Stop me if this sounds familiar. Your company offers a high-deductible plan with an HSA, a traditional PPO, a telehealth carve-out for mental health, and a voluntary critical illness policy. Employees love the choices. Finance loves the cost control. But beneath the surface, something is quietly breaking the system.
Think of your risk pool as a single reservoir. Every employee contributes premiums, and the collective fund covers claims from the few who need expensive care. That’s the actuarial magic that makes employer health plans work. But when you add too many separate buckets-HSAs that draw healthier workers, voluntary benefits that attract those with chronic conditions, and carve-outs that siphon off specific types of claims-you’re no longer managing one reservoir. You’re managing several disconnected ponds, each with its own leakage and its own risk profile.
This fragmentation is almost invisible. No single vendor report shows it. Your TPA sees only medical claims. The voluntary carrier sees its own enrollments. The PBM sees pharmacy data. But as the employer holding all the risk, you’re left with a blind spot that can quietly drive up costs year after year.
How the cracks form
HSAs and the healthy exodus
When you offer an HDHP alongside a traditional PPO, you’re essentially inviting your healthiest employees to leave the main pool. They get lower premiums and tax-free savings in their HSA. Meanwhile, employees with ongoing health needs stay in the PPO, because the high deductible feels impossible. Over time, the PPO pool grows sicker. Your overall costs may stay flat for a while, but the composition of who pays for what has shifted dramatically. And this shift is invisible to traditional underwriting models.
Consider this: if you seed every HSA with the same dollar amount, you’re effectively giving a tax-free bonus to the healthy, while doing nothing to offset the higher costs of the sicker PPO group. That’s not just a cost issue-it can raise nondiscrimination concerns under IRS rules, though few employers ever audit this properly.
Voluntary benefits as a reverse screen
Voluntary benefits like hospital indemnity or critical illness insurance are marketed as low-cost protection. But they act as a reverse selection tool. Employees with a family history of cancer or a chronic condition are far more likely to sign up. Healthy employees often skip them. So the voluntary carrier ends up with a pool of higher-risk individuals-and they price accordingly. Your primary health plan, however, still covers those same employees’ major medical claims, but receives none of the voluntary premium dollars. The result: your primary risk pool is left with the costs, while the voluntary carrier profits on a selectively healthier sub-set.
Carve-outs that create silos
Telehealth, mental health, and pharmacy carve-outs promise better care and lower costs. And they can deliver. But they also fragment your data. When employees use a carved-out mental health provider, those claims disappear from your primary plan’s risk pool. That seems good-until you realize the carve-out vendor now has a concentrated group of high-need behavioral health users. If their utilization management is weak, you pay twice: once for the carve-out fees, and again for residual high-cost claims that spill back into the primary plan.
The bigger issue: no single system tracks this cross-subsidization. Your stop-loss carrier re-underwrites based on incomplete data, and you may see a sudden premium spike that no one can explain.
Why your underwriter misses it
Stop-loss carriers and TPAs typically analyze claims at the aggregate level. They look for large claims, trend factors, and demographic adjustments. What they don’t see:
- HSA funding disparities between plan choices
- Voluntary benefit election rates that correlate with chronic conditions
- Telehealth-only participants who rarely use the in-person network but file large specialty drug claims through a carve-out
This creates a systemic blind spot. Your employer may look like a stable, diversified risk to the carrier, but internally the sickest employees are concentrating in the richest benefit tier, while the healthiest drift toward low-cost options. Eventually, the main pool tilts, and renewal rates jump.
What you can do about it
The fix isn’t to eliminate choice. It’s to bring intelligence to your benefits ecosystem. Here are four practical steps:
- Build a unified data platform. Require all vendors-TPA, PBM, specialty networks, voluntary carriers-to contribute anonymized risk metrics into a single warehouse. Modern benefits platforms can support this. The goal: track how many employees in each sub-pool have chronic conditions, high Rx usage, or high predicted costs.
- Demand sub-pool visibility from your stop-loss carrier. Ask them to factor in HSA elections, carve-out utilization, and voluntary enrollment when setting rates. Employers with balanced sub-pools should get better terms.
- Design benefits to prevent self-selection. Use default HDHPs with automatic HSA contributions for all tiers. Offer layered voluntary benefits that supplement base coverage rather than standing alone. Integrate telehealth into your primary network so virtual visits remain within the same risk pool.
- Advocate for regulatory clarity. The DOL, Treasury, and HHS should issue guidance on whether standalone carve-outs create non-qualified risk pools under ERISA. Until then, treat your entire benefit ecosystem as one plan for risk aggregation.
The bottom line
Your health plan risk pool was never a single bucket. But the fragmentation we see today is different-it’s invisible, systemically embedded, and growing. The employers who will manage costs best over the next decade are those who invest in benefits systems integration-not just for administration, but for true risk intelligence.
Stop treating risk pooling as a static actuarial assumption. See it as a dynamic data and design challenge. Your employees, your stop-loss carrier, and your CFO will all be better off.
Have you seen this fragmentation in your own benefits ecosystem? I’d love to hear about it-I’m gathering real-world examples for a follow-up piece.