The Promise Inside the Immunity

Before Brianne Dressen received the injection, AstraZeneca set an informed-consent form in front of her. On the question that would matter most if the experiment harmed her, the form gave two answers.

The first answer was a promise. The company carried insurance for research injuries. A study doctor would provide treatment or arrange it, and AstraZeneca would pay the reasonable costs of medical treatment for injuries caused by the vaccine, tests, or procedures—so long as Dressen followed the study doctor’s instructions and did not cause the injury herself.

The second answer was a warning. Because the country was living through a public-health emergency, federal law might prevent her from suing the researchers, sponsor, manufacturer, or distributor. A federal compensation program might offer limited benefits instead, but only if Congress had funded it and her injury met the program’s conditions.

Dressen signed the form and received the injection.

Nearly six years later, the Tenth Circuit held that PREP Act immunity covered Dressen’s breach-of-contract suit even though the suit sought to enforce that promise. The court did not decide whether AstraZeneca broke the agreement, whether the vaccine caused her alleged injuries, or whether the promise could ultimately be enforced. It sent the case back for one question: did AstraZeneca give up its immunity by making the promise?

Congress enacted the PREP Act to block most civil suits over covered countermeasures used during a public-health emergency. Dressen’s claim rested on a written commitment—an affirmative, specific promise the company had allegedly used to secure her participation. The dispute is whether a sponsor may take the participation that promise helped obtain, then invoke the emergency itself to keep a court from deciding what the promise meant.

The Injury the Form Was Written For

In November 2020, no one yet knew how the vaccine race would end. The first emergency-use authorizations in the United States had not issued. The people who entered clinical trials that autumn stepped into that uncertainty before the American public was asked to accept any COVID-19 vaccine.

Dressen enrolled in AstraZeneca’s experimental COVID-19 vaccine trial in Salt Lake County, Utah, with Velocity Clinical Research running the study on the company’s behalf. The consent form she signed laid out the known risks, the procedures, the small payments for study visits and phone calls, and the arrangements for injury.

According to her complaint, the symptoms began within an hour of the shot: tingling in her right arm, then her shoulder, then her left arm, followed by blurred vision, ringing in her ears, nausea, and a raw sensitivity to sound and light. The allegations trace a hard path from there—the persistent sensation of an electric shock coursing through her body, emergency rooms, neurological treatment, evaluation at the National Institutes of Health, and diagnoses that came to include post-vaccine neuropathy, dysautonomia, and chronic inflammatory demyelinating polyneuropathy. She alleged that the abnormal, painful sensations never let up; that she could no longer work or train as an athlete; that driving became something she could barely do; that she could not be the parent she had been.

At the motion-to-dismiss stage, all of that remained allegation. Neither the district court nor the Tenth Circuit weighed the medical evidence or found that the vaccine caused any of it. If Dressen’s allegations were true, the situation was exactly what the injury clause addressed: a study participant who needed treatment after the experimental shot and repeatedly asked the sponsor to pay.

According to the complaint, AstraZeneca’s response was limited. Seventeen calls and emails between January and July of 2021 produced a single payment—$590.20. Later that year, study administrators offered $1,243.30 as a final settlement of every claim arising from the study. Dressen’s husband turned it down. "We have a contract stating that they will cover medical bills," he wrote. "We have been waiting this long to be insulted like this?" AstraZeneca later asked for records, said it was evaluating the claim, and then—according to the complaint—simply stopped writing back in September 2022.

Dressen sued in May 2024. Her complaint bypassed defective-design, negligent-manufacture, and failure-to-warn theories and relied on the form itself: breach of contract and breach of the implied covenant of good faith and fair dealing. She asked for past and future medical expenses, lost income, transportation costs, household and childcare expenses, emotional damages, attorney fees, and more. She alleged that the injection caused her injuries and that AstraZeneca later broke its promise to pay. Judge Shelby treated the refusal to pay as a separate legal wrong; the Tenth Circuit held that the claimed losses still arose from the injection.

What Informed Consent Must Disclose

Federal informed-consent rules require disclosure of risk, procedures, alternatives, voluntariness, and—for research carrying more than minimal risk—whether compensation or medical treatment is available after injury, along with its terms or where to learn them.

The regulation requires clarity about whether payment or treatment exists while leaving a sponsor free to offer neither. It also forbids language that makes a participant waive—or appear to waive—legal rights, or that releases the sponsor, investigator, or institution from liability for negligence.

AstraZeneca’s form described an insurance policy, promised treatment or referral, and said the company would pay the reasonable medical costs of research injuries. It answered a practical question for a clinical-trial participant: who will pay if the research causes an injury?

That answer can affect whether someone enrolls. A person may reasonably distinguish between a study that leaves every later medical bill to the participant and one that says insurance will cover research injuries.

The form also warned Dressen about the PREP Act. It said federal law might block a lawsuit over injuries caused by the investigational product and directed her to the Countermeasures Injury Compensation Program. Immunity and payment appeared side by side on the page.

Read together, the two clauses pull against each other. The warning that federal law "may" limit suit reads like a reservation of immunity. The promise that AstraZeneca "will pay" reads like the assumption of a specific duty in spite of it. One clause names the shield Congress supplied. The other seems to define what the company agreed to do while standing behind it. The form asked Dressen to consent to the experiment after reading both. The law must now decide whether the promise survived the warning.

The Emergency Shield

Congress passed the Public Readiness and Emergency Preparedness Act in 2005 to enlist private manufacturers and providers in public-health emergencies at speed and scale. When the Secretary of Health and Human Services issues a qualifying declaration, covered manufacturers, distributors, planners, and providers receive an extraordinary grant of protection for covered countermeasures.

The operative language is sweeping: a covered person "shall be immune from suit and liability" under federal and state law for "all claims for loss" caused by, arising out of, relating to, or resulting from the administration or use of a covered countermeasure.

The statute defines loss as "any type of loss," including death; physical, mental, or emotional injury, illness, disability, or condition; fear of such harms and the cost of medical monitoring; and property damage, including business-interruption loss. Its causal reach extends across design, development, testing, manufacture, labeling, distribution, marketing, sale, prescribing, administration, and use.

Congress made a deliberate tradeoff. In an emergency, companies and providers may hesitate to develop, ship, or administer an unfamiliar treatment if every injury can produce ordinary civil litigation. The Act therefore protects many of those participants from suit.

The immunity protects covered parties from liability and from having to litigate the case. If a company had to complete discovery and trial before appealing, it would lose much of the protection even if it ultimately won.

AstraZeneca appealed the moment the district court refused to dismiss the complaint. Ordinarily appellate review must wait for final judgment. The collateral-order doctrine lets a narrow band of appeals go up early—but only when the question is conclusively settled, separate from the merits, and effectively unreviewable if it waits.

The Tenth Circuit held that PREP Act immunity qualified. Congress had conferred a right not to stand trial. Requiring AstraZeneca to litigate now and appeal later, the court reasoned, would destroy part of the statutory entitlement even if the company ultimately won.

Because immunity protects against litigation itself, the court had to decide it before discovery into the promise, testimony about AstraZeneca’s claims process, or any finding that the agreement had been broken.

The District Court’s Line

Chief U.S. District Judge Robert Shelby drew a line between the subject of a contract and the legal injury caused by breaking it. A vaccine may be the reason the agreement exists. It may be the event that triggers the duty to perform. But administering a vaccine cannot, by itself, break a promise to pay. The breach happens later—when the promisor fails to do what the contract requires.

On that reasoning, the vaccine triggered AstraZeneca’s duty to pay, but the alleged breach occurred later when the company refused payment.

Tort duties come from law; contract duties come from promises parties choose to make. Judge Shelby reasoned that a statute could remove tort liability without erasing a payment obligation AstraZeneca voluntarily assumed.

Judge Shelby tested AstraZeneca’s argument against the form’s separate promise of $125 per completed study visit and $30 per completed phone call. If PREP Act immunity reached every contractual duty related to the trial, he reasoned, the company could refuse those payments as well. He warned that such a rule would allow covered entities to use promises to obtain participation and then invoke immunity against enforcement.

The district court treated breach-of-contract claims as categorically outside PREP Act immunity. That ruling kept the case alive and left Dressen’s alternative waiver argument untouched. The Tenth Circuit rejected the categorical distinction.

All Claims for Loss

Judge Harris Hartz, writing for a unanimous panel, began with the statute’s breadth. The PREP Act protects "all claims for loss" without sorting them by cause of action. A contract action is a claim, and the damages Dressen sought—medical expenses, lost income, loss of services, emotional harm—lined up almost word for word with the statute’s definition of loss.

Business-interruption loss supplied further confirmation. As a familiar contract remedy, its inclusion cut against the idea that "loss" belonged to tort alone. The words "all" and "any" carry their ordinary breadth. The statute, the panel concluded, sweeps past legal labels and stops at causation rather than at the line between tort and contract.

Less than two months earlier, Congress had enacted the Protection of Lawful Commerce in Arms Act—another industry immunity—with an express exception for breach-of-contract and warranty actions. Congress knew how to preserve contract claims inside an immunity statute and wrote no such exception into the PREP Act.

Dressen invoked interpretive canons to narrow "loss," but the court found none applicable. The phrase "any type of loss, including" uses examples to illustrate breadth. "Including" opens a category rather than closing one. The panel declined to add the contract exception Congress omitted.

The statute speaks in claims and losses, not causes of action. Once Dressen’s requested damages fell within the statutory definition of loss, relabeling the cause of action left them inside the immunity. The opinion nevertheless stopped short of immunizing every contract touching a countermeasure because the statute still demands a causal link between the loss and the countermeasure’s administration or use. Here, the same alleged vaccine injury both triggered AstraZeneca’s promise to pay and brought Dressen’s claim within the PREP Act.

Two Causes

Dressen argued that the true cause of her contract damages was the breach, not the vaccine. AstraZeneca’s refusal to honor the promise was what left her bills unpaid. Take away the refusal, and the contract loss simply disappears. The panel answered that an injury can have several but-for causes.

Under ordinary but-for analysis, an event is a cause if changing it changes the outcome. Most injuries have several necessary conditions. A broken promise may be one. The injection may be another. Without the alleged breach, Dressen would have no uncompensated contract damages—true. But without the injection and the injuries she attributes to it, she would have no treatment costs for AstraZeneca to reimburse and no claim under the injury provision at all. The vaccine, then, was also a but-for cause of the very loss she sought to recover.

The district court had said the research injury was not, by itself, enough to make out a breach-of-contract claim; a broken promise was needed too. The Tenth Circuit called sufficiency the wrong test. But-for causation asks whether an event was necessary—not whether it was enough on its own. Congress had reinforced that breadth twice over: with four causal formulations—caused by, arising out of, relating to, resulting from—and by demanding proximate cause only in the statute’s narrow willful-misconduct action. For ordinary immunity, the panel held, bare but-for causation was all it took.

The contract provision applied only if the vaccine, tests, or study procedures caused a research injury. The same causal connection activated AstraZeneca’s alleged duty to pay and the PREP Act’s immunity. The injury was therefore essential to both sides of the dispute. A promise unrelated to a covered loss might fall outside the statute, but this promise applied only when the vaccine, tests, or procedures caused an injury.

The panel accepted the alleged breach as one cause and held that the countermeasure supplied another. Because the immunity provision did not require proximate cause, the vaccine did not have to be the closest cause of Dressen’s unpaid bills. It only had to be a necessary cause. The panel held that it was, so labeling the claim as contract did not take it outside the Act.

Congress Chose Immunity Anyway

Dressen and the district court pressed the consequence hard. If a covered entity can make a promise, obtain the participation that promise encouraged, break it, and then invoke immunity, the consent process shifts risk to the participant without securing the promised protection. The Tenth Circuit treated that consequence as legally rational rather than absurd. Under the absurdity doctrine, departure from statutory text is reserved for results so bizarre that Congress could scarcely have intended them. The panel instead saw a harsh but intelligible choice: emergency response depends on limiting the litigation exposure of those developing and administering countermeasures, and contract labels could otherwise become an easy way to repackage medical-malpractice or product-liability claims.

Medical duties have long traveled through both tort and contract, and product liability has long borrowed from warranty. A categorical contract exception could let pleading choices undo the comprehensive immunity Congress created.

The panel confronted Judge Shelby’s stipend hypothetical directly. Even if refusing the study-visit fee were immunized—a point the panel did not decide—the result would not be legally unthinkable. Congress may decide that the public interest in shutting off litigation is worth leaving even clear harms uncompensated.

The Act’s "sole exception" reinforced that reading. Congress allowed an action for death or serious physical injury proximately caused by willful misconduct, then imposed unusual barriers: particularized pleading, a complaint verified by the plaintiff under oath, an affidavit from a non-treating physician, certified medical records supporting proximate causation, and prior resort to the compensation process. Even willful misconduct causing lesser injury may remain immune. If Congress accepted that result, the panel reasoned, ordinary breach of contract could hardly lie beyond its contemplation.

"Perhaps a grant of immunity in such circumstances is bad public policy," the opinion said, "but it is congressional policy." Congress supplied the immunity. The consent form may show that AstraZeneca voluntarily gave up part of it by promising payment. That is the waiver question the district court must decide.

The Remedy Behind the Warning

The consent form paired its immunity warning with a reference to the Countermeasures Injury Compensation Program.

Congress created the CICP to compensate some people injured by covered countermeasures while the PREP Act protects manufacturers and others from suit. It can reimburse certain unreimbursed medical costs, replace some lost wages, and pay death benefits when a covered countermeasure directly causes serious injury or death. As a payer of last resort, it generally requires filing within one year of administration and provides no attorney-fee reimbursement. Its causal standard is unforgiving. HRSA demands compelling, reliable, valid medical and scientific evidence; temporal proximity alone cannot satisfy it. HRSA administrators decide the claims, and the available benefits cover far less than ordinary civil damages.

HRSA’s figures show the gulf between the statutory program and a contractual promise to pay reasonable treatment costs. As of July 1, 2026, the CICP had received 14,146 COVID-related claims and decided 7,584 of them. Of those decided, 112 were found eligible for compensation. Sixty-two had been paid. More than 2,600 were denied for missing the one-year deadline; more than 1,400 for failing the proof standard or a covered-injury requirement; more than 3,100 because the required medical records never arrived.

Those figures do not show whether any individual claim deserved approval, but they do show that only a small fraction of decided claims received payment.

The consent form promised a different source of payment. It said AstraZeneca carried insurance and would pay reasonable medical costs for research injuries when stated conditions were met. The CICP appeared separately—as a possibility that might provide compensation if immunity closed the courts. One sentence described what the government might give. Another described what the company said it would give.

If the company’s sentence means nothing unless it qualifies as a waiver, the enforceability of the consent form depends on a doctrine the form never named. The participant must grasp the medical risk, the existence of immunity, and whether the sponsor’s affirmative language was meant as a legally binding surrender of that immunity. A form meant to clarify risk left enforceability to a doctrine that took six years of litigation to surface.

The Remaining Question: Waiver

The Tenth Circuit reversed the district court’s categorical rule and stopped at waiver. Dressen had argued that AstraZeneca gave up PREP Act immunity through its words and its conduct. Because the district court had held that issue in reserve, the panel handed it back to Judge Shelby: whether this statutory immunity can be waived at all, and if so, whether the consent form did it.

Whether PREP Act immunity can be waived may depend on whose interests it protects. A right belonging chiefly to the covered party can usually be surrendered; a safeguard embedded in the national emergency response may resist private bargaining. The panel’s collateral-order analysis treated freedom from litigation itself as part of the national emergency response.

If the immunity can be waived, a court must then decide how much clarity waiver demands. Waiver ordinarily requires a clear surrender of statutory rights. The company will point to the form’s express warning that federal law may bar suit, and argue that no document carrying that warning can be read as a clear surrender of the same protection. Dressen will emphasize the promise’s specificity: AstraZeneca stated that it had insurance for research injuries and will pay reasonable medical costs on stated conditions. The promise does almost no work if the company retained an unqualified right to block enforcement the instant the promised contingency arrived.

Any reading must give effect to both clauses. The warning cannot be waved away, and "will pay" cannot be reduced to aspiration without explaining why a participant should have understood it that way.

If AstraZeneca assumed a duty while preserving immunity from suit, "will pay" may describe voluntary administration rather than an enforceable obligation. If the promise is enforceable, the company stepped out from behind the shield for the narrow class of costs it chose to cover. The district court must decide how clearly a sponsor must speak before a promise waives immunity. After the Tenth Circuit’s ruling, Dressen can enforce the promise only if the district court finds that AstraZeneca gave up its immunity.

What Future Forms Will Say

The decision’s practical consequences will appear first in future consent forms. If the district court holds that research-injury language like AstraZeneca’s waives PREP Act immunity, sponsors will rewrite their consent forms to keep the shield in express terms. They will separate voluntary benefit programs from contractual obligations, state that no promise is enforceable in court, condition payment on their own internal findings—or simply offer less.

That response could preserve the legal certainty Congress wanted while corroding the practical trust research depends on. Emergency studies rely on people willing to take risks before those risks are fully known. A form that speaks in maximal legal self-defense may be entirely accurate and still make participation harder to win.

If the language did not waive immunity, the form could state that AstraZeneca "will pay" and that insurance exists while leaving no judicial forum to test a refusal. Payment would then depend on insurers, internal processes, ethics review, and reputational pressure rather than contract enforcement.

The case exposes a drafting trap built into the emergency statute itself. Federal rules require greater-than-minimal-risk studies to explain whether compensation and treatment are available. The PREP Act may make the most affirmative version of that explanation legally dangerous for the sponsor: the clearer the promise, the stronger the case for waiver. Yet the harder the form disclaims enforceability, the emptier the assurance becomes for the person being asked to sign.

Future research-injury clauses will have to state whether the sponsor retains federal immunity or waives it.

The Remaining Waiver Question

The Tenth Circuit left breach, medical causation, ultimate enforceability, and waiver unresolved. It held that PREP Act immunity can encompass contract claims and that Dressen’s alleged losses had the required causal connection to the injection.

The PREP Act was written to encourage a rapid national response to an emergency by limiting civil liability. AstraZeneca’s consent form addressed a different decision: whether Dressen would receive an experimental vaccine after being told who would pay if the research injured her.

The Tenth Circuit held that PREP Act immunity covers Dressen’s contract claim. Judge Shelby must now decide whether AstraZeneca’s own words surrendered that protection. Dressen signed a form that both warned her she might not be able to sue and said the company would pay if the research injured her. Nearly six years later, the enforceability of that promise remains unresolved.