The fascinating, and ultimately frustrating, pair of decisions in Guinan v. A.I. duPont Hospital for Children, ___ F. Supp.2d ___, 2009 WL 307019 (E.D. Pa. Feb. 6, 2009) (“Guinan I“), and Guinan v. A.I. Dupont Hospital for Children, ___ F. Supp.2d ___, 2009 WL 311113 (E.D. Pa. Feb. 6, 2009) (“Guinan II“), demonstrate how judicial sympathy for a plaintiff attempting to pound a square liability peg into a round legal hole can cause serious damage to the law. If there had been a third opinion, maybe we’d call this a trilogy. As it is, we’ve only got a parody.
Guinan is another of those lawsuits in which what are really malpractice or informed consent cases have been stretched to try to make out product liability claims. Levine and Riegel come to mind as other examples, as we’ve discussed here (Levine) and here (Riegel). To some extent this phenomenon comes from plaintiffs’ lawyers looking to involve deeper pocketed defendants, something that’s been around since the beginning of legal time. But to a large extent we see it as something more recent – an unintended consequence of tort reform efforts that have made it more difficult and less remunerative for plaintiffs to pursue actions against medical professionals.
We don’t bear any grudges against the docs for that, but this does make things more difficult for our clients at times, and Guinan is one of those times.
Guinan involved some innovative surgery designed to correct a set of congenital heart defects (“tetralogy of Fallot,” “complete common atrioventricular canal defect,” and “severe pulmonary stenosis” – don’t ask us about these, we’re simply reporting what we’re reading) that frequently accompany Downs Syndrome. Guinan I, 2009 WL 307019, at *1. The patient was a new born. To deal with resultant circulatory problems, the doctors who ultimately ended up as defendants tried something called a “Fontan procedure.” That required two rounds of open heart surgery to complete. Unfortunately it couldn’t be completed because “severe leakage of [the] atrioventricular valves.” Id.
There was a dispute (decently prepped plaintiffs always deny informed consent) whether the parents were informed about this complication or the change in surgical plans that resulted. Id.
Instead of the normal method of surgical completion, the doctors inserted a medical device – a “covered stent” – via catheter. This “relatively new” procedure avoided a second open-heart procedure. Id. at *2. Once the device gets involved, we get interested.
And it does get interesting.
For one thing, the device that was implanted wasn’t FDA-approved – for anything. Id. The manufacturer had been making these stents available under a number of (it thought) exceptions to FDA approval: (1) as a “humanitarian use device” (for treatment of rare conditions), (2) as an “investigational” device for use in clinical trials, and (3) as a “custom device” made to fit the patient-specific “order of an individual physician.” Guinan II, 2009 WL 311113, at *2-3. This particular stent was supplied as a custom device. Id. at *3.
However the docs got their hands on this unapproved stent, it was supposed to be used in accordance with procedures overseen by the defendant’s hospital’s institutional review board (“IRB”). Remember, this is pediatric surgery, and very few companies sell either drugs or devices specifically intended and approved for use in children. Guinan I, 2009 WL 307019, at *2.
So in went the stent. It seemed to work for a little while, but the “cold” that the infant got about a month later turned out not to be a cold at all, but the onset of two more “rare and potentially life-threatening conditions”: “protein losing enteropathy” and “plastic bronchitis.” Id. at *2 (again, just because we blog about this stuff, don’t assume we know what it is).
At this point, plaintiffs changed doctors and hospitals – ending up at the best children’s hospital in the country. The new doctors decided they couldn’t risk removing the stent. But to do what they thought needed to be done (lower the venous blood pressure), they had to punch a hole in the stent, something called “fenestration” (not to be confused with “defenestration” of the sort the Czechs practiced in the Middle Ages). The new docs also found stenosis (narrowing) in another artery and implanted a second stent. Guinan I, 2009 WL 307019, at *2.
But the hole that the second set of doctors punched in the stent wasn’t big enough, because the plastic covering on the stent (the first one) started covering it back up. To deal with that, the docs (second set) decided to implant another stent (the third one) at the hole in the first stent. Id. at *3. Got that?
Didn’t work. The plastic bronchitis persisted. That’s where things stood at the time of the litigation. The only thing left to do, if anything could be done, would be to remove everything (except maybe the second stent, we’re not sure) and start over – something called a “take-down Fontan procedure.” Id. at *3.
So plaintiffs sue everybody. There are factual disputes all over the place, as the plaintiff parents claimed they weren’t told anything, and the defendant doctors swore they told the parents everything. One of these disputes was whether the doctors used an informed consent form provided by the defendant manufacturer, which explained what the first stent was all about. Id.
After the inevitable winnowing down of the plaintiffs’ extravagant pleading, including dismissal of class action allegations, id. at *4, all the defendants moved for summary judgment. The judge wrote two opinions, one involving the medical defendants (Guinan I) and the other involving the medical device manufacturer and its CEO (Guinan II). The court threw out a lot of claims but not all of them.
We’re going to go through Guinan I first, but quickly (yeah, right, we know you guys), because it’s necessary to understand Guinan II – the one we’re most interested in.
The remaining claims in Guinan I were malpractice (medical negligence), informed consent, fraud/misrepresentation, and medical monitoring.
The first question was choice of law. Plaintiffs were residents of New Jersey, and the complained-of surgery occurred in Delaware. That would be too easy, so plaintiffs sued in the Eastern District of Pennsylvania, making for a three-ringed circus. Guinan I, 2009 WL 307019, *5. There were bunches of real conflicts that we won’t bore our readers or ourselves with because they involve the medical claims. Suffice it to say that, if plaintiffs had wanted either Delaware or New Jersey law to apply, they would have sued there.
Because the state where the complained-of medical treatment occurred had the most meaningful contacts, the court applied Delaware law. Id. at *6-8.
That caused big problems with plaintiffs’ medical monitoring claim, because while New Jersey and Pennsylvania recognized medical monitoring in some situations, no Delaware court ever had. Worse than that, the Delaware Supreme Court more or less rejected medical monitoring as a separate cause of action in Mergenthaler v. Asbestos Corp., 480 A.2d 647, 651 (Del. 1984), an asbestos case (obviously), but that was as much because of lack of exposure as due to any evaluation of the viability of the underlying claim.
Moreover, Guinan I stated that not only was medical monitoring itself uncertain under Delaware law, but plaintiffs advocated a “novel theory that a tort claim for medical monitoring can be applied to medical procedures and devices.” 2009 WL 307019, at *7 (emphasis added).
We’ll skip over the lengthy discussion of the Delaware Health Care Malpractice Insurance and Litigation Act. Suffice it to say that the vague report filed by plaintiffs’ expert didn’t come close to supporting a claim for medical negligence. Id. at *9-14. Next up was informed consent. Because the plaintiff parents signed an informed consent form, their claims sounded in negligence. To prove negligence they needed the expert whose report had already been thrown out. Bye-bye informed consent. Id. at *14-15. Plaintiffs didn’t even get close to establishing fraud. Id. at *15.
So that left medical monitoring, and that’s where things go off the rails.
Plaintiffs had some good facts. Once they found out the first stent hadn’t been approved, they complained to the FDA. The FDA got on the IRB’s case and demanded a report. The IRB concluded that, in violation of procedures, the defendant doctors “had implanted an unapproved device in several patients, and had handled Plaintiff’s parents’ complaints poorly.” Id. at *16. After that, the FDA went to the device manufacturer (which it directly regulated) and “urged” (the court’s word) it to tell similarly situated patients that they should “seek long-term medical follow-up.” Id. The manufacturer not only did what the FDA suggested, but sent a letter to every doctor and hospital that had ever used one of its devices.
That letter (which came back to haunt everyone) stated that “[t]he safety and effectiveness of the C.P. stent is unknown at this time; therefore [the manufacturer] is concerned that patients receive appropriate follow-up care.” It went on to suggest that patients who had these stents implanted should be evaluated on an annual basis “until full growth is reached” and that the evaluation should “include a history and physical exam, a chest x-ray, an echocardiogram, an EKG, and an MRI or spiral CT.” Guinan I, 2009 WL 307019, at *16. There was no evidence (at least none mentioned in either opinion) that the manufacturer had any idea that the defendant doctors weren’t following the procedures established by their own hospital’s IRB.
So in the hallowed tradition that, in litigation, no good deed goes unpunished, Guinan I used this recommendation as the jumping off point in concluding that plaintiffs had a viable claim for medical monitoring.
Even though the Delaware Supreme Court had never recognized such a claim in any context.
Even though application of medical monitoring to medical procedures and products was admittedly “novel.”
The court, of course, had to make an Erie prediction (that means take a guess at what a state court might do with an undecided issue under principles first established in Erie Railroad Co. v. Tompkins, 304 U.S. 64 (1938)). We lawyers like to name our shorthand jargon after cases.
First, it reviewed the Mergenthaler decision and quite correctly determined that the Delaware Supreme Court’s dismissal of that particular monitoring claim wasn’t conclusive on whether such a claim could ever be brought. Guinan I, 2009 WL 307019, at *17-18. The court didn’t bother to delve further into Delaware precedent, thus it did not consider a second dismissal of a medical monitoring claim by a Delaware trial court. See Alderman v. Clean Earth, Inc., 2007 WL 1930664, at *2 (Del. Super. June 26, 2007).
Instead, Guinan I found “several considerations” that “militate[d] in favor” of medical monitoring in the case:
- “Plaintiff has a Class III medical device in her body.”
- “[I]t is undisputed that the device did not have premarket approval from the FDA at the time the Medical Defendants implanted it in Plaintiff and was thus considered an ‘adulterated’ device.”
- “Plaintiff had contact with the adulterated device[, which] remains in her body.”
- “[T]he FDA, [the manufacturer], and the Institutional Defendants have all suggested that Plaintiff should receive follow-up care to monitor the CP stent.”
2009 WL 307019, at *18. Relying on a number of decisions from other jurisdictions, the court concluded that “considerations of fairness, efficiency, and deterrence favor recognizing a cause of action for medical monitoring here.” Id.
It looks like there was simply no way that this injured baby was getting thrown out of court – not when one of the defendants’ own IRBs had admitted there was a screw up.
There’s just one little, slight problem with that.
It’s not allowed.
Not only haven’t Delaware state courts recognized medical monitoring, but expansion of medical monitoring to the use of prescription medical products is admittedly “novel.” Guinan I, 2009 WL 307019, at *7.
Under universally accepted Erie principles, it’s beyond the proper power of a federal court exercising diversity jurisdiction to predict a novel expansion of state-law liability because, as in Guinan I, the court thinks such an expansion is a better approach. We’ve discussed this before, but Guinan I is the most egregious violation of this principle we’ve seen, at least since the decisions that prompted our earlier post.
It’s also against the Third Circuit’s repeated admonitions to the district courts. Lexington National Insurance Corp. v. Ranger Insurance Co., 326 F.3d 416, 420 (3d Cir. 2003):
[A] federal court in a diversity case should be reluctant to expand the common law. . . . [T]here has to be a limit somewhere. Our duty here is to predict how the [state] Supreme Court would view this case and we will discharge that duty by holding that it would reject [plaintiff’s] claims.
Werwinski v. Ford Motor Co., 286 F.3d 661, 680 (3d Cir. 2002):
[E]ven if we were torn between two competing yet sensible interpretations of [state] law. . ., we should opt for the interpretation that restricts liability, rather than expands it, until the Supreme Court of [the state] decides differently.
City of Philadelphia v. Beretta U.S.A. Corp., 277 F.3d 415, 421 (3d Cir. 2002):
[I]t is not the role of a federal court to expand state law in ways not foreshadowed by state precedent. Instead, a federal court follows the precedents of the state’s highest court and predicts how that court would decide the issue presented. [State] precedent does not support the. . .claim plaintiffs advance here, and we cannot predict that the [state] Supreme Court will choose to expand state. . .law in the manner plaintiffs urge.
Camden County Board of Chosen Freeholders v. Beretta, U.S.A. Corp., 273 F.3d 536, 541-42 (3d Cir. 2001):
[This] is a matter of state law, and the role of a federal court ruling on a matter of state law in a diversity case is to follow the precedents of the state’s highest court and predict how that court would decide the issue presented. It is not the role of a federal court to expand or narrow state law in ways not foreshadowed by state precedent. . . . While it is of course conceivable that the Supreme Court. . .may someday choose to expand state  law in the manner that the County urges, we cannot predict at this time that it will do so.
Northview Motors, Inc. v. Chrysler Motors Corp., 227 F.3d 78, 92 n.7 (3d Cir. 2000):
[W]e have stated that a federal court in a diversity case should be reluctant to expand state common law. . . . [R]egardless of the source of its jurisdiction, the principle is valid because a federal court is applying state law.
Leo v. Kerr-McGee Chemical Corp., 37 F.3d 96, 101 (3d Cir. 1994):
[F]ederal courts may not engage in judicial activism. Federalism concerns require that we permit state courts to decide whether and to what extent they will expand state common law. . . . Our role is to apply the current law of the jurisdiction, and leave it undisturbed.
City of Philadelphia v. Lead Industries Ass’n, 994 F.2d 112, 123 (3d Cir. 1993):
A federal court in a diversity case is not free to engraft onto those state rules exceptions or modifications which may commend themselves to the federal court, but which have not commended themselves to the State in which the federal court sits. . . . In a diversity case. . .federal courts may not engage in judicial activism. Federalism concerns require that we permit state courts to decide whether and to what extent they will expand state common law. Our role is to apply the current law of the appropriate jurisdiction, and leave it undisturbed.
Bottom line, if plaintiffs in Guinan had wanted to argue for this sort of doubly novel expansion of Delaware law, they shouldn’t have played games with choice of law by filing in a federal forum having nothing to do with their cause of action. They should have filed in a Delaware state court. Basically, the Guinan plaintiffs played forum-shopping games – and the court ended up rewarding them for it.
So Guinan I leaves the plaintiffs hanging on by their fingernails against the medical defendants, able to pursue only a medical monitoring claim that has no basis in relevant state law. In Guinan II, the court decided to visit the sins of the medical defendants upon the device manufacturer.
But not before making some interesting and useful rulings concerning the plaintiffs’ other claims.
The rulings are interesting because the manufacturer was also investigated by the FDA. Even though undisputed evidence established that the manufacturer had not affirmatively promoted the use that supposedly injured the plaintiff, see Guinan II, 2009 WL 311113, at *3 (the doctors “approached” the manufacturer), the FDA charged it with “marketing an adulterated device” – presumably because it had run afoul of the “custom device” process by which it sold the device. Id. at *4. In the end, both the company and its principal officer pleaded guilty to misdemeanors. Id.
The stent, however, wasn’t found to be unsafe. Quite the contrary. As part of the manufacturer’s plea agreement, the FDA required it to fund a clinical trial for another pediatric use of the stent, and on top of that “to supply it to health care providers who request it for [the studied] treatment . . . free of charge.” Id.
The Agency would never have demanded, as part of a legal settlement, that the manufacturer distribute the device for free if it thought the device was actually dangerous. So from a product liability perspective, Guinan deals with claims brought against a convicted violator of the FDCA over the unapproved use of a safe product.
As to the manufacturer as well, the plaintiffs played the same choice-of-law games, only this time there was a fourth state involved – New York, where the manufacturer was incorporated.
So the possible choices were: (1) New Jersey, where the plaintiffs lived, (2) Delaware, where the doctors used the device in surgery, (3) New York, the state of incorporation (and presumably principal place of business), and Pennsylvania, where suit was filed. We’ll spare you the details, but the court held that the place where the medical treatment using the device occurred, rather than either the plaintiffs’ residence or the defendant’s principal place of business, had the greatest interest in the application of its laws. Guinan II, 2009 WL 311113, at *5. That is, except for plaintiff’s attempt at piercing the corporate veil, which, being a matter of corporate law, was governed by the state of incorporation. Id. at *6.
So that’s the first interesting and useful ruling – on choice of law. It’s yet another rejection of principal place of business as a dispositive contact, something we’ve denounced before.
Plaintiffs brought three types of negligence claims against the manufacturer: “failure to test, failure to warn, and negligence per se based on violation of the FDCA.” Id. at *6. They’re all interesting, but regular readers know of our near obsession with FDCA based negligence per se. Rather than clutter up the post with links, we direct your attention to the “negligence per se” label at the bottom of this post. By clicking on that, you’ll get a list of each and every one of our prior posts on that subject (even us Luddites recognize the value of these labels).
But unfortunately, Guinan II doesn’t get into the really juicy stuff.
That’s because, once again, the plaintiffs’ experts let them down. The first ground on which plaintiffs’ negligence claims failed was lack of expert testimony on causation. The expert “offers no opinion regarding the causal connection between” the “stent and Plaintiff’s claimed injuries.” Id. at *7. A win is a win is a win, but that’s not a rationale we find very interesting.
But things look up from there. The next cause of action is for fraud/misrepresentation. The first issue is whether the learned intermediary rule applies to fraud. The court held that it does:
This reasoning [for the learned intermediary rule] certainly makes sense in the context of this case, where Plaintiff could not have obtained independent access to the [device], where the Medical Defendants had to order the [device] from [the manufacturer] by prescription, and where the Medical Defendants inserted the [device] into Plaintiff during a medical procedure.
Guinan II, 2009 WL 311113, at *9.
Interesting and useful ruling number 2: the learned intermediary rule applies to fraud claims.
Plaintiffs then claimed that, because the manufacturer didn’t give the doctors any information at all about the safety of the device for use in the Fontan procedure, it was necessarily liable. Id. at *10. But that misstated the facts. Instead, the manufacturer (through its principal) “did not assure the doctors that the [device] was safe to use in the Fontan completion.” Id. The device was unapproved (with evidence establishing that its FDA status was stated on the informed consent form the company gave the doctors, id. at *11), and the doctors, not the manufacturer chose to it for this purpose. Therefore, no fraud.
We’ll skip the battery (informed consent) discussion in Guinan II because it dodged the interesting question (whether a manufacturer’s provision of an informed consent form could be a voluntary undertaking to obtain informed consent), and dismissed the claim for the same reasons as Guinan I. We’ll also skip strict liability, which was dismissed on the very Delaware-specific ground that the state’s law has never adopted the doctrine. 2009 WL 311113, at *13-16.
Delaware never adopted strict liability because the legislature had already provided warranty causes of action. Express warranty is usually the weak sister of the two, and was so here as well. That claim was dismissed because, while plaintiffs claimed they relied on the device being “custom” when in fact it was unapproved for anything, there was no evidence that they even knew of that description:
[W]e will grant summary judgment in favor of [the manufacturer] for the same reason: Plaintiff has not proffered evidence that the nature of the stent was “part of the basis of the bargain” to constitute an express warranty. To the extent that [the manufacturer] made any representations about the stent prior to the sale of the stent, Plaintiff was not aware of them. . . . Plaintiff’s argument that she relied on [the manufacturer’s] description of the stent as “custom” is without merit.
Guinan II, 2009 WL 311113, at *20.
Interesting and useful ruling No. 3: In express warranty, if the plaintiff was not told about the alleged express warranty by the prescribing doctor, the claim fails for lack of reliance/benefit of the bargain.
Implied warranty is where the real action is, since that claim substitutes for strict liability in Delaware. To establish a product defect, plaintiffs hauled out an FDA expert – not a design expert – who testified essentially that the device was “not safe for use in humans” because it had “not been shown to be safe.” Guinan II, 2009 WL 311113, at * 21. That, of course, was a blatant attempt to use FDA-speak to flip the common-law burden of proof from the plaintiff to the defendant. Whatever a manufacturer’s burden might be before the FDA, in a product liability action, it’s the plaintiff’s job to produce evidence affirmatively establishing a defect:
The expert testimony on which Plaintiff relies fails to support an inference that the stent was defective. . . . Plaintiff maintains that [expert] testimony that the [device] was not “shown to be unsafe” supports an inference that [it] was unsafe or defective. Plaintiff therefore relies on an absence of evidence of defect as support for the proposition that a defect exists. Plaintiff commits the logical fallacy of argumentum ad ignorantiam, that is, an argument from ignorance. . . . Plaintiff mistakenly assumes that an absence of evidence to support its proposition establishes the proposition. However, an absence of evidence that the [device] was defective does not establish defectiveness. Plaintiff’s logical fallacy underscores her failure to carry her burden of proffering evidence that allows an inference of defect.
Id. at *21 (emphasis added) (citations to five Bone Screw cases omitted).
And there’s more along these lines. The Guinan II court similarly rejects the proposition that lack of FDA approval for a use means that a product is not “fit” for that unapproved use:
Implicit in Plaintiff’s argument is the proposition that Fontan completion is an “ordinary purpose” for which the [device] is used. Plaintiff therefore relies on the [device’s] lack of FDA approval as evidence of its defective design or manufacture. A lack of FDA approval is not a per se defect of design or manufacture under Delaware law. . . .[T]he lack of FDA approval of the [device] here does not constitute a defect under the applicable UCC provision. Plaintiff has identified issues regarding FDA approval. . ., but Plaintiff has not pointed to evidence of a manufacturing or design defect in the stent itself . . . . The lack of FDA approval, without more, is not enough to create an inference of a manufacturing or design defect.
Id. at *22 (emphasis added) (citations to three more Bone Screw cases and a footnote citing one of Bexis’ articles omitted).
Interesting and useful ruling No. 4: Resolved: Lack of FDA approval is neither evidence of, nor supports an inference of, a defect.
To us that’s the most interesting and useful ruling in Guinan. We’ve seen plaintiffs try to flip the burden of proof like that quite a bit over the years, mostly in litigation over off-label uses, and the devastating refutation in Guinan II is as thorough and well-reasoned as we’ve ever seen – and we’ve done this issue to death.
We’ll pass over the discussion of the implied warranty of fitness for a particular purpose, id. at *23, and piercing the corporate veil, id. at *24-15, because they’re both dependent upon the case-specific factual record in the case. Also, we’re getting tired.
That brings us back to the kicker – medical monitoring. Once again, the court denies summary judgment. There’s no independent review of the claim, only a reference to the discussion in Guinan I, and a statement that the Erie prediction is “based. . .in large part on the fact that. . . the FDA [and all the defendants] have advised that medical monitoring is necessary because, inter alia, the safety of the [device] is unknown.” Guinan II, 2009 WL 311113, at *23.
Sorry, we don’t buy that in Guinan II any more than in Guinan I. We’ve already addressed the infirmity of the court’s Erie prediction, so we’ll close with the other question we have.
Where’s the tort?
There’s no negligence – that claim was dismissed for lack of any viable expert opinion on either liability or causation. There’s no product defect either. There’s no fraud. Despite the defendant’s regulatory transgressions, the FDA didn’t order the product off the market, and instead required the manufacturer to make it even more available than it had been.
There’s just no tortious conduct upon which to predicate medical monitoring – even if the court had the power to recognize such a novel cause of action in the first place.
And there’s no such thing as medical monitoring in the air. Every state we know of that allows medical monitoring requires, as an element of the claim, proof of tortious conduct. Between them, Guinan I and Guinan II cite medical monitoring cases from Pennsylvania and New Jersey. In Pennsylvania one of the seven essential elements of medical monitoring is an increased risk “caused by the defendant’s negligence.” Redland Soccer Club, Inc. v. Department of the Army, 696 A.2d 137, 145 (Pa. 1997). In New Jersey, medical monitoring is not allowed at all in product liability actions. Sinclair v. Merck & Co., 948 A.2d 587, 594-95 (N.J. 2008). See also, e.g., Potter v. Firestone Tire & Rubber Co., 863 P.2d 795, 823 (Cal. 1993) (“as a result of a defendant’s tortious conduct”); Meyer v. Fluor Corp., 220 S.W.3d 712, 717 (Mo. 2007) (“consequences of the defendant’s tortious conduct”); Hansen v. Mountain Fuel Supply Co., 858 P.2d 970, 979 (Utah 1993) (“which exposure was caused by the defendant’s negligence”); Bower v. Westinghouse Electric Corp., 522 S.E.2d 424, 432 (W. Va. 1999) (“through the tortious conduct of the defendant”); Petito v. A.H. Robins Co., 750 So.2d 103, 106 (Fla. App. 1999) (“caused by the defendant’s negligence”).
Thus, in the end, we have to conclude that the pair of Guinan decisions produced a parody of established medical monitoring law. Delaware hasn’t recognized medical monitoring at all, and certainly hasn’t applied it to drug/device product liability cases. A federal court sitting in diversity has no business making that kind of decision for a state. Beyond that, even if Delaware were to recognize this very controversial tort – it’s still a tort. Medical monitoring isn’t free health insurance – the law doesn’t make a product manufacturer pay for somebody’s medical tests simply because it admits that those tests might be a good idea. There has to be a tort – some form of wrongful conduct – and in Guinan there simply wasn’t. The court recognized that, by dismissing every other cause of action that the plaintiffs had, whether based on negligence, strict liability, warranty, or fraud, but decided to let the plaintiffs proceed anyway.
We think Justice Cardozo said it best:
The judge, even when [s/]he is free, is still not wholly free. [S/]He is not to innovate at pleasure. [S/]He is not a knight-errant roaming at will in pursuit of [her/]his own ideal of beauty or of goodness. [S/]He is to draw his[/her] inspiration from consecrated principles. [S/]He is not to yield to spasmodic sentiment, to vague and unregulated benevolence.
Benjamin N. Cardozo, “The Nature of the Judicial Process,” at 141 (1921) (sexist language from the period cleaned up).
Simply stated, the tort law should not give away defendants’ money like Christmas presents, and judges should not play Santa Claus.