Health insurance and prevention: Using customer behavior study and targeting

Health insurance and prevention: Using customer behavior study and targeting

179 Views
Added in HEALTH LIFE

Thanks! Share it with your friends!

URL

You disliked this video. Thanks for the feedback!

Sorry, only registred users can create playlists.
URL

Description:

in order to maximise the effectiveness of prevention programs          

Speaker(s): Céline Blattner (ACTUARIS), Jean-Louis Rulliere (ISFA / Lyon1 University)

Using customer behavior study to improve the rate of accession to prevention programs, and targeting policyholders with machine learning methods : two essential means in order to maximise the effectiveness of prevention programs provided by health insurers.

For several years, most health insurers have been providing access to prevention programs included into their health products. On the one hand, they want to reduce health expenses by managing the risks. On the other hand, they offer prevention as a service aimed for improving the health status of policyholders, and therefore their quality of life.

During our presentation, we will discuss two topics which we believe to be crucial in order to maximise the effectiveness of prevention programs provided by health insurers :

  • Customer behavior study to improve the rate of accession - and adherence over time - to prevention programs. We will first present the different cognitive bias, and how they affect the sector of health insurance. In a second part, we will focus on how consideration of these bias affects the rate of accession to preventive health measures, and why we cannot neglect them. Finally, we will explain how to handle two particular bias of our policyholders : temporal inconsistency and base rate fallacy.
  • Machine learning methods to pinpoint homogeneous policyhoder groups, and then to provide them with specific prevention programs (according to their current or potential health risks and pathologies). The results we will present have been obtained by applying a matrix factorization algorithm, combined to data visualization and clustering techniques. By comparing two anonymized health expenses databases, we have been able to highlight similarities or disparities in regard to policyholders appetite for specific preventive health measures. We have also used more conventional algorithms in order to validate the model.  

Post your comment

Sign in or sign up to post comments.
Be the first to comment