Using Artificial Intelligence to Combat Billing Fraud in the Care Sector

Fraud in the billing of outpatient care services has become a criminal phenomenon with increasing case numbers in recent years. This causes great damage to the social insurance system and results in enormous costs of several billion euros annually for the community of solidarity. Together with the Public Prosecutor's Office in Dresden and the Commissariat for Economic Crime of the Leipzig Police Department, researchers of Fraunhofer Institute for Industrial Mathematics ITWM are taking action against this. They are developing AI software against billing fraud in the »PflegeForensik« project. The work is funded by the German Federal Ministry of Education and Research (BMBF).

Up to now, it has been very time-consuming to accurately check the accounts of nursing services and contract physicians; detecting fraud has involved a great deal of complex, manual paperwork. At the same time, the special situation in nursing care (patients with dementia, many »small« services) makes it difficult to prove a complaint about individual services. In addition, the greater the need for care, the more difficult it is for patients to check the accounts themselves and to report care services that have not been provided or have been provided incorrectly. In addition, the lack of transparency in the billing system makes it susceptible to manipulation. There are many aspects and levels that present challenges and hurdles.

Machine Learning Method Supports Smart Fraud Detection

The joint project »PflegeForensik« (Care Forensics) with the subtitle »Effective prosecution of care fraud by automated image processing« is facing these challenges and is funded by the BMBF within the framework of the program »Research for Civil Security«. Researchers of Fraunhofer ITWM support law enforcement with modern algorithms of Artificial Intelligence (AI) in the field of image as well as text recognition.

The core objective of the project is to develop algorithms for the automatic reading and intelligent evaluation of the mountains of paper. Until now, every nursing service has had its own paper documents, which are structured differently and usually have little digital content. Some of them are completed by hand, some are tables, others are continuous text. This makes automated checking very complex.

AI Sleuth for Billing in Nursing Care

Machine Learning (ML) methods are used in the research project, especially in the digital capture of the various document types. A combination of image processing methods and modern deep learning methods is used. Various algorithms learn from a mixture of artificial and anonymized, real data to first recognize crucial information and then to detect anomalies. In order to train these AI algorithms, a database is filled by the ITWM team and the Leipzig police department. This means that in an initial training step several hundred documents must have been created by humans and marked with properties in order to make the algorithm intelligent in the first place and to enable case-wise analysis with a much more limited reannotation effort. The algorithms are programmed and repeatedly tested and improved with data from real investigations. The analysis of the documents then forms the basis for the evaluation and detection of conspicuous features.