Written by xScion
on June 26, 2018


Regulatory changes, new business models and consumerism are change forces that collectively are transforming the healthcare industry. Payers and providers are responding to these changes by leveraging data and information assets in ways unlike ever before. Payers are using many types and sources of data to proactively improve Payment Integrity and zero in on specific business challenges, such as claims denials due to missing data, to automatically increase claims first pass rate.

Designing and implementing advanced techniques in data science (e.g. Machine Learning, Modeling and Simulations) can help maximize the value of data assets. Predictive models that address Payment Integrity issues by preempting erroneous payments provide an opportunity to correct issues in the data before claims are processed.


Benefits of Advanced Analytics & Machine Learning for Payers

Payers can achieve several benefits from the application of these advanced techniques and solutions, including:

  1. Reduction in lost revenues associated with “Pay and Chase”
  2. Reduction in costs and time associated with manually reviewing claims for erroneous payments
  3. Improve first pass rate by preemptively resolving claims data issues.

The Cost of Doing Nothing

Improper payments contribute more than $200 billion to the annual cost of healthcare in the United States. Payers estimate that up to 30% of revenue is lost each year and only a fraction of these dollars are ever recovered using the “pay and chase” method. In addition to lost revenue through improper payments, this method offers diminishing financial returns due to the operational cost of trying to recover payments once a claim has been processed and paid.

The good news is that data and techniques exist today that can improve payment integrity by preventing these losses in the first place, making it a financial imperative for payers to proactively take steps to contain these costs and increase revenue. xScion has developed a proven approach that leverages existing, traditional analytics and then supplements these with predictive and preemptive models, and machine learning to improve the bottom line.


Case Study: Moving Beyond Descriptive Analytics to Predictive, Preemptive Analytics and Machine Learning

For a healthcare client, xScion built a descriptive model, referencing three months of historical claims data, to identify actual erroneous claims payments. To supplement this information, a predictive model was then built to identify significant trends, or “predictive factors”, in the data and to discover how these trends correlated to improper payments. This model accurately “predicted” erroneous payments with an accuracy of 82%. Each month, historical claims data is fed into the predictive model, enabling a continuous “learning“ cycle that is resulting in a consistent increase in the accuracy rate month over month. This is the value of machine learning.

xScion then developed a preemptive model by leveraging the predictive factors in the data to identify the attributes of certain types of claims at high risk of being paid erroneously. Then, by flagging claims preemptively (before they enter the claims adjudication process), the client:

  • Achieved the ability to prevent erroneous payments before they occurred
  • Developed an opportunity to review and adjust claims data before a claim is processed
  • Developed automated edit resolution solutions to improve claims first pass rate from 1-2% to 4%, double the original goal, for certain challenges such as missing provider data on a claim.

How is your organization maximizing data assets and transforming claims to achieve increased revenue, reduced costs associated with manual, erroneous claims payment recovery, AND improve first pass resolution rates? xScion can help you discover how to take advantage of these advanced techniques and machine learning to improve your business and the bottom line.