Seventh Circuit: Failure-to-Warn Causation Requires Non-Speculative Proof That an Alternative Label Would Have Changed Medical Feeding Decisions

I. Introduction

This bellwether appeal arises from multidistrict litigation concerning cow’s-milk-based preterm infant formulas and necrotizing enterocolitis (“NEC”), a severe gastrointestinal disease affecting premature infants. Ericka Mar, as administratrix of the Estate of RaiLee Mar, sued Abbott Laboratories after RaiLee—born 12 weeks prematurely—was fed Similac Special Care 24 when maternal milk became unavailable and the hospital had no donor milk. RaiLee was diagnosed with NEC the next day and died shortly thereafter.

The central issue on appeal was not whether NEC is associated with formula feeding, but whether—under West Virginia failure-to-warn law—Mar produced sufficient evidence that Abbott’s alleged labeling deficiency caused RaiLee’s death. Mar proposed an alternative warning: “human milk has a lower risk of NEC than formula.” The Seventh Circuit affirmed summary judgment for Abbott, holding that Mar’s causation theory depended on speculative counterfactuals about hospital and parental decisions.

II. Summary of the Opinion

The Seventh Circuit affirmed the district court’s grant of summary judgment to Abbott on Mar’s failure-to-warn claim and also affirmed denial of Mar’s Rule 59(e) motion.

• Failure-to-warn causation: Even assuming Abbott should have used Mar’s proposed warning, Mar did not show that the alternative warning would have “made a difference” in preventing RaiLee’s death, as required under West Virginia law.
• Speculative counterfactuals rejected: The court held that Mar’s theories (hospital would create/accelerate a donor-milk program; hospital would accept informal milk sharing contrary to policy; Mar would have acted differently) were either contradicted by the record or too speculative.
• Warning-reading requirement: Because Mar did not read the existing warning, the court applied West Virginia authority requiring plaintiffs to have read the warning to show causation based on an alternative label.
• Rule 59(e): The court upheld denial of reconsideration because the “new” witnesses were available earlier and, in any event, would not have changed the outcome.

III. Analysis

A. Precedents Cited

1) Standards at summary judgment: permissible vs. speculative inferences

• Lewis v. Ind. Dep't of Transp. — Cited for de novo review and the obligation to view facts in the non-movant’s favor. The court nonetheless emphasized that this deference is bounded by reasonableness.
• FKFJ, Inc. v. Village of Worth and Osborn v. JAB Mgmt. Servs., Inc. — Used to distinguish reasonable inferences from those supported only by “speculation or conjecture.” These cases underwrote the court’s refusal to bridge evidentiary gaps about what a hospital board, staff, or parents would have done in a counterfactual world.
• In re Greenpoint Tactical Income Fund LLC — Quoted for the proposition that courts “will not draw inferences that are supported only by speculation,” reinforcing the court’s treatment of Mar’s donor-milk program theory as conjectural.
• Waukegan Potawatomi Casino, LLC v. City of Waukegan — Invoked for the idea that a “string of inferences” cannot defeat summary judgment; it framed Mar’s causal chain as multi-link and unsupported at crucial steps.
• Waldridge v. Am. Hoechst Corp. — Acknowledged the general caution that summary judgment is not a substitute for trial, but the court used it as a transition point to explain why trial is not warranted when the proffered inferences are too speculative.

2) West Virginia failure-to-warn framework

• Morningstar v. Black and Decker Mfg. Co. — Cited for the foundational West Virginia principle that manufacturers may be liable for distributing products without adequate warnings.
• Church v. Wesson — Supplied the first element: it must be “reasonably foreseeable” to the manufacturer that the product would be “unreasonably dangerous if distributed without a warning.”
• Tracy v. Cottrell ex rel. — Provided the second, dispositive element here: the failure to warn must be cause in fact; another warning must have “made a difference.” The court treated this “made a difference” requirement as the controlling causation lens for evaluating Mar’s counterfactuals.

3) The role and limits of counterfactual evidence

• Groh v. Ramirez — Cited for Mar’s broad proposition that reasonable inferences belong to the jury. The court did not reject that principle; it limited it by demanding that the inference be reasonable and grounded in record evidence.
• Cloutier v. GoJet Airlines, LLC — The court’s main comparator on counterfactual proof. It read Cloutier to allow juries to decide counterfactuals only when they are “sufficiently grounded in the record” and “absent concrete evidence to the contrary.” Here, unlike Cloutier, Abbott had “strong evidence” undermining Mar’s proposed alternative sequence of events (existing physician knowledge, institutional barriers, and unchanged policies).

4) Warning-reading requirement under West Virginia law

• Howard v. Eaton Corp., Meade v. Parsley, and Shanklin v. Allis-Chalmers Mfg. — Cited for the proposition that West Virginia law requires a plaintiff to have read the existing warning to establish causation in a failure-to-warn case. This requirement defeated Mar’s argument that she would have acted differently (e.g., refused formula or kept using blood-tainted milk) had the label been different.

5) Rule 59(e) reconsideration principles

• Reilly v. Will Cnty. Sheriff's Off. — Provided the abuse-of-discretion standard of review for the Rule 59(e) ruling.
• Oto v. Metro. Life Ins. — Supplied the Rule 59(e) standard: the movant must show “newly discovered evidence” or “manifest error of law or fact,” and may not use reconsideration to introduce evidence that could have been presented earlier. This precedent controlled the outcome because Mar’s witnesses were available during discovery.

B. Legal Reasoning

1) Causation hinged on “made a difference,” not on general risk information

The court assumed, for purposes of its causation analysis, that Mar’s proposed warning was “comparatively more effective and adequate.” Even so, West Virginia law (as quoted from Tracy v. Cottrell ex rel.) required Mar to show the alternative warning would have “made a difference” to RaiLee’s outcome. The court treated this as a practical, decision-focused inquiry: would the alternative label have altered what decisionmakers (clinicians, hospital administrators, or parents) did in time to avoid the harmful exposure?

2) Donor-milk program theory failed on institutional reality and contemporaneous physician knowledge

Mar’s first causal path asserted that the label would have spurred the hospital to obtain donor milk by implementing a donor-milk program. The court rejected the inference as speculative because:

• Dr. Maxwell’s full testimony emphasized donor milk implementation was “very expensive,” involved multiple logistical hurdles, and required hospital board/foundation approval—making it implausible that a label would produce an immediate program in the short window relevant to RaiLee’s feeding.
• Critically, the record showed RaiLee’s physicians already knew of the NEC/formula association from the American Academy of Pediatrics guidance and medical literature. The court reasoned that if that knowledge did not produce donor milk availability in 2014, it was not reasonable to infer that Abbott’s label would have done so.

This reasoning illustrates the opinion’s core move: it evaluates “difference-making” causation against real-world institutional constraints and what the relevant actors already knew, rather than treating a warning as automatically behavior-changing.

3) Informal milk-sharing theory failed because policy barriers were not shown to be malleable

Mar’s second path contended that a different warning would have led clinicians to accept unpasteurized, untested breast milk offered by another patient—contrary to hospital policy designed to prevent disease transmission (e.g., HIV, hepatitis). The court held Mar produced no evidence that a label would have changed the hospital’s policy. Without such evidence, the court treated the causal theory as conjecture.

4) Parental choice theory failed because the plaintiff did not read the warning

Mar’s third path posited she would have refused formula and refused to discard blood-tainted milk. The court rejected this because Mar did not read the existing warning; under West Virginia authority (as cited), that defeats causation based on label content. Put differently, an alternative warning cannot be said to “make a difference” to a person who did not read the warning in the first place.

5) Counterfactuals are permissible, but only when the record supports them and the opponent lacks concrete rebuttal

Responding to Mar’s reliance on Cloutier v. GoJet Airlines, LLC, the court clarified its approach: counterfactual inferences may go to the jury only when sufficiently grounded and not undermined by concrete contrary evidence. Here, Abbott pointed to multiple record facts undermining Mar’s hypotheticals—physicians’ preexisting knowledge, time/resource barriers, administrative approvals, and fixed safety policies—so Mar’s counterfactuals could not survive summary judgment.

6) The court avoided deciding broader medical causation (“but for” cause of NEC)

Abbott urged the court to go further and hold Mar lacked proof that the formula was a “but for” cause of NEC. The court declined, noting the issue could depend on measurement of RaiLee’s feeding schedule. Instead, it affirmed on narrower grounds: irrespective of the formula’s causal role in NEC, Mar failed to show a different warning would have prevented the death.

C. Impact

• Higher evidentiary demands for warning-causation in medically constrained settings: In cases where clinicians had limited or no practical alternatives, the decision signals that plaintiffs must produce concrete evidence of an available substitute course of care (or a realistic pathway to it) that would have been implemented in time.
• Institutional decisionmaking matters: The opinion underscores that hospitals are not single-actor decisionmakers; board approval, funding, logistics, and policy constraints can break the causation chain unless plaintiffs can show those constraints would likely have been overcome.
• Preexisting knowledge can defeat “label would have changed practice” theories: Where physicians already knew the relevant risk relationship from medical literature or guidelines, plaintiffs may face difficulty showing a manufacturer warning would have changed treatment.
• Reinforcement of the warning-reading requirement under West Virginia law: The ruling strengthens defendants’ ability to challenge causation where plaintiffs did not read the label—especially in medical contexts where patients may not interact with packaging at all.
• MDL bellwether significance: As a bellwether affirmance, the decision may shape settlement valuations and litigation strategy in the broader preterm infant formula/NEC docket by highlighting a potent summary-judgment pathway focused on “difference-making” causation.

IV. Complex Concepts Simplified

• Failure to warn: A claim that a product is defective because adequate warnings/instructions were not provided. Liability typically requires both foreseeability of danger without a warning and proof that the warning deficiency caused the injury.
• Cause in fact / “made a difference”: A practical causation test: would the injury likely have been avoided if the warning were different? The court treated this as requiring more than possibility—there must be record support for the alternative sequence of actions.
• Counterfactual inference: An inference about what would have happened in an alternative scenario (here, what clinicians/hospital/parents would have done if the label said something else). Courts permit these only when grounded in evidence, not speculation.
• Summary judgment: A pretrial ruling where the court decides there is no genuine dispute of material fact for a jury. Even if facts are viewed favorably to the non-movant, speculative theories cannot create a triable issue.
• Rule 59(e) motion: A post-judgment request to alter or amend the judgment. It is not a “do-over” for evidence that could have been submitted earlier; it generally requires truly newly discovered evidence or a clear legal/factual error.
• Bellwether case: A test case in multidistrict litigation used to inform parties about how juries and courts may treat recurring issues, often influencing settlement dynamics.

V. Conclusion

Ericka Mar v. Abbott Laboratories clarifies—through the lens of West Virginia law applied in federal court—that failure-to-warn claims live or die on non-speculative proof of “difference-making” causation. The Seventh Circuit required Mar to connect a proposed alternative warning to an evidence-supported change in real-world decisions by clinicians, hospital administrators, or parents within the relevant timeframe. Because the record showed significant institutional barriers, preexisting physician knowledge of the risk association, unchanged hospital policy on milk-sharing, and a warning-reading defect as to Mar, the court deemed the causal chain too conjectural to reach a jury. The opinion’s broader significance lies in its disciplined treatment of counterfactuals and its insistence that warning-based liability requires more than generalized risk information—it requires proof that the warning would have changed outcomes.