Description: The SOA, with SCOR as the researcher, completed an experience study of post-level term (PLT) lapse and mortality. In the first part of this session, the presenters will cover the key findings of the traditional analysis regarding the shock lapse, the PLT lapse and mortality experience. The remainder of the session will focus on the results of the predictive modeling, which will provide further insights into policyholder behavior in the post-level term period.

Description: 2021 Life Meeting - Session 8B: Proxy Discrimination in Life Insurance 9.1.21 Recording

Description: Data science, predictive analytics and AI are certainly the buzzwords these days. They are heavily used and developed in underwriting and marketing but for most actuaries, they might/may as well be science fiction. The presenters will demonstrate how a normal actuary can utilize data science techniques in their daily job function without forming an official data science team within a company. It can be done in a gradual fashion so that normal business is not disrupted and improvements are made over time. The usage of data science techniques will include demonstrations of the following areas: - Regression analysis - Economic scenario generator - Experience study - Data quality checks - Inforce compression for explicit hedging

Description: Regression to the mean (RTTM) is inflating your health care intervention evaluations if you are not carefully managing it's omnipresent and significant impact RTTM is the statistical phenomenon which arises when comparing sequential data points when an initial observation is extreme or an outlier, and states that with no other intervention, future observations will be closer to the average. Members who had high costs last year will, on average, have lower costs this year, even in the absence of any intervention or clinical program, simply by 'regressing' to the mean. This means that if your clinical programs are simply targeting members who had adverse medical experience last year, you’re targeting the wrong people. And if your savings methodologies are simply last year’s costs minus this year’s costs, the savings are inflated. By attending this session, you will be able to identify situations where RTTM will be a significant factor, how to eliminate it's impact when performing retrospective evaluations, how to identify prospects for clinical intervention with RTTM in mind, and more.

Description: Renal disease is a very heterogenous medical condition whose end state is dialysis, a recurring cost intensive medical condition. Dialysis cost is one of the major medical cost drivers in USA. Insufficient and improper care management in the early stages of CKD (chronic kidney disease) leads to early incidence of end stage renal disease (ESRD) among CKD patients which not only leads to high medical costs, but also reflects poor health quality.

Description: Following her first 100 days as the CMS Administrator, Chiquita Brooks-LaSure released her six strategic pillars for the agency. The Center for Medicaid and CHIP Services (CMCS) followed these pillars and laid out their priorities for states under the Biden-Harris administration. With these priorities established, recent 1115 submissions from several large states, including California, Massachusetts and New Jersey, may give insight to how CMS will operationalize their vision under this program authority. This session will include presentation from state policy staff who have led recent 1115 waiver proposals.

Description: "Explore a general introduction to the opportunity of applying a relatively newer area in machine learning called uplift modeling, which has been successfully used in retail (Wayfair etc). The concepts are parallel to the intervention programs in health care, where we are interested in measuring the treatment effects and ultimately directing the right members to the most effective programs, driving the highest value for the programs both from a financial and health outcomes perspective. Uplift modeling can also be viewed in context of causal modeling where we are interested in assigning cause to a treatment relative to a member not receiving the treatment. Check out an introduction on the concept, seehow it relates and builds on existing work, and use a demo to showcase the potential of integrating this ML approach to current work."

Description: "As digitization and automation rapidly progress in today’s life and health industry, issues with misrepresentation during the insurance application process are also becoming a growing concern for insurers around the globe. A recent survey finds that misrepresentation in application forms costs the average insurance customer an estimated 5% to 10% in higher premiums. In order to reduce this risk, insurers must find and correct areas of misrepresentation, consult post-issue sampling, review the automated underwriting processes, and monitor the quality of financial advisory firms. Learn how predictive disclosure modeling detects and reduces misrepresentation by harnessing readily available underwriting data to analyze issues, helping insurers understand various drivers for disclosure rates and areas of misrepresentation.Advantages of this modeling to various stakeholders in the life insurance journey include improved risk selection, claims experience, and reinsurance terms for insurers. For customers, it leads to increased certainty of coverage, payout of the claim, and ultimately lower premiums on average. By the end of the session, attendees will understand: The disclosure modeling process at a high level The technical details of a standard approach to disclosure modeling: data cleaning, standardization, visualization, and predictive modeling Success stories about how predictive disclosure modeling has impacted insurers"

Description: "A crucial importance in forecasting is to be able to distinguish cases when temporary slowdowns in health care utilization are followed by a “rebound” period of higher utilization. Come explore an analysis of a new mathematical model describing the return-of-care process. By means of simulation and case studies using the R programming language, the relationship between the visibility of these effects and the endogenous variance of the utilization time series is explored. After attending this session, you will be able to: Describe the parameters and mathematical form of the model Apply the model to real-world utilization data"