Data processing: financial – business practice – management – or co – Automated electrical financial or business practice or... – Health care management
Reexamination Certificate
1996-08-14
2001-06-26
Millin, Vincent (Department: 2164)
Data processing: financial, business practice, management, or co
Automated electrical financial or business practice or...
Health care management
C705S001100, C705S003000, C705S007380, C705S014270, C705S014270
Reexamination Certificate
active
06253186
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to a computer method and apparatus for analyzing the billing patterns of providers and suppliers of goods and services and, more particularly, to a computer method and apparatus for differentiating between the billing patterns of fraudulent and non-fraudulent providers and suppliers.
BACKGROUND OF THE INVENTION
In many industries, large numbers of providers and suppliers of services and goods are employed. For example, in the health care industry, a large number of providers of medical services and suppliers of medical and related goods are involved in the delivery of health care to a given population. Certain entities, such as insurors, government payors and others must often process and pay large numbers of claims submitted by these providers and suppliers in relatively short time periods. The existence of, and the potential for, abuse of existing administrative systems by fraudulent providers or suppliers is a problem which exists in the health care setting, and in other analogous settings as well.
An objective of the present invention is to provide an automated system for processing a large number of claims submitted to a payor to identify patterns in the claim data which may be indicative of a fraudulent provider or supplier. Another objective of the invention is to provide such a system which is capable of processing claim data and identifying a potentially fraudulent provider or supplier before payments of suspect claims are made.
A preferred embodiment of the invention utilizes two popular classes of artificial intelligence: neural networks and expert systems. A neural network is a computer program which attempts to model, albeit crudely, the workings of the human brain. Neural networks are trained by example and generally excel at tasks involving pattern recognition. Expert systems are rules-based systems that deduct a solution or answer to a problem based on a series of “if . . . then” statements. An expert system attempts to mimic the deductive reasoning that an expert would employ in solving a given problem.
SUMMARY OF THE INVENTION
The present invention achieves the above-stated and other objectives by providing a computerized method and apparatus for detecting potentially fraudulent suppliers or providers of goods or services. An embodiment of the subject method comprises the steps of: a) collecting data on a plurality of suppliers and providers, including data relating to claims submitted for payment by the suppliers and providers; b) processing the data to produce a fraud indicator for at least one of the suppliers and providers; and c) determining, using the fraud indicator, whether the selected supplier or provider is a potentially fraudulent supplier or provider.
The step of collecting data further includes the steps of: periodically accessing claim data submitted by a selected supplier or provider; determining whether the claim data is of a type which can be processed in the processing step; and extracting selected claim data and storing the data in a claim file. The step of extracting selected claim data includes the step of omitting data relating to claims that would in any event not be paid. The step of determining whether the claim data is of a type which can be processed may include the step of examining an HCPCS (Health Care Procedure Code System) code associated with the claim data. This step may also include the step of determining whether a neural network used in the processing step has been trained to process the claim data.
The step of processing the data preferably includes the steps of: selecting elements of information from the data stored in the claim file; encoding the selected elements of information to produce an encoded claim file; and storing the encoded claim file. The encoded claim file is preferably sorted by supplier or provider code to produce a sorted, encoded claim file. The processing step further comprises the steps of reading data from the sorted encoded claim file, and analyzing this data by means of a neural network to produce the fraud indicator for the selected supplier or provider. In a preferred embodiment, the analyzing step includes producing a plurality of fraud indicators based on a plurality of claims submitted by the selected supplier or provider, and computing a composite fraud indicator from the plurality of indicators. In at least one embodiment of the invention, the composite fraud indicator is computed by averaging a plurality of fraud indicators for the selected provider or supplier.
The determining step preferably includes the step of comparing the composite fraud indicator to a predetermined threshold indicator value. Alternatively, one or more of the individual fraud indicators may be compared to predetermined threshold indicator value(s) as part of the determining process.
The individual or composite fraud indicators may be stored in at least one of a database file and a statistics file. An optional report may be produced to document the results of the collecting, processing, and determining steps.
In one embodiment of the invention, the processing step includes the step of analyzing the data by means of a neural network to produce the fraud indicator for the selected supplier or provider. The determining step may also include at least one of the following additional steps: a)performing an analysis of previously stored statistical information relating to the subject supplier or provider; b) performing a neural network analysis of the subject supplier or provider physical characteristics; and c) performing an analysis of statistical utilization data relating to the subject supplier or provider. An expert system may be used in one or more of these analyses. The method preferably includes the additional step of periodically refining a set of rules associated with the expert system in response to actual data relating to fraudulent suppliers or providers detected by the computerized method. The subject method may also include the step of periodically updating the neural network used to perform the analysis of supplier or provider physical characteristics, as well as the neural network used to produce the fraud indicator for the selected supplier or provider.
One embodiment of an apparatus for detecting potentially fraudulent suppliers or providers of goods or services in accordance with the invention comprises a processor, a storage device, input means for communicating data from an input device (such as, a keyboard or mouse) to the processor and storage device, and output means for communicating data from the processor and storage device to an output device (such as, a screen or printer) . The storage device is preferably configured to contain at least a claims data file for storing information relating to a plurality of claims submitted for payment by a selected supplier or provider, encoding lookup tables for use with the claims data file to produce an encoded claims data file, and a neural network program, and means for processing the data in the encoded claims data file to produce an indicator of potentially fraudulent activity by the selected supplier or provider. The apparatus may further comprise means for comparing the indicator produced by the neural network program to a predetermined threshold value. The subject apparatus may also include means for sorting the encoded claims data file by supplier or provider prior to analysis by the neural network program.
In a preferred embodiment, the storage device is further configured to include a neural network data base file for storing data relating to potentially fraudulent suppliers or providers. The storage device may also include a statistics file for storing statistical data relating to the selected supplier or provider, and a program for producing a statistical screening file from data contained in the neural network data base file and the statistics file.
The subject storage device may further be configured to contain a supplier/provider application file for storing data relating to the ph
Barnes & Thornburg
Blue Cross Blue Shield of South Carolina
Kanof Pedro R.
Millin Vincent
LandOfFree
Method and apparatus for detecting fraud does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for detecting fraud, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for detecting fraud will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2535517