Leverage Fraud Analytics to Detect and Prevent Fraud

Digitization does not eradicate healthcare malpractice. But it helps shift the odds. Cases of fraud are many and varied, such as those perpetrated by doctors, pharmacists and care providers. For example, services that were never provided are billed to health insurance companies.

Seeing through the scammers’ tricks is highly complicated business that is generally left to professionals.
metafinanz has a better answer: concentrated professional competence as well as technical skills and know-how regarding data and the use of predictive instruments.

Benefits Delivered by Fraud Analytics

Data-driven fraud detection using multi-level analytical methods in the pharmaceutical industry:

  • Identify fraud
  • Recognize unknown fraud patterns
  • Automated processing
  • Optimized processing
  • Reduce costs
  • Gain the time to focus on customer interactions

Project Insights

Our success is our customers’ success.

Fraud Detection in the Healthcare Industry

Detecting fraud by combining multiple methods

More information

The project objective was to develop a model for the automated detection of suspicious invoices submitted by healthcare service providers.

Four different methods were combined for this purpose:

1. Business rules

2. Forecast model

3. Detection of outliers (anomaly detection)

4. Network analysis

Thanks to combining various methods, it was possible to detect fraud cases by identifying recurring conspicuous billing patterns of individual service providers.

Automated evaluations on the patient level, hospital level and diagnostics level performed across several hospitals were specifically applied to chronic diseases.
In addition to predictive models and network analysis in R, text mining techniques were applied to diagnostic descriptions to uncover potential fraud patterns. The methods applied included latent semantic indexing, latent dirichlet allocation and correlated topic modeling.
Using the R programming language, we also benchmarked various data mining techniques for fraud prediction.

“Automatic fraud detection allowed us to cut our costs noticeably.”

Would you like more information? 

 
Your contact

Stefan Schwesig

metafinanz Informationssysteme GmbH
Leopoldstrasse 146 | 80804 München, Germany

Tel. +49 89 360531-0

Mail: stefan.schwesig@metafinanz.de