The OIG Shines Light on “history of” in MA Programs

Learn how providers can avoid falling into the trap of making clinically suspect diagnoses.

The U.S. Department of Health and Human Services (HHS) Office of Inspector General (OIG) is charged with oversight and protecting the integrity of our governmental agencies and is supposed to prevent and detect fraud, waste, and abuse.

Their reports are publicly available, and one came out mid-March which piqued the interest of Dr. Ronald Hirsch who brought it to my attention. It is called Medicare Advantage Compliance Audit of Specific Diagnosis Codes that Geisinger Health Plan Submitted to CMS (Medicare Advantage Compliance Audit of Specific Diagnosis Codes That Geisinger Health Plan (Contract H3954) Submitted to CMS (A-09-21-03011). (hhs.gov)).

There has been concern that Medicare Advantage (MA) programs have been crediting patients with hierarchical condition category (HCC) conditions which may not be clinically valid, which results in higher risk adjustment scores and, thus, higher payments. The OIG audit focused on nine groups of high-risk diagnosis codes and then extrapolates alleged overpayment. The institution, which is being investigated can, and usually does, dispute the charges.

I am going to shed some light on those diagnoses which were targeted, and why, and how providers can avoid falling into the trap of making clinically suspect diagnoses.

First, I need to explain that the verbiage, “history of” means something different to a coder than it does to a clinician. Clinicians are often indicating that “the patient’s medical history includes X, Y, or Z,” and they are likely not trying to convey whether the condition is currently present or historical. In contradistinction, “history of” leads a coder to a set of personal history codes in the Z85-87 categories. This signifies to them that the condition is resolved, old, no longer actively being treated.

Conversely, when a provider is selecting an ICD-10-CM code for a “history of” condition, they often choose with convenience. They are not— and do not aspire to be, coders— and they are trying to pick a code. Any code. Just something to put in that field so they can submit a bill and move on to their next patient. This can lead to error because practitioners are not coders. Did I already mention that? They may pick something that approximates what the patient has and then they don’t look back.

Here are the conditions which the OIG cited:

• Acute stroke: This is not the first occurrence of an OIG report finding erroneous claims of acute stroke. There are 3 basic code sets regarding cerebrovascular disease:
  • Acute cerebrovascular accidents. These are subcategorized into type of stroke (e.g., nontraumatic subarachnoid hemorrhage, nontraumatic intracerebral hemorrhage, cerebral infarction), and maximal specificity of the code includes site or artery involved. These are found in ICD-10-CM codes I60-I63.
  • Sequelae of cerebrovascular disease, found in I69. These codes detail the specific deficit (e.g., aphasia, frontal lobe and executive function deficit, monoplegia, hemiplegia) and relate it to the specific type of stroke as noted in the bullet above. It should be noted that paralytic conditions stemming from stroke carry a higher risk adjustment score than acute stroke because they predict future resource consumption.
  • Codes that designate occlusion and stenosis of arteries not resulting in cerebral infarction (I65-66), Z86.73, Personal history of transient ischemic attack (TIA), and cerebral infarction without residual deficits, or Z86.79, Personal history of other diseases of the circulatory system (e.g., history of subarachnoid or other hemorrhage, presumably without sequelae).

Providers are not aware that for patients with a previous stroke, they should be choosing either an I69 code for sequelae of stroke or Z86.73/Z86.79. If they have the urge to click on a code for an acute stroke, they should be reaching for the phone to dial 911 to transport the patient to the emergency department. If the patient is not having an acute stroke, the provider should not use an I60-I63 code (maybe they need a sign in their office: Acute stroke code, call 911!).

• Acute heart attack: This one is trickier. If a patient has an acute myocardial infarction right there in your office, again, you should call 911. However, there is a rule that after a patient has an acute myocardial infarction (MI), it continues to be coded as acute (I21.-) for four weeks as long as it meets the definition of “other diagnosis.” Isn’t a provider likely to evaluate and assess or do medication management regarding a recent MI upon follow-up within that first month? After four weeks, if the patient is no longer receiving active treatment for the myocardial infarction, it gets relegated into I25.2, Old myocardial infarction.
  • Acute MI ➡️ I21.- code and call 911
  • Subsequent (Type 1) MI within 4 weeks of first one➡️ I22.- code and call 911 (you would also code the I21 code)
  • Follow up within 4 weeks of an acute MI and monitoring, evaluating, assessing, or treating➡️ I21.- code and treat appropriately
  • Complications following STEMI/NSTEMI within 28 days➡️ Complication code from I23 or other (e.g., E24.1, Dressler’s syndrome)
  • Follow up after 4 weeks has elapsed since a previous acute MI and no longer providing active treatment for index MI ➡️ I25.2
• Embolism:
  • Is there still a clot there (pulmonary embolism or deep vein thrombosis)? – provider needs to document and code acute or chronic PE/DVT. Providers often forget to specify chronic conditions; they don’t realize there may be a different code.
  • Has the clot resolved, been resorbed, been removed? Whether the patient is taking anticoagulation to prevent a future clot or not, this would now be a Z86.7- personal history code.
  • Is the patient still taking anticoagulation? There should be a Z79 code for long term (current) use of anticoagulants. This is not included in the HCC list.

The rest of the conditions had similar objections by the auditors. In the service year, was there any evidence that the conditions were being treated?

• Vascular claudication: auditor consideration: was there a prior diagnosis of this condition within the preceding 2 years and was the patient prescribed medication to ameliorate the symptoms or condition? I would add that if the provider documented a history consistent with claudication currently (e.g., Patient states that when they walk short distances, they experience severe cramping in their legs which causes them to have to stop and rest until it resolves. Will refer to vascular medicine.), it would be acceptable to make a clinical diagnosis. Otherwise, it may require some medication management or monitoring to validate the diagnosis.
  
• Major depressive disorder (MDD): October 1, 2021, we got a new code for the documentation of “depression,” F32.A, Depression, unspecified. Until then, it was coded as F32.9, Major depressive disorder, single episode, unspecified, which often did not accurately represent the clinical scenario. However, neither of these two codes is included in HCC 59, Major Depressive, Bipolar, and Paranoid Disorders. In order to accrue risk adjustment factor, a single episode of major depression must have severity specificity or be noted to be in remission. MDD specified as being recurrent does not need to be as specific; the unspecified version, F33.9, is included in the HCC.
  Again, in order to be considered clinically valid, the auditors have an expectation that a patient will be on medication. This is not necessarily accurate – some patients opt for therapy, and if the MDD is in remission, the patient may no longer be medicated. Is the depression actively being addressed?

• Lung, breast, colon, and prostate cancer: The criterion that the OIG used for determining if cancer was a clinically valid diagnosis during the service year was whether there had been surgery, radiation, or chemotherapy administered within a 6-month period before or after the diagnosis. I would include immunotherapy or other cancer-specific treatment as well. Additionally, if a patient declined therapy but still had a tumor, that would also count as having cancer as opposed to a history of cancer. Finally, prostate cancer (and possibly other cancers) can be managed with active surveillance or “watchful waiting.” This would also render it a valid diagnosis, but best practice would be to explicitly document it that way.
  The ICD-10-CM coding guidelines advise using a history of cancer, Z85 code, if “a primary malignancy has been previously excised or eradicated from its site and there is no further treatment directed to that site and there is no evidence of any existing primary malignancy at that site.” Although the OIG uses 6 months, there is no hard and fast rule for length of time cancer-free mandated in the coding guidelines, and clinician discretion may be indicated.

Providers must think in ink to support the diagnoses they make. They also must be instructed to select appropriate codes if they are required to choose ICD-10-CM codes for themselves. They must be taught the difference between chronic conditions and historical ones.

It would probably behoove systems to have coders review diagnosis codes to ensure that the coding is compliant prior to submitting the claim. It seems like it is a ton of work to fight an OIG investigation and beaucoup bucks to give back if the provider wasn’t coding correctly, even if there was no malevolent intent.

Programming note: Listen to Dr. Erica Remer Tuesday mornings as she cohosts Talk Ten Tuesdays with Chuck Buck, 10 Eastern.