Background: Healthcare fraud entails great financial and human losses; however, there is no consensus regarding its definition, nor is there an inventory of its manifestations and factors. The objective is to identify the definition, manifestations and factors that influence health insurance fraud (HIF).Methods: A scoping review on health insurance fraud published between 2006 and 2020 was conducted in ACM, EconPapers, PubMed, ScienceDirect, Scopus, Springer and WoS.Results: Sixty-seven studies were included, from which we identified 6 definitions, 22 manifestations (13 by the medical provider, 7 by the beneficiary and, 2 by the insurance company) and 47 factors (6 macroenvironmental, 15 mesoenvironmental, 20 microenvironmental, and 6 combined) associated with health insurance fraud. We recognized the elements of fraud and its dependence on the legal framework and health coverage. From this analysis, we propose the following definition: “Health insurance fraud is an act of deception or intentional misrepresentation to obtain illegal benefits concerning the coverage provided by a health insurance company”. Among the most relevant manifestations perpetuated by the provider are phantom billing, falsification of documents, and overutilization of services; the subscribers are identity fraud, misrepresentation of coverage and alteration of documents; and those perpetrated by the insurance company are false declarations of benefits and falsification of reimbursements. Of the 47 factors, 25 showed an experimental influence, including three in the macroenvironment: culture, regulations, and geography; five in the mesoenvironment: characteristics of provider, management policy, reputation, professional role and auditing; 12 in the microenvironment: sex, race, condition of insurance, language, treatments, chronic disease, future risk of disease, medications, morale, inequity, coinsurance, and the decisions of the claims-adjusters; and five combined factors: the relationships between beneficiary-provider, provider-insurance company, beneficiary-insurance company, managers and guānxi.Conclusions: The multifactorial nature of HIF and the characteristics of its manifestations depend on its definition; Identifying the influence of the factors will support subsequent attempts to combat HIF.

Health insurance fraud (HIF) places a substantial economic burden on global health systems. While supervised machine learning (SML) offers a promising solution for its detection, most approaches are ad hoc and lack a systematic methodological framework that ensures replicability, adaptability, and effectiveness, especially in contexts with severe class imbalance. We developed PDHIF (Phases for Detecting Fraud in Health Insurance), a six-phase systematic methodology that introduces a holistic focus that integrates fraud theory, actors, manifestations, and factors with the complete SML lifecycle. We applied this methodology in a case study using a dataset of 8.5 million claims from a public health insurance system in Peru. We trained and evaluated three SML models (Random Forest, XGBoost, and multilayer perceptron) in two experimental scenarios: one with the original, highly unbalanced dataset and another with a training set balanced via the K-means SMOTE technique. When PDHIF was applied, the results revealed a stark contrast: in the unbalanced scenario, the models were ineffective at detecting fraud (F1 score < 0.521) despite high accuracy (>98%). In the balanced scenario, the performance improved dramatically. The best-performing model, RF, achieved an F1 score of 0.994, a sensitivity of 0.994, and an AUC of 0.994 on the test set, demonstrating a robust ability to distinguish suspicious claims.

To describe the monitoring model of the Health Care Service Institutions (HCSI) of the National Health Authority (NHA) and assess the factors associated with risk-adjusted normative compliance (%RANC) within the Peruvian Health System (PHS). We carried out a case study of the experience of the NHA in the development and implementation of a monitoring program based on the ISO 31000-2009. With HCSI as the units of analysis, we calculated the %RANC (a scorein continuous scale ranging from 0 to 100) for comprehensive monitoring (CM) and for specific evaluations made from 2013 to 2015. A higher score in the %RANC means lower operational risk. Also, slope coefficients (β) and their 95% confidence intervals (95% CI) were estimated using generalized linear models to estimate the association between %RANC as outcome, and health subsector, region, level of care and year, as explanatory variables. The NHA made 1444 evaluations. For CM, only the Social Security Administration had higher %RANC than private centers (β=7.7%; 95% CI 3.5 to 11.9). The HCSI of the coastal region (β=-5.2, 95% CI -9.4 to -1.0), andean region (β=-12.5; 95% CI -16.7 to -8.3) and jungle region (β=-12.6, 95% CI% -17.7 to -7.6) had lower %RANC than those located in Lima Metropolitan area. %RANC was higher in 2015 than 2013 (β=10.8; 95% CI 6.4 to 15.3). The %RANC differs by health subsector, region and year of supervision. For CM, the HCSI in the Social Security Administration and in the Lima Metropolitan area had better scores, and scores improved over time. The implementation of actions aimed at improving %RANC in order to foster the full exercise of health rights in the PHS is suggested.