Data mining application to healthcare fraud detection: a two-step unsupervised clustering method for outlier detection with administrative databases

Abstract

Background: The healthcare sector is an interesting target for fraudsters. The availability of a great amount of data makes it possible to tackle this issue with the adoption of data mining techniques, making the auditing process more efficient and effective. This research has the objective of developing a novel data mining model devoted to fraud detection among hospitals using Hospital Discharge Charts (HDC) in Administrative Databases. In particular, it is focused on the DRG upcoding practice, i.e., the tendency of registering codes for provided services and inpatients health status so to make the hospitalization fall within a more remunerative DRG class. Methods: We propose a two-step algorithm: the first step entails kmeans clustering of providers to identify locally consistent and locally similar groups of hospitals, according to their characteristics and behavior treating a specific disease, in order to spot outliers within this groups of peers. An initial grid search for the best number of features to be selected (through Principal Feature Analysis) and the best number of local groups makes the algorithm extremely flexible. In the second step, we propose a human-decision support system that helps auditors cross-validating the identified outliers, analyzing them w.r.t. fraud-related variables, and the complexity of patients' casemix they treated. The proposed algorithm was tested on a database relative to HDC collected by Regione Lombardia (Italy) in a time period of three years (2013-2015), focusing on the treatment of Heart Failure. Results: The model identified 6 clusters of hospitals and 10 outliers among the 183 units. Out of those providers, we report the in depth the application of Step Two on three Hospitals (two private and one public). Cross-validating with the patients' population and the hospitals' characteristics, the public hospital seemed justified in its outlierness, while the two private providers were deemed interesting for a further investigation by auditors.

Fourteen years of manifestations and factors of health insurance fraud, 2006–2020: a scoping review

“…(a) (Francis, 2020 ) Upcoding Intentionally coding a health claim based on an inaccurate use of codes to obtain greater economic value. (b) (Gasquoine & Jordan, 2009 ; Massi et al, 2020 ; Palutturi et al, 2019 ; Phillipsen et al, 2008 ; Sheffali & Deepa, 2019 ) Unperformed or billing for services not provided Known as phantom billing, claims are presented for medical services, medications, medical devices not delivered to the patient. (c) (Aral et al, 2012 ; Bauder & Khoshgoftaar, 2020 ; Bayerstadler et al, 2016 ; Brooks et al, 2012 ; Dolan & Farmer, 2016 ; Gasquoine & Jordan, 2009 ; Jou & Hebenton, 2007 ; Lee et al, 2016 ; Li et al, 2008 ; Palutturi et al, 2019 ; Perez & Wing, 2019 ; Phillipsen et al, 2008 ; Smith et al, 2013 ; Yang, 2003 ) Misrepresenting the diagnosis or procedure to justify payment Manipulation of procedures, diagnoses, requests, complaints, dates, frequency, duration or description of the services provided.…”

“…We identified 26 studies that explained 15 factors, the most referenced factors are audit, supervision and control, with 8 studies, while 6 studies explain the general characteristics of the provider. The factors supported by applied studies that have shown influences in favour of the HIF occurring are the general characteristics of the provider (Herland et al, 2018 ; Kang et al, 2010 ), in favour of the HIF decrease (Vian, 2020 ), ambivalent (Massi et al, 2020 ); the management and policy of complaints show results that contribute to HIF (Vian, 2020 ) and ambivalent results (Lesch & Baker, 2013 ); while that the reputation shows ambivalent influence (Tseng & Kang, 2015 ), the audits, supervision and control contributes to reducing the HIF (Kang et al, 2010 ).…”

“… (Musal, 2010 ) (Manocchia et al, 2012 ) Mesoenvironmental Factors General characteristics of the provider It includes their legal nature for profit or not, location, competitiveness index, services they provide, schedules, payment statistics, history of their production. (Bauder et al, 2017 ; Herland et al, 2019 ) (Wan & Shasky, 2012 ) (Massi et al, 2020 ) (Herland et al, 2018 ; Kang et al, 2010 ) Responsibility of the provider If the provider maintains responsible conduct in its administrative and medical actions. (Kerschbamer & Sutter, 2017 ) Measures of the administrative authority The guidelines given by health care authorities influence payment for fraud and abuse, including medical records.…”

“…iii) the plan looks like insurance, but the agent avoids calling it "insurance" and instead uses evasive terms like "benefits" (Thornton et al, 2015) v) A third-party administrator who processes claims on behalf of Medicare signed an integrity agreement with the Department of Justice in response to a number of allegations, including the fact that he made incorrect payments for claim filings (Liu & Vasarhelyi, 2013) control, with 8 studies, while 6 studies explain the general characteristics of the provider. The factors supported by applied studies that have shown influences in favour of the HIF occurring are the general characteristics of the provider (Herland et al, 2018;Kang et al, 2010), in favour of the HIF decrease (Vian, 2020), ambivalent (Massi et al, 2020); the management and policy of complaints show results that contribute to HIF (Vian, 2020) and ambivalent results (Lesch & Baker, 2013); while that the reputation shows ambivalent influence (Tseng & Kang, 2015), the audits, supervision and control contributes to reducing the HIF (Kang et al, 2010). Each Macroenvironmental and Mesoenvironmental factor was briefly described, and one explanation was provided.…”

Developing an Instrument for Assessing Self-Efficacy in Data Mining and Analysis

“… Zhao et al (2020) used data mining to study the risk factors that can predict IHD during pheochromocytoma surgery, and observed that data mining techniques are increasingly being used in clinical and medical decision-making to provide continuous support for the diagnosis, treatment, and prevention of disease. Massi et al, 2020 noted that the healthcare industry is an interesting target for fraudsters. The availability of large amounts of data makes it possible to solve this problem through the use of data mining techniques, thereby making the review process more effective.…”