December 2020- In this inaugural edition of Gibson Talks, Executive Recruiter Trey Wheless consults with Fahad Ahmed on recommendations for the most current and pertinent issues in Risk Adjustment.
1. What are some of the benefits and hurdles to transitioning from RAPS to EDS?
In 2012, CMS began collecting encounter data through the Encounter Data System (EDS) in parallel to the traditional Risk Adjustment Processing System (RAPS). CMS intends to fully retire RAPs by the submission year 2020 (payment year 2022). Until then, health care systems will continue to submit the same diagnosis data in parallel to both RAPS and EDS.
EDS is more complex and requires 150+ data elements to be submitted per encounter to CMS, whereas RAPS requires only 5 data elements. A more comprehensive dataset enables CMS to perform advanced data analytics and improve the methodology to predict future medical expenditures. However, for provider groups and MCOs, an expanded data set means that they need to invest additional capital and operating expenses to develop and maintain an infrastructure that can provide advanced data processing capabilities. In addition, the technology platforms will need to include data filtering logic to identify records that are likely or unlikely to get accepted by CMS for risk adjustment.
CMS accepts encounters for risk adjustment based on a data filtering logic, which determines the minimum required fields based on the type of submission being made (e.g., institutional vs. professional). The logic also includes validations on an acceptable combination of data elements to ensure quality. In the RAPS process, each MAO applies the filtering based on CMS guidance and determines which encounters to submit for risk adjustment. However, in the EDS process, CMS performs the data filtering universally for all MCOs and performs error checks on data elements. It would be imprudent to assume that full responsibility and accountability for accurate interpretation of the logic has been shifted to CMS. There have been many issues at CMS that adversely impacted the accurate calculation of the risk score. Therefore, MCOs need to implement systems that can validate EDS filtering rules and predict which encounters may get rejected. This will provide invaluable insights on data quality issues and opportunities to ensure accurate payment from CMS.
CMS has indicated that additional member-HCCs sourced from chart reviews can be submitted for appropriate risk adjustment. In the RAPS process, these HCCs can be submitted unlinked, meaning that the supplemental record containing this HCC does not have to link to an original claim; it can be submitted as a stand-alone record. However, when submitting this member-HCC to EDS, CMS recommends that the chart review record be linked to an original claim. Linking occurs when two records can be joined on a member, provider, and date of service. In December 2019, the Office of Inspector General (OIG) issued a report, “Billions in Estimated Medicare Advantage Payments from Chart Reviews Raise Concerns.” This report indicates that CMS will be stringent in evaluating unlinked encounters from chart reviews in the future. To mitigate the risk of compliance issues, health organizations need to be one step ahead of CMS and ensure that providers are consistently submitting claims, and data systems are not rejecting valid encounters. In addition, in-depth data analytics will likely show that provider NPI is the main reason for records not linking. For example, the claim encounter may contain the primary provider’s NPI, but the chart review records may contain the physician assistant’s NPI. This causes records to be unlinked and may raise concerns with CMS.
2. What are the best practices for risk adjustment coding? To follow-up on that, how do you manage operations and oversight over a program?
Accurate coding is mission-critical to improve the quality of care patients receive. It also enables accurate reimbursement rates so that medical groups and MCOs are appropriately compensated for the care provided. Over-coding or under-coding not only jeopardizes the mission to deliver value-based care but also increases the risk of a RADV audit. Provider and coder training is vital to mitigate compliance risk.
Develop provider education. When providers and coders understand risk adjustment rules, the quality of documentation and coding improves. For example, providers have incorrectly used the terminology “history of” in the medical note to code a chronic condition as if the condition no longer exists. Such documentation errors may lead to coding errors and inaccurate submissions to CMS. In addition, knowledge of HCC rules (e.g., HCC18 trumps HCC19, HCC17 trumps HCC18 and HCC19) and the impact of each HCC on RAF score will ensure that providers accurately code to the highest level of disease specificity. If a provider mistakenly selects a lower HCC, when a higher one is more accurate, the RAF score will be inaccurate, and reimbursement from CMS will decrease.
Start member outreach early in the year. Develop a member outreach program that assists providers in scheduling appointments, sending reminders, and ensuring that patients have the means to travel to the clinic. Identify members who are unlikely to visit their physician based on past history, and develop a targeted outreach plan. Leverage telehealth services as necessary, particularly as we continue to face COVID-19 challenges. In a recent study conducted in July 2020, CMS indicated that up to 21% of the beneficiaries reported needing health care for something other than COVID-19 but not getting it because of the pandemic (CMS, October 2020). Much of the care missed included regular checkups and treatment for ongoing conditions. This will likely impact the RAF score, unless the provider groups and MCOs can leverage telehealth as one way to deliver care until the risk associated with the pandemic have subsided. In addition, establish plans for a post-pandemic environment for these patients. Offer wrap-around services such as House Calls to connect with patients who generally do not come in for their appointments. [View Images]
Align Operations. From an operational perspective, many organizations have multiple coding teams, some at local provider groups and some at the national corporate level. It is ideal to have all the teams aligned under unified leadership, including any vendor-based coders, to enable better oversight on quality, productivity, and compliance. Aim for 100% coding accuracy by leveraging first level and second level quality reviews on both addition and deletes of diagnosis.
In December 2019, the OIG report pointed out that MAO organizations only added diagnosis 99% of the time from chart reviews (Chiedi, 2019). This sends a signal to the industry to anticipate additional oversight when submitting a diagnosis from chart reviews. Therefore, creating a unified organizational structure for all coding teams will enable an organization to implement chart audits programs efficiently and effectively.
Focus on Data Integration and Interoperability. Integrate claims, supplemental, and medical record data set to improve coder’s ability to enter accurate diagnosis and other required data elements. Access to multiple data sources in a unified manner will enable coders to validate medical records and confirm diagnosis accuracy. EMR interoperability will enable providers to leverage care management tools to see patient history of chronic conditions and treatments being provided by other physicians. This will make it easier for providers to confirm chronic conditions based on prior year history.
Establish multi-level metrics. Establish executive, management, and operational level metrics on coding quality and productivity. The idea is to inspect what is expected so that team behavior and actions are aligned to the overall strategy and mission. Ensure that the metrics on each level of reporting are integrated and enable decision making. Some of the key metrics to measure include errors per member, duplicates charts coded, and charts coded per hour by chart type (e.g., primary care provider, specialty, hospital). Leverage data analytics to manage operations and identify new opportunities.
3. What are some steps that payers can take to ensure accurate coding practices and proper reimbursements?
Improve documentation quality. Perform quality audits based on stratified random samples on different types of charts. In this audit, attempt to identify HCCs coded or submitted by the provider, but not supported in the medical records. This will help to understand the current level of risk so that a benchmark can be established. Determine the root causes of quality issues and address them accordingly. The issue may be due to a lack of provider education, inefficient front-end tools, or an inexperienced coding team.
Implement coding guidelines. Develop internal coding guidelines to align all stakeholders toward a higher quality of documentation that exceeds expectations. The guidelines should clearly specify coding requirements for medical charts, discharge summaries, progress notes, or other acceptable sources. Identify minimum data elements needed to be extracted so that downstream systems can properly assemble the record for CMS submission. To develop the guidelines, use the Official ICD-10-CM Coding Guidelines and best practices from the American Health Information Management Association (AHIMA).
Conduct end-to-end data reconciliations. Ensure that all valid and risk adjustable encounters submitted by providers and identified through chart reviews are submitted to and accepted by CMS. Leverage the MAO-002 and MAO-004 reports from CMS to identify which encounters are missing or were rejected. Prioritize the encounters based on new and unique HCC impact and remediate them prior to the CMS final sweep deadline. Because the data submission process has numerous handoffs and enrichments, there is room for errors to occur. For example, it is acceptable to code from a stand-alone discharge summary when it is part of the provider documentation. Do you submit it as a professional or institutional record? Are all the necessary data elements accessible (e.g., for risk adjustment, professional records require a procedure code, institutional records do not)? Therefore, an end-to-end reconciliation will shed light on potential submission issues in the process and reduce the risk of inaccurate reimbursements from CMS.
Develop RAF Score Dashboard. Develop a reporting dashboard on RAF score by the provider and ensure that it can be rolled for higher-level reporting. Compare year-over-year scores by chronic conditions and HCCs re-captured. Perform analytics on disease prevalence by provider group, by region, and other criteria to enable cross-sectional comparison among providers to identify best practices and lessons learned. Include other metrics such as members with a visit and members without a visit to drive annual wellness checkups in a timely manner – this will reduce the last-minute rush to complete all the visits by year-end.
4. How can payers best prepare and mitigate risks associated with RADV audits?
In December 2019, OIG published a report that sets up the potential for more RADV audits in the coming years. The study findings showed that MAOs almost always used chart reviews as a tool to add diagnosis – over 99% of the chart reviews in the audit added diagnosis code. For 2017, CMS estimated $2.7B in risk-adjusted payments on chart review diagnosis that MAOs did not link to a specific service provided to the beneficiary (Chiedi, 2019). Therefore, it is more crucial now to be prepared for RADV.
Perform provider and coder training. Train the team on risk adjustment model and coding guidelines. Establish a feedback loop to focus on continuous improvement. After training is provided on specific HCC modules, perform random audits at the 30-60-90-day mark to ensure that providers and coders have improved documentation. If quality does not improve to 100% documentation accuracy, then provide a refresher of the prior lessons.
Conduct random mock audits similar to RADV. Establish and proactively execute random internal audits as part of operations. Mimic the actual RADV process so that there is an added internal layer of control prior to RADV occurring. In addition, perform targeted audits on disease areas of higher risk for the particular provider group or MAO. Diligently prevent unsupported HCCs in the population. If an unsupported outlier HCC gets included in the 201 RADV audit sample, it can significantly skew the results and lead to a higher error rate. When this error rate is multiplied by the total population based on the CMS extrapolation method, the financial penalties can increase significantly.
Perform targeted audits on selected HCCs. CMS also provides a list of miscoded HCCs. For example, Diabetes with Chronic Conditions was listed on the top in the 2018 RADV results (Department of Health and Human Services, 2020). Make a list of top miscoded HCCs and conduct an audit on provider documentation and coding results to ensure alignment between the two groups. If any are identified during the self-administered audit, then expand the sample sizes to conduct additional audits and focus on identifying the root cause of the issue. Identify all records that are not supported and perform appropriate deletes.
Submit deletes in a timely manner. Ensure that deletes identified through the audit are submitted to CMS in a timely manner. Submit deletes to the MAO at the appropriate level and clarify whether it is a data issue delete or medical record delete. If a data transformation led to incorrect submission, it would constitute a data delete. In contrast, if a diagnosis is not supported in the medical record, that would be a medical record delete. Mismatching the delete types can lead to over or under delete, which also increases RADV risk. CMS has recently proposed limiting the “safety-net” that can result from under-coding; therefore, even under-reporting HCCs may have an adverse impact on the overall RADV score in the future.
Chiedi, J. M. (2019). Billions in Estimated Medicare Advantage Payments from Chart Reviews Raise Concerns. U.S. Department of Health and Human Services. Office of Inspector General. Retrieved Nov 16, 2020, from https://oig.hhs.gov/oei/reports/oei-03-17-00470.pdf
CMS. (October 2020). COVID-19 Experiences Among the Medicare Population. Medicare Current Beneficiary Survey. Retrieved Nov 16, 2020, from https://www.cms.gov/files/document/medicare-current-beneficiary-survey-covid-19-data-snapshot.pdf
Department of Health and Human Services. (2020). 2018 Benefit Year HHS Risk Adjustment Data Validation Results. CMS. Retrieved Nov 11, 2020, from https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs/Downloads/2018_BY_RADV_Results_Memo.pdf
About the Participants
Fahad Ahmed has ten years of immense experience in value-based care and end-to-end encounter data management. He’s built in-house data platforms that support risk adjustment data oversight and controls. He has developed numerous data reconciliation and audit systems to ensure complete and accurate submission and payment from CMS. Fahad has been recognized for ideating and conceptualizing numerous innovations to help organizations navigate complex challenges. He’s received Innovation of the Year recognition from two different fortune 500 companies for his contribution.