Claim Adjudication Analytics for a Large Medical Billing Services Company

Industry: Healthcare

Introduction

Client is a leading provider of technology-enabled, data-driven solutions ensuring accurate payment integrity across the healthcare and property and casualty industries. Healthcare payer organizations and third-party administrators examine each claim minutely at itemized bill level to adjust the billed claim amount. Itemized bill received from different hospitals has their own wording of item description and hence it’s hard to categorize them into different charge types. It makes adjudication process manual, labour intensive, costly and inefficient. Client approached Abzooba to get an Automated solution to this unique problem.

Project journey

Abzooba, leader in providing data science and machine learning based solution, used Natural Language Understanding and Advanced Machine Learning to capture domain lingo and came out with the claim adjudication automation process.

Objective:

To automate the adjudication process by classifying each item description into a charge type accurately in-order to apply business rules of adjudication.

Approach:

  • Abzooba used natural language processing and advanced machine learning algorithm to understand the context of the item description and prepared a continuous learning system to classify each item into a charge type (e.g. laboratory changes, monitoring charges etc.).
  • Each item is classified with a confidence score and re-evaluated from domain experts. All discrepancies and low confidence items are fed back into the application to do the incremental learning. This application gave accuracy of more than 90% with minimal supervision.
  • The workflow consists of a classification engine which performed
    • Pre-processing, i.e., standardize the item description (spell correction and abbreviation expansion)
    • Concept identification i.e. domain similarity of the item with different charge type.
    • Advanced machine learning to classify item into one of charge class with a confidence score.
  • Classified item is then passed to a decision node which on basis of threshold will send it either to business rule engine or for a review from domain expert. Rule engine with the help of this classification and contract guidelines will decide adjudication.

Flow Diagram:

Business benefits:

  • Increased the Assessment speed by more than 40%
  • Provided Higher Throughput
  • Reduced Manual Intervention
  • Brought down Cost.