Completeness and Accuracy in Risk Adjustment

Completeness and Accuracy in Risk Adjustment Robin Lloyd

In mid-December, the Health and Human Services Office of the Inspector General (HHS-OIG) announced that an investigation into Medicare Advantage organizations (MAO) had found $6.7 billion in potentially unsupported risk adjustment payments, generated from chart review programs that overwhelmingly yielded additional diagnoses. The study further found that 99% of chart reviews resulted in added diagnoses—not deletions—and that $2.7B of the $6.7B in risk adjustment payments involved diagnoses for which no related patient encounter was referenced. These findings have heightened concerns that MSOs’ reliance on chart reviews is resulting in inappropriate reimbursements and reverberating through the Medical Advantage and Risk Adjustment solution markets. So, what should those of us building and supporting value-based care programs make of these findings?

 

Some background: In May 2019, the first Encounter Data Processing System (EDPS) data was made available for 2016 dates of service. CMS commissioned the HHS-OIG study to determine whether chart reviews (the retrospective review of patient records by MAOs, with or without an associated clinical encounter), were being used to generate inappropriate risk adjustment payments. The study set out to quantify a. the total impact of chart reviews on risk-adjustment payments to MAOs; and b. what portion of risk adjustment payments to MAOs from chart reviews were not substantiated by a linked patient encounter. The findings have done little to assuage worries about how MSOs are utilizing chart reviews, though much more analysis will be required to assess what, if any, changes should be made.

 

Concern over risk adjustment practices is not new. There has already been litigation regarding the demands on MSOs to substantiate the diagnosis codes and Risk Adjustment Factor (RAF) scores used to calculate reimbursements for Medicare Advantage beneficiaries. That litigation and numerous other studies have found that MAOs are, in the aggregate, likely not being overpaid for their MA populations. That said, it is essential that CMS, MAOs, and vendors internalize the study’s findings and work collaboratively toward the most transparent, compliant, efficient value-based care models possible. And further research into the practical outcomes of risk adjustment practices is critical for continued effective, compliant operation.

 

This need for completeness and accuracy lies at the core of any effective risk adjustment solution and surrounding program. At Health Fidelity, our Lumanent solution suite and professional services support for risk adjustment are designed to ensure both complete diagnosis capture and coding compliance.

 

First, addressing the initial concern of the investigation on chart-linking, all Lumanent modules, in payer or provider settings, include a linked chart to review. While there is significant customization available, the supporting evidence is always there. The Lumanent modules for provider-driven HCC review take this one step further, identifying potential conditions at the encounter level, working in conjunction with existing solutions and processes for condition capture/recapture. Presenting and assessing suspected condition evidence (for both additions and deletions) within pre-claim workflows dramatically reduces the prevalence of risk adjustments without associated clinical encounters.

 

Addressing the additional concern that chart reviews overwhelmingly result in conditions being added: two-way review (addition and deletion) is fully enabled and encouraged within the Lumanent solution suite. Ideally, MAOs would conduct a comprehensive “two-way” review (looking for both additions and deletions) across all members, but the coding resources needed to do so are often prohibitively limited. By selecting, prioritizing, and presenting only charts with a suspected, evidence-supported opportunity for correction, Natural Language Processing (NLP) can mitigate this resource challenge. The sharp improvement in coding efficiency delivered by the NLP technology within Lumanent Post-Encounter and Retrospective Review tools lowers the barrier to expanded two-way reviews.

 

Effectively applying technology within a robust compliance framework can, in fact, deliver both optimal revenue capture and reduced audit risk. One of our largest provider clients using Lumanent Post-Encounter Review for rigorous, two-way coding review, achieved a $1.4m upside from 22k MSSP population (a net +0.8% RAF improvement) despite 40% of the coding actions being deletions. Contrast these results with the Healthcare Dive finding that the $6.7 billion in chart review-driven adjustments was offset by only $200 million in deletions. Clearly there is scope to achieve a more balanced outcome in risk adjustment.

 

A final way in which effectively utilizing risk adjustment technology can address the concerns raised by the HHS-OIG study is the facilitation of outreach to patients whose conditions warrant a clinical intervention. The study’s structure did not yield insight into why some of the risk-adjusted members did not have an identified clinical encounter during the review period. It is safe to assume, however, than some portion of these members should have been seen for evaluation and treatment, if only they had been identified, contacted, and scheduled more proactively. One key benefit of an NLP-powered pre-encounter workflow is the generation of prioritized lists of patients whose conditions warrant care. Our Lumanent Insights and Lumanent Pre-Encounter products, for example, enable providers to allocate resources efficiently and offer care for high-need patients previously slipping through the cracks.

 

The ultimate success of value-based care initiatives in the US will depend, in part, on rigorously maintained trust between providers, payers, regulatory agencies, and the public. Studies like this one by the HHS-OIG, while perhaps inconclusive, can help guide us all toward a more sustainable, trusted approach to aligned incentives and fair, effective payment models. Whether by supporting organizations’ financial goals or ensuring completeness and accuracy across risk adjustment programs, technology partners like Health Fidelity owe it to ourselves and the industry to deliver thoughtful, forward-looking solutions and approaches that, above all, aid the greater good.