600 Patients, $360K, 25% Overhead: Atlas MD's Founder Lays Out the DPC Math

Most doctors assume the reason DPC works is that patients pay cash. That’s half of it. The other half sits on the expense side of the ledger, and Atlas MD founder Josh Umbehr spelled it out this week in a way that’s hard to argue with.

His example is deliberately plain. Six hundred patients, each paying about $50 a month, comes to $360,000 a year. One staff member covering one or two physicians. One or two exam rooms. Overhead in that setup runs somewhere between 20 and 25 percent, which means the doctor keeps most of what comes in (Medical Economics). Umbehr has been running this model since 2010, so the numbers rest on a decade of receipts.

He borrows a line from The Lean Startup to describe what a clinic actually needs on day one: a shoestring and a stethoscope. It sounds like a slogan until you compare it to what a traditional practice carries.

The overhead gap is the whole argument

A conventional primary care office typically runs 55 to 65 percent overhead. Billing staff, coders, denial appeals, the software that talks to payers, the people who follow up when the payer doesn’t. All of it exists to collect money from insurers, and none of it touches a patient.

Strip the insurance relationship out and most of that expense goes with it. That’s the part residents tend to miss when they hear “$50 a month” and assume the model must be a grind. The membership covers one side of the ledger. The absence of a claims department carries the rest.

Think about what that gap does to the same $360,000. A fee-for-service practice keeping 35 to 45 cents on the dollar is running hard to clear what a DPC clinic clears while keeping 75 to 80. You don’t need a bigger panel to earn a living. You need a smaller cost structure, and DPC hands you one by design.

This is why the model rewards restraint. The physicians who struggle in DPC are often the ones who rebuild the overhead they left behind, adding staff and systems before the panel justifies them. Umbehr’s version stays deliberately thin.

Small panels, direct access

The economics only hold because the panel stays small. Most DPC physicians care for somewhere between 400 and 800 patients, against the 2,000-plus a typical employed doctor carries (DPC Alliance and DPC Frontier practice data). A smaller panel is what makes 30 and 60-minute visits possible, along with same-day access and messaging that doesn’t route through a portal nobody checks.

Umbehr also points to the perks that come baked into a cash model: unlimited visits, telemedicine at no extra charge, no co-pays, and wholesale pricing on labs and medications that he says can run a fraction of retail, in some cases discounts he puts as high as 95 percent. Treat that top figure as his claim rather than a guarantee. Wholesale savings vary widely by practice, state, and what a physician is licensed to dispense, and dispensing rules differ from one state to the next.

The membership price itself is worth a caveat too. Fifty dollars is Umbehr’s example, not a benchmark. Most DPC practices land between $70 and $100 a month for an individual, with families higher (DPCA and DPC Frontier). What stays constant across all of them is the overhead math, not the sticker.

What This Means

For a resident weighing an employment contract against hanging a shingle, the useful takeaway isn’t the revenue number. It’s the cost structure. DPC lets one physician and one staffer run a functioning practice because the model deletes the expense of dealing with payers. That’s the thing that makes the panel size, the visit length, and the schedule possible. Everything people like about DPC traces back to the same source.

For physicians already in DPC, Umbehr’s breakdown is a reminder that discipline on the expense side is the quiet engine. The temptation to scale staff and tools early is real, and it’s how a lean practice slowly turns back into the thing its founder left.

Umbehr expects insurance-free primary care to become the assumed model for a lot more people within three to five years, partly because patients in 2026 are dropping coverage mid-cycle when they can’t afford it or no longer see the point. Whether that timeline holds is anyone’s guess. The arithmetic underneath it, though, has been sitting in plain view since 2010. More doctors are finally doing the math.