50% Reduction in Time and Cost by Using Machine Learning to Pay the Right Provider
A national healthcare payer that manages a data warehouse consisting of 50 million claims - and growing- was challenged to correctly pay claims due to missing provider data. In theory, providers should be identified by their unique National Provider ID (NPI), but in practice, missing NPIs commonly challenge a payer’s claims adjudication process and its ability to pay the right providers.
The client used a large team of claims processors to manually review claims yet was unable to correctly pay providers who did not have an NPI in their system. The current manual corrective process was inefficient and inaccurate resulting in additional costs associated with incorrectly paying providers. To improve payment integrity, the client needed an automated solution that would pay the right provider, allowing staff to focus on high value activities.
xScion leveraged advanced data science tools and techniques to develop machine learning algorithms and predictive models that identified the correct providers more accurately than human claims processors could. Machine learning and predictive analysis were utilized to learn from human claims processors and existing claims data to accurately predict the correct recipient of claims payments.
Machine learning predictions were evaluated against manual reviews, which could then be bypassed when the automated predictions surpassed the accuracy of the human reviewers. An ongoing feedback loop was created to continually update and improve the predictive models.
xScion’s solution leveraged machine learning and automation to reduce the need for manual review and enable the client to devote human resources to higher value activities such as claims audits. A 98% accuracy rate in resolving missing NPIs was achieved, reducing time spent on remediation and associated costs by 50%.