By Mary Butler
One of the most famous cases in medical history that led to regulation of the number of hours that resident physicians are allowed to work is also a case study in clinical documentation failures.
In 1984, a college student in New York, NY named Libby Zion was admitted to a Manhattan emergency room (ER) with a high fever and agitation. After consulting with her family physician, the residents who evaluated Zion in the ER administered a sedative and painkiller. What none of the caregivers knew—because Zion didn’t disclose the information at the point of care—was that she was taking an anti-depressant that was dangerously contraindicated with the drugs the physicians gave her in the ER. The drug combination ultimately proved fatal and Zion died from cardiac arrest.1
While the legal fallout from this case centered on the hours that doctors work and how closely Zion was monitored by the ER staff, it serves as a cautionary tale for physicians who frequently treat patients in the ER without having full, up-to-date medical histories where and when they need it. Even though electronic medication reconciliation practices that weren’t even possible in 1984 are standard today, an incomplete patient record can have the same disastrous consequences now that it did then.
Patient safety is just one of the many reasons that formal clinical documentation improvement (CDI) programs are flourishing. The growth of CDI is also being driven by the increased specificity needed for ICD-10-CM/PCS, as well as the transition to pay-for-performance versus fee-for-service payment methodologies. Payers and outside auditors such as recovery audit contractors (RACs) and Medicare audit contractors (MACs) are scrutinizing claims and health record documentation for proof of medical necessity and quality indicators.
The best way to prevent a documentation disaster is by recognizing the most frequent kind of documentation errors and putting procedures in place to stop them before they can strike. According to several HIM experts, the top four documentation mistakes are:
- Mixed messages from a physician vis á vis misunderstood dictation or illegible handwriting
- Misuse of copy and paste or copy forward functions in the electronic health record (EHR)
- Incomplete or missing documentation
- Misplaced documentation
Documentation Disaster #1: Mixed Messages
Among the arguments for adopting technologies such as EHRs and the dictation program Dragon was the notion that they would eliminate problems like illegible handwriting and hard-to-understand dictation. Sheila Burgess, RN, RHIA, CDIP, CHTS-CP, director of CDI at Sutherland Global Healthcare Solutions, says that despite recent advances old habits die hard.
Common transcription mistakes include typing “hyper” instead of “hypo” or vice versa, and typing “he” instead of “she.” This is a problem Burgess has seen since the 1990s. Or physicians that are in a hurry will dictate a note into the wrong patient’s chart, and the coder doesn’t catch the error because they are just coding without checking to see if the note matches the rest of the chart. Most times, if the CDI staff is well trained, this kind of mistake will be caught—but it can be costly when it’s not.
“If a patient comes into the ER unconscious they could be emergently treated for a condition they don’t have if a physician pulls up their record and the documentation is incorrect or it’s the wrong patient’s information,” Burgess says. “That’s one thing that could be detrimental.”
She notes that sometimes when a wrong diagnosis gets embedded in a patient’s chart, that can get carried on when that patient changes insurance companies and a pre-existing condition is noted. The patient can be mistakenly put on a waiting period for insurance due to that kind of mistake if it’s not cleared up.
Burgess says electronic prescribing has helped eliminate some of the problems related to sound-alike medications, such as Xanax and Zantac, but it can be a problem for nurses who take orders over the phone. On the whole, she’s seen this problem diminish. Lots of hospital HIM departments post lists of frequently confused medical terminology and medications to prevent those kinds of slipups.
Many have argued that EHRs also eliminate the problem of illegible physician handwriting, but Dr. Jon Elion, MD, FACC, president and CEO of ChartWise, and associate professor of medicine at Brown University, argues that electronic records can also be difficult if not impossible to interpret.
“Just because you can read the letters doesn’t mean you can decipher what they’re saying,” Elion says. “In this world of Twitter and text messages, people are using horrendous shorthand. And very, very illegible notes are coming out of that. So that’s certainly a danger.”
Widely used abbreviations can cause problems in paper and electronic documentation. For example, “q.i.d.,” which means that medication must be taken four times a day, can look and sound a lot like “q.d.,” which means that a medication has been prescribed to be taken once a day. Elion says doctors are discouraged from using the abbreviation “MS” because it has multiple meanings. It can be used to refer to the drug morphine sulfate, to refer to the disease multiple sclerosis, to note altered mental status, or to denote the cardiology term mitral stenosis.
“To type the note more quickly, they [physicians] might abbreviate things that only they know what they really mean. So that hampers documentation,” adds Susan Wallace, MEd, RHIA, CCS, CDIP, CCDS, FAHIMA, director of compliance at Administrative Consultant Service. She says another frequent shortcut is physicians who document “multi-organ failure” rather than citing the specific organs, which can fail to reflect the severity of an illness, thereby impeding accurate reimbursement and appropriate patient care.
Natural Language Processing Boosts CDI Capabilities
Even with a robust remote clinical documentation improvement (CDI) workforce, HIM leaders at Baystate Health, based in Springfield, MA, decided they wanted members of their CDI team to query far more than one out of every five records they review at their flagship facility. Walter Houlihan, MBA, RHIA, FAHIMA, director of health information management (HIM) and clinical documentation at Baystate Health, and Jennifer Cavagnac, CCDS, Baystate’s assistant director of clinical documentation improvement, wanted CDI staff to boost their query rate from 20 percent to 75 percent. More specifically, they wanted to be able to identify cases with the most “opportunity” for clarification, Houlihan and Cavagnac say, such as patients with anemia, congestive heart failure, COPD, certain types of infections, and patients who meet various criteria for sepsis.
The tool they’ve chosen to help meet this goal is natural language processing (NLP), which is computer software that uses algorithms to look for a variety of different variables including words and phrases that can be sent to CDI specialists in real time for validation. NLP is similar to computer-assisted coding (CAC) applications, which searches charts for individual words and assigns a code. According to Cavagnac and Houlihan, the learning curve is far steeper for CAC, which can take coders a couple months to get used to. With NLP, CDI specialists had one day of training and were easily able to use it handily within days.
The NLP application works like a search engine, scanning charts as soon as they are entered into Baystate’s electronic health record (EHR) system then sending the findings to the online portals of CDI specialists for review. Since the NLP system runs 24 hours a day, CDI specialists log in to their portals when they start their shift at 6 a.m. and the application will have already created their work list of charts to review.
At press time, Baystate only had NLP up and running for a little over a month, but Houlihan and Cavagnac say they are already pleased with the results. NLP can help CDI identify common documentation errors such as those generated through copy and paste.
“NLP can identify instances in which something may be noted multiple times. We’d look at that in conjunction with a CAC tool,” Cavagnac says. “Once a patient’s been identified we can use the CAC tool to actually get a true count on a phrasing or condition or a lab value that’s used repeatedly throughout the record.”
It also can help weed out unapproved abbreviations doctors might use in their notes, which can prevent medication errors.
“I’d strongly encourage other hospitals to look at these types of applications, just for the benefit of being able to locate information in an EMR,” Cavagnac says.
Disaster #2: Copy and Paste Carelessness
The easiest way to cause a ruckus at a CDI conference is to utter three words: “copy and paste.” And for good reason. It’s the function in many EHRs that allows users to re-use all or parts of the detailed narrative portion of a health record, which is seen by others who may access the record down the line. While it can save physicians time, it’s a breeding ground for documentation errors. When it’s used incorrectly, copy and paste can make it difficult to track the progression or resolution of an illness, which can result in over- or under-reimbursement and send up red flags for auditors looking for fraud. Misplaced or incorrect documentation in a patient’s record can also cause severe physical harm and even death.
An example of a copy and paste error that Burgess came across recently involved a newborn that had a slight hematoma of the head right after delivery.
“It was really minor [injury] but the doctors kept cutting and pasting so it looked like they were really monitoring this baby, this neonate, for head trauma the whole time. But they weren’t. And that was because of copy and paste and lazy documentation,” Burgess says.
Wallace says that one time she observed a case where “for a whole week, the patient appeared to be on ‘post-operative day three’ because the entire note was just copied forward, copied forward, including the heading [on the chart], when that clearly was not the case.” That kind of mistake implies that a patient has made no progress from their surgical procedure when in fact they have.
Elion, however, notes that there are two specific circumstances under which physicians can use copy and paste to win the appreciation of coders and CDI teams. The first is when documenting the findings of a radiology report in a progress note. Since you can’t code from a radiology report, Elion says, the doctor should copy the whole note, paste it into the progress note, and then add a line or two that says “I’ve personally reviewed the X-ray and discussed the findings with the radiologist. I agree with his description of the location and nature of the fracture.” Doing this will make a coder “fall in love with you,” he says.
“Ditto with anatomic pathology. How many times do we see the note that says ‘brochial biopsy was positive. I’ll schedule oncology to see the patient as an outpatient?’ That’s useless. You can’t code from that. Copy and paste the anatomic pathology report, document stage 2 bronchial carcinoma, whatever it is. Those kinds of things are very important,” Elion says.
Not only can copy and paste perpetuate errors, it can add pages upon pages to a patient’s record which can slow down any number of processes a chart is used for—from being able to read through it quickly in an emergency to coding from it.
According to Dr. William Haik, MD, FCCP, CDIP, a practicing physician and director of DRG Review, EHRs generate so much data that finding the pertinent information when he needs it is tedious.
“That’s a problem for auditing records, which I do, and I imagine it’s an enormous problem for coders. As a physician, I’m not about to dig through 10 pages of a progress note. Or worse, the nursing notes. The day is gone when you can look at a nursing note and figure out what happened to a patient in one day’s worth of nursing notes,” Haik says.
Disaster #3: Incomplete Documentation
In a perfect world, caregivers would chart every patient encounter as if that record had to stand alone 10 years down the road or more, and in such a way that it tells the patient’s whole story. But the reality is that too many clinicians use vague terminology like “unspecified” and “not otherwise specified,” or records get passed off to so many different parties that something critical goes missing. Having incomplete documentation can have patient safety impacts like those discussed at the beginning of this article, as well as negative reimbursement impacts.
CDI specialists are well positioned to fix incomplete documentation errors when it comes to clarifying whether certain conditions were present when a patient was admitted, says Haik. A good example is when a patient is admitted to the hospital with complications related to chronic obstructive pulmonary disease (COPD), but the admitting physician doesn’t know that upon admission that same patient already had a decubitus (or pressure) ulcer. If a patient develops an ulcer like that while they’re an inpatient, both the physician and hospital can receive a reimbursement penalty. Therefore, making sure this is noted as “present on admission” is important.
Jamie Wilding, MBA, RHIA, coding compliance manager at Esse Health, says incomplete documentation to support a diagnosis can invite calls from Medicare auditors and insurance companies looking for documentation that supports medical necessity for an overnight stay. From an auditor’s perspective, if it’s not documented, it didn’t happen. A frequent documentation weak spot, says Wilding, is not recording a patient’s chief complaint.
“Chief complaints are sometimes an issue because ancillary staff is entering in chief complaints, so sometimes it’s missing,” Wilding says. Wilding notes that chief complaints are critical because they help support medical necessity. “Without the proper chief complaint, Medicare must view that as invalid and known, and I’ve read reports that Medicare will recoup money if chief complaint is missing,” Wilding says.
Elion says there are two simple phrases that could go a long way in improving progress notes if CDI specialists could get physicians to do it: “due to” or “manifested by.” With reimbursement changes and ICD-10 on the way, writing “due to” instantly forces the documenter to add the needed specificity. “You can’t just say ‘anemia and GI bleed,’ you have to say the ‘anemia is due to a GI bleed.’ You can’t just say ‘GI bleed,’ you have to say the ‘GI bleed is due to a bleeding gastric ulcer,’” Elion says. “‘Due to’ is really a game changer in the way doctors approach documentation. It’s more complex and comprehensive than that—if you can get the doctors to use those two words, the notes just got massively better.”
The phrase “manifested by” offers similar improvement. For example, rather than generically saying a patient has diabetes-related problems, a better note would say “the patient has complicated diabetes manifested by neuropathy and nephropathy, and retinopathy,” according to Elion.
Getting physicians to make these changes requires communication with CDI teams, says Cortnie Simmons, MHA, RHIA, CCS, CDIP, managing director of education services at himagine solutions inc.
“You have a lot of physicians that code, whether they run a practice or even at hospitals they’re assigning codes through the EHR systems, and you want to make sure their job is more about the documentation and less about the coding, and be available to them,” Simmons says.
She says the best way to do this is to sit down with physicians and show them examples of where they’re struggling. CDI specialists also need to be willing to match their own communication style with the physician. “Because, usually, there’s been pushback from docs on queries, we have to find new ways to jump in front of physicians,” Simmons says.
Disaster #4: Misplaced or Conflicting Information
From data entered into the wrong field in an EHR to scrounging for missing pieces in a hybrid health record environment, misplaced information is still a problem that haunts health information management.
A frequent example, according to Burgess, is that procedure notes end up in the progress note portion, or physicians who are unhappy with a current EHR’s set up start bringing in their own templates.
“It is common that things will end up in the wrong place in the record, and you have to search for it. For coders and for CDI, you have to think, ‘What did this patient come in for?’ If you cannot find a procedure note where it’s supposed to be, look in the rest of the chart,” Burgess advises.
That kind of misplaced information, while frustrating and time consuming, is less harmful than health information spread across a variety of formats. Elion worries that until a given provider is 100 percent electronic, the risk to patients is significant since portions of the complete record may go unnoticed. Many hospitals are still transitioning.
“My own hospital was 90 to 95 percent electronic for several years, which was very dangerous, because I got in the habit of only looking at the electronic [records] and not looking at the paper,” Elion says. “Some documents were still in a three-ring binder that was not available electronically.”
Because of this, physicians and nurses often are working with incomplete information—lab work, radiology reports, family histories—that aren’t available when the patient is in front of them. Hybrid environments, that is, providers using both paper and electronic records, can also breed conflicting information.
For example, the problem list presents its own set of problems, Wallace says. Physicians have been making a solid effort to try and incorporate that information into the body of the patient’s medical record. However, Wallace says she often sees records where the problem list has been brought forward, but the status of the problem has not been updated.
In one case Wallace witnessed, the problem list was copied into the progress note and “the coder assigned a code of a stroke for that patient when in fact they had a stroke six months ago on a prior visit, but because the problem list listed CVA (cerebrovascular accident), they thought it was a current condition.”
Conflicting information is often a side effect of having a large care team collaborating on the same patient, which is the case in hospitals that have a lot of hospitalists. When each one updates a chart, they might use different terminology to explain the same condition. For example, one physician might document that a patient has acute renal failure, and the physician who does the discharge summary might write that the patient has acute renal insufficiency.
“That’s a huge issue because it affects how you code it, how you get paid, and it affects what condition is going to be the principal diagnosis in some instances,” Wallace says.
Outpatient Coding Challenges in CDI
Getting physicians to document in a way that proves medical necessity for a given visit or treatment is a CDI battle waged in inpatient and outpatient coding, but the outpatient side offers a few extra challenges.
Jamie Wilding, MBA, RHIA, coding compliance manager at Esse Health, says that in the outpatient setting physicians do more of their own coding than hospital physicians, which requires more training on coding and electronic health records (EHRs). To keep the documentation up to snuff, “We have to be a little more invasive with our training for our physicians and mid-level staff,” Wilding says.
Outpatient coding also raises the stakes on specificity of documentation. In outpatient coding, “We are not to code for conditions that are deemed unconfirmed or presumed, so if a physician said, ‘I suspect there to be pneumonia,’ he would not be able to code that per outpatient [guidelines] as opposed to inpatient [guidelines] for pneumonia. They code for the finer symptoms until that definitive X-ray comes back,” Wilding says.
One technique that has improved outpatient documentation is the use of medical scribes, according to Wilding. Scribes can significantly reduce a physician’s time spent on paperwork and documentation.
Ironically, though, the same documentation errors that run rampant without scribes are the same ones that occur with scribes, including copy and paste errors and incomplete documentation.
“The errors we find are an increase in the same errors a doctor would make. So whenever labs are ordered they’ll often forget to jot down the planning piece for that diagnosis; ‘Lab ordered today, lipid panel ordered’ under a diagnosis for hyperlipidemia,” Wilding says. “Without that piece of the puzzle” external auditors and insurance companies would argue that there’s no documentation to support the services rendered, she says.
Preparation the Best Way to Avoid or Manage Disaster
There are several things CDI specialists can do that will prevent many, if not all, of these four documentation disasters. There are technology tools such as natural language processing (see sidebar above), computer-assisted coding, scribes (see sidebar at left), and software, some of which offer an electronic physician query platform and helps provide suggestions for missing diagnoses.
Elion says that in a recent presentation he advised CDI specialists to “train your doctors to write a good note wherever they are and whoever they are seeing. You don’t write one style of note for a Medicare patient and another style for a BCBS patient. You don’t write a different style of note for an outpatient than you do for an inpatient… don’t guess what the payer is looking for.”
Having a physician champion as part of the CDI team can work wonders in getting physicians to take documentation improvement seriously—and understand the disastrous consequences if they don’t.
“With physician champions or advisers, you can have physician to physician communication which is sometimes a lot easier of a position,” Burgess says. “I know some hospitals can’t afford to hire physician champions. But possibly take your physician who is marked for case management and utilization review, and make this part of their role.”
1 Lerner, Barron H. “A Case That Shook Medicine.” Washington Post. November 28, 2006. www.washingtonpost.com/wp-dyn/content/article/2006/11/24/AR2006112400985.html.
Mary Butler (firstname.lastname@example.org) is associate editor at the Journal of AHIMA.
"Preventing Healthcare's Top Four Documentation Disasters"
Journal of AHIMA