Risk Adjustment Takes the Stage at RISE Nashville

Key Observations and How Prognos Can Help

By Frank Jackson


I attended the RISE Nashville conference again this year, which is dedicated to risk-adjustment, clinical quality, and care management.  Although attendees were curious about the national healthcare issues and the future of the ACA, they were mostly focused on the “street-level” practical way to improve risk-adjustment and clinical quality measures.  There was a diverse set of content and workshops, but here are five key observations from the conference:

    1. It was clear in various keynote presentations and specific risk-adjustment presentations that it was imperative for health plans to identify and receive the correct amount of reimbursement so they are able to allocate the right levels of investments in care programs.  If health plans weren’t allocating the right amount of resources into managing disease states dictated by the health risk, they would not be successful in the new healthcare world. An overarching sense of purpose across payers was evident.  One presenter stated “If you don’t receive the right risk-adjustment payments corresponding to your member risk, payers can’t invest into member programs necessary to remain viable.”
    2. The programs to support risk-adjustment, clinical quality, and care management are being viewed in a more integrated way.  Increasingly, one function such as risk-adjustment is responsible for the data sources that feed quality and care management.  Payers are increasingly investing in expand the data supply, improve data quality and integrating data programs across all of the clinical disciplines.
    3. Payers need more clinical data sources – this was not a revelation but as risk-adjustment and quality have become more important to the viability of the business, claims data just isn’t enough. There were many presentations that continued to emphasize the problem with claims – e.g. missing diagnosis and lack of clinical granularity.  The attendees were reminded in one presentation that  providers bill by CPT code and are only required to complete one diagnosis, which paint an incomplete patient picture.
    4. The RISE conference include many topics on the chart chasing processes.  Many vendors had booths that addressed various parts of this industry challenge.  As I talked with payers, they consistently complained about the inefficiency of this process, the large cost, and the difficulty finding the data they needed to be successful.
    5. Over the last several years at the RISE conference, I have seen increasing use of advanced analytics such predictive analytics, but the use of these techniques is still limited.  For example, although artificial intelligence has become an important technology in other parts of healthcare, it was virtually non-existent in the RISE conference and the solving of the Risk-adjustment and clinical quality problem.

Payer Challenges:  A Street-Level View

I talked with many payers at large plans and small plans across all business types.  I talked with clinical leaders, front-line analytic professionals, IT, and vendors.  I was impressed by payers’ thoughtfulness and motivation to improve their risk-adjustment and quality processes.  These payers ranged from star ratings of 2.5 to 4.5.  Here are some further observations from the street-level:

  • Many clinical leaders are highly motivated to improve their data supply.  I heard from one health plan last year that they estimate that 30% of health conditions are NOT identified by claims.  This leaves significant work to identify and find clues to these suspects. 
  • Many clinical leaders felt that their programs were solid but they were leaving opportunities on the table because suspects weren’t identified or documentation was poor. Payers are seeking better and more timely clinical data sources. 
  • There was a consensus from payers that clinical lab results were a good clinical source to identify risk, gain clinical specificity, and understand their members; however, their programs under-utilized lab data. Many payers cited the difficulty in getting labs to cooperate and send this data.  Another reason was the poor quality of this data, and payers did not have the in-depth lab data expertise to accomplish this task.  Lastly, payers admitted that they have not committed resources to clinically interpreting lab results to make them actionable and embed them into their analytic solutions. 
  • The amount of vendors have increased in this space over the last 5 years.  Payers were generally overwhelmed by the number of vendors.  Payers are looking for vendors who solve their business issues and provide more action insight regarding risk and clinical quality.


Introducing Lab Analytics

At Prognos, we specialize in lab analytics and integrating these additional insights into solving risk-adjustment, quality improvement and care management so I asked health plans many lab specific questions.  

First, health plans were not comprehensively and effectively using lab data.  They generally estimated around 30%-50% member coverage with their existing lab connectivity. They emphasized the difficulty of getting lab data from labs.

Second, they universally complained about the quality of the lab data.  Of the lab data that they were receiving, the data was inconsistent, incomplete and difficult to standardize.

Third, when asked about the about of clinical resources devoted to interpreting the myriad of lab tests and results, they admitted to other priorities for these high-demand resources.

In summary, although health plans agreed on the potential of the clinically rich lab results, they were grossly under-utilizing this data source because of the challenges.  If you are a health plan, here are five self-assessment questions that you can ask yourself:

  1. Are you receiving clinical lab results for over 80% of your members? 
  2. Are you successfully standardizing, normalizing, augmenting, and harmonizing this data at a necessary level for analytics? 
  3. Are these processes scalable by using machine learning, natural language processors, etc.? 
  4. Do you have clinical expertise to translate lab results to health conditions?  Have you mapped the health conditions to the relevant HCC codes? 
  5. Have you (or your vendors) integrated lab results into risk-adjustment and stars processes? E.g. connected lab analytics to prioritize and pinpoint you chart chasing activities?


Lab Analytics is a Goldmine

At Prognos, we believe that lab analytics is a clinically rich goldmine.  Approximately 70% of medical decisions are based on lab results. Clinical lab results are the most timely insights in the healthcare environment.  It is not “after the fact” and is generally available with very low latency.  Although the broader lab dataset can have data quality issues, the actual lab test results is of one of the highest quality in the industry. The aggregate health profile told by lab results can be extremely insightful.  

If clinical lab results are a goldmine, why are they so under-utilized in the payer market?  The reason is partially answered with the challenges outlined above.  In short, it’s “hard.” Until the recent few years with increasing payer challenges, payers did not see the risk-reward.  We are seeing a perfect storm.  The payer business imperatives around medical costs, identifying health risk, optimizing risk-adjustment in both the Exchange and MA markets, and higher expectations in clinical quality score have changed the calculus.  Payers now are waking to the fact that they need additional clinical data sources and they need to innovate.  Lab analytics using advanced analytics and AI techniques have advanced.  Although the problem to harmonize and interpret the lab data is difficult, Prognos has solved the problem and reduced the cost and improved analytics while the business ROI has dramatically increased.



Clinical lab analytics have a strong ROI for payers, but payers have a heavy workload with no spare resources.  My recommendation is perform a quick assessment before committing to a plan.  These simple questions will help…

  1. What is my lab member coverage across the business?  By using lab claims, for what percent of members am I receiving clinical lab data? 
  2. What is the data quality of the lab data that I receive?  Talk with IT staff and assess the quality and usability of this data.  What percent of data is not used? 
  3. Have I committed clinical resources to properly interpreting the clinical lab results? 
  4. Have my analytic teams integrated actionable lab insights into their processes such risk-adjustment, care management, clinical quality improvement?

If you would like assistance to assess your capabilities or discuss ways to take a leap forward and leverage this data goldmine then please contact fjackson@prognos.ai.

Up Next

Terms of Use | Privacy | © 2018 Prognos. All rights reserved