Pharmacy Benefit Claim Pricing-Error Detection With Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Pricing errors in pharmacy benefit management systems result in overcharging patients, and the process to identify and rectify impacted claims is slow and tedious, especially for large-scale adjustments.
Innovation Solution
A machine learning system that includes a data store, front end, database service, and modeling processor to generate predictive models that identify and adjust claims impacted by pricing errors, using a composite multi-algorithm approach and active re-training to ensure accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional manual processes are used to identify and rectify pricing errors, then accuracy in identifying impacted claims can be maintained through careful review, but the time required to complete the process becomes extremely long (weeks or months)
Solution Approach 1:
The patent replaces manual mechanical review processes with machine learning algorithms and automated computing systems. The system uses trained models to automatically identify impacted claims, substituting human analysts with computational processes that can process data much faster while maintaining or improving accuracy through systematic pattern recognition.
Solution Approach 2:
The patent introduces machine learning models as intermediaries between the pricing error detection and the actual claim adjustment process. These models serve as a bridge that automatically processes large volumes of claims data, identifies patterns indicating pricing errors, and generates lists of impacted claims for verification, thereby accelerating the overall process.
2Reliability
If manual review processes are used to investigate pricing errors, then thorough investigation can be conducted, but the complexity and tediousness of the process increases significantly
Solution Approach 1:
The patent divides the complex investigation process into distinct modular components: data retrieval module, feature extraction module, machine learning inference module, and verification module. Each module handles a specific aspect of the investigation, making the overall complex process manageable and systematically executable with reduced manual intervention at each stage.
Solution Approach 2:
The system enables self-service investigation where the machine learning models automatically perform data analysis, pattern recognition, and impact identification without requiring continuous human guidance. The models self-train on historical data and autonomously investigate pricing errors, reducing the burden on investigators while maintaining thoroughness.
3Productivity
If static predictive models are used, then model complexity is reduced and processing is faster, but the accuracy deteriorates over time as pricing patterns change
Solution Approach 1:
The patent implements dynamic machine learning models that continuously adapt to changing pricing patterns. The system periodically retrain s models on new data, allowing them to evolve with changing pricing strategies and patterns. This dynamic approach maintains high prediction accuracy over time while preserving the computational efficiency needed for rapid processing.
Solution Approach 2:
The system incorporates feedback loops where prediction results and actual pricing outcomes are fed back into the model training process. This feedback mechanism allows the models to learn from their predictions and improve accuracy over time, while the established feedback channels maintain efficient processing through automated model updates rather than continuous manual retraining.
Data Source
AI summary
A machine learning process for use with a pharmacy benefits management system. The machine learning process identifies a first predicted set of drug benefit claims impacted by a pricing error, reprices a sample of the first predicted set of drug benefit claims to adjust for the error, and trains a predictive model as a function of the repriced sample. Based on the trained model, the machine learning process predicts a second predicted set of drug benefit claims impacted by the error and initiates automatic repricing.