In the life insurance industry, the pressure is on to accelerate underwriting for as many applicants as possible. And by “accelerate,” we mean to assess risk without requiring the applicant to undergo additional labs. (Vocabulary matters.)
According to a recent GenRe study,1 68% of life insurers have a fully implemented accelerated underwriting program, 10% have a partially implemented one and 13% plan to implement one in the near future. That means more than 90% of the industry accelerates or plans to accelerate some portion of its applicants.
Why is accelerated underwriting such a strategic priority? Three reasons:
As a wholly owned subsidiary of MassMutual, LifeScore Labs works closely with their underwriters and data scientists to align our strategic priorities. Like the rest of the industry, MassMutual is leaving no stone unturned when it comes to reducing the number of applicants who are kicked out of accelerated underwriting.
In conducting a thorough evaluation of applicants who had been kicked out of the accelerated process, MassMutual determined the presence of psychiatric conditions was one of the top reasons why.
LifeScore PRISM-P helps reduce kickouts due to psychiatric conditions, and our latest testing proves it.
We recently conducted a simulation that applied LifeScore PRISM-P to the MassMutual book of business.
The goal was for the algorithm to reach the same conclusion as the underwriter when psychiatric conditions were relevant and “ratable” — or worthy of underwriting debits.
We brought together our top data scientists, product specialists and underwriters to tackle the problem.
First, we examined a universe of 18,756 total applications that included the required data:
Of those applicants, 3,194 of them, or 17%, triggered psychiatric rules and were kicked out of an accelerated program for manual underwriter review.
We applied LifeScore PRISM-P to those 3,194 applicants who triggered kickouts. The result: 845 applicants — or 26% of total psychiatric kickouts — could be returned to an accelerated process without the need for additional medical evidence or underwriter review.
That’s 845 applicants whom an underwriter didn’t have to touch.
Not every carrier has the dataset required to train a model of this specificity.
The LifeScore PRISM models, including LifeScore PRISM-P, are a series of sophisticated machine-learning predictive models that were built, trained and tested on over 300,000 applicants. In these cases, underwriters manually evaluated traditional underwriting evidence, including a medical exam and/or attending physician statements and the following data:
In more than 80,000 (27%) of the cases, applicants had significant psych-related conditions that resulted in a declination or change in their underwriting rating.
The flip side of that analysis meant that 73% had potential psychiatric conditions that did NOT affect their rating, as many psychiatric conditions are well managed and don’t significantly affect mortality risk.
Trained on this data and the outcomes, the model can identify the presence and severity of psychiatric conditions.
We’re going to preempt any legal warnings that, of course, results vary depending on your threshold for risk, your data and your underwriting ecosystem. That’s why we feel strongly that conducting a simulation or pilot of this nature is crucial for any client we work with.
For more information, visit the product page for the LifeScore PRISM family, shoot us an email or schedule some time.