Lei de Benford e regras de associação no Power BI: ferramentas estatísticas aplicadas à auditoria

Abstract

The life expectancy of the population has been increasing and, consequently, the health- care system ends up undergoing major transformations. The healthcare business involves

a lot of money in both the public and private sectors, hence the demand for good hospital management gains importance. In Brazil, a country that is always among the worst in the world ranking of corruption, it is not easy to run a business. It is in this context that audits gain strength. In order to maximize profits, the auditor’s main function is to guarantee the accuracy of a corporation’s accounting and equity information. As a result of the audit process, frauds are often detected. There are several techniques used by auditors to detect fraud. In this work, we highlight Benford’s Law. This technique was propose by Benford (1938) and uses the frequencies on which the digits appear to guarantee the authenticity of the information. Benford concluded that there is a pattern in the frequencies with which digits appear, when they arise from random events. The second method used in this work to assist in auditing comes from the field of data mining: rules of association. This procedure searches for behaviors within the database. The joint application of these techniques was made in two real databases referring to the billing and paid bills of a private hospital in the city of João Pessoa. During the analysis, we found some anomalies, however, upon closer inspection, there was no evidence of fraud. All the techniques were implemented in the Power BI and R project software. The Business Intelligence platform proved to be a promising tool for this type of analysis, providing support for decision making by auditors.