By Robert S. Gold, MD
In my last column, I wrote about some educational opportunities I've found in chart reviews for clients, and I thought I'd provide some more.
Sequencing viral gastroenteritis
We're probably all familiar with the Coding Clinic advice regarding sequencing of viral or unspecified gastroenteritis, especially where the evidence exists that, without other diagnoses, this could be treated at home with over-the-counter preparations and chicken soup.
Coding Clinic started with July-August 1984, p. 19-20, where the advice suggested that dehydration should be the principal diagnosis because IV therapy trumps chicken soup.
It continued with Second Quarter 1988, p. 10, where the advice stated ICD-9-CM code 008.8 (intestinal infection due to other organism not elsewhere classified) or 009.0 (infectious colitis enteritis and gastroenteritis) should follow the code for dehydration. The sequencing advice continued to First Quarter 2008 where, again, dehydration superseded viral gastroenteritis.
Well, what I've been seeing is the attempt to sequence ICD-9-CM code 584.9 (acute renal failure [acute kidney injury]) as the secondary diagnosis and the nonbacterial gastroenteritis code as principal, even though the acute renal failure is caused by severe dehydration from the patient's diarrhea and vomiting.
Why? Because sequencing the gastroenteritis first gives a higher relative weight and because the physician doesn't specifically call it "dehydration." Sequencing for dollars rather than truth is inappropriate and just plain wrong to do. The acute renal failure is the reason for admission, not the lesser condition, just as it says in many other Coding Clinic citations.
Coding past medical conditions
With the plethora of copy-and-paste notes making analysis of medical records a nightmare and the advent of computer-assisted coding making some parts better and other parts worse, I'm finding more diagnoses in the final code list that no longer exist.
After a renal transplant, a patient does not have end-stage renal disease--unless the transplanted kidney has failed.
After a liver transplant, a patient does not have end-stage liver disease--unless the transplanted liver has failed.
After an aortic valve replacement, the patient does not have aortic stenosis--unless the transplanted valve is functioning improperly.
After replacement of a left knee for osteoarthritis, a patient no longer has osteoarthritis of the left knee, but may have it in the right knee or the hip.
After a pancreatic transplant for type I diabetes, the patient no longer has diabetes when a functioning pancreas is creating all the insulin the patient needs, but the patient may still have all of the manifestations of diabetes.
When the patient was in the hospital for an infection (or sepsis from an infection) and is no longer on antibiotics, the patient doesn't have an infection.
Watch out for this copy-and-paste disease. This is something that has to be fought intensively and critically.
Assignment of neonatal codes
Codes in the 770 series (other respiratory conditions of fetus and newborn) are frequently used in children's hospitals when a child either develops a problem after birth or is admitted to the hospital within 28 days of being born.
The 770 series of codes is intended to identify problems in a child that originated in the process of being born, during the intrauterine period or the trip down the pathway to the brave, new world. It is not--let me repeat not--intended to be used for something that happened to a child after birth, even within the first 28 days of life, if it was not directly related to the process of being born.
The direction of not using these codes if it's evident that the condition was community acquired is misperceived and grossly overlooked by coders who are delighted in getting a DRG with a relative weight of about 3.5 rather than one with a relative weight of 0.6.
Neonatal and newborn have definitions in the coding world. But the intent of the 770 series of ICD-9-CM codes has been internationally defined for conditions related to being born and congenital conditions.
Cardiac arrest of the newborn (779.85) is not intended for a child who had a SIDS event when the child vomited and aspirated and stopped breathing and living. The birth process did not cause that.
Don't assign neonatal sepsis (771.81) when a child gets an infection from a dog bite and becomes septic. This is not related to the birth process.
Other disturbances of temperature regulation of the newborn (778.4) is not the code for a hot baby whose aunt left the child in a car while she went shopping and the child got heatstroke. That is not a congenital condition.
The ICD-9-CM Official Guidelines for Coding and Reporting state:
If a newborn has a condition that may be either due to the birth process or community-acquired and the documentation does not indicate which it is, the default is due to the birth process and the code from Chapter 15 should be used. If the condition is community-acquired, a code from Chapter 15 should not be assigned.
The ICD-10-CM Official Guidelines for Coding and Reporting contain the same guideline, except that the codes exist in Chapter 16.
Use your head. Don't default to congenital when it's obviously not and code it just because the doc didn't say "community acquired." It doesn't take much to make the right choice.
Open, closed, or no reduction
The problem of open and closed reduction persists despite two statements of advice in Coding Clinic and a zillion talks I've given on the subject. Basically, don't code what the doc said he did. Code what the doc actually did. This means reading the operative report.
Femoral fractures can be open or closed; that's easy. Closed fractures can be displaced or nondisplaced. The doc can reduce the fracture before or after the incision.
Sometimes, when a doc performs a reduction before the incision, the alignment still isn't good enough when inside the thigh, so the physician must perform further manipulation and reduction.
Sometimes the alignment is so good that no reduction has to be accomplished at all. This is almost always stated quite well in the body of the operative report. And the docs often call them all open reduction internal fixation (ORIF) because that's how they bill for their CPT® code.
The last word
Valid coding related to presence or absence of the diagnosis in the discharge summary causes meltdown in a lot of hospitals, in both their coding and CDI departments.
Some of it started when Recovery Auditors denied claims when a particular diagnosis was not parroted in the discharge summary, whether it was a CC or MCC or even the principal diagnosis. Then, a few years ago, a note came out in a Medicare Learning Network® Matters article stating that it was inappropriate to assign a code for a diagnosis not present in the discharge summary. Boy, did that raise hackles.
Through considerable discussion with MLN, AHIMA, Coding Clinic, and National Center for Health Statistics leaders, that advice was thrown out. It is always appropriate to assign codes for diagnoses that are not present in the discharge summary when those diagnoses are truly established in the body of the medical record.
When a patient is admitted for pneumonia and, on the way out the door, falls and breaks a hip and is operated on by orthopedics for repair and is then discharged, what's the orthopedist's diagnosis? Fractured hip. Think there's any reference to pneumonia? Not a chance. Was the patient admitted and treated and got better from the pneumonia? Sure enough. Is it codeable? Absolutely. In fact, it's the principal diagnosis even though it was totally absent from the discharge summary.
A consultant is called in to check on a patient's heart on a congestive heart failure (CHF) admission. An echo shows an ejection fraction of 32% and the consultant calls it "decompensated systolic heart failure." The attending or the resident just puts CHF in the discharge summary. Is that a discrepancy? No, it is not. Is that a disagreement? No, it's not. The specificity was supplied by the consultant and the supplied specificity is validated in the results of the echo in the body of the medical record. Is it codeable? Yes.
When a diagnosis is not established but gets copied and pasted throughout the medical record because somebody got lazy and then that diagnosis doesn't appear in the discharge summary, is that codeable? Not a chance! So who's to know? That raises more hackles. And the docs have to learn that they are causing bad data to be entered into their patients' databases and that's unethical and just plain bad medicine. Let them know it.
Editor's note: Dr. Gold is CEO of DCBA, Inc., a consulting firm in Atlanta that provides physician-to-physician CDI programs, including needs for ICD-10. Contact him at (770) 216-9691 or rgold@DCBAInc.com.