Dr. Robert S. Gold founded DCBA, Inc., in Atlanta, a consulting firm that provides physician-to-physician programs in CDI. The goals are data accuracy, profile management, and compliance in the inpatient and outpatient arenas. He can be reached by phone at 770/216-9691 or by e-mail at DCBAInc@cs.com.
I recently attended the annual meeting of the ACDIS in Orlando where I heard a variety of perspectives during many informative
A common thread in many discussions was how to query physicians using information in the medical record and when it’s appropriate to do so. Coders and CDI specialists must remember that a valid diagnosis must be supported by evidence in the body of the medical record. What does this mean?
Let’s consider the often difficult decisions associated with three common diagnoses: Anemia, acute renal failure, and sepsis.
Q: Is systemic inflammatory response syndrome (SIRS) an inherent part of an infection? For example, when a patient comes in with pneumonia, and clinical indicators are present for SIRS but the physician did not specifically write SIRS, should coders or CDI specialists query physicians?
A: SIRS is a response by the body to dead or injured cells. It is inherent in infections and should not be sought after when the condition is an obvious infection. Pneumonia, for example, is an obvious infection.
Certain advice from coding authorities has CDI specialists/coders to lead physicians down the garden path of identifying and documenting SIRS whenever a patient has two of the four criteria described in the Journal of Critical Care Medicine. However, the criteria for SIRS referenced in the article were actually established in 1992 as part of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference. The conference concluded that the manifestations of SIRS include, but are not limited to:
- Body temperature less than 36°C or greater than 38°C
- Heart rate greater than 90 beats per minute
- Tachypnea (i.e., high respiratory rate) with greater than 20 breaths per minute and/or an arterial partial pressure of carbon dioxide less than 4.3 kPa (32 mmHg)
- White blood cell count less than 4000 cells/mm3 (4 x 109 cells/L) or greater than 12,000 cells/mm3 (12 x 109 cells/L), and/or or the presence of greater than 10% immature neutrophils (band forms)
But soon CDI specialists/coders started to query physicians about SIRS when only a couple of the aforementioned criteria were present regardless of how sick the patient was and regardless of whether vital sign changes were even related to the identified infectious process.
Overcoding SIRS leads to inappropriate billing, not to mention the potential for RAC recoupments and tremendous penalties if identified by the Office of Inspector General.
When a patient has sepsis (ICD-9-CM code 038.xx), the patient has sepsis.
The 2001criteria for sepsis from the Journal of Critical Care Medicine emphasizes that. It states that the patient has to be truly sick from the infection. If the patient does not have sepsis in the face of an infection, the presence of SIRS is implied.
Virtually all patients with acute appendicitis meet the criteria for SIRS, and it’s only the rarest cases of appendicitis that have sepsis. Most patients with bacterial pneumonia exhibit signs and findings of SIRS, but most receive outpatient treatment and do not have sepsis.
Editor’s Note: This article first appeared on JustCoding.com. Learn more about the difference between sepsis and SIRS in the following articles:
By Robert S. Gold, MD
If a coder reviews the chart of a patient in the emergency department (ED) or intensive care unit with documented symptoms, such as fever (or low temperature), elevated white cell count (or low white cell count), altered mental status, evidence of an infection (e.g., pneumonia, pyelonephritis, a rigid abdomen with speculation of diverticulitis or perforation of another intestinal organ), and hypotension, remember the following thought process.
The chart may show that the physician gave the patient a bolus of saline (250–500 cc), or Ringer’s lactate and another bolus. If the patient perks up and feels better—and the creatinine drops from 5.4 to 2.7—then the patient likely had hypotension due to severe dehydration (ICD-9 code 276.51).
The rapid change in creatinine levels show that the patient was, indeed, pretty dehydrated and is coming back toward better levels. However, if the creatinine does not return to the patient’s baseline within 24 hours of fluid resuscitation and remains significantly elevated over the patient’s baseline for more than 24 hours, there was likely acute renal damage or acute kidney injury. The mechanism of the damage likely depends on the presence of absence of shock (in giving the physician a clue if it represented acute renal tubular necrosis or ATN).
The patient may have had sepsis (ICD-9 codes 038.9 and 995.92) from that infectious process and metabolic encephalopathy, explaining the altered mental status (ICD-9 code 348.31) due to the sepsis with acute renal failure.
If the patient does not respond to the fluid challenge, and the physician starts the patient on pressors (ICD-9 code 00.17), such as levophed, dobutamine, or dobutrex, coders may assume the patient is probably in shock. The question is, was it hypovolemic shock (ICD-9 code 785.59) or septic shock (ICD-9 code 785.52)? The physicians should document the presence of shock and the etiology of it in this case.
We, in the United States, will soon need to adopt the electronic health record (EHR). Over the past few years, of course, healthcare entities have slowly converted from the traditional handwritten medical record toward one that is totally digital. There have been a myriad of interim steps, from models which simply scan the handwritten documents into a computer system to those which begin with an electronic template complete with check boxes and drop down menus.
Some programs use dictated and transcribed elements for physician or mid-level provider entries. Some programs take the electronic record and utilize fuzzy logic to search for key words, phrases, and abnormal lab results in an attempt to assign ICD diagnosis codes.
But what is the purpose of this transition? What is the stated goal? What is the real goal to be achieved with an EHR? There are two major reasons for the development and adoption of the EHR. One we are close to achieving, the other, in my opinion, just doesn’t exist.
The desire to ensure that an electronic model of a patient’s encounters contains all of the elements needed for somebody’s perception of a “complete” record has led to the currently marketed and sold models. To know that a patient’s problem list is present, that his or her medication list and reconciliation is there, to know that the history and physical (H&P), progress notes, discharge summary, nursing notes, and all the rest of the elements that make up a traditional patient’s chart are present is a great reason to have an electronic record. It forces people to remember to make these entries. Also important is the electronic records ability to:
- Keep the patient’s inpatient and outpatient encounters available in one compendium
- Track crossover treatments provided by different specialists
- Monitor for drug interactions
- Track scheduling and follow-up visit dates
The potential to facilitate practice guidelines and order sets based on best practice is great—if the physician knows how to set them up and use them or, in specific cases, elect not to use them for the benefit of the outlier patient. So, for these reasons, I say good for the companies who have put such an important functionality together.
But, there are still problems with implementation of the so-called complete EHR. In the January 2011 edition of the Journal of AHIMA, Genna Rollins writes about the experience of Barbara Drury, president of Pricare Consulting. According to the article, Drury experienced difficulties with the EHR recording inappropriate dates at inappropriate times. She also expressed frustration with the inability of tracking patient data when someone does not have access to the electronic record and worried about those who would then develop alternative, manual, work-arounds.
Yes, there are still process issues.
I challenge you to find one textbook of medicine, physiology, or pathology that refers to a category of hematologic abnormality as “acute blood loss anemia.”
But coders are very familiar with that term. They may ask a physician whether a patient has acute blood loss anemia. Or they may ask, “What kind of anemia is this?” And physicians essentially have no chance of ever guessing that they are after “acute blood loss anemia” or even “chronic blood loss anemia.”
What coders really want to know is the cause of the patient’s anemia. Was it anemia due to acute blood loss from a ruptured esophageal varix or from a fractured femur? Was it anemia due to chronic blood loss from hematuria from a patient’s multiple bladder polyps or from menometrorrhagia? Was it anemia due to a chronic infection such as chronic osteomyelitis or chronic hepatitis C? Was it anemia due to the patient’s CKD 4?
That’s why the CDI specialists need know how to take that basic coder impulse and help to ask the right question, and obtain the most accurate documentation for the care the physician provided. After all, it’s the end result that counts—and makes everyone happy.
Many of the nation’s hospitals now have clinical documentation improvement (CDI), management, or integrity programs. They are designed to help physicians improve the documentation of diagnostic or procedural information in inpatient medical records so that the documentation meets the needs of the coding process. There are good things that can come out of these programs, but there can also be bad things.
The first thing for the physician to know is that, if done correctly, a CDI program is one of your best friends. When coders correctly assign the ICD-9-CM (and soon ICD-10) diagnosis and procedure codes, physicians get the following benefits:
- Proper and accurate representation of patient morbidity and mortality statistics
- Proper and accurate representation of physician’s thought processes and actions in treating the patient so that physician billing—as well as the hospital’s—is accurate
- Increased communication among the healthcare providers in dealing so that the most expeditious care can be rendered
- Improved documentation of physician’s daily thought processes in treating patients or those for whom physician’s are consulting
Let me tell you about some things to look out for, however, because you should probably not participate in the following scenarios:
- Acute blood loss anemia. After a hip fracture repair, you may see that the 76-year-old female patient’s hemoglobin dropped from 14 to 12.5 and stabilized. If someone asks the physician to document “acute blood loss anemia,” that’s a problem. First, the patient is not anemic, so it’s unethical. If the patient’s hemoglobin drops from 14 to 8.5 and she becomes tachycardic and physician decides to either transfuse or provide an iron supplement after discharge, call it “anemia due to acute blood loss from the hip fracture.” That tells the story properly.
- Acute respiratory failure. If a patient comes to your service on the cardiovascular ICU after a coronary artery bypass graft and is on a ventilator because of morbid obesity, and early extubation would likely lead to problems, don’t document “acute respiratory failure.” The patient is sedated and maintained basically anesthetized on purpose, not due to a disease. Use “ventilator management” in your documentation. However, if the patient develops a disease and must be reintubated, acute respiratory failure becomes appropriate.
- Systemic inflammatory response syndrome (SIRS). The criteria for a significant infection that the Society of Critical Care Medicine put out in 1991 have undergone considerable modification. Can you recall any patient with acute cholecystitis or acute appendicitis who has not had two of the four criteria of SIRS? Have any of your children had a middle ear infection and did not have a fever and tachycardia because it hurt? Did your child need to be on a critical care unit with a central line, or did he simply need some amoxicillin? SIRS is built into bacterial infections. Don’t describe abnormalities in vital signs or lab results in a patient with a simple bacterial infection as SIRS. When a patient has sepsis, you can tell from across the room. Call that sepsis! Reserve SIRS for noninfectious conditions such as significant burns or necrotizing pancreatitis.
- Acute kidney injury (AKI). The Acute Kidney Injury Network has written an exhaustive treatise on what is still often referred to as acute renal failure. When a patient has simple dehydration and small bumps in the creatinine level and everything else returns to normal within six hours of fluid resuscitation, it was not AKI. Should an increase in creatinine level from dehydration persist after initial rehydration, then it was AKI. If it resolves, it was dehydration, prerenal azotemia, or acute renal insufficiency.
- Acute on chronic systolic heart failure. You probably have many patients in your practice who are chronic heart failure (CHF) patients. The specificity of left, right, or biventricular heart failure is important. If you see a patient in your office with left heart failure, differentiate whether it is characterized by systolic or diastolic dysfunction or both. When a patient is admitted to the hospital for acute CHF, document it as “decompensated” plus the chronic state that you know from your office and the echo performed a month ago. If it’s apparent, also document the cause of the acute decompensation. In the hospital, “decompensation” or “exacerbation” is sufficient.
- Encephalopathy. This is simple. If the patient has a condition that caused encephalopathy, call it whatever specific encephalopathy you can. A person who is simply drunk does not have encephalopathy. A person who took a dozen Elavil as a suicide attempt is tranquilized. These are not encephalopathy.
The issues above mean a lot for you, your patients, and the nursing staff. They’re important in the office as well as the hospital. When a CDI staff member asks a good question that improves informational value, provide all the help you can. But when you are asked to document something you don’t believe in or something unsupported in medical literature, think twice.
Editor’s Note: This article first appeared in the January 2011 edition of Medical Records Briefing.
In March 2008, the National Kidney Foundation (NKF) sent a letter to the National Centers for Health Statistics and asked that the term “acute kidney injury” (AKI) be an optional phrase for assigning ICD-9-CM code 584.9 (i.e., that it was the current terminology for what used to be called “acute renal failure.”) The request was granted.
CDI specialists and coders looked at the definitions published through the Acute Kidney Injury Network and noted the criteria of stages 1, 2 and 3 of AKI. They began to ask physicians to document AKI whenever a patient’s labs or urine output met those criteria. But in so doing we hurt ourselves—and we hurt “acute renal failure.” And it’s a shame. Now, with CMS’ resulting reduction of acute renal failure to CC, it now has no more severity attached to it than a simple urinary tract infection.
Let’s see where we went wrong and where we have to go to get it right.
What AKI was
Long ago, in a galaxy far, far away, we had the terms “acute renal insufficiency” and “acute renal failure,” and we had advice as to when
each was appropriate terminology to use in order to assign the codes that were designed for those conditions. These codes included 593.9 for renal insufficiency and 584.9 for acute renal failure (notice I didn’t use the term “acute” in the description of 593.9—it was a nonessential modifier then).
In the medical textbooks and within the medical community there was confusion in the terminology. For example, one medical textbook called the disease “renal insufficiency” and talked about “progressive renal failure,” and another textbook had a chapter entitled “renal failure” and spoke of “anemia of renal insufficiency.”
AHA’s Coding Clinic for ICD-9-CM, First Quarter 1993, p. 17, gave us the following definitions as guidance:
“Renal insufficiency: It is generally accepted that renal insufficiency (code 593.9, Unspecified disorder of the kidney and ureter) refers to the early stages of renal impairment, determined by mildly abnormal elevated values of serum creatinine or BUN or diminished creatinine clearance. Clinical symptoms or other abnormal laboratory parameters may or may not be present but are usually minimal.
The treatment of renal insufficiency depends to a very large extent on the underlying cause, with much attention given to the possibility of preventing progression to renal failure.”
And on p. 18, Coding Clinic states:
“Renal failure: Renal failure (code 584, Acute renal failure, code 585, Chronic renal failure, code 586, Renal failure, unspecified) is a progression of renal insufficiency where renal function is further impaired and overt clinical consequences, such as anemia, have developed. In essence, renal insufficiency is more of an abnormal laboratory assessment, while renal failure incorporates both abnormal laboratory and clinical findings.”
Back then, we looked for more than dehydration. We looked at more than a minor bump in creatinine. We looked for a sick patient. Sure, there may have been obstruction due to prostate cancer associated with a high creatinine level which returned to normal soon after suprapubic tube insertion.
Yes, there may have been a patient found down for three days at home who came in with depressed mental status and rhabdomyolysis and responded to a couple of liters of IV fluids in a couple of days. And these patients came back, at least measurably, to normal. But they were sick and they had acute renal failure. And if a patient had two days of diarrhea “altered mental status” and responded to a glass of water to return the creatinine level to normal, we were happy with dehydration or prerenal azotemia or acute renal insufficiency—because that’s what is was.
During this time, the renal world recognized the problems with lack of consistency and did some studies of acute decrease in renal function. It came up with mortality rates, rates of need for Renal Replacement Therapy (RRT – or dialysis) and identified three pretty distinct levels of change in creatinine, change in Glomerular Filtration Rate (GFR), and change in urine production. It also identified two levels of long-term outcome: Return of measurable function, or no return of measurable function. They published their results and findings and called the system RIFLE. And all rejoiced.
The reason for assignment of a certain principal diagnosis is occasionally incomprehensible. First, consider what a principal diagnosis is. It’s the condition or group of conditions that, after workup, the physician determines led to the patient requiring inpatient hospital care. Essentially, it’s why the patient is in the hospital. Sometimes it’s the same as what got him or her off the couch, but not always.
Here’s one example: A patient may come to the ED for a crushed finger after dropping something heavy on it. But when the physician works up the patient in the ED, he or she may determine that the patient dropped the heavy thing because of weakness caused by a stroke.
The stroke is what needed inpatient care—the finger could have been taken care of in the ED.
Or consider the following: A patient is admitted with acute renal failure with rhabdomyolysis and a potassium level of 6.8. The folks in the ED work with the physician on treating the patient urgently with Kayexalate, insulin, and glucose. Soon after, the physician learns that the acute renal failure actually represented an organ failure from the patient’s severe pneumonia that he or she could not see on the admitting chest film because of the extent of dehydration. The physician and the infectious disease consultant agree that the patient was septic due to the pneumonia with acute tubular necrosis causing the acute renal failure.
When coding this scenario, coders will assign this case to the category (i.e., DRG) of sepsis. The principal diagnosis code will be septicemia, followed by pneumonia and acute renal failure and a code for severe sepsis. Why? Because those are the coding rules. Coders can’t deviate from certain priorities for sequencing of disease codes. And hang the fact that the physician never called it “septicemia.”
Another situation that could be confounding is the admission of a patient who has cancer. Take a patient who is severely anemic and weak; this presentation leads the physician to admit and work the patient up and administer a few units of packed cells so that he or she would feel stronger. Well, if the patient does okay and is discharged, then coders will report anemia as the principal diagnosis. Hopefully, the physician will specify whether that anemia was due to chronic blood loss from the cecal neoplasm or chemotherapy that the patient had been receiving or whatever the physician determines the cause of the anemia to be.
Consider also that if the patient dies of terminal cancer, the principal diagnosis might be the cancer. However, consider if that same cancer patient is stabalized and is about to be discharged but has an acute ST-elevation myocardial infarction (MI) which puts him or her into cardiogenic shock. The physician uses a balloon pump and discusses with the patient’s family the decision to let him or her pass away according to the patient’s wishes. The patient then dies of the acute MI. But the principal diagnosis will not be the acute MI, even though it may have been the cause of death.
See how it’s not so simple? Sometimes the rules don’t make a lot of logical sense, but they are the rules.
Editor’s Note: This post is an excerpt from JustCoding.com premium content.
Some people go into an endeavor full bore, knowing that they’re always right. Some people go into an endeavor learning as they go, knowing that they don’t know everything, eager to learn and always open to learning.
I found at the 2010 ACDIS conference in Chicago that most of the attendees were eager to learn and were sopping up knowledge like a sponge. Some–a very few–thought that the attention turned by a few of the presenters toward the compliance and regulatory agencies was too heavy-handed. However, the fact is that organizations are out there challenging what is done in CDI circles. It’s unhealthy and potentially dangerous to ignore the warning signs.
Back in the mid-1990s, hospitals were told that it’s right to change the code for documented “pneumonia” to the code for “gram negative pneumonia” when a patient was identified as residing in a nursing home and being diabetic with COPD or CHF. The reasoning was that “Coding Clinic tells us that this is so,” referring to an enducational citation on patients at risk of gram negative pneumonias (Coding Clinic 3Q 1988, p. 11).
Hospitals were told that these are the rules. Hospitals believed the consultants. Hospitals got fined billions of dollars when it was identified that the patients really didn’t have sufficient evidence in the medical record that the physician thought it was really gram negative pneumonia and really treated the patient with shotgun therapy for “CAP.”
I am not saying that it’s wrong to have confidence in what you’re doing as a CDI specialist. I’m saying that not asking questions, validating what you’re told, checking the resources of folks who have been bitten by the OIG and the RACs and the QIOs, and learning from the mistakes made in history, will lead you to repeat the mistakes done by others.
Ostriches are pretty birds. They can be easy prey when they stick their heads in the sand. The message isn’t doom and gloom — it’s check your resources and mind the warnings.