Wiadomości

[Webinar OpenFloor] Predictive Modelling of Medical Claim Costs Using Administrative Claims Data – 29 lipca 2020

W imieniu prelegenta, Leonela Lopesa, Zarząd Polskiego Stowarzyszenia Aktuariuszy serdecznie zaprasza swoich członków na wykład-webinar z cyklu Open Floor pt. „Predictive Modelling of Medical Claim Costs Using Administrative Claims Data″. 

Podsumowanie: Crucial insurance operations rely upon estimates of future claim amounts. Healthcare insurers, in particular, are mainly interested in predicting the financial impact of medical claims of policyholders based on their current health status. The advances of big data and statistical learning methods allowed investigations of the potential use of administrative claims data for that purpose. This is a very rich and detailed source of historical medical data that are formed by the invoices paid by the insurers to the healthcare providers. In this talk, I will discuss the application of predictive methods, such as model-trees, to the administrative claims data, show how their accuracy compares to the traditional regression models and present the main insights from the data.

O prelegencie: Leonel Lopes is a final year Ph.D. candidate at Cass Business School. He finished his Bachelors in Actuarial Science in Brazil at Universidade Federal de Minas Gerais in 2009. He has 4 years of experience in the healthcare insurance market, being responsible for providing analyses of strategic data and advising on decision-making. In 2013 he concluded his Masters in Actuarial Science at Cass Business School, in London, to where he returned in 2015 to initiate his Ph.D. His research aims to investigate alternative predictive models using administrative data for forecasting health care claims costs. In parallel to his studies, Leonel is part of the Institute and Faculty of Actuaries COVID-19 Action Taskforce, where he collaborates in different projects.

Termin i miejsce: 29 lipca (środa) w godzinach 17.00-18.00, spotkanie Zoom. Wykład odbędzie się w języku angielskim.