Before SOUTHWICK, WILLETT, and OLDHAM, Circuit Judges. DON R. WILLETT, Circuit Judge: IntegraNet sued a competitor and its two owners (one of whom was a former IntegraNet employee) in state court. They removed to federal court. Then IntegraNet sued the former-employee owner in state court for another claim. And the owner again removed. The district court consolidated the cases. The so-called consolidation order sua sponte barred IntegraNet from filing another case against Appellees without the judge’s permission. On appeal, IntegraNet argues that the district court lacked jurisdiction and that consolidation was improper, as was the sua sponte pre-filing injunction. We agree with IntegraNet. We reverse the consolidation order, vacate the injunction, and remand both claims with instructions to the district court to remand the Texas state law claims to the Texas state court from which the cases were removed. I. Background IntegraNet is a Texas independent physician association (IPA), a network of independent physicians who’ve banded together to reduce overhead costs and negotiate contracts with payers. IntegraNet employed Robert Vincent Roth as Vice President of Contracting and Director of Financial Analytics. A big part of Roth’s job was to collect money owed to IntegraNet from insurers. Roth allegedly dropped the ball. In 2014 he sought recovery of only $337,784 of roughly $4.3 million owed to IntegraNet by an insurer. IntegraNet asserts that Roth knew about the underreporting prior to the claim deadline and failed to correct it. Roth quit IntegraNet that same year. About three years later, IntegraNet supposedly caught the mistake, but managed to collect only around $550,000 of the roughly $4.3 million. So IntegraNet is coming after Roth for the rest. After Roth quit IntegraNet in 2014, he started a competitor IPA called Texas Independent Providers (TIP). TIP serves patients through Medicare Advantage plans. Roth and his partner, Dr. Carlos Palacios, are the principal owners. IntegraNet says that Roth, Palacios, and TIP conspired to harm IntegraNet’s business, improperly prying physicians and insurers away from IntegraNet. For example, IntegraNet points to the insurer WellCare (a now-dismissed defendant). WellCare was an IntegraNet insurer, but withdrew in 2018.[1] WellCare notified its members of the termination with IntegraNet. The parties agree that federal regulations and the contract between IntegraNet and WellCare required IntegraNet to abstain from interfering with patient enrollment upon WellCare’s termination of the contract.[2] Put differently, IntegraNet was not supposed to pressure WellCare enrollees to switch insurance plans. IntegraNet insists it didn’t. But in April 2018, WellCare sent cease-and-desist letters to IntegraNet and some of its providers. The letters state that WellCare received “a complaint relating to unsolicited contacts with one or more of [the doctors'] patients who are Medicare-eligible beneficiaries.” IntegraNet says these letters were “thinly-veiled threats” to its doctor-members motivated by the desire for them to leave IntegraNet for another IPA—like TIP—or to contract directly with WellCare. IntegraNet also alleges that, around the same time, many of its doctor-members received marketing materials from TIP touting the benefits of their membership. These materials supposedly contain mischaracterizations about IntegraNet “that only Roth would have the knowledge to make.” Appellees Roth and Palacios co-own TIP. Palacios is also a Medical Director at WellCare. IntegraNet believes that WellCare acted in concert with TIP, Roth, and Palacios. IntegraNet filed two suits in state court. Suit #1 was against TIP, Roth, Palacios, and WellCare for interfering with IntegraNet’s business (TIP lawsuit). Suit #2 was against Roth alone for breach of fiduciary duty and negligence (Roth lawsuit). But WellCare and Roth removed to federal court. Here’s the timeline: April 18, 2018—IntegraNet files TIP lawsuit in state court. IntegraNet petitions for, and is granted, a temporary restraining order against defendants. April 29, 2018—WellCare removes TIP lawsuit to federal court under 28 U.S.C. §§ 1441 (removal), 1442 (federal officer), and 1446. May 10, 2018—After WellCare had been dismissed from the TIP lawsuit, IntegraNet moves for remand. June 18, 2018—IntegraNet files Roth lawsuit in state court. August 30, 2018—Roth removes the Roth lawsuit to federal court based on supplemental jurisdiction and 28 U.S.C. § 1446(b)(3), and he moves to consolidate with the TIP lawsuit. September 17, 2018—IntegraNet moves for remand of the Roth lawsuit. September 18, 2018—The district court consolidates the TIP and Roth lawsuits. The consolidation order forbids IntegraNet from “fil[ing] another case against the defendants on the facts in this action without [the district] court’s permission.” September 24, 2018—IntegraNet appeals the consolidation order under 28 U.S.C. § 1292. June 5, 2019—The district court denies remand of the Roth lawsuit. In this appeal, we review the interlocutory order consolidating the two cases and the denial of the motions to remand. II. Jurisdiction and Standard of Review A. Statement of Jurisdiction The preliminary issue is whether we have jurisdiction to hear the appeal of this interlocutory order consolidating the two cases. Appellees argue that we lack jurisdiction. They are mistaken. Under 28 U.S.C. § 1292(a)(1), we may review interlocutory orders “granting, continuing, modifying, refusing or dissolving injunctions, or refusing to dissolve or modify injunctions, except where a direct review may be had in the Supreme Court . . . .” The relevant question is whether the district court’s order amounted to an injunction. The order forbade IntegraNet from “fil[ing] another case against the defendants on the facts in this action without [the district] court’s permission.” We have previously held that restrictions similar to those imposed by the district court’s order constitute an order granting an injunction under § 1292(a)(1).[3] Baum, for example, had to do with a similar command from the same district judge as here: The plaintiff couldn’t “fil[e] claims, directly or indirectly, in courts or with agencies in the state of Texas without the express written permission of Judge Lynn N. Hughes.”[4] We determined that the command was an injunction, and thus jurisdiction existed under § 1292(a)(1).[5] And in our 2010 Qureshi decision, we considered a similarly worded dismissal order from the same judge: “Because of his persistent abuse of the judicial process, [plaintiff] may not file papers in Texas federal courts without written permission of Judge Lynn N. Hughes.”[6] We did not explain why the order constituted an injunction. But for taxonomical purposes, the bottom line is that the court cited Baum and called the order a “pre-filing injunction.”[7] Returning to this case, the “Order of Consolidation” bars IntegraNet from filing “another case against the defendants on the facts in this action without [the district] court’s permission.” Under Baum, this command not to sue qualifies as an injunction, which is immediately appealable. So we have jurisdiction to review the order. Appellees contend the order isn’t an injunction because the district court merely ordered the consolidation of two cases, as its title—”Order of Consolidation”—suggests. But the order’s title is not dispositive.[8] Rather, “[i]n determining what is an appealable order under 28 U.S.C. § 1292(a)(1), courts look not to terminology, but to the substantial effect of the order made.”[9] Here, like in Baum and Qureshi, the district court’s order prevents a party from filing future lawsuits without that judge’s permission—meaning it’s a pre-filing injunction. Thus, we have jurisdiction under § 1292(a)(1) to review the order. As such, we also have jurisdiction to review the district court’s denial of IntegraNet’s motion to remand to state court. An order denying a motion to remand standing alone would not be appealable.[10] But, where there’s an independent basis for jurisdiction, we must review the district court’s subject-matter jurisdiction over the lawsuits.[11] Here, the district court’s pre-filing injunction provides the requisite, independent basis for review under § 1292(a)(1). B. Standard of Review When a district court remands (or denies remand) based on supplemental jurisdiction, we review for abuse of discretion.[12] And when a district court grants or denies an injunction, we review that for abuse of discretion too—though we review underlying questions of law de novo.[13] III. Discussion A. Appellees Failed to Satisfy the Federal Officer Removal Statute. A court must have jurisdiction to decide a case.[14] So we turn to the threshold inquiry of whether the district court had jurisdiction over the TIP and Roth lawsuits.[15] By statute, federal courts have original jurisdiction over federal-question and diversity cases.[16] They can also exercise supplemental jurisdiction over “all other claims that are so related to claims within [federal- question or diversity cases] that they form part of the same case or controversy under Article III of the United States Constitution.”[17] By statute, federal courts also have removal jurisdiction in some cases originally filed in state court. For example, the federal officer removal statute, 28 U.S.C. § 1442(a)(1), allows a defendant to remove a state-court action against “any officer (or any person acting under that officer) of the United States.” We have held that to satisfy § 1442, a defendant must show “(1) that it is a person within the meaning of the statute, (2) that it has ‘a colorable federal defense,’ (3) that it ‘acted pursuant to a federal officer’s directions,’ and (4) ‘that a causal nexus exists between its actions under color of federal office and the plaintiffs claims.’”[18] Importantly, the defendant bears the burden of showing that it meets § 1442′s requirements.[19] TIP asserts that WellCare satisfied the criteria for removal under § 1442(a)(1),[20] thus creating jurisdiction. We disagree. Appellees didn’t show that WellCare acted under federal-officer directions. The Supreme Court has noted that the words “acting under” are “broad . . . but broad language is not limitless.”[21] The Court’s 2007 Watson[22] decision demonstrates one of those limits. The Watson plaintiffs sued cigarette manufacturer Philip Morris, alleging it manipulated test results to show lower nicotine and “tar” content in cigarettes marketed as “light.”[23] The Eighth Circuit held that Philip Morris was “acting under” a federal officer and could remove pursuant to § 1442 because the Federal Trade Commission mandated that Philip Morris adopt the specific testing method plaintiffs challenged and enforced that mandate with “ongoing monitoring,” laboratory inspections, independent verification of results, publication of nicotine ratings, and enforcement actions against manufacturers.[24] The Supreme Court reversed, holding that “the fact that a federal regulatory agency directs, supervises, and monitors a company’s activities in considerable detail” is not sufficient to satisfy the “acting under” requirement.[25] The Court further explained that: [A] highly regulated firm cannot find a statutory basis for removal in the fact of federal regulation alone. A private firm’s compliance (or noncompliance) with federal laws, rules, and regulations does not by itself fall within the scope of the statutory phrase “acting under” a federal “official.” And that is so even if the regulation is highly detailed and even if the private firm’s activities are highly supervised and monitored. A contrary determination would expand the scope of the statute considerably, potentially bringing within its scope state-court actions filed against private firms in many highly regulated industries.[26] Citing Watson, we recently addressed a similar issue in City of Walker.[27]There, appellants asserted that a construction company was “acting under” a federal officer because its work—including the hydraulic design that allegedly exacerbated the flooding damage at issue—was “subject to inspection and approval by the federal government,” which “retained ‘oversight responsibility’” for and contributed funding to the contracting job.[28] But we said even if that were true, that didn’t amount to being a government contractor “acting under” a federal officer when it designed and built the project.[29] Likewise, WellCare wasn’t acting pursuant to a federal officer’s directions. First, Appellees’ assertion that federal regulations required WellCare to notify its members of the termination with IntegraNet is simply beside the point.[30] It’s a strawman argument. IntegraNet takes issue with the cease-and-desist letters, not the notification letters. After the notification letters, WellCare sent cease-and-desist letters to IntegraNet and some of its providers. The letters state that WellCare received “a complaint relating to unsolicited contacts with one or more of [the doctors'] patients who are Medicare-eligible beneficiaries.” Appellees argue that WellCare acted “at the direction of [the federal government] when it sent letters to IntegraNet and its physicians demanding that they stop violating federal rules in their efforts to coerce their patients into switching Medicare Advantage plans.” The parties disagree about whether any such interference occurred. But that’s irrelevant to this jurisdictional question. What matters is that Appellees cite no evidence showing that the government told WellCare to send those cease-and-desist letters. Appellees merely argue that WellCare’s letters urged compliance with applicable regulations.[31] But that’s not acting under officer direction. Appellees quote an unreported Western District of Texas case for their insistence that “Medicare insurance carriers act under federal direction for purposes of 28 U.S.C. § 1442(a)(1).”[32] That case cites a published Fifth Circuit opinion from 1975 called Peterson.[33] Again, Appellees misapply caselaw. Peterson dealt with a doctor who sought damages for supposed “wrongful suspension under the Medicare program.”[34] In Peterson, the federal government received a complaint that “questioned the payment to Dr. Peterson for physical therapy services not rendered by him.”[35] The government investigated, and uncovered possible fraud.[36] So the government “issued a letter to Blue Cross/Blue Shield ordering suspension of further payments to Dr. Peterson under the Medicare program.”[37] Without analysis, the district court concluded that “it is indisputable” that the insurance carrier acted under federal direction.[38] This case is different. The federal government did not order WellCare to send the cease-and-desist letters to IntegraNet and its providers. WellCare wasn’t acting under federal direction. So the district court should have remanded to state court. B. The District Court Abused its Discretion in Exercising Jurisdiction. Let’s assume, for the sake of argument, that WellCare satisfied the federal officer removal statute.[39] Even with that assumption, IntegraNet dismissed WellCare from the suit. Taking our cues from the Supreme Court, we have long applied the “general rule” that “a court should decline to exercise jurisdiction over remaining state-law claims when all federal-law claims are eliminated before trial.”[40] That said, district courts may decide remaining state-law claims if the relevant statutory and common-law factors favor retaining jurisdiction.[41] These statutory factors come from 28 U.S.C. § 1367(c).[42] As we stated in Enochs, the factors are: “(1) whether the state claims raise novel or complex issues of state law; (2) whether the state claims substantially predominate over the federal claims; (3) whether the federal claims have been dismissed; and (4) whether there are exceptional circumstances or other compelling reasons for declining jurisdiction.”[43] The Supreme Court articulated the common-law factors—judicial economy, convenience, fairness, and comity—in its 1988 Carnegie-Mellon University decision.[44] We applied these factors in Enochs and held that the district court abused its discretion by refusing to remand the case to state court following the dismissal of all federal claims.[45] Today’s case has numerous similarities with Enochs. And Enochs governs this appeal.[46] Here, as in Enochs, the statutory and common-law factors weigh against supplemental jurisdiction. First, we analyze whether the state claims raise complex or novel state-law issues. No party has argued that the cases present novel state-law issues. But, we agree with IntegraNet that there are complex state-law issues—particularly in the TIP lawsuit and considering Appellees’ desire to consolidate the Roth lawsuit, which involves different facts and subject matter than the TIP lawsuit. Together, these lawsuits involve claims for breach of fiduciary duty, unfair competition, theft of trade secrets (under the Texas trade secret statute), and negligence. There are also complex facts arising out of the relationship between myriad entities, including multiple IPAs, an insurance company, a doctor, and a former executive. This complexity weighs in favor of remand. A Texas state court is best positioned to address a complex array of issues relating to Texas law—especially under complex factual circumstances. Regarding the second and third factors, IntegraNet never brought any federal claims. Plus, the party invoking federal jurisdiction was dismissed before filing its answer. This of course means that, as in Enochs, “the Texas state law claims predominate over the non-existent federal claims.”[47] Again tracking Enochs, “[t]he fourth factor also favors remand, as the heavy balance of the common[-]law factors in favor of remand constitutes another compelling reason to decline jurisdiction.”[48] Let’s turn to the common-law factors: judicial economy, convenience, fairness, and comity. The suits are in their infancy. No discovery. No Rule 26(f) conference. No scheduling order. Like Enochs, “hardly any federal judicial resources, let alone a significant amount of resources, ha[ve] been devoted to the district court’s consideration of the Texas state law claims.”[49] So remand wouldn’t harm judicial economy. And remand would be convenient. The Enochs court’s reasoning applies here as well: There would be little—if any—need for any party to duplicate work.[50] Plus, it would be convenient for the case to be heard in Harris County state court where the parties, witnesses, and evidence are located.[51] Next, remand is fair. Again paralleling Enochs, this case deals with “purely Texas state law claims,” making it “certainly fair” to hear them in Texas state court.[52] And nothing indicates that any party would be prejudiced by remanding to state court. Finally, comity favors remand here. Federal courts have limited jurisdiction and must respect the “important interests of federalism and comity.”[53] As the Supreme Court has emphasized, “[n]eedless decisions of state law should be avoided both as a matter of comity and to promote justice between the parties, by procuring for them a surer-footed reading of applicable law.”[54] Federal courts are “not as well equipped for determinations of state law as are state courts.”[55] As in Enochs, “Texas state courts have superior familiarity with, and heightened interests in developing, Texas state law” related to these claims.[56] Because all of the remaining claims concern only state law, comity weighs in favor of remand. In sum, statutory law and common-law interests suggest that the district court abused its discretion in denying IntegraNet’s motion to remand. Yet Appellees argue that federal jurisdiction is “proper and appropriate”—even after WellCare’s dismissal from the lawsuit. Appellees seem to rest their argument on the “purely discretionary” nature of a district court’s decision to exercise supplemental jurisdiction “after dismissing every claim over which it had original jurisdiction.”[57] IntegraNet does not dispute that district courts have this discretion. They do.[58] The question is whether the district court abused its discretion. Appellees appear to misunderstand the Supreme Court’s holding in Carlsbad. Despite Appellees’ representations, Carlsbad was about whether 28 U.S.C. § 1447 bars appellate review when a district court declines supplemental jurisdiction over state-law claims.[59] The Supreme Court considered “whether a federal court of appeals has jurisdiction” to review a district court’s remand.[60] The remand “was not based on a jurisdictional defect but on [the district court's] discretionary choice not to hear the claims.”[61] Thus, the Court reasoned, appellate review was possible.[62] But Carlsbad didn’t affect the standard by which appellate courts review remand decisions. Appellees also cite our 2016 decision in Heggemeier to support their argument that “[d]istrict courts enjoy wide discretion in determining whether to retain supplemental jurisdiction over a state claim once all federal claims are dismissed.”[63] Heggemeier does acknowledge a district court’s “wide discretion.”[64] But that is not a blank check, as Appellees seem to suggest. Rather, we explained that this “wide discretion” required it to “review a district court’s refusal to exercise supplemental jurisdiction [over remaining state law claims] . . . for abuse of discretion.”[65] In fact, we quoted Enochs and applied the common-law factors (albeit briefly).[66] Appellees failed to apply any of these factors. And nothing suggests that the district court considered them either. “[W]e are bound to consider and weigh all the factors.”[67] As explained, the relevant statutory and common-law factors disfavor the district court’s maintaining jurisdiction over the state-law claims at issue in this case. The district court’s refusal to remand these claims was an abuse of discretion under Enochs. Consolidation was Inappropriate. Under Federal Rule of Civil Procedure 42(a), courts may consolidate actions if they “involve a common question of law or fact.” According to the Supreme Court’s decision in Peacock, this language means that “[a]ncillary jurisdiction may extend to claims having a factual and logical dependence on the primary lawsuit.”[68] So long as “that primary lawsuit . . . contain[s] an independent basis for federal jurisdiction.”[69] Generally, “[d]istrict courts enjoy substantial discretion in deciding whether and to what extent to consolidate cases.”[70] But, because the primary lawsuit here lacks a basis for federal jurisdiction,[71] consolidation was not appropriate. The Pre-filing Injunction Was Improper. Similarly, the lack of federal jurisdiction[72] also renders the pre-filing injunction improper. Plus, in Baum, we held that “[n]otice and a hearing are required if the district court sua sponte imposes a pre-filing injunction or sua sponte modifies an existing injunction to deter vexatious filings.”[73] We reaffirmed this holding two years later in Qureshi.[74] Here, there was no basis for federal jurisdiction, no notice, and no hearing—any one of which is sufficient to make the district court’s injunction improper. IV. Conclusion We REVERSE the district court’s denial of remand and hold that it abused its discretion when denying IntegraNet’s motion to remand. We REVERSE the Order of Consolidation and VACATE the district court’s pre- filing injunction. Finally, we REMAND both cases with instructions to the district court to remand the Texas state law claims to the Texas state court from which the cases were removed.
