The Employee Benefit Research Institute published Issue Brief #651 on February 12, 2026, providing the first comprehensive utilization data for cell and gene therapies (CGTs) in employer-sponsored health plans. The headline number: 9.2 per 100,000 enrollees received a CGT in 2022, up from 7.9 per 100,000 in 2018, a 16.5% increase over four years. Those users represent fewer than one-tenth of one percent of covered lives but account for half a percent of total plan spending.
For self-insured plan sponsors, this is no longer a theoretical exposure. It is a quantifiable base rate.
Who It Affects
Self-insured employer health plans of all sizes carry this risk, but the math is most punishing for mid-market groups (500 to 3,000 covered lives). A single CGT claim ranging from $373,000 for CAR-T therapies to $4,250,000 for Lyfgenia (sickle cell disease) can consume 5% to 15% of a mid-sized plan’s annual expected claims. Plans with specific stop-loss attachment points in the $250,000 to $500,000 range will see every CGT claim breach the specific layer immediately. Stop-loss carriers know this: many are now excluding conditions eligible for approved CGTs, shifting the full cost back to the plan sponsor even when stop-loss is nominally in force.
With 48 FDA-approved CGTs as of 2025 and a pipeline increasingly targeting more common conditions, the eligible patient population is expanding beyond the rare-disease cohort that kept utilization low through 2022.
The Reserve Mechanism
The exposure is extreme severity outside historical experience. Among the top 1% of health spenders in employer plans, EBRI found that 0.58% used CGTs, and those users represented 1.6% of high-cost-claimant spending. Because CGTs barely existed in employer plans before 2018, historical claim triangles contain no credible development data for this claim type.
Actuaries projecting IBNR for self-funded plans must model a discrete probability of million-dollar single-claim shocks that fall outside the triangle’s historical range. At 9.2 per 100,000 enrollees, a 2,000-life plan faces roughly an 18% annual probability of at least one CGT claim. That probability is high enough to require an explicit loading factor in the expected claim ratio, not a contingency footnote.
The USC Schaeffer Center’s March 2026 policy paper reinforces this structural problem: employer plan horizons run three to five years, but gene therapy benefits span decades. The plan pays the full multi-million-dollar cost upfront for a cure the employee may carry to a future employer.
What This Means for Your Next Review
Put CGT probability modeling on the agenda for your next reserve study or stop-loss renewal. Ask your actuary whether they are modeling a discrete CGT claim probability in your IBNR estimate, and if so, what utilization rate per 100,000 enrollees they assume for plan year 2026. At 9.2 per 100,000 and trending upward, any assumption below that figure needs justification. Confirm whether your specific stop-loss attachment point has been stress-tested for a $3M to $4M single claim, and whether your aggregate corridor accounts for the scenario where two CGT claims occur in the same plan year.
Sources
- EBRI Issue Brief #651: Cell and Gene Therapies in Employment-Based Health Insurance (February 12, 2026)
- USC Schaeffer Center: High-Stakes Medicine: Rethinking Policies for the Cost and Value of Cell and Gene Therapies (March 19, 2026)
- BenefitsPro: More Self-Insured Employers Want to Hit Health Care Costs in the Gut (June 3, 2026)