent on us to devise ways to improve physicians’ interface with the EMR to make the EMR work for the physician rather than the physician work for the EMR.

April 26, 2017

Electronic Medical Records

Stepping Up Your Documentation Game

Once upon a time, long, long ago, physician progress notes were a way of recording the patient’s condition and your treatments in the patient’s chart in order to communicate and ensure optimal care to the patient. You can lament that those days are gone but that will do you about as much good as lamenting that Homo sapiens are no longer Austrolopithecus. The progress note has evolved into a component of the hospital bill. Often times, what we write in the chart is not what we need to take care of the patient but rather what the billing department needs in order for the hospital to get paid.

In a previous post, I described the documentation game, where we use key words in the chart as adjectives to describe the patient’s severity of illness in order to lower the hospital’s mortality index. In this post, I’m going to show you how to use key words in the progress note to increase hospital reimbursement for inpatient care.

When a patient is admitted with an illness, the hospital gets paid by the insurance company (or Medicare/Medicaid) based on the patient’s primary diagnosis and not on how long the patient was in the hospital or how many charges the hospital accrued during the hospital stay. The hospital then gets paid more depending on the number of various comorbid conditions that the patient had at the time of admission – in other words, the sicker the patient, the more the hospital gets paid for any given primary diagnosis. So, in order to get paid as much as possible, the hospital has to be sure that all of those comorbidities are captured in the medical record. The hospital that documents comorbidities the best, wins the game. But there are a few rules to the documentation game:

  1. The comorbid conditions have to be documented by a physician. Conditions noted by a nurse, dietician, respiratory therapist, or social worker don’t count.
  2. The comorbid conditions have to be documented as being “present on admission”. If these conditions are in the emergency room physician’s note or in the admission history and physical, then they are considered as present on admission. However, if a comorbid diagnosis is identified and/or documented later in the patient’s hospital stay, then the physician has to state in a progress note that the condition “was present on admission”, otherwise, it doesn’t count.
  3. Diagnoses that only appear in a test result (lab result, pathology report, x-ray report, etc.) do not count. The diagnoses have to appear in an H&P, progress note, op note, procedure note, or discharge summary.
  4. Signs and symptoms don’t count.
  5. Lab values don’t count – only the physician’s interpretation of lab values count. So, you can’t just document that “the patient had a potassium of 2.5 so I will order potassium replacement”, instead you have to document that “the patient had hypokalemia so I will order potassium replacement”.
  6. The more specific you make the diagnosis, the more points you get and therefore the more you get paid.
  7. Using the words “likely, “suspected”, or “probable” in front of a diagnosis counts in the documentation game. You don’t have to know for sure that the patient has that specific condition.
  8. Certain words diagnoses that mean everything to you from a clinical management standpoint may mean nothing from a coding standpoint. So (as crazy as it sounds), urosepsis doesn’t count but sepsis does count.

In order to be sure that all of those comorbid conditions that were present on admission are captured in physician documentation, hospitals employ clinical documentation improvement specialists who comb through the charts to hunt for evidence that a patient had a particular comorbid condition that wasn’t documented by a physician. When they find one, they send a query to the attending physician or resident asking if the physician agrees that the condition was present and if so, asking the physician to add documentation to that effect in a progress note. Each of these diagnoses translates to a “risk adjustment factor” (RAF) number and when you add up all of the RAFs, the total score determines how much additional the hospital gets paid for any given admitting diagnosis.

Here are some concrete examples of how you can improve your hospital’s score in the documentation game:

  • A patient is admitted with a right lower lobe infiltrate and has type II diabetes with an elevated glucose. The patient also has chronic pain and has failure to thrive. Those two sentences get you 0 points. Now lets see what happens if you use some diagnosis words rather than signs and symptom words in your note: A patient is admitted with a suspected aspiration pneumonia (0.292 points) and has poorly controlled type II diabetes (0.496 points). The patient also has opioid dependence secondary to chronic pain (0.055 points) and has moderate protein calorie malnutrition (0.409 points). Now, those two sentences get you an RAF of 1.242 points.
  • A patient is admitted with suspected gram negative pneumonia and hypoxemic respiratory distress  with an increased lactate will get you 0.7028 RAF points and the hospital will be paid for a 2.8 day length of stay. However, if the patient is admitted with suspected gram negative pneumonia and chronic respiratory failure with an increased lactate, you now get 0.9469 RAF points and the hospital gets paid for a 3.6 day length of stay. Even better, if the patient is admitted with suspected gram negative pneumonia and acute respiratory failure with hypoxemia and lactic acidosis, you now get 1.3860 RAF points and the hospital gets paid for a 4.6 day length of stay.
  • A patient similar to the last one but has sepsis from a urinary source:
    • Urosepsis with SIRS, acute hypoxic respiratory distress, and an increased lactate” gets you 0 RAF points and no length of stay days.
    • Sepsis, acute hypoxic respiratory distress, and an increased lactate” gets you 1.0283 RAF points and 3.8 days length of stay payment.
    • Sepsis, acute respiratory failure with hypoxemia, and lactic acidosis” gets you 1.7660 RAF points and 4.8 days length of stay payment.

However, Medicare and insurance companies have determined that the documentation game wasn’t challenging enough so they added some more rules to make it harder to get RAF points:

  • For patients admitted with pneumonia, you get more RAF points if you can specify the type rather than just documenting “pneumonia”. So for example, “aspiration pneumonia” or “gram negative pneumonia” gets you more points. Don’t forget, that you are allowed to use the word suspected. So, for example, if you are prescribing vancomycin for a patient with pneumonia on the chance that they might have MRSA, put in your note “suspected MRSA pneumonia” rather than just writing “pneumonia”.
  • “Altered mental status” doesn’t count for any RAF points (it is a finding and not a diagnosis). So, instead, that patient has “acute encephalopathy” in your progress note.
  • “Unresponsive” doesn’t count but “unconscious” does.
  • “Drug-induced delirium” is a psychiatric code and doesn’t get you very many RAF points. It is better to document “toxic encephalopathy”.
  • If someone has a body mass index > 40, you get more points for documenting “morbid obesity secondary to excess calories” than you get for documenting “morbid obesity” alone.
  • For patients with heart failure, you get more points if you add the words “acute” or “chronic” and you also get more points if you add the words “systolic” or “diastolic”.
  • Avoid the word “dysfunction” as it doesn’t count for any points. So, for example, a patient does not have left ventricular dysfunction, they instead have left ventricular failure.
  • “Do not resuscitate” gets you no points but “palliative care” gets you lots of points. So, if you have a discussion with a patient and the patient decided to not undergo resuscitation, don’t document: “I spoke with the patient and he wants to be DNR“. Instead document: “I had a palliative care discussion with the patient and he wants to be DNR“. The terms comfort care, end-of-life care, and hospice care are all considered synonymous with palliative care when it comes to RAF points.
  • To maximize your RAF points, nobody should have “dementia requiring a sitter”. They should always have “dementia with behavioral disturbance”, “dementia with aggressive behavior”, “dementia with violent behavior”, or “dementia with combative behavior”.

The documentation game is by necessity a team sport and the team that wins will have both the strongest physician and the strongest documentation specialist. But it is not enough to be individually good, the physician and the documentation specialist has to work well as a partnership. Better cooperation between the doctor and the documentation specialist = more RAF points = more money for the hospital = better donuts in the physician lounge.

April 13, 2017


Electronic Medical Records

Using 50 Words To Express A 6-Word Thought

My college freshman English professor told me: “You are not writing for yourself, you are writing for your reader. Always put the reader first.” It is a great message… and one that electronic medical records (EMRs) have made us totally forget. That professor would once a week have us critique an article by a particular reporter for the Cincinnati Enquirer – our job was to completely tear it apart from the perspective of the reader. If he could get a hold of a single progress note from our hospital’s electronic medical record, he would have a enough material for an entire semester.

In a previous post, I mentioned that there are fundamentally 3 reasons to have a progress note:

  1. To communicate with other healthcare providers
  2. To provide documentation for medial-legal purposes
  3. To provide documentation for billing purposes

Electronic medical records make it really easy for us to document for billing but as a communication tool to other healthcare providers, they often fall woefully short. Our hospital uses the electronic medical record program, Epic. It is an incredibly powerful program that allows us to import all kinds of information into a progress note: lab test results, vital signs, past medical history, vaccination records, etc. As a physician, you can go wild with documentation – you can generate a 10-page note full of data just for an office visit note for a patient who comes in to get ear wax removed.

I’m on both the sending and the receiving end of electronic medical record progress notes. From other hospitals, I’ll get printed copies of notes mailed to me – pages and pages of data for a single office visit with a doctor who is co-managing the patient with me. From our own hospital, I’ll get notes from a colleague in a different specialty routed to my electronic medical record “Inbasket” and I’ll need to scroll through mounds of regurgitated data just to find the physician’s impression and plan.

But here’s the thing. If I’m treating the a person’s COPD, I don’t really need to know nor care about what their chloride level was in 2011 or the results of their last 6 normal EKGs. If I need that information, I can go to the lab or EKG section of the electronic medical record and I don’t need that information clogging up a progress note sent to me by another specialist seeing the patient for an unrelated problem.

The electronic medical record is a great documentation tool but we make it a poor communication tool.

Fear has taught us that more is better. If we include everything in the patient’s chart in our progress note, then surely we’ll get all of the right elements to withstand a Medicare billing audit. And shouldn’t we include every lab test the patient has had in the past 3 years just in case the person reading the note wants to see them? As a referring physician, I don’t really care about reading a consultant’s 11-point review of systems and re-statement of the patient’s family history since I already know this information and it is just a click away on the EMR if I need it. When it comes to really communicating in an EMR, less is usually more.

Don’t make the reader have to work to read your note.

There is a reason that we speak in sentences and that books are written using sentences. It is the most efficient way that we process the communications that are given to us. Tables and lists can help support the sentences but it takes a lot of time and brain energy to analyze tables and lists and in an EMR, most of them are just unnecessary fluff for the reader. When I get a communication from another doctor, what I really want to know is what their impression of the patient’s disease is and what they plan on doing about it. Sometimes that impression and plan is at the beginning of a long note of otherwise tabular documentation garbage, sometimes it is at the end, and sometimes it is in the middle – causing us to scroll up and down the notes trying to find out the information that we really want to know. The reader shouldn’t look at reading the progress note as a chore. The absolute best communications I get are from one of our surgeons who sends me a 1-paragraph letter saying what he thinks the patient’s problem is and what he is going to do about it. He communicates more in 2 sentences than most routed EMR notes communicate in 6 pages.

An inpatient progress note is often a communication to the outpatient physician.

My return outpatient visits are scheduled every 15 minutes. That means that when I see a patient in the office after a hospitalization, I have 15 minutes to review the record, talk to the patient, do an examination, order tests and medications, and complete my office note. I don’t have 7 or 8 minutes to spare to hunt through the recent inpatient chart to figure out what was going on with the patient and what I need to specifically follow up on. So I go to the last pulmonary consult note in the inpatient chart to see what pulmonary problem the patient was in with and what I need to be focusing on. If the patient needs a follow-up chest x-ray, or pulmonary function tests, or a decision about when to stop a pulmonary medication, I’m expecting that note to have that information. In this sense, the inpatient progress notes are not just a communication to the other physicians and practitioners seeing the patient in the hospital, they are also a communication with the outpatient physicians to insure a smooth transition from inpatient to outpatient care. Often, I’m both the pulmonary consultant in the hospital and the outpatient pulmonologist for a given patient and in this situation, my inpatient notes are notes to me and the nurse practitioner that I work with.

Don’t use a $25 word in a 25¢ sentence.

William F. Buckley was the conservative host of the TV show Firing Line and was famous for using 5 syllable words that he would pull from deep in the bowels of the English dictionary. He sounded extremely learned and intelligent but the average person couldn’t understand half of what he said. He used his extensive vocabulary like a weapon in political debates. That is great for entertainment but terrible for trying to communicate critical medical information to a wide audience of healthcare providers who will be reading your note. So don’t describe a person as macrosmatic, valetudinarian, or pauciloquent in an EMR note just because it showed up this morning on your “word of the day” app – no one else knows what it means any more than you did yesterday.

January 27, 2017

Electronic Medical Records

The PATH Audit That Almost Was

There are fundamentally 3 reasons to have a progress note:

  1. To communicate with other healthcare providers
  2. To provide documentation for medial-legal purposes
  3. To provide documentation for billing purposes

In the 1980’s, communication with other healthcare providers reigned king. We marked up medical student H&Ps with red ink if they didn’t have perfect grammar and we carefully dictated referral letters and then edited them before sending them out. Nobody cared about billing documentation back then – it just wasn’t important. All of that changed on June 21, 1996, when the Office of the Inspector General announced the PATH audits – “Physicians At Teaching Hospitals”. Although designed to be a way to protect Medicare against billing fraud, the PATH audits turned into the medical billing equivalent of the Spanish Inquisition. Federal Inspectors would do probing chart reviews of physicians at academic medical centers and if they found charts that didn’t have the right documentation elements, they would swoop in and do a massive audit of all of the physicians, often resulting in fines of 10’s of millions of dollars.

It was easy for these investigators – Medicare had established billing rules and required that the each progress note contain sufficient elements to justify different levels of billing. So, for example, a given level of billing for a new patient visit had to have at least 3 symptoms, a past medical/social/family history, 9 different systems documented in a review of systems and at least 11 different body parts examined. If the inspectors (who were not physicians) did not find all of those elements, the physician had to pay back the money from that particular bill and was also susceptible to an additional fine for each progress note that didn’t pass muster.

Although there were examples of clear fraud, for example, a surgeon who billed surgical procedures in Minnesota but had credit card receipt documentation that he was in London, England at the time, most of the cases labeled as “fraud” were really just good doctors trying to take care of patients but not documenting every part of their physical exam or forgetting to list all of the patient’s previous surgeries in their H&Ps. There was also a dark side of the PATH audits because they could also be vindictively.

You see, if a person called the Inspector General’s whistleblower hotline and it resulted in an audit of a physician (or better yet, a large academic medical center group practice), then the whistleblower got to keep up to half of all of the fines that the government collected. You could become a multimillionaire simply by calling in the dogs of the Inspector General.

In the late 1990’s, I was the subject of such a vindictive investigation. We had recently cut the salary of some of the physicians in our group because of low productivity and I was in charge of the clinical/financial management of the group. That year, 3 of our physicians left and one particular physician left the University, harboring a lot of anger and resentment, mainly directed to me since I was perceived as the one who cut his salary. So, he called the Inspector General’s whistleblower hotline alleging that I had fraudulently billed millions of dollars to Medicare. Because of the enormous dollar amount alleged, the OIG descended on our hospital’s medical records department and pulled every single progress note, procedure note, and H&P I had written or co-signed over a 3-month period.

It took them months to comb through thousands of my notes and with each note, they had a scorecard that they would check whether or not I had enough review of systems documented and enough body part examinations documented. It must have cost the OIG a small fortune to send investigators to review all of these notes. We hired an attorney to represent me through this process because of the fear that if I didn’t have the right documentation, it could open the door to a dreaded PATH audit that could essentially wipe-out the physician faculty ranks at Ohio State University. After the completion of the audit, here is what they found:

  1. I had over-billed by one level about 5 times (I had to pay back the difference – total was less than $100)
  2. I had under-billed by one or two levels about 25 times (total was about $2,000 but they don’t give you any money back in this situation)
  3. I had a few notes that I had written when billing ventilator management charges (used by pulmonologists rather than the more documentation-intensive return visit charges). Unfortunately, in addition to my notes, my medical students also had notes in the chart that referred to the ventilator settings and ventilator weaning. Because Medicare classifies ventilator management as a procedure and because any procedure done by a medical student, even with full supervision by the attending physicians, cannot be billed to Medicare, they asked for all of the money back for these notes – total about $200. Although I had done all of the actual ventilator analysis and written all of the orders, since the students also documented the ventilator settings, they said that the “procedure” of ventilator management involved medical students so I couldn’t bill it.
  4. I had come into the hospital one night to see a patient with acute respiratory failure in the ICU – I intubated her, put a central line in her, and did a bronchoscopy. In addition to the procedures, I also billed a critical care charge – for this charge, you have to document that you spent at least 31 minutes providing critical care services independent of any procedures. I had my documentation right there – 11:40 PM to 12:25 AM and I had each procedure note timed with a different time either before or after the critical care times. However, the auditors stated that since the time crossed midnight, 20 minutes of the time had to be accounted to one calendar day and 25 minutes to the next calendar day and since neither of the days’ time was >31 minutes, I couldn’t bill any critical care charges and in fact, was not allowed to bill anything. I had to give back $175.

Before this, I had been a documentation freak so my notes were actually very well-documented and so the amount I had to pay back to Medicare was truly trivial. The OIG spent many times that much in salary, hotels, food, and airfare for the inspectors that they sent to Columbus. My ex-partner got nothing as a “whistle-blower” and our University did not get a resultant PATH audit.

Our electronic medical records now make it simple to ensure that each note has the required number of body parts examined and the correct minimum number of systems documented to be reviewed in the review of systems. The PATH audits have gone away because there just isn’t any money in it for the OIG anymore. However, the legacy of the PATH audits is the topic of my next post – the cluttering of physician progress notes with excess documentation by doctors who do electronic medical record documentation overkill to avoid even a chance of not having enough documentation to support a Medicare bill in the event of an audit.

January 23, 2016