“Spend more, save more” sounds like a used-car pitch. In healthcare, it turns out to be true — and now there’s an at-scale study to prove it.
For years, employers have tried to bend the cost curve by doing less — reducing benefits, cost shifting, and patching the cracks. Every one of those moves trades a short-term saving for a long-term bill. Real Talk Season 2, Episode 2 made the opposite case: spending more on primary care actually lowers total cost. I sat down with Tory Wolff, Founder & Managing Partner of Recon Strategy, whose new analysis puts numbers behind it, and Dr. Raj Shah, President of Centivo Care, who lives it every day.
More primary care, less cost
Primary care is having a moment — Premise and Crossover merging, health systems chasing revenue beyond fee-for-service, local direct primary care starting to pick up steam. Tory’s read: the market has run out of band-aids, and primary care is the tool at hand. His study makes it concrete. Across a nationally representative commercial population, the more frequently higher-risk patients engage their primary care team, the lower their net cost — not from the visit, but from what it prevents downstream. Structure that rhythm across a population, and you take 3–4% off total medical spend.
“It’s rare to get a result as clean as this one. The data supports what intuition would tell you.”
The logic is almost boring: high-risk members drive the spend, so shaving their downstream costs outweighs the extra you invest in primary care.
Where the money actually goes: decompensation
Dr. Shah put a name to it. More than 90% of the country’s $5 trillion in health spend traces to chronic disease — and the expensive moments aren’t the disease, they’re the decompensation: the hypertensive who goes uncontrolled and lands in the ER with a stroke, the diabetic who was never managed. Put primary care at the front end, and you prevent the crisis that writes the big check.
“But won’t free primary care get overused?”
It’s the objection I hear constantly — Centivo’s $0 primary care must invite overuse. The paper says the opposite: the downstream savings far exceed the incremental spend. The best Medicare Advantage models put 6–7% of total medical expense into primary care; commercial employers sit around 4–5%. We think the right number is 10–12%. And the right cadence isn’t set by benefit designers — it’s set by the patient and the physician.
Can you do this virtually? Meet John.
The next objection is always the same: you can’t manage complex disease through a phone or computer screen. Dr. Shah’s answer — about 20% of complex virtual primary care is clinical decision-making; the other 80% is relationship and logistics. Then he told us about John. John, 43, used urgent care for everything and logged into CentivoCare wanting an antibiotic for a red leg. A holistic intake surfaced uncontrolled diabetes — an A1C of 14 — and high blood pressure he had no idea about. He came for a prescription and left with a medical home, managed through same-day chat and text because he couldn’t take time off work.
“John didn’t know he needed advanced primary care. John just wanted an antibiotic.”
You have to earn the right to treat complex patients, Dr. Shah says — and the great virtual PCP doesn’t replace in-person care, they orchestrate it when it’s needed.
Pay for it like it matters
Tory’s study was conservative — it counted only billed visits, none of the texts and async touches modern primary care runs on. In a virtual model, you have to redefine the unit of care: a single text can save $5,000. As Tory put it, you can’t pay for primary care in a way that strangles it into a referral engine. The obstacle down-market isn’t employers — it’s stop-loss carriers that won’t reimburse anything not tied to a claim, which is exactly backwards, since the point is to prevent the claim.
AI: make primary care more human
On AI, Dr. Shah was clear-eyed: more than half of patients already bring it to their health questions, so resisting it won’t work. The opportunity is a paradox — use AI to make primary care more human, handing agents the administrative load so clinicians can spend their time building the trust that measurably improves outcomes. The risk is AI’s overconfidence; when a patient arrives with a stack of AI printouts, the move isn’t to co-sign it — it’s to acknowledge the effort, slow down, and decide together.
The bottom line
The cheapest move in healthcare is often to invest a little more, a little earlier, in the relationship that keeps someone out of the expensive end of the system. Tory’s version is 3–4% off total spend. Raj’s is a guy named John who came in for an antibiotic and left with his diabetes caught. Same story, told two ways.
More care, less cost. It’s not a slogan — it’s the math.
Watch the full conversation with Tory Wolff and Dr. Raj Shah here.
– Ashok Subramanian
