dissenting:
Under the controlling law of Mississippi, “the operative time [the trigger of limitations] is when the patient can reasonably be held to have knowledge of the injury itself, the cause of the injury, and the causative relationship between the injury and the conduct of the medical practitioner.”1 The majority concludes that this medical malpractice case should never have gone to trial because the patient, on learning of her injury, should have connected her medical condition with a particular drug given over a course of treatment. In doing so, my colleagues misread the law and unwittingly sanction a trial stratagem of defendants. Respectfully, I dissent.
First there is underbrush to be put aside. Nothing supports any implication that Mississippi law allocates to the judge, and not the jury, all questions of when a patient should have known of a causative link between injury and conduct. Nor is there a basis for concluding that application of the statute always presents only a question of law. Indeed, the language of the limitations trigger in the Mississippi statute — “shall or with reasonable diligence might have first been known or discovered”2 — poses a fact question. And of course conceded knowledge of a patient that he has an injury and that he has been treated does not mean that the ultimate questions of reasonableness and diligence are questions of law.
There is another wrinkle here. For transparent reasons, defendants did not seek a jury determination of their limitations defense, preferring to relinquish all of the defense except a contention that the case was time-barred as a matter of law. It follows that the entry of judgement on the verdict is supported by an implied jury finding of any question of fact necessary to rejection of the defense of limitations.3
The failure of defendants to seek a jury determination on the limitations defense was no accident. The fight at trial was causation. Defendants maintained that the prescribing doctors could not have known that Terbutaline was a cause of Huss’s condition because there was no link between the two; the heart condition was idiopathic. It would have been awkward at best to also ask the jury to find that Huss, an untrained layperson, should have looked at Terbutaline as a culprit. Defendants elected to pitch to the jury their contention of no causation unimpeded by the conflicting contention that the patient should have known what the defendant doctors did not. And now, on appeal, defendants contend that there is insufficient evidence to support the finding of causation but, faced with conflicting medical opinions expressed to the jury, add the contention that in any event this layperson should have made the link, as a matter of law. This bold contention cannot be sustained under the decisions of the Supreme Court of Mississippi.
The majority relies primarily on Wright v. Quesnel, but that reliance is misplaced. Essentially, the pregnant patient in Wright exhibited troubling symptoms on two occasions, the doctor ordered bed rest both times, and then the child died in útero. Any reasonable person would connect the *210child’s death to the lack of treatment. Here, by contrast, Huss exhibited troubling symptoms of early contractions, the doctors treated those early contractions with Terbutaline (and other treatments), and then the early contractions ceased4 and she gave birth. And then she was diagnosed with cardiomyopathy. In other words, the treatment here did what it was supposed to do; the allegation is that it did more, which the Husses had to figure out. In Wright, the treatment (or lack thereof) did not do what it was supposed to do. And, of course, it is always easier to connect a total lack of treatment to an injury than which of multiple treatments caused an injury. The patient in Wright needed no further investigation; the Husses did.
The other cases cited by the majority are inapposite for similar reasons. In Powe v. Byrd, the doctor mis-diagnosed colon cancer as gastritis and hemorrhoids for two years. As in Wright, any reasonable person would have recognized negligence on the date injury (cancer) was discovered — after all, the plaintiff knew then that the original diagnosis was wrong and its resulting treatment more or less useless.5 In PPG Architectural Finishes, Inc. v. Lowery, where the plaintiff admitted that she knew “when, how, and by whom she had been injured on the night of her exposure” to paint fumes, the court merely held that limitations began to run that night — after she passed out, went to the emergency room, and the doctor’s “impression” was injury caused by paint fumes — not when the plaintiff later received “definitive medical confirmation” that her resulting brain damages was caused by the fumes.6 The difference between that case and ours needs no explanation. In Wayne General Hospital v. Hayes, the plaintiffs had enough information when their child’s death certificate indicated “sepsis” and the child had been hospitalized a second time after the negligence occurred. The court implicitly held that sepsis is strongly correlated with negligence, leading any diagnosis of sepsis to cause sufficient suspicion.7 Cardiomyopa-thy is different, and a reasonable person would need more than diagnosis of it to suspect negligence.
I do not see Erie as simply a rule controlling the source of substantive law where federal jurisdiction rests on diversity of citizenship. Rather, it is an order of constitutional magnitude allocating state and federal power, reinforced by strands of separation of powers, as Professor Mishkin has explained.8 Hence my disagreement with my colleagues goes beyond principled readings of decisions by a state supreme court in discharge of our Erie duty. The legislature of the State of Mississippi, responding to political forces conflicted over the need of doctors to be free of frivolous and stale claims and the protection of patients from bad medicine, enacted the statute at hand. This statute, with its interpretations by the Supreme Court of Mississippi, are the sources from *211which the question before us today must find its answer. We have a reticulated pattern of cases, some of which find no fact issues where it is undisputed that a patient knows both that he is injured and must as a reasonable man then know that the arrow points to the doctor. Errors in diagnosis — for example, you have a fatal heart attack after being told it’s only indigestion — are of this genre. And then there are fact patterns which can sustain differing results.
But what we do not have from the Mississippi courts is the bright line rule announced by the majority — limitations is triggered by knowledge of injury alone, the causative component dropping away because limitations is triggered “as a matter of law.” The majority does this because it says that these facts permit no conclusion but that limitations commenced to run when the patient learned she had a major heart problem because she should have then immediately focused on Terbu-taline, even though the treating doctors swear that the drug had nothing to do with her condition and, significantly, that the cause of her injury was unknowable. This is a fundamental restatement of the balance struck by the Mississippi legislature. Whether it is good policy or needed reform by some abstract measure of social justice I do not know. But it is not in the constitutive order of things for three Texas judges, even as well-intentioned and able as my colleagues, to decide for the State of Mississippi. At the least we should be most hesitant to trim the role of the jury without much clearer direction from the state.
Contrary to the majority’s statement, I do not maintain that limitations here does not “commence to run unless and until causation is established at trial by a jury’s finding.” Rather, I maintain that a jury should have decided the fact question presented by the limitations defense in this case — or, rather, that it was a fact question wisely dropped by defendants given their defense that causation was unknowable by anyone. In many, perhaps most, cases, there are plausible causation arguments on both sides. In such cases, defendants can argue lack of causation and, simultaneously, argue that plaintiffs, under their own theory of causation, should have connected injury to negligence at a certain date, triggering limitations. Here, however, the defense was that causation was unknowable by anyone, doctor or layman, because the injury was idiopathic. Perhaps the Husses nonetheless should have suspected what the doctors say they never suspected. But to conclude as a matter of law that they should have writes a powerful new policy for the State of Mississippi — painless tort reform by decree, not ballot.
. Sarris v. Smith, 782 So.2d 721, 723 (Miss.2001) (quoting Smith v. Sanders, 485 So.2d 1051, 1052 (Miss.1986)).
. Miss.Code. Ann. § 15-1-36(2).
. See Fed. R. Civ. P. 49.
. There is some evidence that Terbutaline did not fully stop the early contractions. That is irrelevant, since the allegation is not that Ter-butaline harmed the child or did not cease the contractions.
. 892 So.2d at 227-28; see also Gentry v. Wallace, 606 So.2d 1117, 1119 (Miss.1992) (similar). Moreover, Powe found the plaintiffs' limitations argument "disingenuous” because the plaintiff's first petition alleging negligence, dismissed for lack of process, was filed before the date plaintiffs were arguing the clock began to run. 892 So.2d at 228.
. 909 So.2d at 48-52.
. 868 So.2d at 1005.
. See generally Paul Mishkin, Some Further Last Words on Erie the Thread, 87 Harv. L.Rev. 1682 (1974).