Merit Medicine says AI underwriting model could cut stop-loss loss ratios

Merit Medicine  released a white paper showing that its AI-based underwriting platform, Merit Predict, could have reduced medical loss ratios and improved profitability for a national stop-loss carrier.

The retrospective study, validated by actuarial firm Axene Health Partners, analyzed 19 employer groups representing 16,823 members. The model identified four of the six groups that ultimately produced significant losses, flagging them as the highest risk before coverage was bound.

According to the analysis, excluding or repricing these high-risk groups would have reduced the portfolio loss ratio from 83.6% to 55% and increased underwriting margins from 16.4% to 45%. The highest-risk groups, representing 21% of the total, accounted for 52% of underwriting losses, including one group that generated a $2.3 million loss.

Merit said the findings highlight the concentration of losses among a small number of high-volatility groups and the potential for predictive risk scoring to improve pricing discipline and capital efficiency. The platform evaluates risk using group- and member-level data, including clinical indicators, high-cost drug exposure, and projected medical and pharmacy spend.

Merit, founded in 2022, operates in the same segment as Verikai, which is now inactive, and Gradient AI, and is backed by LiveOak Venture Partners.