Pre-payment emphasis on transaction-level controls: the assumption here is that illegitimate claims can be distinguished from legitimate ones by their information content, and can therefore be picked out of the claims processing stream using automated rule-based exception criteria. Thus automated “edits and audits” often test for compliance with applicable regulations and medical orthodoxy, and check the procedures provided against the limits of the relevant insurance coverage. Post-payment emphasis on provider profiling: provider profiling is the most familiar and pervasive form of analysis used within the industry. It serves to pick out providers whose service patterns differ markedly from their peers. Usually such systems operate by testing billing profiles against a range of preset parameters, or by comparative ranking of providers, within a peer group, according to a range of variables. Sole Reliance on "Thinking Applications" : some more recent offerings emphasize the application of Artificial Intelligence, with the beguiling marketing slogan that these systems “think for themselves”….and hence, perhaps, the client organization or its analysts don’t need to think so hard about fraud and abuse anymore.

TRAP Systems maintains that any or all of these three assumptions do not, and can never, provide a satisfactory basis for effective fraud and abuse control. In particular, the TRAP Systems’ approach:

• Recognizes the value of up-front edits and audits in guaranteeing billing correctness, but believes these controls do not much help with the control of criminal fraud. All but the most foolish of fraud perpetrators learn to bill correctly. The problem is that they lie: about the services provided, or the diagnoses, or they fabricate entire medical episodes, often complete with supporting documentation. Most fraud perpetrators have learned to “bill their lies correctly.” TRAP observes, through its experience within the industry, that most sophisticated fraud schemes are devised by perpetrators who assume the existence of transaction-level filters, and who therefore design their fraud schemes so that each transaction comfortably fits a legitimate profile and passes through unchallenged.

• Recognizes the value of provider profiling as a form of post-payment analysis, but recognizes a much broader range of analytic methods as equally relevant. TRAP offers a wide suite of analytic approaches (including provider profiling) and pays an unusual degree of attention to the detection of organized scams and higher levels of fraud. TRAP assumes, based on its experience and that of its clients, that provider-profiles (if carefully constructed by fraud perpetrators) will be made to look normal, and that the real detection opportunities in organized crime involve spotting unnatural coordination or coincidences across multiple provider and patient accounts. Such coordination, if spotted, suggests that multiple provider and patient accounts have come under common, and possibly illegitimate, control.

• Recognizes the usefulness of neural networks and other AI approaches, but regards them as one class of tools in a much broader toolkit. TRAP Systems rejects the notion that such systems should replace analysts or substitute for analytic thinking. TRAP’s approach is to provide analysts and investigators with the broadest range of modern analytic tools; i.e., to equip human analysts and human analysis, rather than displace them.