I’m starting a company called Calypta Health to help people navigate the most contentious and confusing health decisions we face. The mission: lower medical spending, ease provider stress and burnout, improve population health.

This is personal

I’ve written about my unexpected personal adventures with our medical system. My cancer was caught early because I was actively choosing preventive care — having annual physicals, asking questions. Even so, I almost didn’t mention the symptom that led to my diagnosis. I had access to excellent care and expert advice from my own family, and I still nearly missed it. Tools like the one I’m building can change those odds.

I’ve spent my career on large-scale behavior change

At Opower I helped build software that’s saved more than 41 terawatt-hours of energy to date. The insight was simple: give people clear, personalized information about their own behavior and how it compares to the norm, and many of them will change it.

More recently, I founded Vote Forward, which has organized volunteers to send more than 40 million handwritten letters to their fellow citizens. We ran dozens of RCTs to figure out what actually moves people to vote.

Now I’m applying that playbook to problems in the health domain. The goal: shift behavior with clarity and empathy, rigorously measure impact, scale up what works.

Our health information landscape is collapsing

We now have a lifelong anti-vaxxer running HHS, methodically undermining trust in vaccines — one of the most miraculous achievements in human history. HHS recently removed seven childhood vaccines from the recommended schedule, moving them to the category of “shared clinical decision making.”

“Shared clinical decision making” sounds reasonable enough. But our public health authorities are actively eroding confidence in vaccination, muddying the waters for those shared decisions. And most clinicians don’t have time for these conversations in a 15-minute appointment.

The stakes are high. Epidemiologists have modeled what happens if vaccination continues to decline, and the picture is dismal. Measles becomes endemic again. A further 50% drop over 25 years would yield roughly 51 million measles cases, 10 million rubella cases, 4 million polio cases, more than 10 million hospitalizations, and 160,000 deaths. Because these diseases are so contagious, the effects of declining vaccination are highly non-linear. Even a 10% drop translates into millions of additional cases and thousands of preventable deaths.

There’s a similar epistemic crisis brewing across other health topics. We have grifters claiming that statins are pointless, that ASCVD can be cured by buying their supplements, and that ivermectin is a miracle cancer treatment. None of this is new, but it’s all getting worse, and it’s alarming to see so many people making decisions based on fear and misinformation.

The cost is staggering

This is one place where the “MAHA” impulse deserves some credit: bad choices really do sacrifice health and well-being, not to mention oceans of cash. Consider just a few behavior-driven gaps: 28 million adults eligible for statins but not taking them. 28 million age-eligible adults who have never been screened for colorectal cancer. 90 million adults with uncontrolled hypertension. More than 100 million obese adults, most of whom aren’t getting effective treatment. These gaps represent roughly 100 million lost quality-adjusted life years and $1.5 trillion per year in total economic burden in the U.S. alone. The scale of harm is so great that even small behavioral shifts would be enormously valuable.

LLMs can help — they have a bias for the truth

The epistemic landscape is splintering into disjoint realities. But LLMs, as Dylan Matthews recently pointed out, are converging media technology. There’s a line that “reality has a well-known liberal bias.” That framing was always too smug, but the underlying idea holds for LLMs. They have a bias for the truth, and they are increasingly well-calibrated to resist extremist nonsense. They’re not infallible, of course! I don’t trust them to make fine-grained truth claims on subjects where there’s substantial human disagreement. But on the big questions, they reliably point in the right direction.

They’re infinitely patient and non-judgmental. They have encyclopedic knowledge baked into their weights and powerful search abilities to augment it. They speak every human language, at any reading level. This is why the major labs are building health products, and clinicians are already making extensive use of these tools.

So far, though, there’s something structurally missing. Patients and clinicians are mostly using these tools in isolation. There’s no connective tissue; it’s ad hoc. What I’m building: curated experiences that can be prescribed for particular situations, that guide people through the evidence, address their concerns without judgment, and then close the loop with the clinician.

Here’s how it works in practice: a clinician sends a patient a link before or after any appointment where a contentious or time-consuming topic is on the agenda. The patient reviews content about the decision. For vaccines, that includes clear explanations of the risks of the diseases, without pulling any punches.¹ Then they have an open-ended conversation with a frontier LLM about their specific concerns — non-judgmental, unhurried, empathetic. With EHR integration, clinicians get a short briefing on what was discussed.

The result for clinicians: patients arrive more informed. They spend less time on lengthy, contentious conversations. Visits are more productive.

I tested this on vaccine-hesitant parents, and it worked

Last year I ran an experiment with vaccine-hesitant parents. Those who reviewed carefully designed content and then chatted with an LLM increased their MMR vaccination intent by more than a full point on a 7-point scale versus control. Nearly two-thirds moved at least one point, and the effect persisted for several days.

That’s an unusually large effect for a brief informational intervention.

The chat transcripts were illuminating. In several cases, correcting a single misconception was enough: “I am more likely to allow my child to receive vaccination knowing it doesn’t contain aluminum.” In others, providing an actionable plan helped: “I think a preparation plan makes me more eased up.” Parents with well-grounded medical concerns (seizure history, anaphylaxis) revealed context that would help a pediatrician prepare for a complicated conversation.

The incentives are warped, but they exist

In the American system, few actors have strong incentives to improve health behaviors at scale. But the incentives aren’t all broken. Medicare Advantage plans need to hit Stars metrics. Integrated systems and self-insured employers benefit when their people stay healthy. New models like Crowdhealth are emerging, where each member has a financial interest in the health of every other member.

Interested?

I have a beta system live at app.calyptahealth.com. You can try it yourself — signup takes a minute.

These conversations are about to get harder, not easier. Patients are more confused than ever, and public health messaging is actively making it worse. This is the right time to try new tools.

If you’re a clinician who spends too much time on difficult conversations about vaccines, screenings, or medications…

If you work at a health system, payer, or public health department and you’re on the hook for preventive care metrics that you struggle to move…

If you’re a researcher who studies health communication, vaccine hesitancy, or behavior change…

If you run a preventive care advocacy organization and want better tools for reaching your audience…

If you’re reading this thinking “I wish this existed for statins” or “for GLP-1s” or “for blood pressure”…

…I’d love to talk. I’m starting with vaccines and CRC screening, but the pattern generalizes to any conversation where the evidence is clear but patients resist or delay. Send me a note: scott@calyptahealth.com.