Accurate entries are a must for emergency department coders. Manual processes not only slow down coding but also leave a lot of room for error which can lead to compliance issues and high costs. The customer wanted an automated error prediction solution for their emergency department to accurately report and predict errors and their resulting costs.

An in-depth analysis of the customer’s existing processes revealed certain challenges.

• The ED company had a strict SLA of continuously meeting the 95% quality threshold across all its sites
• The customer was able to perform an audit of only 50-70% of the medical charts.
• The customer lacked systems that could balance quality with large scale audits
• The customer needed a process to filter out audit samples that could meet the required mathematical rigor while staying compliant.
• The ED company needed to prioritize claim samples according to expected dollar loss

Apexon leveraged our competencies in AI to come up with a tailored solution for this client which could:

• Use deep learning models to predict the likelihood of error at the claim level
• Analyze 100% of the claims to filter out the most probable error cases along with the likely reasons for the errors
• Generate stratified random samples using the RAT-STATS algorithm
• Build a predictive model to generate the expected CPT codes and determine the expected loss due to errors
• Take into account claim as well as code attributes