332 F.3d 479
Lynne STOCKBERGER, both personally and as the representative of Maurice Stockberger, deceased, Plaintiff-Appellant,
UNITED STATES of America, Defendant-Appellee.
United States Court of Appeals, Seventh Circuit.
Argued April 16, 2003.
Decided June 11, 2003.
Stephen L. Williams (argued), Mann, Chaney, Johnson, Goodwin & Williams, Terre Haute, IN, for Plaintiff-Appellant.
Jeffrey L. Hunter (argued), Office of U.S. Atty., Indianapolis, IN, for Defendant-Appellee.
Before POSNER, COFFEY, and ROVNER, Circuit Judges.
POSNER, Circuit Judge.
The plaintiff in this suit under the Federal Tort Claims Act appeals from the grant of summary judgment to the United States. Maurice Stockberger, an employee of the federal prison at Terre Haute, Indiana, was an insulin-dependent diabetic and known to be such by his coworkers — many of whom, indeed, were medical workers. He had hypoglycemic episodes (episodes in which his blood sugar would fall to dangerously low levels), observed by and known to be such by them, in which he would exhibit personality changes, becoming hostile, suspicious, unresponsive, agitated — and sometimes denying that he had a medical problem. When his coworkers noticed that he was in one of his hypoglycemic states, they would urge him to eat, or to drink Ensure, a nutritious liquid food substitute. On the day of his death, one of his coworkers noticed that Stockberger, who was complaining about feeling ill and said that he wanted to go home, was having one of his hypoglycemic episodes, and offered him Ensure, which he drank. This made him feel better but he said he still wanted to go home. His coworkers wanted him to remain at the prison "until he recovered," but he was adamant about leaving. The coworker who had given him Ensure thought that Stockberger was in no condition to be driving, but he did not offer to drive Stockberger or try to take away his car keys; nor did he try to contact Stockberger's supervisor or wife. The prison had often in the past provided transportation for sick employees, including diabetic employees — including in fact Stockberger. But it had no written policies concerning the treatment of sick employees.
Stockberger got into his pick-up truck and began driving home. He drove very erratically, no doubt because of his hypoglycemia, veering off the road and then back onto it, knocking down traffic signs, and eventually colliding with a tree. His truck burst into flames when it hit the tree, and he died.
The plaintiff makes two separate claims. The first is that the federal prison system was negligent in failing to have a policy of providing transportation for employees who become dangerously ill at work. This claim is clearly barred by the discretionary-function exception to the tort claims act. 28 U.S.C. § 2680(a). The question how far an employer should go in providing medical assistance for employees who become ill at work involves an exercise of judgment (concerning for example the responsibility of subordinate employees, such as Stockberger's coworkers, to evaluate symptoms and report to supervisors or the prison doctor) rather than the straightforward, unarguable application of settled principles of tort responsibility. See Fang v. United States, 140 F.3d 1238, 1242 (9th Cir.1998); cf. Williams v. United States, 242 F.3d 169, 175 (4th Cir.2001); Kiehn v. United States, 984 F.2d 1100, 1106-07 (10th Cir.1993).
The plaintiff's second and more substantial claim is that the prison's action (or rather inaction) in allowing Stockberger to drive in his hypoglycemic condition was a breach of the duty of care imposed by Indiana tort law, the law that, in accordance with the Federal Tort Claims Act, furnishes the rule of decision for the plaintiff's claim. The claim invites consideration of the broader question of the tort duty if any to rescue a person in distress. The common law traditionally took a hard line, rejecting any legal duty to be a good Samaritan. If A saw that B was about to be struck on the head by a flowerpot thrown from a tenth-story window, and A knew that B was unaware of the impending catastrophe and also knew that he could save B with a shout, yet he did nothing and as a result B was killed, still, A's inaction, though gratuitous (there was no risk or other nontrivial cost to A) and even reprehensible, would not be actionable. E.g., Mullin v. Municipal City of South Bend, 639 N.E.2d 278, 284 (Ind. 1994); Hurley v. Eddingfield, 156 Ind. 416, 59 N.E. 1058 (1901); Zelig v. County of Los Angeles, 27 Cal.4th 1112, 119 Cal. Rptr.2d 709, 45 P.3d 1171, 1182-83 (2002); City of Douglasville v. Queen, 270 Ga. 770, 514 S.E.2d 195, 198-99 (1999); Rhodes v. Illinois Central Gulf R.R., 172 Ill.2d 213, 216 Ill.Dec. 703, 665 N.E.2d 1260, 1270 (1996); Harper v. Herman, 499 N.W.2d 472 (Minn.1993); Yania v. Bigan, 397 Pa. 316, 155 A.2d 343 (1959); Richard A. Epstein, Torts § 11.3, p. 290 (1999). The common law rule has been changed in some states, see, e.g., Vt. Stat. Ann. tit. 12, § 519(a); Melvin A. Eisenberg, "The Duty to Rescue in Contract Law," 71 Fordham L.Rev. 647, 653-54 (2002), but not in Indiana. Statutory modifications of the common law rule are common, such as requiring a driver who has caused an accident to remain at the scene even if he was not culpable, see, e.g., Fuentes v. Reilly, 590 F.2d 509 (3d Cir.1979) (N.J.law); Brooks v. E.J. Willig Truck Transportation Co., 40 Cal.2d 669, 255 P.2d 802 (1953), or forbidding a hospital emergency room to turn away a patient brought to it before his condition has been stabilized. Thomas v. Christ Hospital, 328 F.3d 890, 893 (7th Cir.2003). But none is applicable to this case.
Various rationales have been offered for the seemingly hardhearted common law rule: people should not count on nonprofessionals for rescue; the circle of potentially liable nonrescuers would be difficult to draw (suppose a person is drowning and no one on the crowded beach makes an effort to save him — should all be liable?); altruism makes the problem a small one and liability might actually reduce the number of altruistic rescues by depriving people of credit for altruism (how would they prove they hadn't acted under threat of legal liability?); people would be deterred by threat of liability from putting themselves in a position where they might be called upon to attempt a rescue, especially since a failed rescue might under settled common law principles give rise to liability, on the theory that a clumsy rescue attempt may have interfered with a competent rescue by someone else. E.g., Jackson v. City of Joliet, 715 F.2d 1200, 1202-03 (7th Cir.1983) (Illinois law); Farwell v. Keaton, 396 Mich. 281, 240 N.W.2d 217 (1976); Stiver v. Parker, 975 F.2d 261, 272 (6th Cir.1992) (also Michigan law).
Whatever the validity of these explanations for the common law rule, they have been held to be overborne in three types of case. The three types are typically said to involve a "special relationship" between rescuer and victim, e.g., Donaldson v. Young Women's Christian Ass'n, 539 N.W.2d 789, 792 (Minn.1995), but that seems to us an unhelpful label because of its vagueness.
The first type of case is where the rescuer had either assumed, explicitly or implicitly, a contractual duty to rescue the victim, e.g., Mastriano v. Blyer, 779 A.2d 951, 955 (Me.2001); Nickelson v. Mall of America Co., 593 N.W.2d 723, 726-27 (Minn.1999); Folsom v. Burger King, 135 Wash.2d 658, 958 P.2d 301, 311 (1998); Randolph's Administrator v. Snyder, 139 Ky. 159, 129 S.W. 562 (1910), or had created in the victim a reasonable expectation that he had assumed such a duty. E.g., Municipal City of South Bend, supra, 639 N.E.2d at 284-85; Board of Commissioners v. Hatton, 427 N.E.2d 696 (Ind.App. 1981); Nickelson v. Mall of America Co., supra, 593 N.W.2d at 726; Stiver v. Parker, supra, 975 F.2d at 271-72; Wilmington General Hospital v. Manlove, 174 A.2d 135 (Del.1961); 3 Fowler V. Harper, Fleming James, Jr. & Oscar S. Gray, The Law of Torts § 18.6, p. 717 (2d ed.1986).
In the second type of case, the victim was in the rescuer's custody and thus without access to alternative rescuers. E.g., Norman v. Turkey Run Community School Corp., 274 Ind. 310, 411 N.E.2d 614 (1980); Donaldson v. Young Women's Christian Ass'n, supra; Murdock v. City of Keene, 137 N.H. 70, 623 A.2d 755, 756-57 (1993); Beach v. University of Utah, 726 P.2d 413, 415 (Utah 1986); Continental Southern Lines, Inc. v. Robertson, 241 Miss. 796, 133 So.2d 543 (1961); Fagg's Administrator v. Louisville & Nashville R.R., 111 Ky. 30, 63 S.W. 580 (1901). Typical cases of this type are ones in which the victim is a prison inmate or a patient in a mental hospital. E.g., Overall v. State, 525 N.E.2d 1275 (Ind.App.1988); Iglesias v. Wells, 441 N.E.2d 1017 (Ind.App.1982); Murdock v. City of Keene, 137 N.H. 70, 623 A.2d 755, 756-57 (1993); Salazar v. City of Chicago, 940 F.2d 233, 237 (7th Cir.1991); Clements v. Swedish Hospital, 252 Minn. 1, 89 N.W.2d 162, 165-66 (1958). These cases are readily assimilated to cases of the first type through the concept of an implicit contractual duty.
The third class consists of cases in which the victim's peril had been caused by the putative rescuer himself — even if he had caused it nonnegligently, e.g., Palace Bar, Inc. v. Fearnot, 269 Ind. 405, 381 N.E.2d 858, 865-66 (1978); L.S. Ayres & Co. v. Hicks, 220 Ind. 86, 40 N.E.2d 334 (1942); Tippecanoe Loan & Trust Co. v. Cleveland Cininnati Chicago & St. Louis Ry., 57 Ind.App. 644, 104 N.E. 866 (1914); South v. National Railroad Passenger Corp., 290 N.W.2d 819 (N.D.1980); Zylka v. Leikvoll, 274 Minn. 435, 144 N.W.2d 358 (1966); Thomas v. Casey, 49 Wash.2d 14, 297 P.2d 614 (1956); Montgomery v. National Convoy & Trucking Co., 186 S.C. 167, 195 S.E. 247 (1938); Maldonado v. Southern Pacific Transportation Co., 129 Ariz. 165, 629 P.2d 1001 (App.1981); Scatena v. Pittsburgh & New England Trucking, 2 Mass. App.Ct. 683, 319 N.E.2d 730 (1974), but a fortiori if he had caused it negligently or otherwise culpably. E.g., Carlisle v. Kanaywer, 24 Cal.App.3d 587, 101 Cal.Rptr. 246 (1972).
Another exception to the common law rule, though not described as such in the cases, is the doctrine of last clear chance, which requires a potential injurer to take measures to avert a peril that it has created, even if nonnegligently. E.g., Kumkumian v. City of New York, 305 N.Y. 167, 111 N.E.2d 865 (1953). Indiana has replaced contributory negligence with comparative negligence and in the course of doing so has discarded the last clear chance doctrine, Miller v. Ryan, 706 N.E.2d 244, 249 (Ind.App.1999), but that does not affect the principle. The last clear chance cases are a subset of the peril-caused-by-rescuer cases.
In short (and setting to one side the doctrine of last clear chance), when the rescuer either has assumed explicitly or implicitly a duty of rescue, or has caused the injury, the reasons behind the common law rule fall away and the rule is bent. And thus had the prison promised to protect Stockberger from the consequences of hypoglycemia, or induced his hypoglycemic episode by denying him access to insulin, or ordered him off the premises knowing that he was in no condition to drive (thus putting him in a zone of danger), the plaintiff would have a good case. But none of these things is true. The fact that the prison had sometimes accommodated the needs of a sick employee by providing him with transportation home or to a hospital did not create a contractual duty or reasonable expectation that a hypoglycemic employee would be restrained against his wishes from driving home. Stockberger was adamant that he wanted to go home and intended to do so the same way he had gotten to work, namely by driving his truck. His coworkers may have been negligent in failing to restrain him, just as A in our flowerpot hypothetical was negligent in failing to shout, where negligence is determined in Hand formula terms by comparing the burden of precautions (close to zero in both cases) with the harm from failure to take them discounted (multiplied) by the probability of injury. See, e.g., United States v. Carroll Towing Co., 159 F.2d 169, 173 (2d Cir.1947); Bammerlin v. Navistar Int'l Transportation Corp., 30 F.3d 898, 902 (7th Cir.1994); Brotherhood Shipping Co. v. St. Paul Fire & Marine Ins. Co., 985 F.2d 323, 327-29 (7th Cir.1993); Dobson v. Louisiana Power & Light Co., 567 So.2d 569, 574-75 (La.1990). Probably the burden to Stockberger's coworkers of delaying his departure was less than the risk of a serious accident. The fact that he wanted to go home need not have weighed seriously in the calculus of negligence. His thinking may have been impaired and his coworkers may have known this on the basis of their experience with his previous hypoglycemic episodes, in which he had exhibited irrational thinking.
For the plaintiff to prevail, however, the exceptions to the rule that there is no "good Samaritan" liability would have to be enlarged, to encompass the case in which an employee becomes ill at the workplace for reasons unrelated to his work and the employer fails to use due care to treat the illness. The exception could not be cabined so; its logic would embrace the common situation in which a customer becomes ill in a store, or a guest in a private home. Indeed, either would be the stronger case for imposing a duty of rescue. Employees learn about conditions (including employer provisions for their safety) on the job to an extent that a shopper is unlikely to learn about the dangers of the stores in which he shops or a guest the dangers of his host's home. Employees, moreover, can negotiate for more safety or demand compensation for risks; and to add to wages that reflect risks compensation in tort if the risks materialize would be double counting.
The rule toward which Stockberger's claim gestures is that the owner of premises has a duty of care toward any invitee in peril, even if the invitee's illness or injury was in no wise caused or aggravated by the owner or his employees. Such a rule could be defended by reference to hypothetical-contract analysis — by asking whether employees, customers, and guests have a reasonable expectation that if they suddenly collapse on the premises of their employer, business invitor, or host, they will not be left to die; that the employer, business invitor, and host have at least a duty to dial 911. The custodial situation in the case just put, as in a prisoner case, thins the ranks of potential rescuers. And the reasons that support the common law rule are attenuated in these settings, although it could be argued that the employer, business invitor, or social host who is burdened with a duty to care for a sick employee, etc., will be reluctant to invite chronically ill people, such as diabetics; we have also suggested reasons for treating employees differently from other invitees.
Hypothetical-contract analysis is a powerful tool for understanding tort law and determining its scope. It is easy to imagine that if drivers and pedestrians, say, could contract with regard to safety, they would agree that drivers would take cost-justified measures to avoid hitting pedestrians and pedestrians would take cost-justified measures to avoid being hit; for that is the form of contract that would minimize all relevant costs — the costs of accidents and the costs of avoiding accidents. And it is possible that such an analysis would lead to the conclusion that when an invitee suddenly becomes helpless and in peril on the premises of his invitor, the invitor has a duty to take at least minimal steps to save him, since that is the solution that minimizes all the relevant costs. This conclusion might make as much sense as the other exceptions to the common law's rejection of good Samaritan liability, all of which by the way can be rationalized on similar hypothetical-contract grounds — for example, one can easily imagine a contract between potential victims and potential injurers whereby the latter would agree to assist the victim even if the injury itself was not culpable; for the injurer will usually be in the best, and often in the only, position to minimize the costs of the injury to the victim.
Indiana, however, has not yet taken the step of imposing good Samaritan liability on invitors, L.S. Ayres & Co. v. Hicks, supra; J.A.W. v. Roberts, 627 N.E.2d 802, 809 (Ind.App.1994), though some other states have. See, e.g., Pridgen v. Boston Housing Authority, 364 Mass. 696, 308 N.E.2d 467, 475-78 (1974); Southern Pacific Co. v. Hendricks, 85 Ariz. 373, 339 P.2d 731, 733 (1959); Carey v. Davis, 190 Iowa 720, 180 N.W. 889 (1921); Tiedeman v. Morgan, 435 N.W.2d 86 (Minn.App. 1989); W. Page Keeton et al., Prosser and Keeton on the Law of Torts § 56, p. 376 (5th ed.1984); see also Hutchinson v. Dickie, 162 F.2d 103 (6th Cir.1947) (admiralty). A dictum in the Roberts opinion instances the relation of landowner to invitee as a "special relationship" that would take a case out of the common law rule, but without elaboration beyond a citation to a case that involved not rescue but the landowner's liability for a dangerous condition that injured a social guest. Burrell v. Meads, 569 N.E.2d 637 (Ind.1991).
Since cases of the type illustrated by the present case arise but seldom, and since the plaintiff had no choice but to bring her suit in a federal court — she had no opportunity to ask an Indiana court to join the trend toward imposing good Samaritan liability on invitors, because a suit under the Federal Tort Claims Act can be brought only in federal court — we cannot refuse the difficult challenge of predicting whether Indiana will join the bandwagon when a suitable case presents itself. This is not a case in which a plaintiff who has a choice between suing in a state court or a federal court chooses the federal court and asks it to adopt a venturesome interpretation of state law, as in Great Central Ins. Co. v. Insurance Services Office, Inc., 74 F.3d 778, 786 (7th Cir.1996).
But the plaintiff in her briefs and at argument has not tried to persuade us that Indiana would adopt the emergent rule. Instead she has tried without success to fit her case into existing Indiana case law. Even if we were so confident of the soundness or inevitability of the rule as to be willing to predict its adoption by Indiana, its application to situations in which due care requires restraining a person's freedom of movement would have to be very cautious, probably too cautious to give the plaintiff in this case any relief. The cases that we have cited that imposed liability on invitors for failures to rescue involved extreme situations, such as, in Hutchinson v. Dickie, the failure of a yachtsman to try to rescue a guest who fell overboard. It would not be sensible either to place employers or other invitors on a razor's edge where they face a suit for false imprisonment if they don't let the ill person leave or a suit for negligence if they do, or to turn employers and other invitors into nannies required, for example, to take the car keys away from an employee or customer or social guest because the roads are icy and the individual is known to be an unskillful driver. Stockberger was emphatic in wanting to drive himself home, said that he was feeling better and appeared to be, and could be presumed (for his diabetes was of long standing) to have known his fitness to drive as well as his skeptical coworker did. Of course in such a case there might be a legal duty to persons injured by the employee's foreseeable accident, Myers v. Quesenberry, 144 Cal.App.3d 888, 193 Cal.Rptr. 733 (1983), but such a case would not present an issue of good Samaritan liability.