Case Studies

Preventing Overpayments on Medical Claims

Estimated reading time: 2 minutes

Client: Insurance benefit coordination company

Challenge: Human analysts review and assess multiple healthcare insurances for each patient, which is a time-consuming process

Solution: IDP model to predict the likelihood of multiple insurances and to prioritize items for analyst review

Challenge:

Reviewing documents for insurance claims

Many Americans are subscribed to multiple types of health insurance. Patients typically pay less out of pocket when they hold multiple insurance plans, but identifying the liable party for a specific medical expense can be complex. Our client coordinates these benefit plans so insurance carriers correctly pay claims to prevent double payments.

To process these claims effectively, human analysts manually review insurance documents to understand which individuals might have multiple insurance plans in their name. Once multiple plans are found, the analysts investigate each policy’s coverage. These policies can vary widely and require a trained eye to spot nuances in coverages.

Solution:

Building machine learning models to predict patients with multiple policies

Infinia ML identified patterns related to patients that required coordination of benefits. Using those insights, we built machine learning models to predict the likelihood that patients had multiple policies, and whether specific claims were likely to be covered by multiple insurance policies. Claims that were likely to be covered by multiple insurance policies were prioritized for human analyst review. Infinia ML used a semi-supervised approach to continuously label new pieces of data to improve outcomes.

Due to the sensitivity and changing nature of medical claims data, Infinia ML built in customized auditing features to alert our client’s stakeholders on important metrics related to their model’s performance. This information tells stakeholders when to retrain models.

Value:

Insurance claims are processed with more accuracy and speed

Improperly billed insurance claims that were previously missed are detected with more accuracy and speed, saving the mis-billed insurance company money. Infinia ML’s solution allows our client to remain efficient and only review claims requiring human oversight.

Applying these techniques elsewhere

  • Predict customer/stakeholder characteristics
  • Prioritize and isolate unique data entries

Related Case Studies