News & Blog
RGS introduces ML
to search for overtreatment in Clinics' Invoices
Insurance Company Rosgosstrakh, in cooperation with developer of artificial intelligence-based IT solutions for insurance MAINS Lab, has automated the verification process that the company applies to invoices from clinics
Moscow, July 09, 2020

This will allow Rosgosstrakh to reduce losses from services provided with no proper grounds under health insurance programs, the insurer's press service said.

Search for overtreatment in clinics' invoices is an important and labor intensive task for any insurance company that develops health insurance in a serious way. According to Rosgosstrakh experts, 2-5% of invoices drawn up by clinics after they are visited by insured persons are fully or partially groundless. This means that the services actually provided do not correspond to treatment procedures and rules approved by the RF Healthcare Ministry. For a major company, this may mean dozens of millions of rubles paid with no proper grounds.

Most of Russian insurers look for overtreatment manually, i.e. their competent medical experts go through invoices submitted by clinics. It is extremely difficult to ensure a high quality of detecting services that were provided with no proper grounds when volumes are huge. Given that Rosgosstrakh keeps on boosting its health insurance customer base, it was important to introduce automation and standardization to the process of reviewing registers of clinics' invoices as soon as possible.
"Good partner relations with medical providers is a prerequisite for long-standing and mutually beneficial cooperation in health insurance", says Galina Talanova, Rosgosstrakh Board Member and Head of Health Insurance. "That being said, control over submitted registers of invoices is an integral and important component of the business process, as efficiency of this review has an effect not only on saving money of the company and its clients, but eventually also on the quality of medical services provided to insured persons."
Introduction of an automated overtreatment detection service trained by a neural network is fully in line with the goals that the insurer is facing at the moment, says Tatiana Demyanenko, Director of the Medical Expert Appraisal Division within the Rosgosstrakh Health Insurance Block. "The rules of search for groundless services embedded in the MAINS Lab solution, machine learning algorithms and our own rules integrated in the system, taken together, make us sure that claim payment control will have a maximum effect at minimal cost," she stresses.

"Working with medical data is complicated, but we managed to launch a solution, integrate it into Rosgosstrakh business processes and adjust to it experts' needs under a challenging timeline," says Yury Kuvshinov, MAINS Lab CEO. "The current economic situation worldwide and in this country presents extra tough requirements to business efficiency, and this solution makes it possible to save insurance companies' costs significantly by not having to pay groundless claims."

Rosgosstrakh is a member of Otkrytie Bank Group. It maintains more than 1,500 offices and representative offices, as well as about 300 claims settlement centers and points throughout the Russian Federation, and has about 50,000 employees and insurance agents.