Closing Care Gaps Through Prospective Risk Adjustment

Sachin Patel, Chief Executive Officer, Apixio

The lack of a complete and comprehensive patient record limits the ability of the physician to improve care and puts organizations at financial and regulatory risk. The Centers for Medicare and Medicaid Services (CMS) estimated that for payment year 2018 alone, it will recover $428.4 million (net) and $4.7 billion from 2023 through 2032, including extrapolation effects. The HHS Office of Inspector General (OIG) recently reported that, from October 2014 through December 2016, 153 audit reports were issued containing 193 overpayment recoveries totaling $648 million, largely due to errors in medical documentation.

To address these concerns and other matters, CMS announced significant regulatory changes to the Medicare Advantage (MA) program beginning in 2024. Changes to MA rates, the MA risk adjustment (RA) model, and star ratings are being implemented, with some already fully in effect and other RA model changes being rolled out over three years. Now is a critical moment for Medicare Advantage Organizations (MAOs) and risk-bearing providers to prepare. At the same time, CMS is implementing changes in how its Risk Adjustment Data Validation (RADV) audits are conducted and how their findings lead to overpayment assessments. Errors identified in already completed audits will be extrapolated more aggressively for audits going back to 2018, and then going forward for new audits.

While CMS has stated that its primary focus will be on MAOs most at risk for improper payments, all MA plans will face more stringent scrutiny, enforcement, penalties, and repayment requirements for conditions submitted for reimbursement. At the same time, the OIG will continue its broader set of audits, referring its findings to CMS for financial recovery. Additionally — and appropriately — the linkage of care delivery programs to risk condition capture is important from a policy standpoint, and for the success of Value-Based Care (VBC) programs. 

The Time Value of Data

Many of these documentation errors stem from traditionally burdensome processes and the inherent nature of the retrospective chart review process. Retrospective reviews are an important part of RA programs given patient enrollment timing, lack of prior documentation on claims, and the need for true encounter data. As a result, reviews commonly take place many months after the patient encounters in question, which creates three potential issues.

  1. Time and complexity are added to the process, as reviewers must look through months of accumulated patient records to reconcile the data.
  2. Less timely accurate reimbursement, with a lag time of 12-18 months, can impact revenue streams for providers and payers to support patient care.
  3. There is a potential impact on patient care programs across a cohort, given data coordination between payers and providers.

In the short term, organizations would be wise to begin auditing their historic claims submission data and ensuring their medical records and claims are in sync. Any errors discovered can be proactively addressed through chart audit reviews for potential deletes and adjusted reimbursement.

Going forward, organizations can implement strategies to close these gaps at the point of care to address the issues above and improve care for patient populations.

Avoid Penalties With a Prospective Approach

Taking a prospective approach to risk adjustment can solve billing inaccuracies and provide better patient insights. However, a lack of standardization and efficient solutions has prevented providers from adopting and scaling the process. Many providers are struggling with clunky spreadsheets and cumbersome processes that add to clinical team burnout.

The use of technology for concurrent reviews continues to gain uptake as the most viable solution for ensuring complete diagnosis capture and accurate documentation by reconciling HCC codes and documentation immediately after the patient encounter, before billing and claims submission.

AI-powered concurrent solutions can reduce the time it takes for providers and health plans to identify and resolve discrepancies from months to just hours, providing real-time reconciliation of diagnoses. Because this prospective approach catches inconsistencies upfront, it drastically reduces the time and expense of chart reviews. Given the timely documentation of conditions, payers and providers can coordinate better quality and depth of care for patients. For example, proactive interventions to support successful VBC delivery programs for patients with chronic conditions can be rolled out. Furthermore, these activities can help prevent high-cost-of-care scenarios.

By proactively managing and analyzing costly chronic conditions against those that are risk-adjustable, organizations can avoid misrepresenting patient risk. Payers and provider groups can leverage automation to conduct risk adjustment on a broader patient population and more accurately capture and quantify risk, even storing these data elements in a centralized manner for better long-term collaboration. Certainly, implementing prospective programs takes time, operating discipline, and analytical prowess, but it can yield improved outcomes as well as reduce downside risk from government audits.

There will always be a need and demand for retrospective review as part of a comprehensive, accurate risk adjustment program. That being said, closing care gaps through concurrent reconciliation will reduce the retrospective burden over time. But most importantly, this proactive approach can drastically improve quality of care and patient outcomes, and increase physician time with patients while lowering the overall cost of care.


About Sachin Patel

Sachin Patel is the Chief Executive Officer of Apixio, an AI platform that improves administrative, clinical, and financial outcomes for health plans and providers. Patel brings broad experience across both healthcare and technology, spanning a variety of leadership roles, including operations, finance, and development.