Another Office of Inspector General (OIG) audit report which targets Medicare Advantage (MA) risk adjustment (RA) diagnoses has been released. This November audit report was focused on University of Pittsburg Medical Center (UPMC) Health Plan, Inc, and “high-risk” diagnosis from 2015-2016. This focused audit may sound familiar as we have seen several audit reports from the OIG in which the accuracy of MA diagnosis codes was the target. Back in September 2021, the OIG also released a CMS recommendation, titled, “Some Medicare Advantage Companies Leveraged Chart Reviews and Health Risk assessments to Disproportionately Drive Payments” (https://oig.hhs.gov/oei/reports/OEI-03-17-00474.pdf) and in September 2020 they also released a report titled, “Incorrect Acute Stroke Diagnosis Codes Submitted by Traditional Medicare Providers Resulted in Millions of Dollars in Increased Payment to Medicare Advantage Organizations” (https://oig.hhs.gov/oas/reports/region7/71701176.asp)

The OIG audit covered UPMC Health Plan, Inc. (UPMC) identified overpayments for MA Hierarchical Condition Categories (HCCs) totaling 6.4 million dollars. The OIG used an independent medical review contractor to review the medical records to determine whether the selected diagnosis codes that UPMC submitted to CMS were supported. If the contractor identified a diagnosis code that should have been submitted to CMS instead of the selected diagnosis code, we included the financial impact of the resulting HCC (if any) in our calculation of overpayments. In the audit report, the OIG organized diagnoses into ten (10) groups and listed the following diagnoses that they considered to be “high-risk” for noncompliance and MA HCC overpayment:

• Acute Stroke: An enrollee received one acute stroke diagnosis (which maps to the HCC for Ischemic or Unspecified Stroke) on one physician claim during the service year but did not have that diagnosis on a corresponding inpatient hospital claim. A diagnosis of history of stroke (which does not map to an HCC) typically should have been used.
• Acute Heart Attack: An enrollee received one diagnosis that mapped to either the HCC for Acute Myocardial Infarction or to the HCC for Unstable Angina and Other Acute Ischemic Heart Disease (Acute Heart Attack HCCs) on only one physician claim but did not have that diagnosis on a corresponding inpatient hospital claim (either within 60 days before or 60 days after the physician’s claim). A diagnosis for a less severe manifestation of a disease in the related-disease group typically should have been used.
• Acute Stroke and Acute Heart Attack Combination: An enrollee met the conditions of both the acute stroke and acute heart attack high-risk groups in the same year.
• Major Depressive Disorder: An enrollee received a major depressive disorder diagnosis (which maps to the HCC for Major Depressive, Bipolar, and Paranoid Disorders) during the service year but did not have an antidepressant medication dispensed on his or her behalf. In these instances, the major depressive disorder diagnoses may not be supported in the medical records.
• Embolism: An enrollee received one diagnosis that mapped to either the HCC for Vascular Disease or to the HCC for Vascular Disease With Complications (Embolism HCCs) but did not have an anticoagulant medication dispensed on his or her behalf. An anti-coagulant medication is typically used to treat an embolism. A diagnosis of history of embolism (an indication that the provider is evaluating a prior acute embolism diagnosis, which does not map to an HCC) typically should have been used.
• Vascular Claudication: An enrollee did not receive a diagnosis related to vascular claudication (which maps to the HCC for Vascular Disease) for 2 years and then, in the subsequent year, received that diagnosis but had medication dispensed on his or her behalf that is frequently dispensed for a diagnosis of neurogenic claudication. In these instances, the vascular claudication diagnoses may not be supported in the medical records.
• Lung Cancer: An enrollee received a lung cancer diagnosis, which maps to one of the Lung Cancer HCCs, but did not have surgical therapy, radiation treatments, or chemotherapy drug treatments administered within a 6-month period either before or after the diagnosis. In these instances, a diagnosis of history of lung cancer (which does not map to an HCC) typically should have been used.
• Breast Cancer: An enrollee received a breast cancer diagnosis, which maps to one of the Breast Cancer HCCs, but did not have surgical therapy, radiation treatments, or chemotherapy drug treatments administered within a 6-month period before or after the diagnosis. A diagnosis of history of breast cancer (which does not map to an HCC) typically should have been used.
• Colon Cancer: An enrollee received a colon cancer diagnosis, which maps to one of the Colon Cancer HCCs, but did not have surgical therapy, radiation treatments, or chemotherapy drug treatments administered within a 6-month period before or after the diagnosis. A diagnosis of history of colon cancer (which does not map to an HCC) typically should have been used.
• Potentially Mis-keyed Diagnosis Codes: An enrollee received multiple diagnoses for a condition but received only one—potentially mis-keyed—diagnosis for an unrelated condition (which mapped to a possibly unvalidated HCC). For example, ICD-9 diagnosis code 250.00 (which maps to the HCC for Diabetes Without Complication) could be transposed as diagnosis code 205.00 (which maps to the HCC for Metastatic Cancer and Acute Leukemia and in this example would be unvalidated). Using an analytical tool that we developed, we identified 832 scenarios in which diagnosis codes mis-keyed because of data transposition or other data entry errors could have resulted in the assignment of an unvalidated HCC.

A couple key take-aways from the above diagnoses findings is that there is definite hospital inpatient and physician office encounter correlation being done for MA beneficiaries. Compliance risk appears with the MA physician encounter diagnosis of the “acute” condition and this is not on the hospital inpatient encounter, which is where one would expect to see the “acute” condition coded. In addition, the three (3) cancer diagnoses (lung, breast, and colon cancer) describe a length-of-time of 6-months before and after the diagnosis in which the OIG reviewed for evidence of therapy, treatment, or drug administration for a cancer diagnosis. This could be a significant direction for education, auditing, as well as documentation and coding process to determine if the timeframe has been 6-month for the treatment, therapy and/or drug administration.

Certainly, this is another wake-up call regarding the documentation and coding of ICD-10-CM diagnoses in Medicare Advantage encounters. Organizations, facilities, and physician practices need to be proactive with regard to on-going education and auditing of the MA encounters to ensure compliance and also identify gaps and discrepancies so they can be corrected. Talk about this OIG report with your HIM Coding staff, Compliance and Revenue Cycle leadership. Develop a plan to look into your own MA risk adjustment diagnoses, using this list of 10 diagnosis groups and initiate an audit to get you started.

This OIG audit report can be viewed at: Medicare Advantage Compliance Audit of Specific Diagnosis Codes That UPMC Health Plan, Inc. (Contract H3907) Submitted to CMS, A-07-19-01188 (hhs.gov)