Before Barksdale, Engelhardt, and Oldham, Circuit Judges. Andrew S. Oldham, Circuit Judge: The petition for panel rehearing is DENIED. Because no member of the panel or judge in regular active service requested that the court be polled on rehearing en banc, the petition for rehearing en banc is DENIED. On our own motion, we withdraw our prior opinion, 37 F.4th 1062, and issue the following in its place. The question presented is whether a district court can deny a motion to dismiss based on qualified immunity through a boilerplate scheduling order. We hold no. I. Gary Lynch was arrested on an outstanding warrant and booked into the Hunt County, Texas jail on February 12, 2019. Seven days later, because of a gas leak, the jail staff evacuated all prisoners from the facility and temporarily housed them elsewhere. Jail staff held Lynch in the Tarrant County jail while repairs occurred and returned him to the Hunt County facility on the evening of February 22. The next morning, Lynch was discovered dead in his cell. Following an autopsy, a doctor concluded Lynch died from aortic valve endocarditis with myocardial abscess. Plaintiff-Appellee Gwendolyn Carswell is Lynch’s mother. She sued Hunt County and numerous county employees under 42 U.S.C. § 1983 and Monell v. Department of Social Services of the City of New York, 436 U.S. 658 (1978). She alleged the individual defendants knew Lynch was suffering from a heart condition but failed to treat him. The individual defendants moved to dismiss under Federal Rule of Civil Procedure 12(b)(6), asserting qualified immunity. On January 25, 2021, the district court denied that motion and entered its “standard QI scheduling order.” That order is the subject of this appeal. It provided, in relevant part: Any pending motions to dismiss on the basis of qualified immunity are denied without prejudice. See Shultea v. Wood, 47 F.3d 1427, 1431–34 (5th Cir. 1995) (en banc) (qualified immunity must be raised by filing answer). Any defendant desiring to assert qualified immunity who has not already done so by way of answer must file an answer asserting qualified immunity within 14 days of the date of this Order. Except as set forth below, all party discovery is stayed as to any defendant who asserts qualified immunity. Discovery is not stayed as to a defendant asserting qualified immunity as to that person’s capacity as a witness to the extent that there is any other defendant not asserting qualified immunity. The individual defendants complied with the order and filed answers and affirmative defenses. But they also noticed an immediate appeal of the scheduling order. Carswell moved to dismiss the appeal for lack of jurisdiction, arguing the scheduling order was not an appealable collateral order because the district court had not ruled on qualified immunity. We carried that motion with the case. In the meantime, back in district court, the individual defendants moved to stay all discovery and all proceedings. They argued that “[a]ll discovery in this matter should be stayed against all Defendants, including Hunt County, and all proceedings in this case should be stayed, pending resolution of the Individual Defendants’ assertions of qualified immunity.” The district court denied defendants’ motion. In its October 8 order, thecourt “address[ed] the motion at greater length than it customarily would devote to what is essentially a motion to stay.” In so doing, it explained how (in its view) the January 25 scheduling order “attempts to follow the choreography of the Fifth Circuit’s QI dance.” Specifically: [The scheduling order] requires any defendant wanting to assert QI to do so by answer, rather than by motion to dismiss; . . . it requires the plaintiff to file a rule 7 reply to any assertion of qualified immunity. If defendants believe QI can be resolved based on the pleadings, there is a deadline for filing a motion for summary judgment on that basis; if the plaintiff believes discovery is necessary to resolve the QI defense, he or she may raise that issue by way of a Rule 56(d) motion for discovery in response to the defendant’s motion for summary judgment. Significantly, unless the Court allows narrowly tailored discovery on QI, party discovery as to the QI defendants is stayed. On October 13, in district court, Carswell filed an “advisory to the court concerning depositions” indicating that, on the Monell claim, she wished to depose all eight of the individual defendants asserting qualified immunity. She reminded the court that she had previously served all eight with deposition notices. She explained she wished “to notice these depositions again and proceed consistent with the Court’s October 8, 2021, Order.” Appellants moved in this court to stay discovery pending appeal. We granted the stay and subsequently heard oral argument. II. “Jurisdiction is always first.”Arulnanthyv. Garland, 17 F.4th 586, 592 (5th Cir. 2021) (quotation omitted). Under the collateral order doctrine, we have jurisdiction to review orders denying qualified immunity. See Backe v. LeBlanc, 691 F.3d 645, 647–49 (5th Cir. 2012); Mitchell v. Forsyth, 472 U.S. 511, 526–27 (1985). Likewise for district court orders “declin[ing] or refus[ing] to rule on a motion to dismiss based on a government officer’s defense of qualified immunity.” Zapata v. Melson, 750 F.3d 481, 484 (5th Cir. 2014). Such orders are “tantamount to . . . order[s] denying the defendants qualified immunity.” Ibid. The collateral order doctrine permits immediate appeals of these orders because a defendant’s entitlement to qualified immunity must be determined “at the earliest possible stage of the litigation.” Ramirez v. Guadarrama, 3 F.4th 129, 133 (5th Cir. 2021) (per curiam). That’s because qualified immunity is more than “a mere defense to liability.” Pearson v. Callahan, 555 U.S. 223, 237 (2009) (quotation omitted). It’s also “an immunity from suit.” Ibid. (quotation omitted). And one of the most important benefits of the qualified immunity defense is “protection from pretrial discovery, which is costly, time-consuming, and intrusive.” Backe, 691 F.3d at 648;see also Helton v. Clements, 787 F.2d 1016, 1017 (5th Cir. 1986) (per curiam) (a “refusal to rule on a claim of immunity” deprives a defendant of his “entitlement under immunity doctrine to be free from suit and the burden of avoidable pretrial matters”). We have jurisdiction over the scheduling order here because the district court refused to rule on qualified immunity “at the earliest possible stage of the litigation.” Ramirez, 3 F.4th at 133. Defendants asserted qualified immunity in their motion to dismiss. That motion was the earliest possible opportunity for the district court to resolve the immunity question. It declined to do so. Instead, it required defendants to assert their qualified immunity defense by way of answer. And it postponed ruling on the immunity issue until summary judgment. That “effectively . . . denied [defendants] the benefits of the qualified immunity defense” and “vest[ed] this court with the requisite jurisdiction to review the discovery order.” Wicks v. Miss. State Emp. Servs., 41 F.3d 991, 994 (5th Cir. 1995). III. We review the scheduling order for abuse of discretion. Backe, 691 F.3d at 649. We hold the district court abused its discretion by deferring its ruling on qualified immunity and subjecting the immunity-asserting defendants to discovery in the meantime. See ibid. Where public officials assert qualified immunity in a motion to dismiss, a district court must rule on the motion. It may not permit discovery against the immunity-asserting defendants before it rules on their defense. See id. at 648 (It is “precisely the point of qualified immunity . . . to protect public officials from expensive, intrusive discovery until and unless the requisite showing overcoming immunity is made.” (second emphasis added)). It is true that to ensure qualified immunity can be decided at the earliest possible time, we have authorized “discovery narrowly tailored to rule on [defendants'] immunity claims.” Backe, 691 F.3d at 649; see also Lion Boulos v. Wilson, 834 F.2d 504, 508–09 (5th Cir. 1987) (first articulating this limited-discovery procedure). We’ve described that limited-discovery process as “a careful procedure,” which permits a district court to “defer its qualified immunity ruling if further factual development is necessary to ascertain the availability of that defense.” Zapata, 750 F.3d at 485 (quoting Backe, 691 F.3d at 648). The district court must first find that the plaintiff has pleaded “facts which, if true, would overcome the defense of qualified immunity.” Ibid. (quotation omitted). If it still finds itself “unable to rule on the immunity defense without further clarification of the facts,” ibid. (quotation omitted), then we allow the district court to order discovery “narrowly tailored to uncover only those facts needed to rule on the immunity claim,” Wicks, 41 F.3d at 994 (quoting Lion Boulos, 834 F.2d at 507–08). Three points about this “careful procedure” bear emphasis. First, its purpose is only to allow the district court to rule on the defendant’s assertion of QI; its purpose is not to provide a backdoor for plaintiffs to circumvent the defendant’s immunity from suit. Backe, 691 F.3d at 649. Second, where the QI-asserting official determines that any pre-ruling discovery sought or ordered in the district court crosses the line from permissible Lion Boulous discovery to impermissible vitiation of the official’s immunity from suit, the collateral order doctrine authorizes an immediate appeal like the one we entertain today. Ramirez, 3 F.4th at 133. And third, Lion Boulos and its progeny must be understood in light of subsequent Supreme Court precedent. The Supreme Court has now made clear that a plaintiff asserting constitutional claims against an officer claiming QI must survive the motion to dismiss without any discovery. On the lattermost point, consider Ashcroft v. Iqbal, 556 U.S. 662 (2009). There the district court gave the plaintiff discovery before ruling on the officials’ motion to dismiss for qualified immunity. See id. at 670. The plaintiff tried to defend that discovery on the ground that “the Court of Appeals ha[d] instructed the district court to cabin discovery in such a way as to preserve petitioners’ defense of qualified immunity as much as possible in anticipation of a summary judgment motion.” Id. at 684 (quotation omitted). The Supreme Court had none of it. Instead, it reaffirmed its prior holding “that the question presented by a motion to dismiss a complaint for insufficient pleadings does not turn on the controls placed upon the discovery process.” Id. at 684–85 (citing Bell Atl. Corp. v. Twombly, 550 U.S. 544, 559 (2007)). And the Court made a point of emphasizing that its “rejection of the careful-case-management approach is especially important in suits where Government-official defendants are entitled to assert the defense of qualified immunity.” Id. at 685. It concluded the respondent was “not entitled to discovery, cabined or otherwise.” Id. at 686. Thus, Iqbal squarely prohibits interpreting our “careful procedure” as allowing tailored discovery before a district court rules on an official’s motion to dismiss. When defendants assert qualified immunity in a motion to dismiss, the district court may not defer ruling on that assertion. It may not permit discovery—”cabined or otherwise”—against immunity-asserting defendants before it has determined plaintiffs have pleaded facts sufficient to overcome the defense. Ibid. The rule is that “a defendant’s entitlement to qualified immunity should be determined at the earliest possible stage of the litigation”—full stop. Ramirez, 3 F.4th at 133 (citing Mitchell, 472 U.S. at 526–27). None of this means, of course, that the QI determination must be made once-and-for-all at the motion-to-dismiss stage. Today we only hold that where the pleadings are insufficient to overcome QI, the district court must grant the motion to dismiss without the benefit of pre-dismissal discovery. Similarly, where the pleadings are sufficient to overcome QI, the district court must deny the motion to dismiss without the benefit of pre-dismissal discovery. In the latter scenario, however, our precedent gives the defendant-official at least two choices after his motion to dismiss is denied. First, the defendant can immediately appeal the district court’s denial under the collateral order doctrine. See Backe, 691 F.3d at 647–49; Mitchell, 472 U.S. at 526–27. Or second—à la Lion Boulos and its progeny—the defendant can move the district court for discovery limited to the factual disputes relevant to whether QI applies, then reassert QI in a summary judgment motion. Hutcheson v. Dallas Cnty., 994 F.3d 477, 481 (5th Cir. 2021) (“Before limited discovery is permitted, a plaintiff seeking to overcome QI must assert facts that, if true, would overcome that defense.”). Why does the defendant alone enjoy this choice? Because only the defendant-official enjoys qualified immunity from suit. See Iqbal, 556 U.S. at 685 (“The basic thrust of the qualified-immunity doctrine is to free officials from the concerns of litigation, including avoidance of disruptive discovery.” (emphasis added) (quotation omitted)). IV. Carswell offers three additional points in defense of the scheduling order. Each is unavailing. A. Carswell first argues the district court did not refuse to rule on qualified immunity. The district court, for its part, admitted that it “require[d] any defendant wanting to assert QI to do so by answer, rather than by motion to dismiss.” But Carswell says this was “the opposite of a refusal or failure to rule” because the “district court clearly indicated it would timely address Individual Defendants’ qualified immunity defense.” All this gives short shrift to the requirement that qualified immunity must be adjudicated at the earliest possible opportunity. See Ramirez, 3 F.4th at 133. “Unless the plaintiff’s allegations state a claim of violation of clearly established law, a defendant pleading qualified immunity is entitled to dismissal before the commencement of discovery.” Mitchell, 472 U.S. at 526 (emphasis added). The Supreme Court has repeatedly made clear that “the driving force” behind qualified immunity is “a desire to ensure that insubstantial claims against government officials will be resolved prior to discovery,” and it has “stressed the importance of resolving immunity questions at the earliest possible stage in litigation.” Pearson, 555 U.S. at 231– 32 (emphasis added). The district court declined to rule on qualified immunity at the motion-to-dismiss stage. It deferred answering that question until the summary-judgment stage. That is, ipso facto, a refusal to rule at the earliest possible stage in litigation. It does not matter that the court promised to rule promptly once it arrived at the next stage of litigation. B. Carswell next defends the scheduling order because it stayed discovery as to qualified immunity. Specifically, the court stayed “all party discovery . . . as to any defendant who asserts qualified immunity,” but not “as to a defendant asserting qualified immunity as to that person’s capacity as a witness to the extent that there is any other defendant not asserting qualified immunity.” So the district court would have allowed Carswell to proceed with discovery on her Monell claim, including by noticing depositions for all eight of the individual defendants asserting qualified immunity. Iqbal squarely forecloses that, too. Responding to concerns about the burdens litigation imposes on public officials, the Court explained: It is no answer to these concerns to say that discovery for petitioners can be deferred while pretrial proceedings continue for other defendants. It is quite likely that, when discovery as to the other parties proceeds, it would prove necessary for petitioners and their counsel to participate in the process to ensure the case does not develop in a misleading or slanted way that causes prejudice to their position. Even if petitioners are not yet themselves subject to discovery orders, then, they would not be free from the burdens of discovery. Iqbal, 556 U.S. at 685–86. In other words, the Court ruled out even “minimally intrusive discovery” against official defendants before a ruling that plaintiff had met his burden to overcome the qualified immunity defense at the pleading stage. Id. at 686. Carswell responds that “Monell discovery presents no undue burden to the Individual Defendants because they would be required to participate as witnesses in discovery even if they had not been named as defendants.” Red Br. at 30. We disagree for three reasons. First, there are significant differences between naming an individual defendant and then deposing him in two capacities (one personal and the other Monell/official) and not suing the individual and deposing him only in his Monell/official capacity. The former puts the individual’s own money on the line. And the dual-capacity defendant must be particularly careful in a deposition about how his answers can be used against him in not one but two ways. So the stakes differ substantially. Carswell cannot elide these differences by saying the defendant would have to testify either way. Second, it’s no answer to say the defendant can be deposed twice— once on Monell issues (before the district court adjudicates the immunity defense) and once on personal-capacity issues (afterwards). It only exacerbates the burdens of litigation to make a defendant sit for two depositions instead of one. And it turns qualified immunity on its head by doubling the “heavy costs” of litigation. Iqbal, 556 U.S. at 685. Third, Carswell conceded at oral argument that bifurcation of discovery would radically complicate the case. Carswell suggested that a special master could be appointed to police the Monell/official-capacity depositions so that no party could cross the line into personal-capacity questions before the district court adjudicated the immunity defense. But the very fact that Carswell can foresee the need for a special master proves that bifurcated discovery imposes unreasonable burdens on the defendants. C. Carswell also argues the scheduling order must pose no problem because it is “obviously a form order” the district court uses frequently in cases like this one. The district court likewise noted that defendants’ motion to stay discovery presented “a frontal attack on [its] standard qualified immunity (‘QI’) scheduling order.” And Carswell points us to similar district court orders permitting Monell discovery against individual defendants whose assertions of qualified immunity remained pending in motions to dismiss. See, e.g., Saenz v. City of El Paso, No. 14-cv-244, 2015 WL 4590309, at *2 (W.D. Tex. Jan. 26, 2015) (declining to stay discovery despite “acknowledg[ing] the force” of defendant’s arguments based on Iqbal). That the scheduling order here is “standard” in qualified immunity cases tells us nothing about whether it correctly understands the governing law. Today we clarify the governing law. And we trust that will harmonize our circuit’s discovery practices with the Supreme Court’s instructions. V. Finally, Carswell argues that any error in the district court’s scheduling order is harmless because she has clearly stated plausible claims sufficient to defeat the individual defendants’ assertion of qualified immunity in their motion to dismiss. But all agree the district court has not yet ruled on that question. We decline to do so in the first instance. Cf. Cutter v. Wilkinson, 544 U.S. 709, 718 n.7 (2005) (“[W]e are a court of review, not of first view.”); see also, e.g., Arnold v. Williams, 979 F.3d 262, 269 (5th Cir. 2020) (remanding for the district court to consider qualified immunity in the first instance “[b]ecause as a general rule, we do not consider an issue not passed upon below” (quotation omitted)). * * * Carswell’s motion to dismiss for lack of jurisdiction is DENIED. We VACATE the district court’s scheduling order and REMAND for further proceedings consistent with this opinion.
