With GLP-1 agonists, we finally have an effective treatment for obesity. Recent data demonstrates that semaglutide reduces major adverse cardiovascular events by 20% — an effect size comparable to statins — and that this class of medication is cost-effective. Yet at the end of 2025, 100 million obese people who could have benefited did not have access to these medications. Instead, most major insurers raced to announce they would no longer cover GLP-1s for weight loss starting January 2026.

The story of GLP-1s is a reminder that American health insurance was never built for health improvement; it was built to pay for catastrophes. Here is a medicine that will demonstrably improve lives and reduce health care costs, yet we will not pay for it. We reimburse stents for stable angina and spinal fusion for back pain despite questionable benefits. But preventing disease? Making someone healthier before they need the ER? Primary care is systematically devalued.

Why? First, prevention is not a one-time cost. Even if there is a clinical debate regarding the benefit of a spinal fusion, to an insurer, it is a discrete financial event. They pay once; the claim is closed. Second, patients "churn" plans frequently, so preventing a heart attack 10 years in the future may not save your insurer money — a classic market externality. Third, we pay for visits and procedures, not for patient outcomes.

The result is a society that says health is our greatest asset while treating it like everything but an asset. What if we took our words literally? What if we treated health trajectories as assets: tradable derivatives of a valuable uncertainty? I propose a vision of health insurance where we do just that. At the core is a novel mechanism I call the Synthetic Health Outcome Contract — or SHOC. Below, I envision how a patient, and the health care system, may find their health, and their bottom line, improved by SHOCs.

A SHOC to the System

Consider, if you will, the following hypothetical scenario, set in a future where SHOCs are standard: Marcus, 52, sits in the waiting room of the newly established "Federal Health Exchange" in Boston. It’s November; renewal season. He’s here for his yearly biometric screening. A nurse takes his vitals: blood pressure 168/94, weight up 12 pounds, obese. The finger prick for the A1C test stings. He already knows it’s bad; he’s been too busy to cook real meals or remember pills.

"All done," the nurse says. "Upload this to the portal. You have until the end of the year to select your Health Insurance and Improvement (HII) plan."

At home, Marcus scans his QR code. His EHR — every missed appointment, unfilled prescription, and ER visit — uploads to the federal insurance portal. The screen prompts him to select an AI foundation model. He picks Epic CoMET. It's the oldest approved model since it first came out in 2025, before the HII-SHOC system was enacted. Behind the loading bar, an AI processes his entire EHR, simulating trajectories for what the rest of his life may look like. Like the control arm of a randomized clinical trial, it calculates what Marcus' health outcomes will be on his current path compared to those most similar to him.

The report loads. Health Score: 32nd percentile.

He brings the report to a representative, who offers him several available plans. Plan A: Maintenance (0% improvement) costs $287/month. Plan B: plus-10 percentile improvement costs $398/month.

"So I have to pay more to get healthier?" he asks.

"Yes, but think of it like paying for a gym membership that guarantees results," the representative says.

Marcus remembers his last ER visit for chest pain — $4,200 out of pocket. He selects Plan B.

Behind the scenes, his de-identified profile is released to the provider market as a SHOC. Hundreds of health care organizations are now analyzing his data.

At Boston General Medicine (BGM), a doctor reviews Marcus' data. She calculates BGM's traditional costs: PCP visits for weight counseling, nutrition consults, and endocrinology referrals. It totals $4,200 in overhead. She analyzes his "Greeks" — the rate of change, stability, and volatility of his health. She submits a bid on behalf of BGM with a standard profit margin, priced at $6,500.

Across town, however, GutInc, a clinic focused on weight loss through GLP-1 agonists, reviews Marcus' data. They see BGM's bid as mispricing the asset. GLP-1 agonists can produce high delta (large change in the underlying health metrics per dollar spent) without other visits. Plus, their clinic believes that with proper incentives — even a direct cash payment — they can ensure adherence. Since HII-SHOC was enacted, providers like GutInc have sprung up offering innovative techniques to achieve better outcomes on the contract. This lets them submit a lower bid of $6,000.

The SHOC market assigns their clinic as the winner.

Marcus gets an email: "You have been matched with GutInc. Your welcome kit will arrive December 10."

Two weeks later, a box arrives with pre-filled pens and a smart case. No appointment to schedule. No lecture about vegetables. A non-traditional intervention. He takes the first GLP-1 injection that night.

Twelve months later, Marcus has lost 42 pounds. He returns to the Exchange and sees his new Health Score: 44th percentile. He’s exceeded the target by two percentile points. GutInc is especially pleased because SHOCs include an outcomes-based bonus structure. The insurer pays the bonus, which increases Marcus' premium slightly next year to recoup costs, but because his baseline health has improved, his base premium actually decreases. If GutInc had failed, they would have refunded part of the SHOC value plus penalties. This swiftly removes bad players from the market.

Marcus doesn't see this calculation. He just knows he feels better.

Benefits: Bringing Financial Markets to Patient Outcomes

Once you have this pricing mechanism, market behaviors become fascinating. Imagine a provider organization that specializes in high-volatility patients: homeless populations with diabetes or those cycling through EDs. Everyone else sees them as expensive risks. This provider sees massive positive delta and focuses on stabilizing housing and diet. They go long on volatility, bringing private financial capital to vulnerable populations, enabling targeted interventions, and aligning profit motives with genuine health improvement.

Or consider end-of-life care. A patient with Alzheimer's could be offered a "palliative care trajectory" contract. If a provider can keep the patient comfortable and out of acute care through home interventions, beating the natural history of the disease, they profit from their good clinical judgment to avoid unnecessary care.

When innovation happens, it propagates instantly. A provider discovers that certain care patterns work better? They underbid everyone else on those contracts. Competitors either adopt the innovation or get priced out. Markets bring a race to the bottom on price, but a race to the top for patient outcomes and societal wellness.

Critiques: On the Health Score Algorithm

Transformer models like CoMET already show astonishing accuracy in predicting future medical encounters. There is understandable concern about the error that could result from entrusting an AI model to quantify someone's health. I alluded to factors that mitigate this in competing models and centralized federal data, but some worry about bias. For example, AI learning that people from a certain ZIP code always have worse health. However, bias exists in our current system: we've just hidden it behind insurance adjusters and private models. With AI, we can audit the bias. Furthermore, the system creates a market incentive to counteract bias. If a provider finds that people from a specific area are consistently given worse Health Scores than justified, that means their cost for improvement is actually lower. The provider can underbid those contracts, eventually resulting in market correction.

Others may worry about the "black box" nature of AI. In fact, this is a strength: it prevents the effects of Goodhart's Law: "when a measure becomes a target, it ceases to be a good measure." By using a transformer model like CoMET, which learns non-linear, complex interactions, the only way to "game" the system is to actually improve the patient's health.

Critiques: On the Ethics

We tend to think of people being punished for genetic or environmental luck as unfair. The Affordable Care Act helped eliminate the effects of such luck on coverage and cost for pre-existing conditions. HII-SHOC takes this further. The AI establishes your health trajectory in comparison to those most similar to you, and payment is based on standard deviations from your baseline. This means a quadriplegic trying to improve one standard deviation and an Olympian attempting the same relative gain face similar contract prices.

I don't pretend SHOCs would be perfect. Implementing this approach requires sophisticated infrastructure and federal policy. Patients will always surprise us. Yet I believe the asset-ification of health is a world where preventing disease is profitable and where the financial incentive is to make people healthier, full stop.

As we enter 2026, I anticipate the advent of many new medications such as GLP-1s which not only improve patients' lives, but prevent health catastrophes that cost our system and society outsized harm. Our health insurance system is not up to the task of financing this new era of life improvement through medicine. Physicians can take a leading role here through roles in health insurance companies. I urge readers considering alternative careers to help reform our insurance system into a health improvement system — improving both patients' lives and the business of health care.

Aditya Jain is an MD student at Harvard Medical School, and a researcher on applications of AI in medicine at the Broad. He is interested in the business of health care and its intersection with technology and policy. More of his writing can be found on Substack @adityajain42. Aditya was a 2023–2024 and a 2024–2025 Doximity Op-Med Fellow, and continues as a 2025–2026 Doximity Op-Med Fellow.

Image by rob dobi / Getty Images