The renewal letters started landing in late April. By the time you’re reading this, every mid-market self-funded employer in the country has either seen their 2026 number or knows it’s coming. The headline is the same everywhere: up 9% to 13% on medical, up 12% to 18% on specific stop-loss, and another double-digit bump on the pharmacy carve-out.
This isn’t a one-year blip. The 2024 trend was 8.5%. The 2025 trend was 9.2%. The 2026 trend is landing closer to 11%. Three consecutive years of double-digit medical cost growth — after a decade of “trend will normalize next year” — is now a structural pattern, not a forecast error.
If you’re a CFO or HR leader at a 100-to-1,000 employee self-funded company, this post is for you. We’ll do four things:
- Decompose the 11% — what’s actually inside the number.
- Walk through the four levers mid-market employers are actually pulling.
- Explain why three of those four don’t compound, and one does.
- Give you the questions to ask your broker before October when next year’s RFP cycle starts.
What the 11% actually is
Trend isn’t one thing. Carriers and consultants quote “trend” as a single percentage because that’s how the renewal letter is formatted, but inside the number are five separate pressures growing at different rates:
| Driver | Share of 2026 trend | Growth rate | What’s driving it |
|---|---|---|---|
| Specialty pharmacy (GLP-1, oncology, autoimmune) | ~30% | 18–22% | GLP-1 utilization tripled 2024→2026; biologics and oral oncology growing 12%+ per year |
| High-cost claimants (>$100K) | ~25% | 14–16% | More million-dollar claims; carriers pricing lasers tighter at renewal |
| Hospital outpatient & facility fees | ~20% | 8–10% | Site-of-service shift from inpatient to outpatient; hospital pricing power in consolidated markets |
| Primary & specialist office visits | ~10% | 6–8% | Fee schedule updates, modest utilization growth |
| Stop-loss premium | ~15% | 12–18% | Reinsurer hardening + GLP-1/specialty pharmacy retention shifts |
A few takeaways that should drive how you read your own renewal:
Specialty pharmacy and high-cost claimants are 55% of the problem. If your renewal conversation is dominated by network discounts and PCP fee schedules, your broker is solving the wrong third of the equation.
Stop-loss is harder to fix than the line item suggests. When specific stop-loss premium goes up 15%, employers tend to raise their specific deductible to compress premium — which works for one renewal cycle and then sends them straight into a claims tail they hadn’t planned for. The structural answer is fewer claims hitting the laser, not a higher laser.
Office-visit primary care is roughly 10% of the problem. That’s important — because as we’ll see, it’s the only lever in this stack that compounds.
The four levers mid-market employers are pulling
Walk through the average self-funded HR director’s 2026 playbook and you’ll see the same four moves, in roughly this order:
1. Plan design — raise deductibles, narrow the formulary, exclude GLP-1s
This is the first lever because it’s the fastest. Re-tier the formulary, raise the in-network deductible from $1,500 to $2,500, exclude GLP-1s for weight loss while keeping them for type 2 diabetes, push more services to coinsurance instead of copay.
Net effect on renewal: roughly 2–4% of trend absorbed.
What it costs: employee satisfaction, and the recruiting/retention drag of a benefits package that visibly degraded year-over-year. In tight labor markets that’s real money — just not money that shows up in the medical line.
2. Stop-loss restructuring — raise the spec, change the carrier, add a captive
The second lever because it’s the second-fastest. Move from $150K specific to $200K specific, shop three reinsurers, look at a group captive arrangement.
Net effect: 5–15% reduction in stop-loss premium, depending on claims history.
What it costs: a one-time savings that doesn’t recur. The reinsurer prices the new structure at the new claims experience. If your underlying high-cost-claimant trend is still 14% per year, you’re back to a hard renewal in 24 months.
3. Vendor proliferation — point solutions for diabetes, mental health, MSK, fertility, navigation
The third lever, and the one most likely to be sold to you by a benefits consultant earning a placement fee. Add a diabetes management vendor. Add a mental-health virtual platform. Add a musculoskeletal point solution. Add a navigation vendor on top to coordinate the other vendors.
Net effect: claimed PMPM (per member per month) savings of $30–$80 across the stack. Validated PMPM savings in the academic literature: closer to $5–$15, often statistically indistinguishable from zero.
What it costs: $15–$40 PMPM in vendor fees, plus an HR team that now manages six vendor relationships and a fragmented member experience. For 500 lives that’s $90K–$240K a year in fees against $30K–$90K in real savings.
4. Onsite or near-site primary care — restructure where care happens
The fourth lever, and the only one that targets the actual cost stack instead of the premium. Put a primary care provider in your building (or in a shared near-site space) at zero member cost-share. Same-day visits. Real chronic-disease management. Pharmacy steerage. Specialty referral redirection.
Net effect: net-of-cost reduction of 8–15% of total medical spend at mature utilization (typically year 2 onward).
What it costs: roughly $80,000–$250,000 per year for the 200–500 employee range, depending on shared vs. dedicated and days-per-week. See the math at 200, 500, and 1,000 employees for the unit economics.
Why three of these four don’t compound
Levers 1, 2, and 3 are all what economists call one-shot plays. You pull them and you save money for a year. The underlying cost engine — chronic disease incidence, specialty drug utilization, fragmented care, ER overuse — keeps growing at its own pace underneath.
- Raising the deductible saves money in year one. It does nothing about the diabetic who’s still going to develop kidney disease in year four.
- Restructuring stop-loss saves money in year one. It does nothing about the cancer claim that will hit your laser in year two.
- Adding a diabetes vendor with a $20 PMPM fee and a 4% engagement rate saves you nothing. Even if it worked perfectly, it would only address one condition.
Lever 4 — onsite primary care — is structurally different. It changes where care happens, who sees the member first, and how often they’re seen. The math compounds for three reasons:
Earlier diagnosis bends the curve. A diabetic caught at HbA1c 6.8 in year one of an onsite program never becomes the $180K nephrology claim in year five. We’ve written about this elsewhere — the diagnosis that came too late — but the principle is structural, not anecdotal: the longer a condition is undiagnosed, the more expensive every subsequent year becomes. Primary care moves the diagnosis upstream by 6–18 months on average.
Utilization compounds engagement. Year one engagement is 50–65% of eligible members. Year two is 75–85%. Year three approaches 90%. Every additional member who builds a relationship with the onsite provider is one fewer member relying on an ER, an urgent care, or a Google search. That engagement curve is the asset.
The data layer gets smarter. With three years of longitudinal clinical data on your population, the next year’s interventions are targeted rather than scattershot. You know who’s at risk for what, you know which conditions are actually driving your claims tail, and you can stop spending money on vendors that don’t move your specific population.
Where the 8–15% net savings actually comes from
For the CFO reading this, here’s the line-item breakdown of where onsite primary care recovers cost. These are the figures we model at the headcount-and-claims level — not industry averages, not “estimated” — for clients running 200–1,000 employees through our reporting layer:
| Recovery source | Year 1 | Year 3 (mature) |
|---|---|---|
| ER & urgent care deflection (30–50% of UC, 10–20% of ER) | 1.5–3% of medical | 3–5% |
| Specialty referral redirection (60–70% of presenting complaints resolved in PCP) | 1–2% | 3–4% |
| Pharmacy steerage (generic + mail-order, 8–15% chronic-med reduction) | 1–2% | 2–3% |
| Earlier chronic-disease management (HbA1c, BP, lipids caught at lower acuity) | <1% | 2–3% |
| Total net (cost of program already netted out) | 3–6% | 8–15% |
Year-one savings rarely cover the program cost. That’s not a flaw — it’s the shape of the curve. The fully-loaded ROI argument is a 3-year argument, not a 12-month argument, which is why employers who evaluate onsite primary care on year-one ROI alone almost always pass on it and almost always regret it three renewal cycles later.
What to ask your broker before October
If you’re staring at a 2026 renewal letter and trying to figure out what to do about 2027, here are the five questions worth asking before the next RFP cycle opens:
-
What share of our 2026 trend was driven by specialty pharmacy and high-cost claimants? If your broker can’t break this out by line, they’re working from the carrier’s summary, not your data. The decomposition is in the claims file — ask for it.
-
What percentage of our eligible members had a paid primary care office visit in the last 24 months? Pull CPT codes 99211–99215 and 99381–99397. The number is almost always under 45%, often under 30% for blue-collar populations. That’s the baseline an onsite program rebuilds.
-
What’s our ER-to-PCP visit ratio? Healthy populations run roughly 1:8. Populations with poor primary-care access run 1:3 or worse. The ratio is one of the cleanest leading indicators of where money is leaking.
-
What’s our 3-year stop-loss laser projection if our high-cost-claimant trend stays at 14%? This is the conversation most brokers don’t want to have because the answer is uncomfortable: the laser will move every year until something changes upstream.
-
What would a real-time clinical data layer cost us, and what would it tell us about our population that the carrier doesn’t? Strata-style longitudinal data is now table-stakes for any onsite vendor worth evaluating. If your prospective vendor’s reporting is a quarterly PDF, they’re operating on the same 60-to-120 day lag as your carrier.
If you want the same conversation with the right format already filled in, we maintain an RFP template that benefits teams can send to vendors — including the data-access language you’ll want in any contract.
The structural argument
If trend is going to be 9–11% for the foreseeable future — and the demographic + specialty-pharma drivers say it will be — then the question for every mid-market self-funded employer is not “how do we survive 2026?” It’s “what’s our structural answer for the decade?”
Plan design is a one-year drag. Stop-loss restructuring is a one-year reset. Vendor stacking is mostly theater. Onsite primary care is the only one of the four levers that gets cheaper per member every year your population engages with it, and the only one that captures clinical data you can compound against.
That’s not a marketing claim. It’s the unit economics of where the cost stack actually sits, and where it doesn’t.
FAQ
How much is the 2026 health insurance trend rate for self-funded employers?
Aon, Mercer, and Willis Towers Watson are all reporting projected 2026 medical trend in the 9–11% range for self-funded mid-market plans, with specific stop-loss premium running 12–18% higher than 2025. Specialty pharmacy (especially GLP-1s) and high-cost claimants (over $100K annual spend) together account for roughly 55% of the increase.
Why are healthcare costs going up in 2026?
Five drivers: specialty pharmacy growth (18–22% per year, led by GLP-1s and oncology biologics), more million-dollar claims hitting stop-loss, hospital outpatient pricing power in consolidated markets, modest fee-schedule growth, and stop-loss reinsurers hardening their pricing in response to the prior four. The first two together are roughly 55% of the increase.
What can a self-funded employer do about a 10%+ renewal?
Four levers, in order of how fast they save money: (1) plan design changes — raise deductibles, narrow formulary, exclude non-medical GLP-1s; (2) stop-loss restructuring — higher specific, group captives, three-bidder RFP; (3) point-solution vendors for diabetes, mental health, MSK; (4) onsite or near-site primary care. The first three are one-shot savings that don’t compound. The fourth restructures the underlying cost stack and gets cheaper per member every year of mature utilization.
Is onsite primary care realistic for a 200-employee self-funded plan?
Yes — through a shared near-site model with one to two in-person days per week. The realistic floor is roughly 150 lives. Dedicated full-time onsite still needs 1,500+ lives, but the shared model collapses the cost side of the equation and works at mid-market headcount.
How long until onsite primary care pays for itself for a self-funded plan?
Year one rarely covers full program cost — typical year-one recovery is 3–6% of medical spend against a program cost that may run 1–3% of medical spend, depending on headcount and structure. Mature utilization at year 3 typically returns 8–15% net of program cost, and the curve continues to bend for chronic-disease populations as conditions are caught and managed earlier.
What’s a “high-cost claimant” and why does it matter so much?
A high-cost claimant is any member whose annual claims exceed your stop-loss specific deductible — typically $100K–$250K depending on plan design. They’re roughly 1–3% of members but 25–35% of total medical spend. They’re also the population where primary-care relationships matter most: a managed diabetic who never becomes a dialysis patient is the difference between a $4K claim and a $250K claim across a five-year horizon.
What’s the difference between trend and renewal?
Trend is the projected per-member medical cost growth for the next plan year. Renewal is what your carrier or TPA actually quotes you, which includes trend plus adjustments for your own claims experience, demographic shifts, and stop-loss pricing. A 9% trend year can translate to a 14% renewal for a plan with a bad claims year, or a 6% renewal for a plan with a good one. Underwriters compress the difference; the underlying engine is still trend.
Is excluding GLP-1s a good idea?
Probably not as a stand-alone move. Excluding GLP-1s for weight loss while keeping them for type 2 diabetes is the most common middle path, but it generates appeals and burdens HR. The bigger lever is running the math on coverage: a thoughtful step-therapy protocol with an onsite or near-site PCP managing initiation is far more cost-effective than either blanket coverage or blanket exclusion.
Want the 3-year structural math for your population? Archer Health builds shared onsite and near-site primary care for self-funded employers in the 100–1,000 employee range. We’ll decompose your 2026 trend by driver, model the recovery curve at your headcount, and show you the real-time clinical data layer you’ll have on day one. Talk to our team.