March 17, 2014

Is ICD-10 Ready for Use?

By Robert S. Gold, MD

EDITOR’S NOTE: In this article, Robert S. Gold, MD, explains problems associated with ICD-10 and answers questions he is frequently asked.

Going from a healthcare coding system with 14,000 diagnosis codes to one with 73,000 diagnosis codes is undoubtedly a challenge.

But why is our ICD-10 different from that of any other country?

Our professional associations have learned that the information that really has to be tracked for many patients requires more precise choices. The American College of Cardiology had asked for some revisions in the codes denoting a heart attack, for example, because sometimes specific information as to what artery was involved in the MI is only found at cardiac catheterization (rather than what wall was involved). The organization also asked for some specifics regarding chronic atrial fibrillation and atrial flutter, as these have significant differences in presentation and require different treatment – and one code for each doesn’t cut it.

Orthopedists and the American Academy of Family Practice recognized that it’s important to have information about a fracture that just cannot be captured with our current system – whether the fracture was in the proximal, midshaft, or distal end of a long bone is significant. Whether it was an open fracture or a closed one is also significant to track because of differences in outcome. Likewise, whether a fracture was displaced or non-displaced can have an impact on treatment.

Virtually all of the expansion in volume of ICD codes, from the 13,700 used by parts of the rest of the world to the 73,000 we are starting with this fall, is happening because our doctors wanted it.

But the folks who created and voted on the codes made some significant changes by themselves – and some of them came along with significant errors. Following are a series of questions we’ve been posed recently, plus our answers.

You’ve been yelling about hypostatic pneumonia for a number of years in various publications. Are there any differences in this diagnosis?

Once upon a time (now, actually), hypostatic pneumonia and pulmonary congestion had in ICD-9 a code that groups to pulmonary edema and heart failure. The problem is that the people who classified and defined this condition didn’t take the time to look into the history of it, so they took the words straight from older versions of the ICD classification system. The condition was actually listed in ICD-1. Back then it was called hypostatic pneumonia, pulmonary apoplexy, or congestion of the lungs. It was well-described in the medical literature of the 1850s by the pathologists who examined the patients who died with it.

It was only found during post-mortem examination in people who had been resting in one position for weeks to months – they were often severely malnourished, with no protein stores to hold fluid in the bloodstream, and their lungs had turned to the consistency of the liver (called hepatization of the lungs) due to settling of fluid in the dependent portions of the lungs. There typically was no indication of infection, no indication of heart failure.

In ICD-9 this was found under pulmonary edema or congestion of the lungs because of the original wording – the description that the lungs became severely congested with blood and debris. This led DRG coders to group the diagnosis along with pulmonary edema and respiratory failure, and consultants taught that this was the right thing to assign when there was documentation of “pulmonary congestion.”

Who knows what the instigation of the change was, but in ICD-10, it is now found in with the pneumonias – and it’s not pneumonia. Still, in the current medical literature, you will see descriptions of cases of patients with terminal cancer, presenting as cachectic and severely malnourished with ascites and low oxygen saturations due to consolidation of lung tissue with surrounding leakage of pleural effusion. This is real hepatization of the lung in a terminal patient. This actually represents what the intent of the code was when it was initially described, because back then they didn’t have X-rays. Now, we do, and we can identify the condition before a patient dies. It’s even described in the federal government’s medical literature website at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2967681/.

Again, it’s not pneumonia and should not be grouped there. People took advantage of the phrase of “hypostatic pneumonia” because in France, where the term came from, they called diseases of the lung “pneumonias” whether it was an infection of the lung or not. In the U.S., we have had consultants who traditionally have insisted that any patient who gets atelectasis from lying still for a couple of hours while in the hospital (with the area of atelectasis ultimately becoming infected) should have it documented as “hypostatic pneumonia” in spite of the fact that it’s not pneumonia. They did so because it grouped to a DRG that had a 30–40 percent higher relative weight than pneumonias so they could make more money out of it.

People from the Centers for Medicare & Medicaid Services (CMS) eventually indicated that if a pneumonia is treated with antibiotics, whether it’s called hypostatic or not, it should be classified as ICD-9 code 486. Now it is grouped with pneumonia, but it’s still not pneumonia. It’s now different, but certainly not right.

You’ve mentioned in the past that coding advice on assignment of codes for cardiomyopathy has been erroneous because, as you were told by one of the guardians of ICD codes, the intent of ICD-9 code 425.4 was to address primary cardiomyopathies – and virtually all patients in the Medicare age group have secondary cardiomyopathies. Is this better in ICD-10?

Hardly. With ICD-10, we go back to the dark ages in some areas of cardiomyopathy coding, and that’s so sad. In ICD-9, virtually all of the codes in the 425 series identified causes of sick heart muscles and, whether they were primary or secondary causes, they could address one of the specific diseases in the 425 series (or identify a cardiomyopathy caused by a different disease, or even state that there is no identifiable cause). The whole intent of ICD is to identify and indicate causes of disease.

Now we have two codes that describe what the heart looks like but have absolutely no indicators that the physician has any idea whatsoever as to the cause of the patient’s sick heart muscle – and that’s just plain wrong.

ICD-10 code I42.0, dilated cardiomyopathy, indicates what the left ventricle looks like and gives no indication why it looks that way. There are many causes of dilated left ventricles that don’t work properly, and they can be alluded to along with documentation of “dilated cardiomyopathy” leading to a useless ICD code. Ischemic heart disease (now I25.5), peripartum cardiomyopathy (now O90.3), cardiomyopathy due to coxsackie viral myocarditis that presents in children (now B33.24), toxicity from chemotherapeutic drugs (now included in I42.7), alcoholic cardiomyopathy (now I42.6), cocaine use (also I42.7, listed as a toxin), TakoTsubo syndrome, and stress cardiomyopathy (now I51.81) – all of these lead to dilated hearts. With so many possible choices, as long as we get to I42.0, who cares, right?

We also have an ICD-10 code, I42.5, for other restrictive cardiomyopathy. This describes how the heart works (or doesn’t work) and leads to heart failure – but, like I42.0, it gives absolutely no indication as to why the patient’s heart muscle is faulty. This should never happen,

You’ve discussed the inappropriate definitions of a code for ARDS in the respiratory failure categories of 518.8 codes and the 518.5 codes as well. Under ICD-10, will this get better?

In truth, the 518.8 issue has been fixed – almost. Semantics has hit us many times in the past, and “other respiratory insufficiency,” the title of 518.82, has now disappeared. Instead of having a representation of this in the ICD-10 respiratory failure categories, ARDS has a code of its own, J80. It is defined properly. It is a breakthrough. Unfortunately, some pointers still exist in the index, potentially sending the code assignment to ARDS when the patient just has a symptom of acute respiratory distress, not actual acute respiratory distress syndrome– and that’s poison. Medicare will continue to pay a billion inappropriate dollars if it’s not fixed. And it’s an easy fix. You don’t have to change the code or the definition at all – just dump a couple of pointers in the index and it’ll be perfect.

Unfortunately, having almost fixed the equivalent of “other respiratory insufficiency” from the 518.8 series, the equivalent of 518.52 “other respiratory insufficiency” after surgery has been retained – and its severity has been assigned to a major comorbid condition. The federal government is going to pay multiple billions of dollars in inappropriate payments for a listed condition that is undefinable, unmeasurable, irreproducible, and just plain nonsensical. CMS knows it. They’re having trouble working out if they can fix it and how to get it fixed. NCHS can fix it in a second, but they won’t – unless the whole pulmonary world comes after them for not doing their homework.

All respiratory insufficiency means, as listed in any medical publication produced since the 1850s, is that the lungs aren’t doing all they should to oxygenate the blood or to clear carbon dioxide. To have codes for this is unfathomable. And to create a severity status of MCC when you can’t measure it is equally inappropriate.

The intent of the code was to indicate ARDS after surgery or trauma – but they took away the trauma piece. I’ve gained the support of the most influential physicians in the world on ARDS, yet NCHS won’t even respond to their communications.

If a patient develops acute respiratory failure as a complication of a surgical procedure, then he or she certainly deserves to have ICD-10 code J95.821 assigned, indicating acute post-procedural respiratory failure. That’s a major complication.

Respiratory or pulmonary insufficiency? It’s a joke.

About the Author

A graduate of Hahnemann Medical College in Philadelphia and spending his professional career in the Navy, Dr. Gold and his company, DCBA, Inc. in Atlanta have become leaders and educators in Clinical Documentation Improvement nationally.  His insights have led to the uncovering of several errors in the ICD coding system and upcoming needs for more change in ICD.

Contact the Author

To comment on this article please go to