Betting on sports playoffs is one thing. Betting on whether a cancer drug will work is another — and it is now possible.

Prediction-market company Kalshi, with data firm AppliedXL, has launched markets on clinical trial outcomes and FDA decisions. The pilot opened with about a dozen contracts tied to late-stage trials, letting people wager on whether a drug will hit its primary endpoint, win approval, or when a company will file.

What you can bet on

Early contracts cover closely watched programmes including Takeda’s oveporexton, Intellia Therapeutics’ lonvo-z, Eli Lilly’s retatrutide and VERVE-102, and Compass Pathways’ psilocybin therapy COMP360.

Why blinding makes this different from stock trading

Trading on pharmaceutical outcomes already happens through equities, and the structure of a clinical trial makes a direct contract materially different.

A trial is deliberately blinded so nobody involved knows which patients received the drug, which prevents expectations influencing how outcomes are assessed. Certain people necessarily see unblinded data: independent monitoring committees reviewing safety, statisticians preparing analyses, and staff at sites handling drug supply.

Those individuals hold information that resolves the contract, and the number of them is not small — a Phase 3 trial involves hundreds of people across dozens of sites, in several organisations, in multiple countries.

“At best it incentivizes predictably bad behavior from folks with insider access,” said Shashi Shankar, CEO of Novellia.

Why the enforcement problem is harder here

Insider trading in securities is illegal and prosecuted, and the framework depends on a defined relationship: someone owing a duty to shareholders trading on material non-public information.

Whether that framework extends cleanly to a prediction contract is unsettled, and the people with access are frequently not employees of the sponsoring company. A statistician at a contract research organisation, or a nurse at an investigator site, may owe no obvious duty to anyone whose shareholders are affected.

The concerns are not hypothetical — prediction markets have already produced eye-catching payouts, including a reported six-figure win on Kalshi and a $400,000-plus haul on a rival platform tied to non-public information.

The subtler worry

Amy Bucher, chief behavioral officer at Lirio, raised a different concern: public predictions could nudge how researchers interpret results or how funding flows, through cognitive bias and social influence.

That is worth taking seriously because clinical research involves many judgement calls that are not fully specified in advance. Whether an adverse event is treatment-related, how a protocol deviation is handled, which analysis is treated as primary when several were pre-specified — each is decided by people who could be aware of a market price.

The mechanism does not require anyone to act dishonestly. Being aware of an expectation shifts interpretation of ambiguous evidence, which is exactly why blinding exists in the first place.

The case on the other side

Prediction markets aggregate dispersed information and have outperformed polls and expert forecasts in some domains, and a public probability that a trial will succeed would be genuinely useful to patients deciding whether to enrol, to investors, and to the field.

The difficulty is that the mechanism producing accurate prices is people trading on information — and here the most valuable information is the confidential result itself.

Why it matters

How trial results are normally protected

The existing safeguards around unblinded data are worth setting out, because they were not designed with a public market in mind.

Access is deliberately compartmented. An independent data monitoring committee reviews accumulating results to decide whether a trial should stop for harm or overwhelming benefit, and its members are chosen to have no financial interest in the sponsor. Statisticians preparing those analyses are typically separated from the study team.

Everyone involved signs confidentiality agreements, and companies enforce blackout periods restricting employee trading around anticipated readouts. Those measures address securities trading by people the company employs.

What they address less well is a distributed workforce with partial information. A site pharmacist who notices which patients received an active-looking preparation, or a coordinator who observes patterns in adverse events, holds fragments rather than results — not enough to trade shares confidently, potentially enough to move a binary contract.

Prediction markets are also anonymous and small in individual position size, which makes detection harder than in equities, where unusual trading in a single stock ahead of an announcement is a familiar pattern regulators actively monitor.

Clinical trials are meant to protect patients and produce trustworthy science. Turning outcomes into tradable contracts tests whether that integrity survives a profit motive — a debate regulators and drugmakers are only beginning to have. Business news, not investment advice.