Rodas v. Seidlin

656 F.3d 610, 2011 U.S. App. LEXIS 18082, 2011 WL 3836460

• Court: Court of Appeals for the Seventh Circuit
• Decided: August 31, 2011
• Docket: 09-3760
• Status: Published

Opinion

FLAUM, Circuit Judge.

Gloria Rodas’s appeal presents an important question about the meaning of a provision of the Illinois Good Samaritan Act, 745 ILCS 49/1 et seq., in a case that was removed to federal court under the federal officer removal statute, 28 U.S.C. § 1442. The state-law question is whether and under what circumstances the protections of the Good Samaritan Act turn on the business model physicians use to charge patients for emergency services. The law shields from liability physicians who render certain services “without fee.” Rodas was charged a fee for the services she received, but two of Rodas’s physicians were not paid from that fee. Instead, they received a salary from their medical practice. The district court agreed with the defendants that the salary-versus-fee distinction cloaked the physicians with statutory immunity. Rodas appeals.

The United States, a party to the underlying action but not to this appeal, has filed a brief as amicus curiae urging that we lack jurisdiction. The argument is that the doctrine of derivative jurisdiction created a latent jurisdictional defect that persists on appeal. The doctrine provides that a federal court acquires no jurisdiction upon removal where the state court lacked jurisdiction over the subject matter or the parties. That principle, the government maintains, robbed the district court of subject matter jurisdiction and deprives us of appellate jurisdiction.

We disagree with both the government’s position about our jurisdiction and the defendants’ interpretation of the Illinois statute. As to jurisdiction, we join every other circuit to have considered the question and hold that the doctrine of derivative jurisdiction is cabined by the principles announced in Grubbs v. Gen. Elec. Credit Corp., 405 U.S. 699, 92 S.Ct. 1344, 31 L.Ed.2d 612 (1972), and Caterpillar Inc. v. Lewis, 519 U.S. 61, 117 S.Ct. 467, 136 L.Ed.2d 437 (1996). As to the merits, we conclude that the defendants’ interpretation of the Good Samaritan Act is inconsistent with the plain language of the statute and the case law interpreting it. Therefore, we reverse the judgment of the district court and remand for further proceedings.

I. Background

In 2001, Gloria Rodas had been receiving prenatal care at the Crusaders Central Clinic Association (“Crusader Clinic”) in Rockford, Illinois. The Crusader Clinic is a community health center and recipient of federal funds under 42 U.S.C. § 254b. Doctor William Baxter was one of the physicians who provided care; he was a family practice physician and clinic employee. In a limited sense, however, Baxter had two masters: because of the relationship between the Crusader Clinic and the federal government, both the clinic and Baxter were deemed to be employees of the United States for purposes of medical malpractice liability. See 42 U.S.C. § 233(g)-(n). As a consequence of Baxter and the Crusader Clinic’s (deemed) federal status, the United States could be substituted as a party if either were ever sued. 28 U.S.C. 2679(d)(1). Claims against them would be governed by the Federal Tort Claims Act, and neither would face liability. See 28 U.S.C. § 2679(b)(1); Osborn v. Haley, 549 U.S. 225, 229-30, 127 S.Ct. 881, 166 L.Ed.2d 819 (2007) (explaining the statutory interplay that produces this result). The statutory scheme is designed to allow certain health care providers serving underserved populations to save money on malpractice insurance and direct funds toward patient care. See H.R.Rep. No. 104-398, at 6, 1995 U.S.C.C.A.N. 767 (1995).

Baxter was not the only physician providing care to Rodas and other Crusader Clinic patients. For medically complex situations, the Crusader Clinic contracted with the University of Illinois College of Medicine (“UIC”) to provide specialist services. The arrangement called for UIC obstetrician/gynecologists to provide care at the Crusader Clinic, to provide on-call backup to clinic doctors, and to be on call when clinic patients were treated at local hospitals. In exchange, the Crusader Clinic paid UIC a fixed amount each year. The UIC physicians would fill out billing forms for their work and submit them to the Crusader Clinic; the Crusader Clinic had the right to collect the fees.

The two salaried UIC physicians implicated in this appeal are Doctor Ana-Maria Soleanicov and Doctor John Seidlin. Unlike Baxter, they were not deemed to be employees of the United States, and so they generally faced the specter of individual liability for any medical malpractice they committed. The alleged malpractice in this case occurred on August 2, 2001. During that day’s pre-dawn hours, Rodas went into labor. Following previously provided instructions from the Crusader Clinic, she went to the delivery floor at Swedish American Hospital. There, she was seen by a UIC resident until Baxter — our deemed federal employee — arrived at the hospital. Baxter took over care for Rodas, along with another UIC resident. Baxter’s understanding was that Soleanicov and Seidlin would be available to assist if needed. And they were needed, because Rodas’s delivery became an emergency. After about eight hours of labor, the baby’s heart tones had dropped, Rodas was not pushing effectively, and the baby was not descending. Soleanicov and Seidlin arrived to provide assistance. After multiple unsuccessful attempts to deliver the baby using a vacuum extractor and then forceps, Andrea Rodas was delivered via Cesarean section. She would not live very long. Less than two weeks later, she died from hypoxic ischemic encephalopathy, a condition in which the brain does not receive sufficient oxygen. Soleanicov prepared a bill for her work. Seidlin, deviating from his common practice, did not. The Crusader Clinic was reimbursed for the services Soleanicov provided.

Gloria Rodas filed a tort suit in state court in 2003, naming as defendants Baxter, Soleanicov, Seidlin, the Crusader Clinic, and Swedish American Hospital. The complaint alleged that the defendants negligently managed Rodas’s labor in a variety of ways, resulting in Andrea’s death. The United States removed the case to federal court, substituting itself as a party in place of the Crusader Clinic and Baxter. See 42 U.S.C. § 233(c) (removal); 28 U.S.C. § 2679(d)(1) (substitution). Because the action was against the United States and could proceed only under the Federal Tort Claims Act, see 28 U.S.C. § 2679(b)(1), Rodas was required to seek relief directly from the appropriate federal agency before filing suit, 28 U.S.C. § 2675(a). Therefore, she voluntarily dismissed her claims against the United States and the remainder of the case was remanded to state court. Rodas v. Swedish Am. Health Sys. Corp., 3:03-cv-50483, ECF No. 4 (N.D.Ill. Nov. 19, 2003). Rodas then filed an administrative claim with the United States Department of Health and Human Services.

The agency denied Rodas’s claim for damages, ruling that she had not established that the death of Andrea was caused by a negligent act or omission of a federal employee acting within the scope of em ployment. The letter informed Rodas that, if dissatisfied with the determination, she could appeal within the agency hierarchy or file suit against the United States “in the appropriate federal district court within six (6) months from the date of mailing this determination (28 U.S.C. § 2401(b)).” Rodas, however, did not seek to return to federal court.

Instead of returning to federal court, Rodas amended her state-court complaint against Swedish American Hospital, Soleanicov, and Seidlin. And instead of re-adding Baxter and the Crusader Clinic as defendants to that complaint, she named the United States directly (along with Soleanicov and Seidlin). The United States again removed the case to federal court, this time invoking in its notice of removal the federal officer removal statute, 28 U.S.C. § 1442. Section 1442(a)(1) provides that a “civil action ... commenced in a State court” may be removed to federal court if the action is against “[t]he United States or any agency thereof or any officer (or any person acting under that officer) of the United States or of any agency thereof, sued in an official or individual capacity for any act under color of such office.... ” After removal, the case proceeded through the motion to dismiss phase and discovery was taken. By the time defendants Seidlin and Soleanicov separately moved for summary judgment in June 2008, the case’s second life in federal court was more than three years old.

Seidlin and Soleanicov filed their summary judgment motions in late May and early June 2008. Both contended that they were shielded from liability by the Illinois Good Samaritan Act provision governing emergency care. The Act shields from liability physicians “who, in good faith, provide[] emergency care without fee to a person.” 745 ILCS 49/25. The district court agreed that Soleanicov and Seidlin fell within the Act’s ambit because fees from Crusader Clinic patients did not go directly to them. Expanding on dicta from the Illinois Appellate Court’s decision in Estate of Heanue v. Edgcomb, 355 Ill.App.3d 645, 291 Ill.Dec. 537, 823 N.E.2d 1123 (2005), the district court reasoned that Soleanicov and Seidlin could not possibly have provided services for a fee within the contemplation of the Good Samaritan Act. Rodas v. SwedishAm. Health Sys. Corp., 594 F.Supp.2d 1033, 1042 (N.D.Ill.2009) (“Put simply, a doctor’s act of submitting documentation of the services rendered to the billing department for Crusader Clinic differs drastically from, and does not equate with, what the Illinois appellate courts consider charging a ‘fee’ for purposes of the Good Samaritan Act.”). The district court entered judgment under Rule 54(b) of the Federal Rules of Civil Procedure. Rodas filed a notice of appeal with us.

After judgment in favor of Soleanicov and Seidlin was entered, the United States moved to dismiss the case, contending that the doctrine of derivative jurisdiction deprived the court of subject matter jurisdiction over the case. The district court entered an indicative ruling, Fed.R.Civ.P. 62.1(a), denying the government’s motion. (The case against the United States proceeds below.) The district court reasoned that although the doctrine of derivative jurisdiction applies to removals under the federal officer removal statute, the government had not properly removed on that basis. In fact, removal had been effected under 28 U.S.C. § 2679(d)(2) and 42 U.S.C. § 233(c). Rodas v. SwedishAm. Health Sys. Corp., 2010 WL 4386678, at *3-4 (N.D.Ill. Oct. 29, 2010). The district court assumed that the doctrine of derivative jurisdiction would not ordinarily sap a court of jurisdiction for removals under those provisions, but held in a footnote that the Supreme Court’s decision in Grubbs would shield its judgment even if the doctrine applied. Id. at *4 & n. 2.

In response to the district court’s ruling, the United States filed a brief with us as a friend of the court, raising the alleged jurisdictional defect. The parties have responded to the government’s arguments in their merits briefs, generally echoing the district court’s reasoning.

II. Discussion

Summary judgment rulings, like matters of subject matter jurisdiction, are subject to de novo review. Kaplan v. United States, 133 F.3d 469, 472-73 (7th Cir.1998). We reach the jurisdictional issue first, as “[w]ithout jurisdiction the court cannot proceed at all in any cause. Jurisdiction is the power to declare the law, and when it ceases to exist, the only function remaining to the court is that of announcing the fact and dismissing the cause.” Steel Co. v. Citizens for a Better Environment, 523 U.S. 83, 94, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998) (quoting Ex parte McCardle, 74 U.S. (7 Wall.) 506, 514, 19 L.Ed. 264 (1868)). These are our conclusions: the criteria for removal under 28 U.S.C. § 1442(a)(1) were satisfied and no one has questioned that the doctrine of derivative jurisdiction applies to such removals. Although the doctrine applies to removals under Section 1442(a)(1), the teachings from the Supreme Court’s decisions in Grubbs and Caterpillar limit the doctrine’s application. Every circuit to have considered the question has so held, and we decline the government’s implicit invitation to create a circuit split.

Proceeding to the merits, the defendants misread the language of the Illinois Good Samaritan Act and overstate the implications of the Illinois Appellate Court’s decision in Heanue. Doctor Soleanicov rendered her emergency services for a fee and thus does not enjoy statutory immunity under the Good Samaritan Act. Although Seidlin did not charge a fee, there is a genuine issue of material fact about whether he did so in “good faith,” a predicate to the Act’s liability shield. See 745 ILCS 49/25; Heanue, 291 Ill.Dec. 537, 823 N.E.2d at 1129 (a doctor who has not billed in order to trigger the Act has not acted in good faith and does not enjoy the Act’s protections).

A. (Derivative) Jurisdiction

Before reaching the merits of this appeal, we must address a jurisdictional issue the United States has raised. When a case is removed from state to federal court, the jurisdiction of the latter is said “in a limited sense” to derive from the former. Edwards v. United States Dep’t of Justice, 43 F.3d 312, 316 (7th Cir.1994) (quoting Minnesota v. United States, 305 U.S. 382, 389, 59 S.Ct. 292, 83 L.Ed. 235 (1939)). Accordingly, “[wjhere the state court lacks jurisdiction of the subject matter or of the parties, the federal court acquires none, although in a like suit originally brought in federal court it would have had jurisdiction.” Id.; see also Gen. Inv. Co. v. Lake Shore & M.S. Ry. Co., 260 U.S. 261, 288, 43 S.Ct. 106, 67 L.Ed. 244 (1922) (‘When a cause is removed from a state court into a federal court, the latter takes it as it stood in the former. A want of jurisdiction in the state court is not cured by the removal, but may be asserted after it is consummated.”); Lambert Run Coal Co. v. Baltimore & O.R. Co., 258 U.S. 377, 382, 42 S.Ct. 349, 66 L.Ed. 671 (1922) (reciting the same language as appears in Edwards and Minnesota above). This doctrine has been referred to as the doctrine of derivative jurisdiction. E.g., Palmer v. City Nat’l Bank, of West Virginia, 498 F.3d 236, 239 (4th Cir.2007). More than one commentator has observed that the justification for the rule is hardly obvious, but that makes it no less entrenched — or binding on us. Compare, e.g., Erwin Chemerinsky, Federal Jurisdiction § 5.5, at 288 (1989) (“This rule has little justification, but has long existed.”), with Arizona v. Manypenny, 451 U.S. 232, 242 n. 17, 101 S.Ct. 1657, 68 L.Ed.2d 58 (1981) (characterizing the doctrine as “well settled”).

The government’s argument can be succinctly summarized: By its terms the doctrine of derivative jurisdiction applies to this case. Rodas commenced her action against the United States in state court, but federal sovereign immunity deprived the state court of subject matter jurisdiction. See Hercules, Inc. v. United States, 516 U.S. 417, 422, 116 S.Ct. 981, 134 L.Ed.2d 47 (1996) (“The United States, as sovereign, is immune from suit save as it consents to be sued.”) (alterations and quotation marks omitted). The United States has waived its sovereign immunity to tort liability only under the Federal Tort Claims Act, which vests exclusive jurisdiction over such claims in federal court. 28 U.S.C. § 1346(b)(1). Moreover, the doctrine of derivative jurisdiction has been abrogated in the general removal statute, see 28 U.S.C. § 1441(f), but not the federal officer removal statute, 28 U.S.C. § 1442(a)(1), which was invoked in this case. Contrary to the district court’s ruling, removal under Section 1442(a)(1) was proper. Therefore, the doctrine of derivative jurisdiction deprived the district court of subject matter jurisdiction over the case, and we should vacate the judgment and instruct the district court to dismiss the United States and remand the claims against Soleanicov and Seidlin to state court. 16 Moore’s Federal Practice § 107.15[l][b], at 107-130.21-130.22 (Matthew Bender 3d ed.2009) (describing the proper procedure). The course is appropriate because pendent party supplemental jurisdiction, see 28 U.S.C. § 1367(a), provides the only basis for hearing the claims at issue in this appeal. Powerex Corp. v. Reliant Energy Servs., 551 U.S. 224, 234-35, 127 S.Ct. 2411, 168 L.Ed.2d 112 (2007) (stating that it is “far from clear” that federal courts would have subject matter jurisdiction on these facts); Rifkin v. Bear Steams & Co., Inc., 248 F.3d 628, 634 (7th Cir.2001) (no subject matter jurisdiction over supplemental claims that were not attached to claims over which the federal court has original jurisdiction). Moreover, the judgment should not be spared by the Supreme Court’s teachings in Grubbs and Caterpillar, because those cases do not apply to jurisdictional defects. See Grupo Dataflux v. Atlas Global Group, L.P., 541 U.S. 567, 574, 124 S.Ct. 1920, 158 L.Ed.2d 866 (2004) (“The resulting holding of Caterpillar ... is only that a statutory defect ... did not require dismissal once there was no longer any jurisdictional defect.”).

The government’s argument has considerable surface appeal, and most of it is without flaw. For our purposes, the only questions are (1) whether removal was proper under Section 1442, such that the doctrine of derivative jurisdiction applies; and (2) whether the doctrine of derivative jurisdiction creates a defect in subject matter jurisdiction that evades the teachings of Grubbs and Caterpillar. We take up each question in turn.

1. Federal officer removal and the doctrine of derivative jurisdiction.

Removal under 28 U.S.C. § 1442(a)(1) was proper. Section 1442 is known as the federal officer removal statute. Section 1442(a)(1) provides that “[a] civil action commenced in a State court against ... [t]he United States or any agency thereof or any officer (or any person acting under that officer) of the United States or of any agency thereof, sued in an official or individual capacity for any act under color of such office” may be removed to federal court. The provision is an exception to the “well-pleaded com plaint” rule, which provides that for non-diversity cases to be removable, the complaint must establish that the case arises under federal law. Kircher v. Putnam Funds Trust, 547 U.S. 633, 644 n. 12, 126 S.Ct. 2145, 165 L.Ed.2d 92 (2006). The forebear of Section 1442(a)(1) was enacted near the end of the War of 1812, in “an attempt to protect federal officers from interference by hostile state courts.” Watson v. Philip Morris Companies, Inc., 551 U.S. 142, 147, 127 S.Ct. 2301, 168 L.Ed.2d 42 (2007). One of its primary purposes, the history “clearly demonstrates,” was to have federal officers’ defenses to state-law actions litigated in federal courts. Willingham v. Morgan, 395 U.S. 402, 407, 89 S.Ct. 1813, 23 L.Ed.2d 396 (1969). “In cases like this one, Congress has decided that federal officers, and indeed the Federal Government itself, require the protection of a federal forum.” Id. at 407, 89 S.Ct. 1813.

At the outset, we observe that the United States generally has the authority to invoke Section 1442(a)(1). When the modern version of the statute was enacted in 1948, it originally allowed for removal only by federal officers themselves. That was the holding of Int’l Primate Protection League v. Admins, of Tulane Educ. Fund, 500 U.S. 72, 79-80, 111 S.Ct. 1700, 114 L.Ed.2d 134 (1991), and in response Congress amended the law to its current form, which allows the United States or one of its agencies to remove to federal court. See also S. Rep. 104-366, at 24, 1996 U.S.C.C.A.N. 4202, 4210 (1996) (commenting on the change and stating that it “fulfills Congress’ intent that questions concerning the exercise of Federal authority, the scope of Federal immunity and Federal-State conflicts be adjudicated in Federal court”). The only question is whether the federal officer removal statute does not apply because the government was not originally named in Rodas’s state-court complaint.

The district court concluded that timing is everything under Section 1442 and that the United States could not invoke the provision because it was not originally named as a defendant in the state-court complaint. We disagree with the district court’s analysis. The district court attached temporal significance to the proviso that Section 1442(a)(1) applies to a “civil action ... commenced in a State court against ” the United States (emphasis added). Here, the case was filed in state court, naming only individuals, removed to federal court, and then remanded to state court. Only then was the complaint amended to add the United States as a defendant.

We do not accept the district courts’s construction of Section 1442(a)(1) for several reasons. First, it requires an unorthodox reading of the provision to conclude that Congress was focused on the identity of named defendants at the time the state suit was commenced; the most natural reading of Section 1442 is that Congress was concerned where the civil action originated. Several other removal provisions also indicate, sensibly if not always in consistent language, that they apply to actions that start out in state court. See 28 U.S.C. §§ 1441(a), 1443, 1444. Had Congress intended to focus on the identity of parties at the time the suit was filed, it would have been simple enough for it to do so. Hamilton v. Lanning, — U.S. -, 130 S.Ct. 2464, 2474, 177 L.Ed.2d 23 (2010) (“[W]e would expect that, had Congress intended ... a specialized — and indeed, unusual— meaning ..., Congress would have said so expressly.”). The plain language of the statute, the best indication of Congressional intent, counsels against the district court’s interpretation. E.g., Smith v. City of Jackson, Mississippi, 544 U.S. 228, 249, 125 S.Ct. 1536, 161 L.Ed.2d 410 (2005) (discussing the plain language of a statute in terms of its “natural reading”).

The district court’s interpretation of Section 1442 is further weakened by the structure of Chapter 89 of Title 28, United States Code, which contains Section 1442 and several other provisions relating to removal. A separate provision in Chapter 89 already specifies the proper course to follow when an action that is not initially removable subsequently satisfies the criteria for removability. Specifically, 28 U.S.C. § 1446(b) provides: “If the case stated by the initial pleading is not removable, a notice of removal may be filed within thirty days after receipt by the defendant ... of a copy of an amended pleading....” See also Caterpillar, 519 U.S. at 69, 117 S.Ct. 467 (stating the timing rules governing removal of amended complaints). Given that Section 1446 sets out the procedures to follow when a civil action does not at first meet the criteria for removal, there are scant grounds to conclude that Congress included a different trigger to govern removals under Section 1442 and that it did so in such a cryptic manner. “Congress, we have held, does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions — it does not, one might say, hide elephants in mouseholes.” Whitman v. Am. Trucking Ass’ns, Inc., 531 U.S. 457, 468, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001).

Finally, our reading of the text and structure of the Act are confirmed by Section 1442’s purpose and history. The Supreme Court has often stated that the policy behind the federal removal statute — ensuring that federal defenses raised by federal actors are evaluated in a federal forum — “should not be frustrated by a narrow, grudging interpretation” of the provision. Willingham, 395 U.S. at 407, 89 S.Ct. 1813; see also Jefferson County, Alabama v. Acker, 527 U.S. 423, 431, 119 S.Ct. 2069, 144 L.Ed.2d 408 (1999). Under the district court’s interpretation of the statute, a person could file suit in state court against private actors. Shortly thereafter, the person could amend the complaint to add the federal officer (or agency, or the United States). It is unlikely that Congress, animated by an approximately 200-year-old concern that the contours of federal power be determined by federal courts, would have intended such an obvious end-around. Judge Friendly reminds us that “Section 1442 ... vindicates ... the interest of government itself; upon the principle that it embodies ‘may depend the possibility of the general government’s preserving its own existence.’ ” Bradford v. Harding, 284 F.2d 307, 310 (2d Cir.1960) (quoting State of Tennessee v. Davis, 100 U.S. 257, 25 L.Ed. 648 (1880)). Nothing about the federal interest turns on whether the plaintiff names a federal actor in his first complaint or in an amended pleading.

In sum, the purpose and history of the statute confirm what its text and structure tell us and counsel against the district court’s interpretation of Section 1442. As to the doctrine of derivative jurisdiction, the parties agree that it applies to removals under this provision. The doctrine provides a background rule against which all of the removal statutes operate; it applies unless abrogated. See Palmer, 498 F.3d at 239. Congress has specifically abrogated the doctrine only with respect to removals under the general removal statute. See 28 U.S.C. § 1441(f). 1 Yet, the doctrine has been criticized a great deal over the course of many years. E.g., Washington v. Am. League of Prof'l Base ball Clubs, 460 F.2d 654, 658-59 (9th Cir.1972) (“This is the kind of legal tour de force that most laymen cannot understand, particularly in a case where the federal court not only has subject matter jurisdiction, but has exclusive subject matter jurisdiction.”); 14B Charles Alan Wright, et al., Federal Practice and Procedure § 3722, at 339-42 (4th ed.2009) (recounting some of the criticism and characterizing it as “deserved”). Given that Congress explicitly abrogated the doctrine of derivative jurisdiction only with respect to removals under Section 1441, it supports the notion that — for whatever reasons — Congress intended to keep the doctrine in place with regard to other removal provisions. See Bob Jones Univ. v. United States, 461 U.S. 574, 599, 103 S.Ct. 2017, 76 L.Ed.2d 157 (1983) (failure of Congress to amend a statute, despite the fact that “adherents advocating contrary views [to an agency’s interpretation of a statute had] ventilated the subject for well over three decades” supported the conclusion that Congress acquiesced to the agency’s interpretation); Palmer, 498 F.3d at 245. For that reason, the district court’s hypothecated alternative grounds for removal do not escape the doctrine of derivative jurisdiction. The district court did not predicate its jurisdiction on Section 1441, nor did either of the statutes it relied on contain language abrogating the doctrine of derivative jurisdiction.

Having determined that the preconditions for removal under Section 1442(a)(1) were satisfied and that the doctrine of derivative jurisdiction applies to such removals, the next question is whether the doctrine is subject to limiting principles.

2. Limitations on the doctrine derivative jurisdiction.

The doctrine of derivative jurisdiction, despite its perhaps improvident name, is best understood as a procedural bar to the exercise of federal judicial power. That is, the doctrine creates a defect in removal, but is not an essential ingredient to federal subject matter jurisdiction. Because the district court would have had jurisdiction over a hypothetical complaint filed at the time it entered the judgment now under review, the fact that the state court lacked jurisdiction over the case when it was removed has no significance. To explain these conclusions, a word about the principles announced in Grubbs and Caterpillar is in order.

In Grubbs, the Supreme Court held that where a district court had “jurisdiction of the parties at the time it entered judgment ... the validity of the removal procedure followed may not be raised for the first time on appeal.” 405 U.S. at 700, 92 S.Ct. 1344. The General Electric Credit Corp. had initiated suit against Grubbs in Texas state court, seeking recovery on a promissory note. Grubbs filed a “cross-action” against General Electric Credit Corp. and also brought the United States, which had a judgment lien against Grubbs, into the case by adding it as a party defendant. The United States removed the case to federal district court, relying on 28 U.S.C. § 1444, which provides for removal of certain foreclosure actions where the United States is a defendant, and added claims of its own. See also 28 U.S.C. § 2410(a) (providing that the United States may be named as a defendant in certain state-law property actions where the United States “has or claims a mortgage or other lien” on the property). All parties litigated the case as if jurisdiction in the district court were proper, and the case proceeded to a bench trial. At the conclusion of the trial, the district court ruled against General Electric Credit Corp. on the promissory note claim and in favor of Grubbs on his counter-claim. The district court dismissed the claims brought by and against the United States. Grubbs, 405 U.S. at 701-02, 92 S.Ct. 1344.

On appeal, the Fifth Circuit held that the priority of the United States’ judgment lien against Grubbs provided a “spurious” basis for removal under 28 U.S.C. § 1444. Because removal had not been authorized, the Fifth Circuit concluded that there was no other basis for the district court’s jurisdiction over the action and the cause had to be remanded to state court. The Supreme Court reversed. It did not take issue with the appellate court’s fix on the removal provision at issue but concluded that it had no bearing on the appeal. “Long-standing decisions of this Court make clear ... that where after removal a case is tried on the merits without objection and the federal court enters judgment, the issue in subsequent proceedings on appeal is not whether the case was properly removed, but whether the federal district court would have had original jurisdiction of the case had it been filed in that court.” Grubbs, 405 U.S. at 702, 92 S.Ct. 1344.

In concluding that jurisdiction was proper on appeal, the court relied on its earlier teachings from Baggs v. Martin, 179 U.S. 206, 21 S.Ct. 109, 45 L.Ed. 155 (1900), and Mackay v. Uinta Dev. Co., 229 U.S. 173, 33 S.Ct. 638, 57 L.Ed. 1138 (1913). In both of those cases, removal had been improperly effected, but the district court could have exercised original jurisdiction had the case been filed in federal court. In Baggs, the lower court would have exercised federal question jurisdiction; in Mackay, the lower court would have exercised diversity jurisdiction. See 179 U.S. at 209, 21 S.Ct. 109, 229 U.S. at 176-77, 33 S.Ct. 638. Applying the teachings to the facts of the case, the Grubbs Court observed that Grubbs and General Electric Credit Corp. were of diverse citizenship. “There was thus diversity jurisdiction in the Federal District Court under 28 U.S.C. § 1332 if the action had been brought in that court originally.” Grubbs, 405 U.S. at 704, 92 S.Ct. 1344. Our Circuit and others have since applied the teachings of Grubbs to cases decided at summary judgment. Gentek Bldg. Prods., Inc. v. Sherwin-Williams Co., 491 F.3d 320, 327 (6th Cir.2007) (equating summary judgment with a trial on the merits); Lively v. Wild Oats Mkts., Inc., 456 F.3d 933, 941 n. 11 (9th Cir.2006) (Grubbs applies to decisions on the merits reached at summary judgment); Hartford Accident & Indem. Co. v. Costa Lines Cargo Servs., Inc., 903 F.2d 352, 359 (5th Cir.1990) (same); Vantine v. Elkhart Brass Mfg. Co., Inc., 762 F.2d 511, 518 (7th Cir.1985) (applying Grubbs); see also Mofitt v. Residential Funding Co., LLC, 604 F.3d 156, 159-60 (4th Cir.2010) (applying Grubbs in the context of an interlocutory appeal).

The Supreme Court subsequently extended the rule in Grubbs to a situation in which the district court lacked subject matter jurisdiction at the time of removal because the parties were not completely diverse. Caterpillar Inc. v. Lewis, 519 U.S. 61, 117 S.Ct. 467, 136 L.Ed.2d 437 (1996). In Caterpillar, the lower court would have had jurisdiction over a hypothetical suit filed at the time of judgment — the party who rendered diversity incomplete when the suit was first filed had dropped out of the case. The Court held that “a district court’s error in failing to remand a case improperly removed is not fatal to the ensuing adjudication if federal jurisdictional requirements are met at the time judgment is entered.” Caterpillar, 519 U.S. at 64, 117 S.Ct. 467. The result obtained even though the plaintiff opposed removal by filing a timely motion to remand. Id. at 70, 117 S.Ct. 467. Although the Court acknowledged that its ruling risked frustrating the rules Congress had imposed for removing cases, the argument ran “up against an overriding consideration.” Id. at 75, 117 S.Ct. 467. “Once a diversity case has been tried in federal court, with the rules of decision supplied by state law ... considerations of finality, efficiency, and economy become overwhelming.” Id.

Like Grubbs, the Court’s opinion in Caterpillar distinguishes between procedural defects in removal on the one hand, which generally must be raised within 30 days, see 28 U.S.C. § 1447(c) (motion to remand on basis other than lack of subject matter jurisdiction must be made within 30 days), and defects related to the Court’s subject matter jurisdiction. If a procedural defect in removal is cured by the time a judgment is entered, then practical concerns — finality, efficiency, and economy — militate in favor of retaining jurisdiction on appeal. See also Korea Exch. Bank, New York Branch v. Trackwise Sales Corp., 66 F.3d 46, 50 (3d Cir.1995) (“[T]he Supreme Court clearly suggested, even if it did not directly hold, that it does not view the removal statutes as imposing independent jurisdictional restrictions on the federal courts. Rather ... the relevant [jurisdictional] inquiry is whether the case cquld have been filed originally in federal court.”).

Put simply, if the doctrine of derivative jurisdiction constitutes a mere defect in the process by which a case reaches federal court, then this court may continue to exercise jurisdiction on appeal because the district court would have had original jurisdiction. 2 If, on the other hand, the derivative jurisdiction doctrine creates a latent and persistent defect in the subject matter jurisdiction of federal courts, then the issue can be raised at any time by any one. E.g., Capron v. Van Noorden, 6 U.S. (2 Cranch) 126, 127, 2 L.Ed. 229 (1804).

Admittedly, grappling with the procedure-versus-subject-matter-jurisdiction question is challenging because the doctrine of derivative jurisdiction is itself difficult to explain as a matter of first principles. The argument is also difficult to evaluate because the doctrine does not fit cleanly into the classifications that Grubbs and Caterpillar discuss. Both cases anticipate that procedural defects will relate to statutory requirements for removal. See 405 U.S. at 702, 92 S.Ct. 1344 (tacitly accepting the Fifth Circuit’s determination that there was a statutory defect in the United States’ removal petition); 519 U.S. at 73, 117 S.Ct. 467 (discussing the statutory flaw in Caterpillar’s notice of removal). And while Caterpillar in particular phrases the jurisdictional inquiry as turning on whether original jurisdiction could have been exercised at the time of judgment, that does not answer the antecedent question of whether the doctrine of derivative jurisdiction is essential to a court’s subject matter jurisdiction such that it cannot be cured. That notion brings us to the Supreme Court’s decision in Grupo Dataflux, 541 U.S. 567, 124 S.Ct. 1920. There, the parties on appeal were diverse at the time of judgment because the citizenship of one of the parties changed during the litigation, but were non-diverse at the time the suit was filed. The Supreme Court ruled that jurisdictional defect was not cured at the time of judgment, because the jurisdictional defect was that there was no diversity jurisdiction at the time of filing. The “time-of-filing rule ... measures all challenges to subject-matter jurisdiction premised upon diversity of citizenship against the state of facts that existed at the time of filing” regardless of when raised. Id. at 570-71, 124 S.Ct. 1920 (characterizing the rule as hornbook law). 3 Implicit in the government’s position is the argument that, like the time-of-filing rule that controls a party’s citizenship for diversity purposes, the presence of subject matter jurisdiction in the state court simply is a prerequisite to federal subject matter jurisdiction over removed actions.

After examining the issue, we reject the argument that the doctrine of derivative jurisdiction constitutes an essential ingredient of federal subject matter jurisdiction over removed actions. To be sure, the argument enjoys superficial appeal based on the name and language of the doctrine, and at least a handful of authorities explicitly describe derivative jurisdiction in subject matter jurisdiction terms. Philadelphia & Reading Ry. Co. v. Sherman, 230 F. 814, 816 (2d Cir.1916); Crowley v. S.Ry. Co., 139 F. 851, 853 (C.C.Ala.1905); see also Nordlicht v. New York Telephone Co., 799 F.2d 859, 863 n. 1 (2d Cir.1986) (assuming that derivative jurisdiction presents a defect in subject matter jurisdiction that may be raised at any time, but concluding that the defect was not present); Daley v. Town of New Durham, 733 F.2d 4, 6-7 (1st Cir.1984).

Nonetheless, it would be too easy to unquestioningly accept a handful of lower court authorities as conclusive merely because they incant the words “subject matter jurisdiction.” The Supreme Court has long warned, even as it confesses occasional complicity, that courts are too quick to affix the “jurisdiction” label. Compare Henderson ex rel. Henderson v. Shinseki, - U.S. -, 131 S.Ct. 1197, 1202, 179 L.Ed.2d 159 (2011) (“Because the consequences that attach to the jurisdictional label may be so dramatic, we have tried in recent cases to bring some discipline to the use of this term.”), with Ayers v. Watson, 113 U.S. 594, 599, 5 S.Ct. 641, 28 L.Ed. 1093 (1885) (courts have used the concept of jurisdiction upon removal “somewhat loosely”). As Judge Randolph of the D.C. Circuit has aptly observed, “ ‘Jurisdiction’ is a word of many, too many, meanings.” United States v. Vanness, 85 F.3d 661, 663 n. 2 (D.C.Cir.1996). And “[n]ot all mandatory prescriptions, however emphatic, are ... properly typed jurisdictional.” Union Pacific R.R. Co. v. Bhd. of Locomotive Eng’rs & Trainmen Gen. Comm, of Adjustment, C. Region, - U.S. -, 130 S.Ct. 584, 596, 175 L.Ed.2d 428 (2009) (quotation marks omitted). Rather, “[sjubject-matter jurisdiction properly comprehended ... refers to a tribunal’s ‘power to hear a case,’ a matter that ‘can never be forfeited or waived.’ ” Id. (quoting Arbaugh v. Y & H Corp., 546 U.S. 500, 514, 126 S.Ct. 1235, 163 L.Ed.2d 1097 (2006)).

The argument that the doctrine of derivative jurisdiction strikes at the heart of subject matter jurisdiction suffers from numerous flaws. At the outset, the textual allure of the jurisdictional argument is not so strong as it at first appears. While the doctrine says federal jurisdiction is derived from state courts, it in the same breath delivers its caveat that this is true only “in a limited sense.” Lambert Run, 258 U.S. at 382, 42 S.Ct. 349. Moreover, the doctrine is expressly concerned with the acquisition of jurisdiction upon removal — “If the state court lacks jurisdiction of the subject-matter or of the parties, the federal court acquires none....” Id. (emphasis added). The acquisition of jurisdiction upon removal speaks to so-called “removal jurisdiction,” Kircher v. Putnam Funds Trust, 547 U.S. 633, 641, 126 S.Ct. 2145, 165 L.Ed.2d 92 (2006); Wis. Dep’t of Corrections v. Schacht, 524 U.S. 381, 385, 118 S.Ct. 2047, 141 L.Ed.2d 364 (1998), not to the distinct concept of federal subject mat ter jurisdiction. Indeed, the modern instinct (in keeping with Henderson’s call for discipline in the use of the word jurisdiction) is to eschew the term removal jurisdiction, “because removal is not a kind of jurisdiction — analogous to federal question jurisdiction and diversity of citizenship jurisdiction. Rather it is a means of bringing cases within federal courts’ original jurisdiction into those courts.” Wright, supra, § 3721, at 27 (emphasis added). See also Merrell Dow Pharm. Inc. v. Thompson, 478 U.S. 804, 808, 106 S.Ct. 3229, 92 L.Ed.2d 650 (1986) (implicitly recognizing the distinction).

Therefore, it is not surprising that federal courts have regularly considered the doctrine of derivative jurisdiction as relating to “removal jurisdiction.” Palmer, 498 F.3d at 248 (“removal jurisdiction is derivative of state court jurisdiction prior to removal”); In re Miles, 430 F.3d 1083, 1087 (9th Cir.2005) (considering derivative jurisdiction in removal-jurisdiction terms); Hollis v. Florida State Univ., 259 F.3d 1295, 1298 (11th Cir.2001) (same); North Dakota v. Fredericks, 940 F.2d 333, 337 (8th Cir.1991) (same); Nishimoto v. Federman-Bachrach & Assocs., 903 F.2d 709, 714 n. 11 (9th Cir.1990) (same); Morda v. Klein, 865 F.2d 782, 784 (6th Cir.1989) (same); W. & S. Life Ins. Co. v. Smith, 859 F.2d 407, 409 n. 4 (6th Cir.1988) (same); Patriot Cinemas, Inc. v. Gen. Cinemas Corp., 834 F.2d 208, 210 (1st Cir.1987) (same); Sorosky v. Burroughs Corp., 826 F.2d 794, 801 (9th Cir.1987) (same); Leach v. Fed. Crop Ins. Corp., 741 F.2d 200, 201 (8th Cir.1984) (same); Witherow v. Firestone Tire & Rubber Co., 530 F.2d 160, 167-68 (3d Cir.1976) (same); see also In re Dutile, 935 F.2d 61, 63 (5th Cir.1991) (characterizing the doctrine as “a judicial gloss on the removal statutes”); Brandon v. Interfirst Corp., 858 F.2d 266, 269 n. * (5th Cir.1988) (characterizing derivative jurisdiction as a bar to removal rather than a substantive limit on federal authority); Dep’t of Revenue of State of Iowa v. Inv. Fin. Mgt. Co., 831 F.2d 790, 792 (8th Cir.1987) (same); Chemerinsky, supra, § 5.5 at 288 (same). And the legislative history of Section 1441 indicates that Congress, too, has conceived of the doctrine merely as a “casela gloss” barring proper removal. See H.R. Rep. 99-423, at 8, 1986 U.S.C.C.A.N. 1545, 1553 (1985).

Thus, the received understanding of the doctrine seems to place it on the procedural side of the Grubbs line. “[A] ‘procedural’ defect is any defect that does not go to the question of whether the case originally could have been brought in federal district court.” Baris v. Sulpicio Lines, Inc., 932 F.2d 1540, 1544 (5th Cir.1991). Courts of Appeals should disregard procedural defects in the removal process, Grubbs and Caterpillar teach, so long as subject matter jurisdiction properly lies at the time the order of judgment under review was entered. Grupo Dataflux, 541 U.S. at 574, 124 S.Ct. 1920 (distinguishing a requirement in the removal statute that a case “be fit for federal adjudication at the time the removal petition is filed” from the concept of subject matter jurisdiction). A procedural defect in removal is waivable. Baris, 932 F.2d at 1543-44.

More critically, subject matter jurisdiction is unyielding, Arbaugh, 546 U.S. at 514, 126 S.Ct. 1235 (subject matter jurisdiction can never be waived), and the earliest cases confirm that the doctrine of derivative jurisdiction is not. The formulation of the rule cited in Lambert Run is an almost verbatim quotation from the Circuit Court of Ohio’s decision in Fidelity Trust Co. v. Gill Car Co., 25 F. 737 (C.C.Ohio 1885). See also Lambert Run, 258 U.S. at 382 n. 3, 42 S.Ct. 349 (citing Fidelity Trust). Indeed, the Fidelity Trust case provides the lengthiest, if at times tortured, exposition on derivative jurisdiction that we have located. In ex plaining the rule, the court was careful to specify that the rule is not without limits: “I do not mean to say that we measure our jurisdiction wholly by that of the state court, and that nothing can be adjudged here which could not have been adjudged there; for cases can be well imagined where this rule should be subject to qualification.... ” 25 F.3d at 739; see also id. (“[A]nd it may be that we should, as the case required, extend or restrict our adjudication, as by our own rule of judgment we should be compelled to do”). If that sort of flexibility inheres in the rule, it is not a matter of subject matter jurisdiction — plain and simple. In that vein, it seems noteworthy that in every case we located in which the Supreme Court discussed the matter of derivative jurisdiction, the matter appears to have been raised promptly upon removal, prior to adjudication on the merits. See Minnesota, 305 U.S. at 384, 59 S.Ct. 292 (government filed motion to dismiss the action after removal); Gen. Inv. Co., 260 U.S. at 284-86, 288, 43 S.Ct. 106 (matter was raised after an initial appeal that did not involve a merits determination; dismissal should have been without prejudice); Lambert Run, 258 U.S. at 380, 382, 42 S.Ct. 349 (matter raised in a motion to dismiss the action after removal); Am. Well Works Co. v. Layne & Bowler Co., 241 U.S. 257, 258, 36 S.Ct. 585, 60 L.Ed. 987 (1916) (matter raised in a remand motion filed by the plaintiff); Cain v. Commercial Publ’g Co., 232 U.S. 124, 129, 132-34, 34 S.Ct. 284, 58 L.Ed. 534 (1914) (defect in personal jurisdiction in the state court raised upon removal, seemingly within 30 days); De Lima v. Bidwell, 182 U.S. 1, 1-2, 174, 21 S.Ct. 743, 45 L.Ed. 1041 (1901) (understanding a demurrer filed after removal to raise the matter of derivative jurisdiction).

The discussion in Fidelity Trust also shows that, unlike the time-of-filing rule at issue in Grupo Dataflux, the derivative jurisdiction rule is not based on the court’s constitutional authority to adjudicate a claim. See also Bidwell, 182 U.S. at 174, 21 S.Ct. 743 (distinguishing the question of derivative jurisdiction from “the jurisdiction of the circuit court as a Federal court”). The Fidelity Trust Court roots its understanding of the doctrine based on whether a lawsuit, brought in the wrong forum, has been properly constructed. Fidelity Trust, 25 F. at 739-40; see also Note, The Supreme Court of the United States During the October Term, 1942: I, 43 Colum. L.Rev. 837, 868-69 (1943) (characterizing the principle as forum regit ac-tum, or “the act of the forum rules”). Unwavering adherence to that notion, even once the case has been brought into the proper forum and judgment has been entered there, strikes us as akin to the common law forms of action, under which a person’s effort to recover for a wrong might have been stymied if he invoked the incorrect writ. E.g., Guille v. Swan, 19 Johns 381 (N.Y.1822) (action successfully characterized as trespass (direct injury), rather than trespass on the case (indirect injury), when a balloonist landed in plaintiffs land and was rescued by a crowd of onlookers who “beat[ ] down his vegetables and flowers”); Scott v. Shepherd, 96 E.R. 525 (K.B. 1773). But the key point is this: subject matter jurisdiction concerns the power of a court to hear a case; the doctrine of derivative jurisdiction is rooted in the idea that there is no case at all. The law has generally jettisoned such artificial notions, and in any event constitutional requirements are distinct from common law precepts, even old ones. E.g., Williams v. Florida, 399 U.S. 78, 86, 90 S.Ct. 1893, 26 L.Ed.2d 446 (1970) (concluding that the 12-person jury requirement was “a historical accident” and finding no support for the notion that the Framers intended to preserve it).

Conceiving of derivative jurisdiction as a procedural defect rather than a subject matter jurisdiction ingredient is consonant with Freeman v. Bee Mach. Co., 319 U.S. 448, 449-52, 63 S.Ct. 1146, 87 L.Ed. 1509 (1943). There, a case was filed in state court for breach of contract. Upon removal, the plaintiff amended its complaint to add a claim for treble damages under the Clayton Act. If the derivative jurisdiction rule truly defined subject matter jurisdiction, then the amendment should not have been allowed: the limits of the federal court’s subject matter jurisdiction should have been derived from the state court’s jurisdiction, and the state court would have lacked jurisdiction over the Clayton Act claim. Instead, the Court allowed the amendment, reasoning that (by statute) once the case was properly in federal court the action could proceed as if it originated there. Id. at 452, 63 S.Ct. 1146; see also People of Illinois ex rel. Barra v. Archer Daniels Midland Co., 704 F.2d 935, 939 (7th Cir.1983) (citing Grubbs and stating that filing an amended complaint in federal court cured any problem with removal brought on by the fact that the original state-court complaint was not removable).

The foregoing analysis demonstrates that the doctrine of derivative jurisdiction, notwithstanding its perhaps improvident name, is a procedural bar to the exercise of federal judicial power. It is not an essential ingredient for a court’s subject matter jurisdiction. Therefore the principles of Grubbs apply. Unfortunately, the government’s brief did not wrestle with the questions we have considered, instead merely asserting that Grubbs does not apply to jurisdictional defects. That presumably is why the government did not inform us that circuits appear to be unanimous in holding that the doctrine of derivative jurisdiction is cabined by the principles announced in Grubbs. See Morda, 865 F.2d at 784 (“Because the District Court would have had jurisdiction had this case origi-

nally been filed there, because there was no other bar to federal jurisdiction at the time of judgment, and because plaintiff never objected to subject matter jurisdiction at any time before or during the three-week trial below, the lack of derivative jurisdiction at the time of removal is irrelevant.”); Foval v. First Nat’l Bank of Commerce in New Orleans, 841 F.2d 126, 129 (5th Cir.1988) (discussing and applying the court’s earlier holding in Lirette v. N.L. Sperry Sun, Inc., 820 F.2d 116 (5th Cir.1987) (en banc)); Sorosky, 826 F.2d at 800-01.

We conclude that because the district court would have had jurisdiction over a hypothetical complaint filed at the time it entered the judgment now under review, the fact that the state court lacked jurisdiction over the case when it was removed has no significance. We proceed to the merits.

B. The Illinois Good Samaritan Act

We review de novo the district court’s decision to grant summary judgment as well as the interpretation of state law on which that decision rests. Salve Regina Coll. v. Russell, 499 U.S. 225, 231, 111 S.Ct. 1217, 113 L.Ed.2d 190 (1991). Summary judgment should not have been granted in this case. The chief question presented is whether Soleanicov charged a fee for emergency services within the contemplation of the Illinois Good Samaritan Act. We reject as unsound the argument that she did not charge a fee because she was paid a salary and because an entity other than her employer derived direct economic benefit from the billing form she prepared and submitted. Likewise, although Seidlin did not submit a billing form for his services, and therefore did not charge a fee, there is a genuine issue of material fact about whether that decision was made in good faith, a predicate for the Act’s protections.

When interpreting state law, a federal court’s task is to determine how the state’s highest court would rule. Konradi v. United States, 919 F.2d 1207, 1213 (7th Cir.1990). Generally, we defer to interpretations offered by state appellate courts unless there is a “persuasive indication[ ] that the state supreme court would decide the issue differently.” Allstate Ins. Co. v. Tozer, 392 F.3d 950, 952 (7th Cir.2004). And no indication is more persuasive than plain language: the Illinois rule is that plain language of a statute is the best evidence of legislative intent. Metzger v. DaRosa, 209 Ill.2d 30, 282 Ill.Dec. 148, 805 N.E.2d 1165, 1167 (2004).

The Illinois Good Samaritan Act spans some 27 sections and applies in a variety of settings. See, e.g., 745 ILCS 49/65 (“choking victim at food-service establishment”); 745 ILCS 49/15 (emergency dental care). Our analysis centers on 745 ILCS 49/25. The provision provides, with omissions not relevant here: “Any ... [physician] who, in good faith, provides emergency care without fee to a person, shall not, as a result of his or her acts or omissions, except willful or wanton misconduct on the part of the person, in providing the care, be liable for civil damages.” The Act immunizes physicians from liability if they (1) provide emergency care, (2) without fee, and (3) do so in good faith. Hernandez v. Alexian Bros. Health Sys., 384 Ill.App.3d 510, 323 Ill.Dec. 460, 893 N.E.2d 934, 941 (2008); Heanue, 291 Ill.Dec. 537, 823 N.E.2d at 1127. Rodas argues that Soleanicov cannot satisfy the second requirement and that Seidlin cannot satisfy the third requirement.

The parties’ primary dispute turns on what it means to provide emergency care without fee. Soleanicov argues that the Illinois Appellate Court answered the question in Heanue and that the interpretive linchpin is the word “fee.” The argument is that the relationship between UIC and the Crusader Clinic, combined with the fact that Soleanicov and Seidlin each drew a salary, means that they could not have charged a fee for their work within the meaning of the Good Samaritan Act. Employing a liberal use of brackets, Soleanicov says that the Heanue Court determined that a fee is “a very specific sort of relationship where the economic benefit is derived [by the physician providing emergency care] directly from the service performed.” Soleanicov Brief at 10 (brackets in original) (quoting Heanue, 291 Ill.Dec. 537, 823 N.E.2d at 1128-29). Seidlin adopts that understanding as well. Seidlin Brief at 12. Unfortunately, the bracketed text does more to advance the argument than the Illinois Appellate Court. With the addition of different bracketed text, the matter is less clear. Observe: a fee is “a very specific sort of relationship where the economic benefit is derived [by a medical practice or its third-party beneficiary] directly from the service performed.” Soleanicov’s brief overstates what Heanue stands for, although we do agree that the court’s use of the word “relationship” in that case did leave some room for confusion.

We read the Heanue Court to say something slightly different than what the defendants assert, and the ease’s discussion leaves us confident in that conclusion: a fee, as the term is used in the Good Samaritan Act, is simply a charge for medical services. In Heanue, a patient’s estate brought suit against a physician who furnished emergency care. The defendant physician was a partner of the medical group who provided treatment for Heanue, but was not the treating physician. After Heanue underwent an elective medical procedure, her post-operative medication was not working adequately. A nurse tried unsuccessfully to locate the treating physician and then paged the defendant physician’s medical practice. The defendant physician stepped in to provide treatment. 291 Ill.Dec. 537, 823 N.E.2d at 1126. The complaint alleged that he did so negligently. The defendant physician’s practice billed for medical services provided to Heanue, but not for the emergency services rendered by the defendant physician. That does not mean he was not paid, however; the compensation structure of the practice ensured that the defendant physician derived economic benefit from the other services that were provided Heanue (and for which charges were issued). Id., 291 Ill.Dec. 537, 823 N.E.2d at 1128. Nonetheless, the Heanue Court concluded that, on these facts, the defendant physician did not charge a fee for his emergency services. The court still reversed the trial court’s decision dismissing the case, because it was not clear that the defendant doctor’s decision not to bill for emergency services was made in good faith. Id., 291 Ill.Dec. 537, 823 N.E.2d at 1130 (“The record in this case allows an inference that the reason no bill was sent for the emergency care was that defendant sought to trigger the Act.”).

In concluding that the defendant physician had not charged a fee for his services, Heanue relies on and quotes the Black’s Law Dictionary definition. (Along with one other dictionary. There is no real difference between the two). Here is the definition of the word fee, which mirrors its ordinary meaning: “A charge for labor or services, esp. professional services.” Black’s Law Dictionary 629 (7th ed.1999) (def. 1). After providing the definition, the court then made the statement that Soleanicov excerpts in her brief: “These definitions do not encompass all relationships where some financial benefit flows to an individual; rather, they envision a very specific sort of relationship where the economic benefit is derived directly from the service performed. In other words, a fee is generated by and tied to the service performed.” Heanue, 291 Ill.Dec. 537, 823 N.E.2d at 1128-29. The use of the word relationship is perhaps a touch misleading. The best reading of the case is that the relationship the Heanue Court was speaking of is the relationship between charges and services.

That charge-for-services understanding of the word fee is the best read of Heanue for several reasons. First, immediately after speaking of “fee” in terms of “relationships,” the Heanue Court said this: “In other words, a fee is generated by and tied to the service performed.” 291 Ill.Dec. 537, 823 N.E.2d at 1129. That is our understanding of the term, too. Second, and more importantly, the charge-for-services definition comports with the ordinary meaning of the term, the best indication of how the Illinois Supreme Court would rule on the question. Metzger, 282 Ill.Dec. 148, 805 N.E.2d at 1167. The defendants make the interpretive mistake of concluding, if only implicitly, that services h'ave been provided without fee simply because the doctor was not compensated from that fee. In fact, the word encompasses both the amount paid in exchange for services and the amount charged for those services. See Webster’s Third New International Dictionary 833 (2002). There is no reason to conclude that the legislature adopted a halfway understanding of the term. Third, the discussion in Heanue bears out our understanding of the case. Recall, the Illinois Appellate Court reversed the case because it was unclear if the decision not to bill was made in good faith. If Heanue stood for the proposition that Soleanicov and Seidlin urge, then the analysis should have started and ended with the question of how the defendant physician was compensated for the services he provided. Instead of standing for the proposition that salaried physicians who provide emergency care automatically enjoy the protections of the Good Samaritan Act, Heanue stands only for the proposition that receiving compensation is not by itself sufficient to vitiate the Act’s protections. 4

Put simply, a fee ties labor to services. The identity of the entity responsible for paying for the services after they have been rendered, like the identity of the entity who has a right to collect on the bill, does not tell us whether there has been a “charge for labor or services.” Black’s, supra, at 629. After all, individuals often contract with insurance companies to reimburse their health care providers. The contract between the insurer and the insured will determine whether the insured must contribute a co-payment. In this case, actual payment was made on the bill from Medicaid, but for others private insurers will provide direct payment to a hospital, medical group, or solo practitioner. See also Villamil v. Benages, 257 Ill.App.3d 81, 195 Ill.Dec. 104, 628 N.E.2d 568, 575 (1993) (suggesting that a doctor would have charged a fee if there had been evidence he submitted a bill to public aid). And on the physician side of things, the fees collected may be carved up in a variety of ways. A physician who is hired into a medical practice may receive a salary, a guaranteed percentage of net proceeds, bonuses based on productivity, and so forth. See also 225 ILCS 60/22.2 (permitting fee-splitting arrangements among physicians). We see no evidence that the legislature, with its use of the unassuming word “fee” intended anything to turn on how a fee is processed or the compensation structures of the physicians who provide treatment. For good reason. The moment the General Assembly makes the coverage of the Good Samaritan Act turn on the business model used to collect physicians’ fees is the moment every medical practice restructures so that every doctor can be a good Samaritan. That outcome would do nothing to advance the enacted purpose of the Good Samaritan Act, which is to promote volunteerism and shield from liability “the generous and compassionate acts” of Illinois citizens. 745 ILCS 49/2.

Multiple eases from the Illinois Appellate Court, though not addressing the question we face here, have framed the inquiry in terms of whether a doctor charged a fee for his or her services, evincing agnosticism on the question of how the physician was ultimately compensated. E.g., Hernandez, 323 Ill.Dec. 460, 893 N.E.2d at 936, 942 (physician described as part of a physician’s group); Heanue, 291 Ill.Dec. 537, 823 N.E.2d at 1126 (physician described as a partner in a medical practice); Neal v. Yang, 352 Ill.App.3d 820, 287 Ill.Dec. 886, 816 N.E.2d 853, 855 (2004) (physician described as an employee of a medical practice). Thus, we have little trouble concluding that Soleanicov charged a fee for her work: when UIC physicians provided services to a Crusader Clinic patient, they prepared specialized billing forms and submitted them to the Crusader Clinic. After performing the delivery of Andrea Rodas, Soleanicov submitted a billing form for the procedure. The form was initialed by Soleanicov, and circled on the form was billing code 59515, “Cesarean delivery with postpartum care.” There is no serious question that she charged a fee for her emergency services.

That addresses the issues related to Soleanicov. The analysis works slightly differently for Seidlin. According to Seidlin, he was in the hospital in street clothes and “just happened to be in the hallway” when Dr. Baxter asked for advice about Rodas. After finding Soleanicov, he provided assistance because there was no time to procure the help of a surgical assistant. Unlike, Soleanicov, he never submitted a billing form for the work he performed. “These facts alone establish beyond question that Dr. Seidlin is entitled to the protection afforded under the Good Samaritan Act.” Seidlin Brief at 16. Incorrect. Seidlin’s conclusion would be sound only if Rodas had not marshaled evidence of her own at summary judgment. Rodas points out that Seidlin’s explanation for why he did not charge for his services was that his services in assisting the delivery were minimal. However, he did not recall ever waiving a fee on that basis in the past. The decision to deviate from his ordinary practice creates a genuine issue of material fact about whether his decision not to bill was made in good faith. Hernandez, 323 Ill.Dec. 460, 893 N.E.2d at 942-43. The matter is best left to a jury.

In sum, there is sufficient record evidence to survive summary judgment that Soleanicov billed for the emergency services she provided Rodas and that Seidlin made a bad faith decision not to bill. Therefore, the grant of summary judgment was not appropriate.

C. Remand

One final matter merits brief attention, and that is the course of proceedings upon remand. Our conclusion that the doctrine of derivative jurisdiction creates a procedural defect in removal suggests that derivative jurisdiction should ordinarily be raised within 30 days, 28 U.S.C. § 1447(c), but by its terms that provision applies only to plaintiffs who seek to remand a case. Likewise, Grubbs and Caterpillar by their terms speak only to the question of whether a court of appeals has power to review the judgment in a case that lands in federal court despite a defect in the removal process. See Grubbs, 405 U.S. at 702, 92 S.Ct. 1344 (original jurisdiction is “the issue in subsequent proceedings on appeal”) (emphasis added). The cases provide no direct guidance on the proper course to follow on remand.

The Fourth Circuit considered the problem in a case in which removal was improper but the district court did not pass on the legal merits of a controversy. In that case, the appeals court reversed a district court’s conclusion that a plaintiffs claim was time-barred. After reversing, the Fourth Circuit noted that no headway had been made on the merits and instructed the district court to remand the case to state court. Marshall v. Manville Sales Corp., 6 F.3d 229, 232-33 (4th Cir.1993). Relying on Marshall, the Ninth Circuit has stated simply that “[t]he Grubbs doctrine does not apply, though, where a circuit court reverses a district court’s grant of summary judgment such that there has been no judgment on the merits.” Emard v. Hughes Aircraft Co., 153 F.3d 949, 962 (9th Cir.1998).

We conclude that “considerations of finality, efficiency, and economy,” Caterpillar, 519 U.S. at 75, 117 S.Ct. 467, warrant the continued exercise of jurisdiction over this case on remand. There is no question that, if this case had been instituted in the proper court, subject matter jurisdiction over the case would have been proper. Moreover, our decision in Barra, 704 F.2d at 939, suggests that any defect in removal created by the doctrine of derivative jurisdiction would be cured if Rodas simply filed an amended complaint. We recently faced a similar issue in City of Joliet v. New West, L.P., 562 F.3d 830 (7th Cir.2009), where an action was improperly removed to federal court, but removal under a different provision would have been proper. In satisfying ourselves that there was no jurisdictional defect, we observed that “[i]t would be pointless to order this suit remanded, only to have [one of the parties] re-remove it in a trice.” Id. at 833. Similarly, it would be pointless for Rodas to file an amended complaint at this point. What is more, if the government could continue to assert the defect in the removal process, the district court would be required to remand to state court the claims against Soleanicov and Seidlin. Our opinion would, in effect, be rendered advisory. Federal courts are not in that business. Hayburn’s Case, 2 U.S. 408, 2 Dall. 409, 1 L.Ed. 436 (1792). Accordingly, the district court on remand shall continue to exercise jurisdiction over the case.

III. Conclusion

For the reasons set forth above, we Reverse the judgment of the district court and Remand for further proceedings.

1 . The provision provides: "The court to which a civil action is removed under this section is not precluded from hearing and determining any claim in such civil action because the State court from which such civil action is removed did not have jurisdiction over that claim.”

2 . The conclusion that the court would have had original jurisdiction is easily reached. Claims against the United States under the Federal Tort Claims Act must be raised in federal court. 28 U.S.C. § 1346(b). Supplemental jurisdiction over the pendent party claims against Soleanicov and Seidlin would have been authorized by 28 U.S.C. § 1367(a).

3 . In Caterpillar, too, the citizenship of the parties was determined at the time of filing. Although the parties were incompletely diverse at the time of removal, the party who created the defect in diversity dropped out of the case by the time of judgment. The feature is important, as it allows Caterpillar and Grupo Dataflux to be harmonized.

4 . Rodas and the United States urge that receiving compensation for medical services is by itself sufficient to fall outside of the Good Samaritan Act's protections. Henslee v. Provena Hosps., 373 F.Supp.2d 802, 812 (N.D.Ill.2005) ("This Court proposes that a reasonable definition of “fee” would be a situation in which either a doctor is paid for ■his services or the client pays a bill for those services.”). The Illinois Appellate Court rejected that proposition in Heanue. This appeal does not require us to provide a (necessarily tentative) resolution of the conflict.