If the proof is in the pudding, grab a spoon

We talk to a lot of carriers about how our mortality risk models support underwriters in making smarter, faster decisions: how they help identify no-touch applicants and bubble up applicants who need a closer look—fast. 

The next question we get is “Can you prove it?” The short answer is “Yes.” Because our data scientists work with one of the largest underwriting data assets in the industry, which has processed 1.5M records over 20 years of applicant data and mortality outcomes, proving it is not a problem. 

The Pudding

LifeScore Med360℠ produced a best class (ultra-preferred) with:

• 9% fewer deaths at 15 years, and
• 10.5% improvement in mortality experience

The Proof

1. We used LifeScore Med360 to score 850,000 applications received by MassMutual from 2000-2016.

2. We created a synthetic book of business to mirror the historical book such that the number of placements by risk class was equal for each year/cohort combination.
3. We compared actual to expected mortality for both books of business. 
   
   Findings: LifeScore Med360 produced preferred risk classes with fewer deaths compared to underwriters’ historical decisions. Standard and sub-standard/decline pools saw higher mortality rates, indicating that the model appropriately placed higher-risk applicants into higher-priced risk pools.

LifeScore Med360 outperformed traditional underwriting, despite having fewer data sources available. Underwriters screen for financial suitability, review prescription drug history, check MIB records, etc. for a more complete view of an applicant, yet LifeScore Med360 still outperformed traditional underwriting in our historical analysis without this additional information. As a result, we feel that these estimates of performance improvement are conservative.

Furthermore, we simulated underwriter decisions using LifeScore Med360 and compared historical results with the actual underwriting decisions. Those comparisons and subsequent actuarial reviews demonstrated the same as above:

LifeScore Med360 produced a best class (ultra-preferred) with 9% fewer deaths at 15 years.

Overall mortality experience improved by 10.5%. 

Cynics are always welcome to contact us. Watch the demo here. 

*Non-tobacco placement comparison. The mortality rate was normalized by the underwriter ultra-preferred A/E at 100. 

CRN202310-284455