Optimizing Risk Adjustment and Quality Measurement With AI

As we’ve covered in the earlier entries in this series, artificial intelligence (AI) is proving useful in streamlining health plan operations of all sorts. AI tools can also augment human expertise for optimal results in less structured but critical processes. We’ll look at how this works in Medicare Advantage (MA) risk adjustments and Healthcare Effectiveness Data and Information Set (HEDIS) performance metrics.

Coding, chart reviews, and data collection and reporting for MA risk adjustment and quality measures are typically manual and labor-intensive tasks. They’re also error prone, which leaves MA plans vulnerable to issues in risk adjustment validation audits by the Centers for Medicare & Medicaid Services (CMS). Manual risk calculations may also miss care gaps, leading to quality issues.

AI tools—combined with experienced coders, data analysts, and compliance specialists—can reshape how MA plans identify and calculate risk adjustment factors (RAFs). This helps ensure appropriate reimbursement, better audit results, and, most importantly, better care and improved outcomes for members.

Here’s what AI brings to the risk adjustment process:

More comprehensive and precise risk adjustment coding. Generative AI tools with language processing capabilities can help ensure that MA plans accurately capture and support the hierarchical condition category (HCC) codes that document the severity of member conditions. AI agents can locate demographic data, medical histories, and diagnostic codes within unstructured data such as clinical notes. The AI agent can then suggest the correct codes based on that information. Similarly, generative AI tools can quickly scan other notes, prescriptions, clinical records, digitized computed tomographic scans, and other diagnostic images to identify International Classification of Diseases, Tenth Rivision (ICD-10) and HCC coding details that might otherwise be overlooked.

Targeted and expedited data retrieval. Using a retrieval-augmented generation (RAG)-enabled AI system allows MA plan coders and auditors to swiftly find supporting documentation among thousands of pages of clinical textual and digitized image data. Health plan analysts can ask plain-language questions, and the system can search and return responses with clear source attribution. RAG-enabled AI eliminate hours of manual record review by extracting only relevant clinical data and context. It can also highlight missing data, inconsistencies, and outliers, enabling MA plans to correct these going forward.

Faster care gap closure. AI predictive models can forecast which members are unlikely to follow up on screening appointment, fill prescriptions, or follow diets and other recommended care. AI copilots can offer support before and during provider visits to help ensure providers recognize these members and take appropriate steps to ensure care is completed and documented.

Individualized member engagement and support. AI agents with intelligent scheduling capabilities can trigger AI chatbots to text or call members to help them schedule mammograms, colonoscopies, lab tests, and other preventive services. AI agents can then track member follow-through, update records, and flag urgent needs for clinician follow-up.

Improved HEDIS measurements. AI tools can help MA plans collect clean, timely, and complete data to ensure HEDIS metrics are as accurate as possible. AI agents can pull structured HEDIS data from thousands of records far more quickly than human reviewers. Generative AI can generate summaries or narratives of the data from records required for HEDIS chart reviews to further expedite the process.

Enhanced provider network management. AI can generate dashboards that show ICD-10 and HCC coding accuracy and care gap closure progress among network providers. AI agents can automatically generate requests to providers, asking for clarification on any ambiguous or inconsistent data. Generative AI agents can also detect patterns and relationships within provider notes, clinical records, and discharge summaries to identify significant risk factors and develop more accurate predictive health models for the managed population. The MA plan can share these with providers so they can better detect these conditions during patient assessments.

Audit readiness and preparation. AI tools can summarize charts and prepare supporting documentation for risk adjustment validation and HEDIS audits. With AI built into the risk adjustment and quality measure processes, an MA health plan can always be audit ready. The tools can prioritize queries from MA plan auditors and coders to streamline the review process by focusing on critical issues.

AI can also conduct risk score audits beyond standard validation, comparing medical record documentation with claims and encounter data. This reduces coding discrepancies and helps calculate the financial impact of risk factors. Enhanced risk prediction can lead to more precise capitation payments to providers and better resource allocation to specific member populations.

Implementing AI tools within risk and quality measurement programs enables MA health plans to build greater accuracy, efficiency, and compliance into their day-to-day operations. These benefits should enable plans to close care gaps and support member wellness for better outcomes, potentially enhancing a plan’s Star ratings. That makes AI in risk and quality monitoring a business advantage for MA plans that adopt the technology.

About the Author

Deepan Vashi is the EVP & Head of Solutions for Health Plans and Healthcare Services at Firstsource with over 27 years of experience in health plan IT, business operations, and consulting. He is renowned for his expertise in developing member-centered digital solutions and building cross-functional teams to ensure successful implementation. In his role at Firstsource, he spearheads solutions and strategy for health plans, including Intelligent Back Office, Health Tech Services, and Platform-based Solutions (BPaaS). Deepan has extensive knowledge of innovative technologies such as Process Mining, Digital Twin, AI, and Blockchain.