The 5,000-Employee Myth: How Mid-Market Companies Get Onsite Primary Care That Actually Works

For three decades, onsite primary care was a Fortune-500 perk. Boeing built one. Cisco built one. Toyota built four. The reason was math, not strategy: a dedicated provider, a build-out, a benefits team to run it — none of that pencils out unless you’ve got 5,000 employees walking past the door.

Then the math changed.

A few of the assumptions baked into the old onsite playbook quietly stopped being true. Build-outs got cheaper. Telehealth normalized as a delivery channel. Self-funded plans pushed downstream into the 100–1,500 employee range. And — most importantly — the shared-clinic model matured to the point where a single primary care provider can serve two or three small employers in the same metro and the unit economics actually work.

If you’re a broker, an HR director, or a CFO at a 200-to-1,500 employee company, the question is no longer can we get onsite primary care. The question is what’s the right shape of the program at our size, and what numbers should we expect in year one.

This post lays out the math.

What “onsite primary care” actually is

Before we get to the numbers, a quick definition. Employer-sponsored onsite primary care means your company contracts directly with a clinical group — bypassing the carrier — to put a provider in or near your workplace. Members access primary care at zero cost-share. The clinic handles same-day appointments, chronic disease management, labs, prescriptions, and often light urgent care.

You can split the model along three axes:

• Onsite (a provider in your building or campus) vs. near-site (a clinic within driving distance, often shared between employers) vs. virtual-first (telehealth-led with periodic in-person days).
• Dedicated (the provider sees only your employees) vs. shared (the provider rotates between a handful of employers in the same metro).
• Full-time (a provider 5 days a week) vs. part-time (the provider runs your clinic 1–2 days a week and other clients the rest).

The cost structure of these three axes is what determines whether onsite primary care is realistic at your headcount. A full-time dedicated onsite clinic costs roughly $400,000–$700,000 per year all-in. A part-time shared near-site clinic costs roughly $80,000–$200,000 per year. The right shape depends on your population, your geography, and how much utilization you need to drive to make the program pay for itself.

Why the 5,000-employee number stuck around

The original Fortune-500 onsite clinic was a full-time dedicated build. A provider, a nurse or medical assistant, a phlebotomy station, a $200,000–$300,000 build-out for the exam rooms, a benefits manager dedicating part of their role to the program. Annual cost: roughly $500,000.

To pay back $500,000 a year through medical-spend reduction, you need:

• 5,000 lives × $100 PMPM impact = $500,000/yr in cost avoidance, OR
• 2,500 lives × $200 PMPM impact = $500,000/yr in cost avoidance

The literature on dedicated onsite clinics consistently shows somewhere between $80 and $150 per member per month in net cost reduction once utilization stabilizes. So the floor was about 5,000 lives. Below that, the program was a benefits expense, not a savings vehicle.

That math is exactly right — for that model. The model just isn’t the only model anymore.

The new math at 200, 500, and 1,000 employees

The shared-clinic model collapses the cost side of the equation. Here’s the back-of-envelope for three realistic mid-market scenarios:

A few things to notice about that table.

Year one isn’t where the savings live. The single biggest mistake employers make in evaluating onsite primary care is judging it on year one. Utilization takes 6–12 months to ramp. Chronic conditions take 12–24 months to bend. The investment-to-payback curve looks more like solar panels than a software contract — flat for the first 18 months, then accelerating.

The savings come from three places, not one:

1. ER and urgent care avoidance. A walking-distance clinic deflects 30–50% of urgent care visits and 10–20% of ER visits. At commercial pricing, that’s the single largest line item.
2. Specialty referral redirection. A primary care provider who actually has time can resolve 60–70% of presenting complaints in-clinic that would otherwise become specialty referrals.
3. Pharmacy steerage. Same-day generic prescribing and 90-day mail-order direction can cut specialty drug spend 8–15%.

The shared model is the unlock. Spreading one provider across three employers in the same metro means each employer pays one-third the fixed cost. Utilization stays high because the provider is physically in each location 1–2 days a week, not virtual-only.

What the program looks like in practice

At 300 employees, a typical Archer-shape program runs like this:

• A nurse practitioner or family medicine physician on-site one day per week at your workplace, plus virtual access (secure messaging + video) the other four days.
• A member portal that handles booking, intake, consent forms, prescriptions, and post-visit messaging — so members don’t need to call a front desk.
• Same-day appointments for sick visits. Walk-in capacity on the on-site day.
• Labs drawn on-site and sent to a contracted reference lab. Results land in the portal within 24–48 hours.
• Annual wellness visits with full screening panels — typically $0 cost-share to the member, often the first time the member has had a real primary care relationship.
• Pharmacy script writing with steerage to the carrier’s preferred mail-order partner for chronic meds.

The thing that surprises HR directors most: utilization is dramatically higher than their carrier’s primary care network. We see 70–95% of eligible members touch the clinic in year one, vs. 30–40% who see their carrier-network primary care doctor in a typical year.

The reason is mechanical, not motivational. The friction is gone.

What’s different in 2026 — the data layer

The other thing that’s changed since the old 5,000-employee floor: employers can finally see what’s happening in real time.

Historically, the only feedback an HR director got about their healthcare program was a claims report 60–120 days after the fact. By the time you knew that visits were down or pharmacy spend was up, the renewal was already half-over. Every wellness vendor in the market sold “we ran X biometrics” and “we sent Y emails” — outputs, not outcomes.

A modern onsite clinic with a real data spine — what Archer calls Strata — captures every visit, every lab, every prescription, every survey, every reward, every consent in real time. The HR director can open a dashboard and see: 218 preventive screenings done this quarter, 43 hypertension cases under management, 35 office visits resolving conditions that would otherwise have become specialist referrals.

That’s the difference between “trust me, this is working” and “watch this dashboard.”

For brokers, the implication is even sharper. At renewal, you stop telling carriers what you think the clinic did. You show them the data and ask them to price the deflection.

Common objections — and what’s actually true

”We’re too small for onsite.”

This is the line that’s most often wrong. If “onsite” means a full-time dedicated provider, sure — that needs 1,500+ lives. But a shared near-site clinic with 1–2 days of in-person service per week is realistic at 150 lives. Several Archer clinics are running at exactly that headcount.

”Our people use their primary care doctors.”

A claims report can answer this in 10 minutes. Pull the percentage of eligible members with a paid primary care office visit (CPT 99211–99215, 99381–99397) in the last 24 months. The number is almost always under 45%. Younger workforces look more like 20–30%. That’s the baseline that an onsite program rebuilds.

”Telehealth is good enough.”

Telehealth-only programs typically run 15–25% engagement. Walking distance gets you to 70%+. The two aren’t competitors — a good onsite program uses both. But telehealth-only is a wellness budget line, not a medical-spend strategy.

”We can’t afford the build-out.”

You don’t need a build-out. Most small-employer onsite programs run out of a 250-square-foot converted conference room or a 600-square-foot near-site space the clinical group rents. Capex is a non-issue at this scale.

”We don’t want to be in the healthcare business.”

You’re not. The clinical group employs the provider, manages credentialing, handles malpractice, manages the EHR, files referrals, and bills (if applicable). You sign a contract, you provide a room (or you don’t, if it’s near-site), and you point your members at the booking page.

Where to start

If you’re a broker writing a proposal for a 200–1,500 employee client, the order of operations is:

1. Pull the 24-month medical-spend claims report. Identify what percentage of paid claims are PCP-resolvable: typically 35–50%. That’s the size of the deflection opportunity.
2. Look at geography. Onsite is realistic if 60%+ of the workforce works from a single primary location. If the population is geographically diffuse, near-site or shared models work better.
3. Frame the cost-comparison correctly. The right comparison isn’t “current PCP spend vs. onsite cost.” It’s “current ER + urgent care + specialist + pharmacy spend vs. onsite cost + projected deflection.”
4. Get a 3-year forecast, not a year-one forecast. Any clinical operator who promises year-one savings is selling — the math doesn’t work that way.
5. Insist on real-time data access. If the clinical group can’t give the HR director a live dashboard, you’ll be back to claims reports 60 days late.

If you’re an HR or benefits leader, the question to ask your broker is: “Has anyone in our headcount range done this successfully in the last 12 months, and can I talk to them?” The market has moved fast enough that 2023’s “this won’t work at 300 lives” advice is already wrong.

FAQ

What’s the minimum employer size for onsite primary care?

For a shared near-site clinic with 1–2 in-person days per week, the realistic floor is roughly 150 lives. Below that, the per-member fixed cost compresses the savings opportunity. For a dedicated full-time onsite clinic, the floor is closer to 1,500 lives. The shared model is what closes the gap for mid-market employers.

How does an onsite clinic save money for a small employer?

Three vectors: ER and urgent care avoidance (a walking-distance clinic deflects 30–50% of urgent care visits and 10–20% of ER visits at commercial pricing), specialty referral redirection (a PCP with time resolves 60–70% of presenting complaints in-clinic), and pharmacy steerage (8–15% reduction in chronic-med spend through generic prescribing and mail-order steerage).

How long until an onsite clinic pays for itself?

Year one rarely covers full program cost. Mature utilization (year 2-3) typically returns 2–4x program cost for fully-engaged employers. The investment curve looks more like solar panels than a software contract.

What’s the difference between onsite and near-site?

Onsite means the clinic is inside your workplace. Near-site means it’s within driving distance, often shared with other employers in the same metro. Walking distance changes utilization by 2-3x. Onsite is the gold standard if your workforce is concentrated in one location; near-site is the practical answer for distributed workforces.

What does an onsite clinic actually do?

Same-day primary care visits, chronic disease management (diabetes, hypertension, lipid disorders, mental health), annual wellness exams with full screening panels, labs drawn on-site, prescription writing with mail-order steerage, secure-messaging access to the provider between visits, and light urgent care for things like rashes, strep, sprains, and UTIs.

Is onsite primary care HIPAA-safe for employers?

Yes. The clinical group is the HIPAA covered entity; the employer never sees individual-level clinical data. Employers see de-identified aggregates (visit counts, condition prevalence, screening completion rates) and program-level metrics, never PHI.

How do I evaluate vendors?

Three questions cut through the marketing fog:

1. What does year-2 utilization look like at our headcount range? Below 60% means the program isn’t pulling its weight.
2. What’s the real-time data look like — can I see what’s happening this week, not last quarter? Vendors who can only produce 60-day-old claims reports are operating with the same lag as your carrier.
3. Show me a reference customer in my headcount range. If the vendor only has Fortune-500 references, the cost structure they’re quoting you probably reflects that.

Want the math for your headcount? Archer Health builds shared onsite primary care for self-funded employers in the 100–1,500 employee range. We’ll run the 3-year ROI for your population and show you the real-time data spine we capture against it — the same dashboard your HR team would see post-launch. Talk to our team.