Red Flag: OIG chart reviews of Medicare Advantage records raise concerns. Billions misspent?

Feb. 14, 2020 / By Barbara Aubry, RN

Audits of Hierarchical Condition Categories (HCCs) have been performed for years as a way for Part C vendors (Medicare Advantage Organizations) to identify and support patients with complex needs. Higher acuity patients translate to increased per member per month (PMPM) reimbursement from CMS.

HCCs are values comprised of specific ICD-10 diagnosis codes. The model is based on chronic and cumulative conditions represented by HCC codes. The HCC rationale supports the idea that diagnoses are clinically related to conditions with similar cost of care. CMS obtains the diagnosis codes from claims and creates a risk adjustment factor for reimbursement.

The idea of risk adjustment is central to the Medicare Advantage Organization (MAO) operating plan since payers need to be incentivized to enroll sicker beneficiaries needing more costly care. This helps ensure continued access to Part C plans.

Sounds fair – so what is the problem?

It’s a fact that not all claims are properly coded; CMS allows the MAOs to perform retrospective audits to determine whether (or not) diagnoses were present in the documentation to support assigning additional ICD-10 codes for the encounter(s). Remember: the sicker the patient, the higher their acuity and the more the MAO is paid to “manage” their care.

The MAOs hire in-house auditors or contract with audit vendors to perform retrospective chart reviews to determine if diagnoses were documented but not included on the claim. Interestingly, most of the audits performed and reported to CMS resulted in increased HCC assignment equaling higher reimbursement. Since the audit volume is high, it stands to reason there would also be a fair number of audits that result in lowering HCC assignment since CMS has specific requirements to assign an HCC. Apparently, that is not the case.

OIG HCC Audit Review Findings reported Dec 23, 2019

In recent reporting of its HCC audit review findings, the OIG stated:

The OIG undertook this study because of concerns that MAOs may use chart reviews to increase risk-adjusted payments inappropriately. Unsupported risk-adjusted payments are a major driver of improper payments in the MA program, which provided coverage to 21 million beneficiaries in 2018 at a cost of $210 billion. To be eligible for risk adjustment, a diagnosis must be documented in a medical record as a result of a face-to-face visit. Although CMS requires MAOs to identify chart reviews in the encounter data, CMS does not require MAOs to link these chart reviews to a specific service associated with the diagnoses. This may provide MAOs opportunities to circumvent CMS’s face-to-face requirement and inflate risk-adjusted payments inappropriately.

My Take

Retrospective HCC auditing should be relatively simple – right? If the diagnosis is documented, add it to the list of any HCC category where it is included. Ah, no, not so simple. CMS requirements include:

  • The(se) diagnoses must be documented annually if the member has a chronic condition, even an amputation; and
  • The diagnosis must be documented properly in progress notes; and
  • The diagnosis must result from a face-to-face visit with a medical doctor or nurse practitioner or physician assistant; and
  • Diagnoses coded last year drive current year payment (prospective model)

There are nuances in accurate HCC auditing. For example, CMS documentation standards apply which means CMS provider signature requirements must be met, etc. Basically, all the CMS documentation requirements must be met for the encounter to “count.” Another nuance: If a patient had a basal cell carcinoma successfully removed from their arm in 1997, and if they have not been treated or tested for this since, it does not support assigning “history of cancer” as a diagnosis to earn an HCC. This situation does not meet the face-to-face visit (for treatment/management) requirement.

HCC auditing has become more complex with increased use of EHRs. Some EHRs drag along a problem list from visit to visit, even when conditions are no longer being treated. My favorite is influenza; when this appears in the problem list for months (in my opinion), it means either the list is bogus, or the treatment is. For the diagnosis to “count” for risk adjustment purposes, it needs to be an ongoing problem that requires treatment or impacts the beneficiary’s functioning. An example would be history of amputation.

It’s equally disturbing from a clinical standpoint when patients have chronic diseases documented in the problem lists but there is no record of any treatment during a calendar year. This brings up a quality of care issue.

The OIG audit’s unsettling results:

  • 99 percent of their chart reviews added diagnoses that were only found on MAOs or vendor chart reviews and not on claims.
  • CMS paid $2.7 billion in risk adjustment dollars that were not linked to a specific service provided to a beneficiary – much less a face-to-face visit.
  • They also found that even though limited to a small number of patients, almost half of MAOs reviewed had received risk-adjusted payments when there was not a single record of a service being provided.

The results of the OIG audit are concerning: Are claims improperly coded? Are retrospective audits improperly supporting services that were never provided? Is CMS somehow reimbursing MAOs based on faulty risk logic? Does the risk adjustment process need a revamp? Has a quality of care issue been uncovered?

CMS concurred with the audit findings but disagreed with the OIG improper payment estimate. CMS believes it to be a billion less. OIG believes “a discrepancy between the RAPS and encounter data would raise serious questions regarding integrity of MA risk adjustment data.” CMS has yet to perform audits from calendar year 2015 claims but intends to begin later this year.In its findings, OIG recommended:

  • Targeted oversight of MAOs that had risk-adjusted payments resulting from unlinked chart reviews for patients who had no service records in the 2016 encounter data.
  • Conduct audits that validate diagnoses reported on chart reviews in the MA encounter data.
  • Reassess the risks and benefits of allowing chart reviews that are not linked to service records to be used as sources of diagnoses for risk adjustment.

What to do in the meantime? Ask audit vendors to supply proof they follow CMS HCC auditing requirements. If internal auditors are used, retrain and document the training. Be sure all audit policies and procedures are properly documented to support internal processes. Expect this to be an ongoing issue – stay tuned.

Barbara Aubry is a senior regulatory analyst with 3M Health Information Systems.


Office of Inspector General, “Billions in Estimated Medicare Advantage Payments From Chart Reviews Raise Concerns

Centers for Medicare & Medicaid Services, Risk Adjustment Data Validation (RADV) Medical Record Checklist and Guidance