Mining for Strong Negative Associations in a Large Database of Customer Transactions
ICDE '98 Proceedings of the Fourteenth International Conference on Data Engineering
Application of Genetic Algorithm and k-Nearest Neighbour Method in Medical Fraud Detection
SEAL'98 Selected papers from the Second Asia-Pacific Conference on Simulated Evolution and Learning on Simulated Evolution and Learning
Discovering patterns of medical practice in large administrative health databases
Data & Knowledge Engineering
Association rule mining: models and algorithms
Association rule mining: models and algorithms
Expert Systems with Applications: An International Journal
Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium
A Decision Support Tool for Health Service Re-design
Journal of Medical Systems
Discovering inappropriate billings with local density based outlier detection method
AusDM '09 Proceedings of the Eighth Australasian Data Mining Conference - Volume 101
AusDM '11 Proceedings of the Ninth Australasian Data Mining Conference - Volume 121
Hi-index | 0.00 |
This paper presents an application of association rule mining in compliance in the context of health service management. There are approximately 500 million transactions processed by Medicare Australia each year. Among these transactions, there exist a small proportion of suspicious claims. This study applied association rule mining to examine billing patterns within a particular specialist group to detect these suspicious claims and potential fraudulent individuals. This work identified both positive and negative association rules from specialist billing records. All the rules identified were examined by a subject matter expert, a practicing clinician, to classify them into two groups, those representing compliant patterns and non-compliant patterns. The rules representing compliant patterns were then used to detect potentially fraudulent claims by examining whether claims are consistent with these rules. The individuals whose claims frequently break these rules are identified as potentially high risk. Due to the difficulty of direct assessment on high risk individuals, the relevance of this method to fraud detection is validated by analysis of the individual specialist's compliance history. The results clearly demonstrate that association rule mining is an effective method of identifying suspicious billing patterns by medical specialists and therefore is a valuable tool in fraud detection for health service management.