Carey Dale GRAYSON, et al 2:12-cv-00316 Plaintiffs, v. WARDEN, COMMISSIONER, ALABAMA DOC, Defendants; Demetrius Frazier, 2:13-cv-00781 ConsolPlaintiff-Appellant, v. Warden, Commissioner, Alabama Doc, Defendants-Appellees; David Lee Roberts, 2:14-cv-01028 ConsolPlaintiff-Appellant, v. Warden, Commissioner, Alabama Doc, Defendants-Appellees; Robin D. Myers, 2:14-cv-01029 ConsolPlaintiff-Appellant, v. Warden, Commissioner, Alabama Doc, Defendants-Appellees; Gregory Hunt, 2:14-cv-01030 Consol Plaintiff-Appellant, v. Warden, Commissioner, Alabama Doc, Defendants-Appellees, 869 F.3d 1204

TJOFLAT, Circuit Judge. L On July 1, 2002, the State of Alabama chose lethal injection, rather than electrocution, as its preferred method of implementing capital punishment. The Alabama Department of Corrections (“ADOC”) thereafter adopted a three-drug protocol to administer the injection. The United States Supreme Court describéd an identical protocol; as implemented by the State of Kentucky, in Baze v.

Rees, 553 U.S. 35, 128 S.Ct. 1520, 170 L.Ed.2d 420 (2008): The first drug, sodium thiopental ..., is a fast-acting barbiturate sedative that induces a deep, comalike unconsciousness when given in the amounts used for lethal injection. The second drug, pancu-ronium bromide ..., is a paralytic agent that inhibits all muscular-skeletal movements and, by' paralyzing the diaphragm, stops respiration. Potassium chloride, the third drug, interferes with the electrical signals that stimulate the contractions of the heart, inducing cardiac arrest.

The proper administration of the first drug ensures that the prisoner does not experience any pain associated with the paralysis and cardiac arrest caused by the second and third drugs. Id. at 44, 128 S. Ct. at 1527 (internal citations omitted). On April 26, 2011, Alabama substituted pentobarbital, “a short-acting barbiturate” sedative, for sodium thiopental, as the first drug in its execution protocol. Powell v. Thomas, 643 F.3d 1300, 1302 (11th Cir. 2011), Then, on September 10, 2014, the State substituted midazolam, a benzodiaze-pine sedative, for pentobarbital. Brooks v.

Warden, 810 F.3d 812, 816-17 (11th Cir. 2016). It also substituted rocuronium bromide for pancuronium bromide as the second drug. Id. at 817. Potassium chloride remained the third drug. Id. In the four cases at hand, the appellants, four death row prisoners awaiting execution, claim that if they are executed in accordance with the lethal injection protocol now in place, they will suffer “cruel and unusual punishment” in violation of the Eighth Amendment. They seek an order under 42 U.S.C. § 1983 enjoining the ADOC from executing them pursuant to that protocol. In Glossip v.

Gross, — U.S. —, 135 S.Ct. 2726, 2737, 192 L.Ed.2d 761 (2015), the Supreme Court made clear that the “controlling opinion in Baze” set forth the two-pronged standard a plaintiff must satisfy “to succeed on an Eighth Amendment method-of-execution claim.” The first prong requires the prisoner to demonstrate that the challenged method of execution presents “a ‘substantial risk of serious harm.’” Id. (quoting Baze, 553 U.S. at 50, 128 S.Ct. at 1531).

That is, the method must “present[ ] a risk that is ‘sure or very likely to cause serious illness and needless- suffering, and give rise to sufficiently imminent dangers.’” Id. (quoting Baze, 563 U.S. at 50, 128 S.Ct. at 1531). The second requires the prisoner to “identify an alternative that.is ‘feasible, readily implemented, and in fact significantly reducefe] a substantial risk of severe pain.’ ” Id. (quoting Baze, 553 U.S. at 52, 128 S.Ct. at 1532). Showing “a slightly or marginally safer alternative” is insufficient to mount a successful challenge to a State’s method of execution.

Id, (quoting Baze, 553 U.S. at 51, 128 S.Ct. at 1631). Appellants contend that the ADOC’s current protocol preseiits a substantial risk of serious harm’that comports .with Baze’s definition. They argue that the risk is substantial because midazolam, a sedative, is not an analgesic like sodium thio-pental and pentobarbital and will consequently fail to create the sustained state of anesthesia necessary to enable them to withstand the intolerable pain that will be generated by subsequent injections of ro-curonium bromide and potassium chloride.

As required by Baze’s second prong, Appellants have also proposed three alternative methods of execution involving single injections of either sodium thiopental, compounded pentobarbital, or a 500-milligram bolus of midazolam. Before us for review is the District Court’s order of October 31, 2016, granting the ADOC’s motion for

summary judgment pursuant to Rule 56 of the Federal Rules of Civil Procedure.

In its order, the Court concluded that Appellants had failed to present probative evidence creating “a genuine dispute of material fact as to [the existence of a feasible and readily implementable] alternative method of execution, an essential prong of the Baze/Glossip test for an Eighth Amendment method-of-execution claim.” Since Appellants’ proof failed to satisfy the alternative-method-of-execution requirement imposed by Baze, the Court dismissed Appellants’ Eighth Amendment claims without addressing the other half of the Baze standard: whether the substitution of midazolam for pentobarbital as the first drug of the three-drug injection protocol created a “substantial risk of serious harm.” Appellants ask that we vacate the judgment because the District Court, rather than determining whether the ADOC had satisfied its Rule 56 burden of showing that there was “no genuine dispute as to any material fact,” improperly assumed the role of the trier of fact and resolved the genuine factual disputes in the ADOC’s favor.

The ADOC, in response, argues that the District Court did not err and, even if it had, the error is of no moment because the law-of-the-case doctrine bars Appellants’ Eighth Amendment claims. If not, they argue the statute of limitations does so: they assert that the switch from pentobarbital to midazolam does not constitute a “substantial change” to the State’s three-drug execution protocol; thus, the two-year limitations period passed years ago. After hearing from the parties in oral argument and considering their briefs, we conclude that genuine issues of material fact preclude

summary judgment. We also conclude that Appellants’ Eighth Amendment claims are not barred by the law-of-the-case doctrine. As to the ADOC’s limitations argument, they did not raise it below and the District Court did not consider it; we thus cannot address it in the absence of a factual determination as to whether the substitution of midazolam for pentobarbital constitutes a substantial change to Alabama’s execution protocol. We accordingly vacate the District Court’s judgment and remand these cases for further proceedings. II.

The operative complaints and answers in these cases are identical with respect to the Appellants’ Eighth Amendment claim now before us. For convenience, we refer only to Demetrius Frazier’s second amended complaint (“Frazier’s Complaint” or “the Complaint”). Similarly, our references to the ADOC’s motion to dismiss and answer correspond specifically to the documents the ADOC filed in response to Frazier’s Complaint.

Frazier’s Complaint, filed on August 25, 2016, alleges — with respect to Baze’s “substantial risk of serious harm” prong — that midazolam will fail to anesthetize Frazier and therefore subject him to the intolerable pain the administration of rocuronium bromide and potassium chloride will cause.

The Complaint alleges — with respect to Baze’s requirement that a prisoner plead and prove the existence of an “alternative” method-of-execution significantly reducing a risk of severe pain — that “a single bolus of [compounded] pentobar-bital .... is read[ily] available, and would entirely reduce the risk of pain associated with administering the paralytic and potassium chloride, because those drugs would not be- used”; that sodium thiopental is available and “would cause death without need of a paralytic or potassium chloride”; that “a 500mg dose of midazolam will likely cause death in under an hour”; and that the ADOC “can obtain midazo-lam.” In addition to this statement of Frazier’s Eighth Amendment claim,- the Complaint also includes facts its drafter' apparently thought would be probative of his Eighth Amendment claim at trial, but that are unnecessary to establish an Eighth Amendment claim sufficient-to withstand a motion to dismiss.

These unnecessary assertions of fact include: that midazolam administered in a multi-drug protocol previously failed to anesthetize prisoners executed in Ohio, Oklahoma, and Arizona; that “[njumerous states have switched from a two or. three-drug protocol to a one-drug protocol”; that “[a] report issued by a bipartisan committee recommended that states discontinue using three-drug lethal injection cocktails and instead use a single large dosage of a barbiturate”; that “[s]tates including Texas, Colorado, Ohio, Georgia, Missouri, Mississippi, Oklahoma, South Dakota, and Pennsylvania have used or intend to use compounded pentobarbital for executions”; that “[s]ince January 1, 20Í4, nearly 40 executions have been carried out using a single bolus of pentobarbital, making it the most common method of execution in the United States”; and that the “Defendant’s own expert endorses” the use of a single 500-milligram bolus of midazolam as a method of execution.

The drafter of the Complaint also apparently thought that portions of-the record in Arthur v. Thomas, a Middle District of Alabama case that presented an' Eighth Amendment claim practically identical to Frazier’s, would be supportive of Frazier’s claim at trial. So, the drafter attached to the Complaint, as Exhibits A and B, the transcript of the evidentiary hearing the District Court held in Arthur on October 18 and 19,2012. Also attached, as Exhibit C, was the State of Alabama’s September 11, 2014 motion requesting the Alabama Supreme Court to set a date for Frazier’s execution.

These exhibits are mentioned in the Complaint by letter, but the Complaint does not incorporate by reference any of their contents. The exhibits were entirely extraneous .to Frazier’s Eighth Amendment Claim. Frazier did not cite them in opposing the ADOC’s motion for

summary judgment, and the District Court made no reference to them in granting the motion. The ADOC moved to dismiss Frazier’s complaint on September 8, 2015. See Consent to Judgment or in the Alternative, Motion to Dismiss at 1, Frazier v. Myers, No. 2:13-cv-781 (M.D. Ala. Sept. 8, 2015). The ADOC argued that in light of the Supreme Court’s decision in Glossip and this Court’s decision in Chavez v. Fla.

SP Warden, the District Court was required to dismiss Frazier’s Complaint for failure to state a claim for relief: As the Eleventh Circuit and now the Supreme Court have found that other litigants cannot show that a 500-milli-gram bolus of midazolam would allow an inmate to feel the effect of the other two drugs — a result that is equivalent to what was achieved when the first drug administered was pentobarbital — Frazier must allege some additional evidence to survive dismissal. Id. at 10.

The ADOC observed that in Glossip, the “Supreme Court held that the fact-findings about the use of midazolam being constitutional were not clearly erroneous,” and, moreover, “noted that every lower court to have considered the question reached the same conclusion: midazo-lam is capable of rendering someone unconscious and, therefore, eliminates any substantial risk of severe pain that might be caused by administering the other drugs in a three-drug lethal-injection protocol.” Id. at 11.

The ADOC further argued that Frazier’s claim had been “time-barred for nearly eleven years” and was thus “due to be dismissed.” Id. at 5. Noting that Alabama’s two-year limitations period applied to Frazier’s claim, it observed that the limitations period begins to accrue either when state review becomes final or when a state makes a substantial change to its execution protocol, whichever is later. Id. at 6.

And since Frazier’s case became final in 2000 and the State adopted lethal injection as its execution method in 2002, it argued that Frazier’s case was time-barred unless a substantial change in its execution protocol was made. It then asserted that no such change was made: with respect to the substitution of midazolam for pentobarbital, the ADOC contended that “Frazier failed to plead any factual allegations or include any affidavits with his complaint that could plausibly show that the use of midazolam is a significant change from pentobarbital.” Id. at 6-7.

Further, it argued that because “the Supreme Court has rejected virtually identical claims about midazolam’s effectiveness,” and because “the factual allegations rejected in Glossip contained much greater detail and specificity than the hypothesized, unsubstantiated allegations” in Frazier’s complaint, Frazier could not claim that the substitution of midazolam constituted a substantial change “as a matter of law.” Id. at 7. Thus, the ADOC argued that Frazier’s claim became time-barred in 2004, two years after Alabama adopted lethal injection. Id. at 6.

Assuming that the Complaint may have stated a plausible Eighth Amendment claim, the ADOC posited an alternative disposition: it would consent to the Court’s entry of a judgment that “suspends the current lethal-injection protocol” and substitutes midazolam as the method of execution. Id. at 4. “After an initial 500-milli-gram bolus of midazolam, the execution team [would] administer an additional 500-milligram dose of midazolam, if needed, until Frazier’s sentence [was] carried out.” Id. at 3.

This alternative disposition had a proviso: that the “Court order Frazier’s counsel to procure a sworn assurance from Plaintiff that he is aware of the alternatives suggested in the latest complaint filed on his behalf, that he understands the implications of the complaint, and that he consents to be executed by a one-drug protocol using midazolam.” Id. at 3. The District Court ordered Frazier to respond to the ADOC’s pleading by September 25, 2015. Order at 1, Frazier v, Myers, No. 2:13-cv-781 (M.D. Ala. Sept. 11, 2015). Frazier did so on that day.

See Opposition to Defendants’ “Consent to Judgment” and Motion to Dismiss at 1, Frazier v. Myers, No. 2:13-cv-781 (M.D. Ala. Sept. 25, 2015). He requested that the Court “reject the ADOC’s offer of a consent judgment, deny Defendants’ motion to dismiss, and issue a scheduling order for motions, hearings and discovery, leading to a trial on the issue of whether Alabama’s present execution protocol is unconstitutional.” Id. at 2.

On October 8, 2015, the District Court denied the ADOC’s motion to dismiss, ordered the ADOC to answer Frazier’s Complaint by October 22, 2015, declared that Appellants’-cases would be referred to as the “Midazolam Litigation,” and scheduled a status conference in the Midazolam Litigation for November 4, 2015, to discuss whether the Court should consolidate the cases for discovery and the final hearing and to determine whether the issues the cases presented differed from those pending in Arthur that would be tried in January 2016. Order at 1-2, Frazier v. Myers, No. 2:13-cv-781 (M.D. Ala. Oct. 8, 2015).

The ADOC answered the Complaint as ordered. The answer responded to the Complaint paragraph by paragraph and denied the material allegations of Frazier’s Eighth Amendment claim, including the allegation that two of the drugs included in Frazier’s proposed alternative method of execution, sodium thiopental and compounded pentobarbital, were available.

The answer admitted that midazolam was available and stated that the “Defendants have agreed to provide Frazier with an execution utilizing a single drug, midazolam, as set forth in his complaint.” The answer also asserted eighteen affirmative defenses, including that Frazier’s Eighth Amendment claim “fail[ed] to state a claim for relief,” was “barred by sovereign immunity,” and was “barred by res judicata.” On November 5, 2015, following the status conference, the Court entered an order consolidating Appellants’ .cases.

On November 20, 2015, the Court issued a “Final Scheduling Order,” scheduling the cases for trial before the Court on April 19-22, 2016; setting deadlines of February 5 and March 18, 2016 for “fact discovery” and “expert discovery,” respectively; and directing-that dispositive motions be filed-by March 25, 2016. On March 16, 2016, the Court granted the parties’ joint motion to extend the deadline for deposing expert witnesses through April 5, '2016. Arthur was tried to the District Court on January 12 and 13, 2016. See Arthur, 840 F.3d at 1278.

The evidence before the Court consisted mainly of the testimony and documents introduced during the evi-dentiary hearing on October 18 and 19, 2012 and deposition testimony taken during discovery conducted after that hearing. See Arthur v. Dunn, No. 2:11-cv-438, 2016 WL 1551475 (M.D. Ala. Apr. 15, 2016), aff’d sub nom., Arthur v. Comm’r, Ala. Dep’t. of Corr., 840 F.3d 1268 (11th Cir. 2016). The District Court rendered its decision on April 15, 2016.

In its findings of fact, the Court found that “[pjentobarbi-tal is not feasible and readily implemented as an execution drug in Alabama, nor is it readily available to the ADOC,. either compounded or commercially,” and that “[s]odium thiopental is unavailable to the ADOC for use in lethal injections.” Id...at *8-9.

In its conclusions of law, the Court held that although the plaintiff “sufficiently pleaded an Eighth Amendment claim,” he “failed to prove” that either compounded pentobarbital or sodium thiopental was “readily available to the ADOC” for use in the single-drug protocols he proposed as alternative execution methods. Id. at *9-10.

Because that failure doomed the plaintiffs claim, the Court did not have to, and accordingly did not, decide whether the ADOC’s implementation of the current three-drug protocol, with midazolam acting as the first drug, created a “substantial risk of serious harm” to a prisoner during execution. On February 26, 2016, while the decision in Arthur was still pending, the ADOC moved the District Court for

summary judgment on Appellant’s Eighth Amendment claims in the Midazolam Litigation. On March 16, the Court granted the parties’ joint motion to extend the discovery deadline to April 5. On April 15, Appellants responded to the ADOC’s motion for

summary judgment. The trial scheduled for April 19-22, 2016 did not take place. Instead, the District Court took the ADOC’s motion for

summary judgment under submission. On October 31, 2016, it granted the motion. It did so because the “Plaintiffs' failed to meet their burden of proof to survive

summary judgment.” It observed that “Plaintiffs clearly bear the burden of proving a known and available alternative method of execution that significantly reduces the risk of substantial pain in the execution” and concluded they failed to meet that burden in that they “failed to present evidence that creates a genuine dispute of material fact as to an alternative method of execution, an essential prong of the Baze/Glossip test for an Eighth Amendment method-of-execution claim.” The District Court reached this conclusion with respect to Appellants’ proposed single-drug protocol based on the testimony of the ADOC’s General .Counsel, Anne Adams Hill.

In deciding to credit Hill’s testimony and then weigh it against Appellants’ proof, the District Court functioned as a finder of fact and ultimate decision maker and therefore erred. See Mize v. Jefferson City Bd. of Educ., 93 F.3d 739, 742 (11th Cir. 1996) (“It is not the court’s role to weigh conflicting evidence or to make credibility determinations; the non-movant’s evidence is to be accepted for purposes of

summary judgment.”). The Court performed the same role when it determined the credibility of testimony and weighed the evidence in summarily disposing of Appellant’s midazolam proposal. In

Part III, we conclude that the District Court’s resolution of credibility issues and its weighing of the evidence requires the vacation of its

summary judgment. In

Part IV, we consider and reject as merit-less the ADOC’s arguments that Appellants’ Eighth Amendment claims were both untimely and barred by the law-of-the-ease doctrine.

Part V relates to the proceedings on remand. III. “Under Rule 56(c),

summary judgment is proper ‘if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.’ ” Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 2552, 91 L.Ed.2d 265 (1986). “The movant has the burden of showing that there is no genuine issue of fact.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 256, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). In deciding whether to grant

summary judgment, a district court “ ‘may not weigh conflicting evidence or make credibility determinations.’ ” Jones v. UPS Ground Freight, 683 F.3d 1283, 1292 (11th Cir. 2012) (citing FindWhat Investor Grp. v. FindWhat.com, 658 F.3d 1282, 1307 (11th Cir. 2011)). With these principles in hand, we turn to the District Court’s

summary disposition of Appellant’s Eighth Amendment claims based on its determination that none of the three alternative modes of execution, single-drug protocols consisting of compounded pentobarbital, sodium thio-pental or midazolam respectively, satisfies Glossip’s requirement that such alternatives be “feasible, readily implemented, and in fact significantly reduce a substantial risk of severe pain.” Glossip, 135 S.Ct. at 2737. We begin with compounded pento-barbital. A.

The District Court began its discussion about the availability of compounded pen-tobarbital by harkening back to the findings of fact and conclusions of law it made following the Arthur trial. The Court recalled that it had found that “the ADOC’s supply of commercially manufactured pen-tobarbital, Nembutal®, expired around November 2013,” and that “compounded pentobarbital” was unavailable to the ADOC as an alternative single-drug protocol. In finding compounding pentobarbital unavailable, the Court had to resolve the conflicting testimony of Gaylen M.

Zentner, Ph.D, a pharmacist, and testimony that Anne Adams Hill, the ADOC’s general counsel, previously gave at the Arthur trial in January 2016. See Arthur, 2016 WL 1551475, at *4-7. Dr. Zentner testified that “pentobarbital sodium for injection is listed in the FDA Orange Book, a publication containing all approved drugs in the United States,” and because no active patents covered the product, “anyone who has the ingredients can make pentobarbital sodium.” Id. at *5.

He performed an internet search for the drug and found a company that in its products list identified the' drug as available for purchase in the United States. Id. at *6. He also noted that “there are overseas suppliers of pentobarbital sodium and that pentobarbital sodium could be produced by drug synthesis labs in the United States.” Id. As for compounding pharmacies in Alabama, Dr. Zentner “testified that he searched the website of the Accreditation Commission for Health Care (ACHC) for accredited compounding pharmacies” and identified nineteen within the State. Id. Dr.

Zentner also indicated that the ADOC should be able to obtain compounded pentobarbital from sources outside of Alabama, opining, because other states have obtained compounded pentobarbital, it should also be available from those pharmacies in the other states and that any pharmacy that is qualified to do sterile compounding and has the necessary equipment could be a source of supply, as the compounding exercise is “very easy and straightforward.” Id. (quoting Dr. Zentner). The District Court seemed skeptical.

Zentner’s carefully qualified answer to the last question put to him on direct examination reveals the tepid and inexact nature of his opinion concerning whether pentobarbital is available to the ADOC: A. My opinion is that the active pharmaceutical ingredient, which is pento-barbital sodium, is available for purchase in the United States and that there are compounding pharmacies that have the skills and licenses to perform sterile compounding of pen-tobarbital sodium. Therefore, the feasibility for producing a sterile preparation of pentobarbital sodium does exist. Id. (quoting Dr. Zentner).

Despite its apparent skepticism, the Court still found Dr. Zentner’s testimony “credible.” Id. at *7 In Arthur, Hill testified as “the ADOC’s party representative.” Id. In the Memorandum Opinion and Order here under review, the Court reiterated its description of her testimony in Arthur. Hill .., testified [that] [s]he was aware that in 2015, Georgia, Missouri, Texas, and Virginia performed executions using compounded pentobarbital [and] that, in her recent efforts to obtain compounded pentobarbital for the ADOC’s use in executions, she had contacted the departments of corrections in at least those four states.

Elaborating, Hill testified that, as part of her job duties, she is routinely in contact with other departments of corrections on a variety of issues, including the subject of lethal injection generally, the availability of compounded pentobar-bital, and, in those states that have been able to obtain compounded pentobarbi-tal, their willingness either to provide it to the ADOC or to provide their source to the ADOC.

She reiterated that she has had these similar, ongoing, conversations not just recently, but “for some time.” In her quest to find a source for compounded pentobarbital, Hill has also contacted all eighteen accredited compounding pharmacies in Alabama, but her efforts were to no avail. “[N]one of the 18 were able to provide the Department of Corrections with compounded pentobarbital.” In all, she has contacted at least twenty-nine potential sources to inquire about obtaining compounded pentobarbital, and all of those efforts failed. After reiterating its description of Hill’s testimony, the.

Court went on to say that, in Arthur, it “credited Hill’s testimony and made the following findings of fact concerning pentobarbital and compounded pentobarbital”: 8. The ADOC has attempted to obtain compounded pentobarbital for use in executions from departments of corrections in at least four states, :Georgia, Missouri, Texas, and Virginia, but those efforts were unsuccessful, 9.

The ADO has contacted all of the accredited compounding pharmacies in Alabama to ascertain whether any of these pharmacies would be willing and able to provide compounded pentobarbi-tal to the ADOC, but those efforts have been unsuccessful, 10. Pentobarbital is not feasible and readily implemented as an execution drug in Alabama, nor is it readily available to the ADOC, either compounded or commercially. In passing on the ADOC’s motion for

summary judgment, the District Court, relying on Hill’s Arthur testimony, adopted the facts depicted' in • paragraphs^8-10 above and examined the record for evidence ‘ indicating that at some time between the conclusion of the Arthur trial on January 13, 2016, and the date the Court took the ADOC’s motion under submission, the availability picture had changed. The Court appeared to ask itself the following question: Had the Plaintiffs presented evidence showing that compounded pentobarbital, though unavailable- to the ADOC as of January 13, 2016, could now be obtained?

In an effort to show that the drug was available, Appellants relied on “the report and testimony of their expert, Deborah L. Elder, Pharm. D,” a “Clinical Associate Professor in the Pharmaceutical and Biomedical Department in the College of Pharmacy at the University of Georgia” who was certified “in compounding” under Georgia law. They also relied on the Complaint’s factual representations that other states had been able to obtain the drug. The Court found Dr.

Elder’s report and testimony unhelpful to Appellants’ cause because, at the end of the day, she was unable to “identify any pharmacist or supplier who could- provide compounded pen-tobarbital to the ADOC.” The Complaint’s representations that other states were able to obtain the drug was also unhelpful to Appellants in establishing that the drug was available to the ADOC.

The Court found it “inconsequential” that in several states “since January 1, 2014, nearly forty executions had been performed ‘using a single large dose of pentobarbital.’ ” Moreover, “that the drug was available in those states at some point over the past two years d[id] not, without more, make it likely that it [was] available to Alabama now.” It was also inconsequential that several states “intend to use compounded pentobarbital for executions.” Given these findings, the District Court concluded that its “earlier finding in Arthur that pentobarbital, either compounded or commercially manufactured, [was] not readily available to the ADOC” was “unchanged by the additional evidence [Appellants] offered.” More to the point, “[plaintiffs’ evidence obtained after Hill’s testimony in the Arthur trial in January 2016[ ] fail[ed] to establish that compounded pentobarbi-tal ha[d] since become available to the ADOC.” * * * The Complaint, which was filed on August 25, 2015, alleges that “a single bolus of [compounded] pentobarbital ... is read[ily] available” to the ADOC.

The District Court, adopting findings of fact it made in Arthur, found that compounded pentobarbital was not available' on August 25, 2015; indeed, the drug was not available as of January 13, 2016, the day the Arthur trial concluded. Thus, to create a genuine issue of fact, Appellants had to demonstrate that the drug became available at some point between January 13, 2016, and the day the Court took the ADOC’s motion for

summary judgment under consideration. The Court concluded Appellants failed to do so. In precluding Appellants from challenging its Arthur finding that compounded pentobarbital was unavailable to the ADOC prior to January 13, 2016, the District Court did not identify the principle of law it was invoking. The Court apparently had two doctrines in mind: issue preclusion and judicial notice. Although superficially each doctrine might seem relevant, we conclude neither was applicable.

Issue preclusion comes into play when: (1) the issue at stake is identical to the one involved in the prior litigation;- (2) the issue was actually litigated in the prior suit; (3) the determination of the issue in the prior suit was a necessary part of the judgment in that action; and (4) the parties are the same or in privity with each other and the party against whom the earlier decision is asserted had a full and fair opportunity to litigate the issue in the earlier proceeding. Baloco v. Drummond Co., 767 F.3d 1229, 1251 (11th Cir. 2014).

Issue preclusion, like claim preclusion, is an affirmative defense under Federal Rule of Civil Procedure 8(c), See Concordia v. Bendekovic, 693 F.2d 1073, 1076 (11th Cir. 1982) (explaining that issue preclusion is “an affirmative defense that should be raised under Rule 8(c)”). It is “incumbent on the defendant to plead and prove such a defense” in answering the plaintiffs complaint. Taylor, 553 U.S. at 907, 128 S.Ct. at 2179-80. The ADOC could not rely on issue preclusion for two reasons. First, it failed to plead the doctrine as an affirmative defense.

The ADOC filed its answer to the Complaint on October 22, 2015, and could not have pled the defense because it was not available. Arthur had not been decided. See Arthur, 2016 WL 1551475, at *1. After Arthur was decided on April 15, 2016, the ADOC could have sought leave to amend its answer, but did not. Second, the defense was not available to the ADOC because the parties in Arthur and Appellants’ cases were not “the same,” and Arthur and Appellants were not in privity, i.e., no exceptions to the rule against nonparty preclusion applied to make Arthur binding in Appellants’ cases.

The general rule, set forth long ago by the Supreme Court, states that “[a] judgment or decree among parties to a lawsuit resolves issues as among them, but it does not conclude the rights of strangers to those proceedings.” Martin v. Wilks, 490 U.S. 755, 762, 109 S.Ct. 2180, 2184, 104 L.Ed.2d 835 (1989) (footnote omitted).

Six exceptions to this rule have been recognized: A court may apply nonparty preclusion if: (1) the nonparty agreed to be bound by the litigation of others; (2) a substantive legal relationship existed between the person to be bound and a party to the judgment; (3) the nonparty was adequately represented by someone who was a party to the suit; (4) the nonparty assumed control over the litigation in which the judgment was issued; (5) a party attempted to relitigate issues through a proxy; or (6) a statutory scheme foreclosed successive litigation by nonlitigants. Griswold v.

Cty. of Hillsborough, 598 F.3d 1289, 1292 (11th Cir. 2010) (citing Taylor v. Sturgell, 553 U.S. 880, 128 S.Ct. 2161, 2172-73, 171 L.Ed.2d 155 (2008)). Collectively, these exceptions are commonly described using the catchall term “privity.” E.g., Taylor, 553 U.S. at 895 n.8, 128 S.Ct. at 2172 n.8. None of those exceptions apply here. Nothing in the record indicates — and the District Court did not find or imply — that Appellants consented to be bound by Arthur's resolution, had any legal relationship with Arthur, were represented by him in any capacity, or had any control over his claim.

Nor is there any suggestion Arthur was attempting to relitigate his claim by using Appellants as proxies. And no statute prevents successive nonparty litigation in this context. Arthur can therefore have no preclusive effect in Appellants’ cases. Nor could the District Court rely on judicial notice to make its findings in Arthur conclusive in Appellants’ cases.

A district court may take judicial notice of an adjudicative fact that is both “not subject to reasonable dispute” and either (1) “generally known within the trial court’s territorial jurisdiction” or (2) “can be accurately and readily determined from sources whose accuracy cannot reasonably be questioned.” Fed. R. Evid. 201(b). “Indisputability is a prerequisite.” United States v. Jones, 29 F.3d 1549, 1553 (11th Cir. 1994) (citing 21 C. Wright & K.

Graham, Federal Practice and Procedure: Evidence § 5104 at 485 (1977 & Supp. 1994)). “If it were permissible for a court to take judicial notice of a fact merely because it has been found to be true in some other action, the doctrine of [issue preclusion] would be superfluous.” Id.; see also Gen. Elec. Capital v. Lease Resolution Corp., 128 F.3d 1074, 1081 (7th Cir. 1997) (explaining “we cannot allow a court to achieve through judicial notice what it. cannot achieve through [issue preclusion]”); Taylor v. Charter Med.

Corp., 162 F.3d 827, 830 (5th Cir. 1998) (noting that instances where a factual finding from another court could satisfy Rule 201’s indisputability requirement for judicial notice would be rare).

In reaching that conclusion, the Taylor Court drew on decisions of the Second, Eighth, and Eleventh Circuits, explaining that all of those circuits had held that “even though a court may take judicial notice of a ‘document filed in another court ... to establish the fact of such litigation and related filings,’ a court cannot take judicial notice of factual findings of another court.” Taylor, 162 F.3d at 830 (citation omitted). Whether compounded pentobarbital was feasible and “readily available” to the ADOC as a single-drug protocol was a factual issue in Arthur.

So was the credibility of Hill’s testimony that the ADOC had been unable to obtain the drug. That the District Court, in Arthur, found for the ADOC on both issues did not transform the findings into indisputable adjudicative facts subject to judicial notice Putting aside the doctrines of issue preclusion and judicial notice and disregarding the Arthur findings of fact, the question becomes whether the record before the Court reveals a genuine factual dispute regarding the availability of compounded pentobarbital.

When we place the evidence proffered by the Appellants and the ADOC on the two sides of the scale, we find a factual dispute ‘the Court could not resolve on

summary judgment. Oh' Appellants’ side are the facts that since' January 1, 2014, forty-two executions had been carried out using a one-drug pentobarbital protocol, and that several States intended to use compounded pento-barbital as a one-drug protocol. The District Court considered these facts “inconsequential.” We disagree. From these facts it can reasonably be inferred that compounded pentobarbital was available, that executions using the drug as a one-drug protocol were ongoing, and that several States contemplated employing the protocol. On the ADOC’s side are these facts.

On January 12, 2016, while testifying in. the Arthur trial, Hill revealed that in 2015, Georgia, Missouri, Texas, and Virginia performed executions using compounded pen-tobarbital.

She contacted “departments of corrections ... in those [and other] states that have been able to obtain compounded pentobarbital [regarding] their willingness either to provide it to the ADOC or to provide their source to the ADOC;” Her conversations with these departments were “ongoing, .., not just recently, but ‘for some time.’ ” What she learned about the availability of compounded pentobarbi-tal during these conversations she did not say. Hill testified she contacted “at least twenty-nine potential sources” of compounded. pentobarbital, inquiring whether they could provide the ADOC with the drug.

Of those sources, eighteen consisted of the accredited compounding pharmacies in Alabama. She reported that none of these sources were able to provide the ADOC with the drug. Hill did not specifically identify any of the sources she contacted, or provide any details regarding the conversations she had with potential suppliers of compounded pentobarbital. What she did say was that all of her efforts failed.

Notwithstanding the lack of specificity provided by Hill’s testimony, the District Court found it credible and, moreover, that it conclusively “rebutted Arthur’s allegation that compounded pentobarbital was an available alternative.” Whether her testimony was credible and whether it tipped the scale in the ADOC’s favor was a matter for a trier of fact. See Strickland v. Norfolk Southern Ry. Co., 692 F.3d 1151, 1162 (11th Cir. 2012) (holding that when credibility is at issue “summary judgment is simply improper”); Mize v.

Jefferson City Bd. of Educ., 93 F.3d 739, 742 (11th Cir. 1996) (explaining “[i]t is not the court’s, role to weigh conflicting evidence or to make credibility determinations; the non-movant’s evidence is to be accepted for purposes of

summary judgment”). Hill’s testimony that she sought compounded pentobarbital in fall 2015, a full year after the State switched to mi-dazolam, is open to multiple inferences. When asked why she was attempting to find a source for the drug after the switch to midazolam, she testified that “[i]f at some point ...

I was able to determine that there was compounded pen-tobarbital available to us, then at that time, we would determine whether or not to amend our protocol and add that as an alternative.” One inference from this testimony is that the ADOC'was concerned about midazolam’s adequacy as a substitute for pentobarbital. Hill did not have final decisionmaking authority over the State’s execution protocol; she was the ADOC’s general counsel and was testifying as a representative of the ADOC.

Therefore, a factfinder could reasonably infer that her efforts to find a new pento-barbital source reflected her or her superiors’ doubts about midazolam’s effectiveness in eliminating pain potassium chloride could cause during executions. Of course, other inferences from Hill’s testimony could be drawn, and that is precisely why her testimony should be been submitted to the trier of fact rather than treated as conclusive on

summary judgment. When we consider the District Court’s reliance on findings of fact it made in Arthur and its determination of the credibility of critical testimony, the conclusion is inescapable that'the Court erred in holding that Appellants failed to create an issue of fact as to whether the ADOC could obtain compounded pentobarbital. B. In Arthur, the District Court made the following findings of fact regarding the availability of sodium thiopental to the ADOC: 11. Per the FDA Orange Book, sodium thiopental is no longer legally available in the United States-. 12.

While sodium thiopental may. be available from an overseas supplier and could conceivably be imported into the United States, such importation requires the approval of the FDA. There was no evidence at trial that any state’s department of corrections had obtained the FDA’s approval to import sodium thio-pental for úse in performing its executions, 13. Sodium thiopental is unavailable to the ADOC' for use in lethal injections. Arthur, 2016 WL 1551475, at *9.

Based on these findings the Court reached the following conclusion of law: Arthur ... failed to carry his burden to establish that sodium thiopental is an alternative drug that is available to the ADOC for use in his execution. Sodium thiopental is not currently an approved drug that is legally available in the United States. Evidence of its availability from an overseas supplier is insufficient to satisfy Glossip’s requirements. Id. at *10. In granting the .ADOC’s motion for

summary judgment on the availability of sodium thiopental in the Midazolam Litigation, the District Court considered whether the evidence Appellants presented in opposition to the. motion was sufficient to overcome the findings of fact it made in resolving the issue in Arthm. The Court determined that the evidence was insufficient.

Plaintiffs’ proof of the availability of. sodium thiopental is similar to the proof offered in the Arthur trial in January 2016, in that Plaintiffs have only provided evidence that sodium thiopental may be available from an overseas supplier and could conceivably be imported into the United States, but only with the FDA’s approval. ...

In short, Plaintiffs’ evidence obtained after the Arthur trial in January 2016, fails to establish that sodium thiopental has since become available to the ADOC. * * * In resolving the sodium thiopental issues, the District Court committed the same error it made in resolving the compounded pentobarbital issues. It relied on its Arthur findings as if it were invoking the doctrines of issue preclusion or judicial notice and, at

summary judgment, weighed those findings against the evidence presented in the case before it. As we have explained, neither doctrine justifies such reliance in this context. In Arthur, whether the ADOC could obtain sodium thiopental was disputed. The Court resolved the dispute by weighing the testimony of Dr. Zentner and Anne Hill. Dr. Zentner testified that although sodium thiopental was “no longer listed in the FDA Orange Book, ... it was his general understanding that sodium thiopental [wa]s available offshore and conceivably could be imported.” Arthur, 2016 WL 1551475, at *6.

The District Court found his “testimony ... credible.” Id. at *7. Hill disagreed with Dr. Zentner. She testified, after the ADOC adopted a drug protocol that included midazolam, she made no effort to obtain sodium thiopental and no effort to determine if it could be legally imported for the ADOC’s use in executions. Further, to her knowledge, sodium thiopental [was] not available in the United States, and she [was] unaware of any state that [had] imported sodium thiopental for use in executions. Id. at *7.

Hill, as the ADOC’s general counsel, was “responsible for developing, proposing, and maintaining the ADOC’s execution protocol.” Id. at 8. But she was not “the final decision and policymaker concerning the ADOC’s execution protocol.” Id. Instead, “the final decision maker [was] the ADOC’s Commissioner.” Id. The Court also found Hill’s testimony credible. Id. We assume that the Court found each witness’s testimony probative of the availability or unavailability of sodium thiopental; otherwise, the District Court would have had no need to make a credibility determination.

Thus, in deciding whether sodium thiopental was available to the ADOC, the Court must have weighed the testimonies of the two witnesses who spoke to the issue, Hill and Dr. Zentner. Dr. Zentner was a pharmacist, an “independent consultant to the pharmaceutical industry.” Id. at 4. Hill was not. Dr. Zentner testified that sodium thiopental was available offshore and conceivably could be imported even though it was no longer listed in the FDA Orange Book. Id. at 6.

Hill testified that after the ADOC adopted the current protocol, she “made no effort to obtain sodium thiopental and no effort to determine if it could be legally imported.” Id. at *7. Nonetheless, she testified that, “to her knowledge,” the drug was not available in the United States. Id. In its findings of fact, the Court found— on the basis of Dr. Zentner’s testimony— that “sodium thiopental may be available from an overseas supplier” and “conceivably” importable. Id. at *9. Hill’s testimony did not negate that fact.

But since “[t]here was no evidence at trial that any state’s department of corrections had obtained the FDA’s approval to import sodium thiopental for use in performing its executions,” the Court concluded that the plaintiff, Arthur, failed to satisfy his burden of proving that sodium thiopental was available to the ADOC. Id. at *9-10. However, the District Court arrived at that conclusion after weighing the conflicting testimony and evidence presented in that case.

The procedural posture of this case demands a different outcome because trial courts are expressly forbidden from engaging in this type of evidentiary weighing. See, e.g., Mize, 93 F.3d at 742 (noting that at

summary judgment, “[i]t is not the court’s role to weigh conflicting evidence”). Thus, the District Court erred in resolving the dispute over the availability of sodium thiopental by judging the credibility of the witnesses’ testimony and. then weighing that testimony, along with the other evidence in the record, to decide the ultimate fact — the availability of sodium thiopental.

See Jones, 683 F.3d at 1292 (“If the record presents disputed issues of fact, the court may not decide them; rather, we must deny the motion and proceed to trial.”) (alteration omitted) (quoting FindWhat Investor Grp., 658 F.3d at 1307). For that reason, the District Court’s determination that there was no genuine issue of material fact regarding the unavailability of sodium thiopental was erroneous and must be set aside. C. Whether the use of midazolam in a single-drug protocol satisfied Baze’s feasible-alternative standard was not before the District Court in Arthur.

The District Court therefore resolved the issue without reference to its Arthur decision.

Frazier’s Complaint expresses Appellants’ alternative proposal as “a single dose of midazolam.” The Complaint states that, according to the finding of a federal district court in Oklahoma, “a 500mg dose of midazolam will likely cause death in under an hour.” The Complaint also states that the “Defendants’ own expert endorses this [alternative] method.” The District Court read the Complaint’s allegation as follows: “Plaintiffs’ third proposed alternative is a 500-mg dose of mi-dazolam administered in a single-drug.protocol.” Since the ADOC had a supply of midazolam on hand, the availability of the drug was not an issue.

Given its

interpretation of the Complaint’s alternative proposal, the Court’s task was to decide whether the record established for

summary judgment purposes that a 600-milligram dose of midazolam would “significantly reduce a substantial risk of severe pain.” Glossip, 135 S.Ct. at 2737 (quoting Baze, 553 U.S. at 52, 128 S.Ct. at 1520).

According to the deposition testimony of Randall Tackett, Ph.D, “a 500-mg dose of midazolam [¡was] too low, presumably to cause death.’* Instead, Tackett explained, if midazolam were used in a one-drug protocol, he would recommend a loading dose of between 2.5 and 3.75 grams, the equivalent of between 2500 and 3750 milligrams- of midazolam, a dose five to eight times larger than the 500-milli-gram dose Plaintiffs have pled as an alternative method of execution. [He] also recommended that this loading dose should be followed by continuous infusion until death.

Based on this testimony, the Court concluded that [t]he opinion of Plaintiffs’ own expert that a 500-mg dose of midazolam is too low refutes the feasibility of Plaintiffs’ proposed alternative, and there is no other scientific evidence of record to support this novel alternative. Dr. Tack-ett’s report and opinion fail to support, and in fact undermine, Plaintiffs’ claim that a 500-mg dose of midazolam in a one-drug protocol would result in significantly less risk of substantial pain than Alabama’s present protocol, as Glossip requires.

Plaintiffs also have failed to raise a genuine dispute of material fact that a 500-mg dose of midazolam in a single-drug protocol is a feasible alternative that is readily implemented. *** The.District Court reached this disposi-tive conclusion after assessing the credibility of Dr. Tackett’s testimony and then weighing it against the contrary evidence the record presented, including testimony by the ADOC’s expert, Dr.

Daniel Buffing-ton, a clinical pharmacologist and faculty member at the University of South Florida Colleges of Medicine and Pharmacy and the President and Practice Director for Clinicial Pharmacology Services in Tampa, Florida. Although Dr. Buffington’s testimony was in the record, the Court overlooked it. Dr.

Buffington testified that “500mg of midazolam in a bolus as provided for in [Alabama’s current three-drug] protocol [would] be- sufficient on its own to cause death,” noting that 500mg is “multiple.times the toxic range for .midazolam and there are deaths [at dosages] as low as 10[mg] to 50 [mg].” In addition to Dr. Buffington’s testimony, the Court overlooked statements the ADOC made in its motion to dismiss Frazier’s complaint. See Frazier v. Myers, No. 2:13-cv-00781 (M.D. Ala. Sept. 25, 2015).

The ADOC stated that it would consent to a judgment that “suspend[s] the current lethal-injection protocol” and utilizes mi-dazolam as the sole drug in Alabama’s execution protocol. Id. at 4. “After an initial 500-milligram bolus of midazolam, the execution team w[ould] administer an additional 500-milligram dose of midazolam, if needed, until [the prisoner’s] sentence [was] carried out.” Id. at 10.

The ADOC represented that it made this proposal in reliance on Frazier’s attorney’s Rule 11 affirmation in good faith that he had “a factual basis to believe” that the use of midazolam in a single-drug execution protocol would result in a “constitutionally acceptable” execution. Id. at 3. In relying on the attorney’s affirmation, was the ADOC implying that an execution protocol consisting of a 500-milligram dose of midazolam with additional doses provided as needed would potentially violate the Eighth Amendment?

Put another way, would the ADOC induce the Court to enter a consent decree that might yield a constitutional violation? The answer to each of those questions must be “no.” The District Court’s error in determining the credibility of Dr. Tackett’s testimony and ignoring Dr. Buffington’s testimony, which was at odds with Dr. Tackett’s, and the significance of - the ADOC’s offer to substitute a single-drug protocol containing midazolam for its current three-drug protocol require vacation of the

summary judgment. Whether that proposal would “significantly reduce a substantial risk of harm,” as Baze requires, cannot be determined in a vacuum. The Court first must determine what risk of harm, if any, the current three-drug protocol creates. IV. A. The ADOC further argues that the District Court’s judgment should be affirmed under the law-of-the-case doctrine. That doctrine holds that “findings of fact and conclusions of law by an appellate court are generally binding in all subsequent proceedings in the same case in the trial court or on a later appeal.” Westbrook v.

Zant, 743 F.2d 764, 768 (11th Cir. 1984) (quoting Dorsey v. Continental Casualty Co., 730 F.2d 675, 678 (11th Cir.1984)). Accordingly, “[a]n appellate decision on an issue must be followed ... unless the presentation of new evidence or an intervening change in the controlling law dictates a different result, .or the appellate decision is clearly erroneous and, if implemented, would work a manifest injustice.” Piambino v. Bailey, 757 F.2d 1112, 1120 (11th Cir. 1985) (citing Westbrook v. Zant, 743 F.2d 764, 768-69 (11th Cir. 1984)). The ADOC points out (1) that this Court, in Brooks v.

Warden (Brooks), 810 F.3d 812 (11th Cir. 2016), and subsequently in Grayson v.

Warden (Smith), 672 Fed.Appx. 956 (11th Cir. 2016), rejected on the merits Eighth Amendment claims “virtually identical” to Appellants’ in that they were “based on the same evidence and allegations” Appellants’ claims “present here”; and (2) that the District Court had consolidated Brooks and Smith with Appellants’ cases “to form the ‘Midazolam Litigation.’ ” Given these circumstances, the ADOC argues (3) that all of the cases are one case, i.e., the same case; (4) that because the appeals in Brooks and Smith were taken in that one case, the Brooks and Smith decisions — specifically their holdings — are binding on the proceedings in this appeal; (5) that Brooks and/or Smith effectively held that the Frazier Complaint failed to allege sufficiently that the current three-drug injection protocol presents a substantial risk of serious harm; assuming it presents such harm, that the protocol does not amount to a significant alteration of the previous protocol (using pentobarbital as the first drug) so as to affect a substantial change in the State’s method of execution and restart the statute of limitations clock; and (7) that the evidence in the

summary judgment record was insufficient as a matter of law to establish that a single-drug protocol consisting of compounded pentobarbital, sodium thiopental, or midazolam was a feasible and readily available alternative execution protocol. Thus, the ADOC argues, Appellants’ challenge to the three-drug protocol’s Eighth Amendment sufficiency is now precluded. We are not persuaded that the Brooks and/or Smith holdings mandate that we affirm the District Court’s judgment on the law-of-the-case doctrine. Two independent bases inform our conclusion.

First, the cases making up the Midazolam Litigation are not one case; hence, the doctrine is inapplicable. Second, even assuming that they are one case, neither Brooks nor Smith mandates rejection of Appellants’ claims, because those cases did not address the same issue raised in this case. 1. We first consider whether Brooks or Smith’s holdings control our decision, in whole or in part, in this appeal on the theory that Appellants’ cases and Brooks and Smith’s cases are one and the same. A review of how Brooks and Smith were processed below makes it clear that neither case is one case with Appellants’ cases.

We begin with Brooks. a. On September 24, 2015, the State moved the Alabama Supreme Court to set Christopher Eugene Brooks’ execution date. On November 2, Brooks, represented by Appellants’ counsel, moved the District Court for leave to “intervene” as a party plaintiff in the litigation of Appellants’ cases. On November 5, 2015, the District Court entered an order consolidating Appellants’ cases for “discovery and trial.” Two weeks later, the Court entered a “Final Scheduling Order” in Appellants’ cases.

That order set the final hearing date for the cases to April 19-22, 2016, and fixed the fact discovery cut-off date as February 5, 2016. On November 23, 2015, the Alabama Supreme Court entered an order scheduling Brooks’ execution for January 21, 2016. On the same day, the District Court granted Brooks’ motion for leave to intervene. Brooks filed his complaint the next day. The complaint essentially mirrored Appellants’ complaints.

On December 1, 2015, the ADOC moved the District Court to dismiss Brooks’ complaint for failure to state a claim and, alternatively, offered to substitute an increased dosage of midazolam as a single-drug protocol in place of the current protocol for Brooks’ execution. Later in the day, the Court ordered the ADOC to file “under seal a proposed alternative one-drug midazolam protocol for lethal injections.” On December 3, 2015, Brooks notified the Court that he had moved the Alabama Supreme Court to stay his execution and that the State had opposed his motion.

On December 4, 2015, the ADOC filed its one-drug midazolam protocol under seal, Brooks moved the District Court to stay his execution, and the Court entered an order scheduling an evidentiary hearing on the motion for December 18, which would continue through the week of December 21, 2015 if necessary.

The order noticed a prehearing conference scheduled for December 15 — at which time the parties would identify their witnesses, scientific reports and other exhibits — and required the parties to submit to the Court on December 18, prior to the commencement of the evidentiary hearing, proposed findings of fact and conclusions. of law. On December 8, Brooks filed separately an opposition to .the ADOC’s motion to dismiss, and a response to the ADOC’s submission of the substitute midazolam protocol.

The response contended that the substitute protocol was a “legal nullity unless and until the present protocol is found unconstitutional, or Defendants eliminate the present three-drug protocol and adopt a single drug protocol using any of the feasible options that Brooks suggested.” On December 11, 2015, the ADOC responded to Brooks’ motion to stay his execution. It contended that the motion should be denied because Brooks had delayed in bringing it and because he was not likely to prevail on the merits of his Eighth Amendment claim.

The ADOO conceded that it could not “adequately address the shortcomings in Plaintiffs presentation of evidence [in support of his motion for a stay] because Brooks [was] not required to provide his evidence until December 15, 2015.” On December 14, 2014, Brooks requested the Court to issue subpoenas summoning four witnesses to appear at the eviden-tiary hearing: Jefferson S. Dunn, Anne Hill, Walter Myers and Kim Tobias Thomas. That same day, the ADOC informed the Court that the Alabama Supreme Court had denied Brooks’ application for a stay of his execution.

And, also on the same day, the Court entered' an order informing the parties that it would resolve Brooks’ motion for a stay “in the absence of a hearing,” cancelling the prehearing conference scheduled for December 15 and the evidentiary hearing that was to. begin on December 18, and declaring “moot” Brooks’ request for subpoenas. On December 22, 2015, the District Court issued a memorandum opinion and order denying Brooks’ motion for a stay.

The Court explained that it denied the stay because Brooks failed to “demonstrate a substantial likelihood of success on the merits,” he “unreasonably delayed filing suit,” and the “balance of the equities did not lie in [his] favor for a stay.” We affirmed the District Court’s denial of his motion for a stay on January 19, 2016. Brooks v. Warden, 810 F.3d 812, 816 (11th Cir. 2016). * * * It. is clear from this series of events that Brooks’ case is affiliated with Appellants’ cases for only one reason: the Court granted Brooks’ motion to “intervene” in Appellants’ cases.

After the motion was granted, Brooks’ cáse followed one route, while Appellants’ cases followed another. Further, the District Court’s decision in Appellants’ cases was based on an extensive evidentiary record. In contrast, and contrary to the ADOC’s representation, the Court decided Brooks’ motion for a stay “in the absence of a hearing” or a proffer of probative evidence, In sum, Brooks’ case and Appellants’ cases were not one case. Whatever our Brooks decision may have held, its holdings can dictate no part of the decision we reach in this appeal. b. We now turn to Smith.

On February 2, 20Í6, the State moved the Alabama Supreme Court to set a date for the execution of Ronald Bert Smith. On April 15, 2016, after discovery had closed in Appellants’ cases and just four days, before Appellants’ cases were set to be tried, Smith filed a complaint against the ADOC. Smith’s, complaint alleged three causes of action.

Only the first cause of action corresponded to the single claim stated in Frazier’s Complaint: “Alabama’s method of execution violates Plaintiffs right to be free from cruel arid unusual punishment, in violation of the Eighth Amendment to the United States Constitution, because midazolam will not sufficiently anesthetize him such that he will not feel the painful effects of potassium chloride.” The complaint proposed a single-drug protocol consisting of compounded pentobarbital, sodium thiopental, or midazolam as an alternative method of execution.

The mi-dazolam description differed from the description in Frazier’s Complaint in that it added these assertions: “The third viable alternative is a one-drug protocol consisting of a large initial dose of midazolam, followed by a continuous infusion.” The complaint further alleged, “Plaintiffs expert, Dr. Tackett, believes that a 500 mg bolus of midazolam, followed by a continuous infusion, would eventually suppress the inmate’s breathing, resulting in death.” Like Frazier’s Complaint, Smith’s complaint included factual assertions unnecessary to state an Eighth Amendment claim, see supra

Part II, and added a description of Christopher Eugene Brooks’ execution that took place on January 21, 2016, concluding that “Mr. Brooks was not insensate at the time he was injected with rocuronium bromide and potassium chloride.” Attached to Smith’s complaint were nine exhibits, A through I. They were cited in the complaint' solely by exhibit letter in footnotes, e.g., “Ex. A.” We assume that the drafter of the complaint intended that the exhibits be incorporated into the complaint in full; otherwise, they were mere extraneous surplusage.

On April 28, 2016, thirteen days after Smith filed his complaint, the District Court entered an order granting the parties’ joint motion to consolidate [Smith’s case] with the Midazolam Litigation provided that a Scheduling Order be entered to accommodate the filing of a Rule 12 motion [to dismiss for failure to state a claim] and if denied, a brief- period of time for limited fact discovery, and the filing of a motion for

summary judgment. The order contained the requested scheduling order, which set the time frames for. filing a motion to dismiss, for discovery in the event a-motion to dismiss was denied, and for

summary judgment. On May 31, the ADOC moved the District Court to dismiss Smith’s complaint. The motion contained the same alternative disposition the ADOC offered Brooks. It would consent to Smith’s execution pursuant to a single-drug protocol consisting of midazolam. The ADOC read Smith’s complaint as a “general challenge” to its three-drug protocol that uses a paralytic and potassium chloride as the second and third drugs.

Thus, it argued, the challenge should have been brought during the two-year statute of limitations period that began to run on July 31, 2002, “after Alabama changed its method of execution to lethal injection.” Smith responded to the motion on June 27, 2016, and the ADOC replied on July 5,2016. On September 14, 2016, the Alabama Supreme Court granted the State’s motion to set Smith’s execution date and scheduled the execution for December 8, 2016. Smith did not respond by moving the Alabama Supreme Court or the District Court to stay the execution.

Nothing occurred until November 9, 2016, when the District Court entered an order for the purpose of “exploring] the midazolam option pled and urged by Mr. Smith and presently offered by Defendants.” The Court explained: Because Mr. Smith has pled it and offered the option as viable, readily implemented and available, Defendants have accepted the offer. The parties all agree that (1) midazolam is available, (2) it is feasible, (3) it is readily implementable, and (4) it is not risky with regard to unnecessary pain and suffering. Mr.

Smith’s complaint proposes a one-drug protocol consisting of a “large initial dose of midazolam, followed by continuous infusion.” The Court therefore ordered the ADOC to submit to the Court for “on or before November 14, 2016, a current one-drug execution protocol and a current three-drug execution protocol for in camera inspection.” It also ordered Smith to show cause, by November 16, 2016, “why the court should not order Defendants to execute him using the method pled in his complaint, viz., a large initial dose of mi-dazolam, followed by continuous infusion.” The ADOC’s reply to Smith’s showing was due on November 18, 2016.

Smith responded to the show-cause order as directed, with this penultimate statement: This Court can and should order the Defendants to use Mr. Smith’s identified single-drug midazolam alternative. Before it can do so, this Court must enjoin the Defendants from using the present protocol, and this Court must be satisfied that Defendants have adopted an adequate protocol, including accounting for all necessary equipment and sufficient training to execute Mr. Smith using his proposed single-drug midazolam alternative. Assuming these prerequisites are satisfied, then this Court can order the Defendants to use Mr.

Smith’s single-drug midazolam alternative identified in his complaint and detailed in Dr. Tackett’s report .... In his report, Dr. Tackett, referring to the midazolam alternative, recommended “a loading dose between 2.5 and 3.75 grams” of midazolam “followed by a continuous IV infusion until death.” The ADOC, replying on November 18, rejected Smith’s proposal that the ADOC employ Dr.

Tackett’s formulation of the midazolam alternative, contending that his proposed formulation was a material departure from the formulation Smith described in his complaint, “a one-drug protocol consisting of a 500-milligram bolus of midazolam followed by a continuous infusion.” On receiving the ADOC’s reply to Smith’s response to its order to show cause, the Court took two steps, both on November 18.

First, realizing that Smith and the ADOC had reached an impasse, the Court entered an order abandoning further consideration of the midazolam alternative, Second, in a Memorandum Opinion and Order, it granted the ADOC’s motion to dismiss Smith’s complaint. Gray-son v. Dunn, 221 F.Supp.3d 1329, 1330 (M.D. Ala. 2016). The Court read Smith’s challenge to the ADOC’s execution protocol in light of the arguments the ADOC advanced in its motion to dismiss, Smith’s response, and the ADOC’s reply to his response.

As a result, the Court agreed with the ADOC that Smith was mounting a time-barred challenge to all three-drug protocols that employ a paralytic as the second drug and potassium chloride as the third drug. In his response, Smith does not address Defendants’ contention that his claim, in actuality, is a challenge against the use of any three-drug execution protocol. And this claim accrued long ago. Id. at 1334 (internal citation omitted).

The Court observed that “Smith could have challenged the ADOC’s use of a three-drug protocol any time after it was implemented fourteen years ago on July 31, 2002, but he failed to do so until 2016.” Id. at 1333. Smith appealed the District Court’s decision to this Court on the same day, November 18. On December 7, 2016, we affirmed the District Court’s decision. Grayson v. Warden, 672 Fed.Appx. 956, 958 (11th Cir. 2016).

In granting the ADOC’s motion to dismiss, the District Court made it clear that its Memorandum Opinion and Order applied only to Smith’s case. * * * Notwithstanding that the District Court’s decision only applied to Smith’s case, the ADOC now represents to this Court that Smith’s challenge to its execution protocol was “virtually identical” to Appellants’ in that both challenges were “based on the same evidence and allegations.” This representation is' incorrect. Smith’s case was decided on allegations in the pleadings, not on evidence as the ADOC represents.

At that stage, discovery was closed in Appellants’ cases and Appellants’ cases were set to be tried. Smith did not proffer the evidence he would introduce at a trial on the merits. The Court thus based its dispositive ruling on the pleadings alone. In contrast, the District Court in Appellants’ cases granted

summary judgment based on a full evidentiary record. Put simply, it cannot reasonably be argued that Smith’s case and Appellants’ cases were one case such that any issue we decided in disposing of 'Smith’s appeal should have a binding effect on our disposition'of this appeal ühder the law-of-the-case doctrine. 2. Assuming that we are wrong in concluding that Brooks and Smith’s cases are not “the same case” as.

Appellants’ cases, our decisions in those two cases do not preclude Appellants’ claims unless those prior decisions decided the same issues Appellants now raise. .The law-of-the-case doctrine “bars consideration of only those legal issues that were actually, or by necessary implication, decided in the former proceeding.” Jackson v. State of Ala. State Tenure Comm’n, 405 F.3d 1276, 1283 (11th Cir. 2005) (quoting Oladeinde v. City of Birmingham, 230 F.3d 1275, 1288 (11th Cir. 2000)).

We therefore consider sequentially what our decisions in Brooks and Smith held and whether those holdings should have a bearing on our disposition of this appeal. a. The question before this Court in Brooks was whether the District Court abused its discretion in denying Brooks’ application for a stay of his execution. Brooks, 810 F.3d at 816. “A district court ¿buses its discretion if it applies an incorrect legal standard, follows improper procedures in making the determination, or makes .findings of fact that are clearly erroneous.” Klay v.

United Healthgroup, Inc., 376 F.3d 1092, 1096 (11th Cir. 2004) (quoting Martin v. Automobili Lamborghini Exclusive, Inc., 307 F.3d 1332, 1336 (11th Cir. 2002)).

The legal standard governing the District Court’s disposition of Brooks’ motion for a stay of his execution is settled law: A court may grant a stay of execution only if the' moving party establishes that: “(1) he has a substantial likelihood of success on the merits; (2) he will suffer irreparable injury unless the injunction issues; (3) the stay would not substantially harm the other litigant; and (4) if issued, the injunction would not be adverse to .the public interest.” Brooks, 810 F.3d at 818 (emphasis added) (quoting Powell v. Thomas, 641 F.3d 1255, 1257 (11th Cir. 2011)).

The District Court denied Brooks’ application for a stay on two independent grounds. First,. Brooks failed to establish a substantial likelihood of success, on the merits of his Eighth Amendment claim. Second, he unreasonably delayed filing suit; moreover, on balance, the equities counseled a denial of a stay. The second ground is irrelevant for the ADOC’s purposes. It is our review of the first ground that resulted in the holdings that purportedly require the affirmance of the District Court’s judgment. The first ground presents a mixed question of fact and law.

It does so due to the way in which a district court determines whether a prisoner challenging a state’s method of execution has established a likelihood of success on the merits of his claim. The court engages in a three-step process. In the first step, the court decides a question of law: whether the plaintiff has stated a claim for relief — here, whether the prisoner’s claim, as alleged, satisfies the Baze standard. On appeal, we review its decision de novo. If the district court concludes that the prisoner’s claim fails to satisfy that standard, it denies a stay and the three-step process ends.

If the claim satisfies the Baze standard, the court takeá the second and third steps. In those steps, the court functions as a fact finder, and we review its findings for clear error. In the second step, the court considers the evidence the prisoner has proffered in support of his claim and finds whatever facts the evidence establishes. In the third step, the. court weighs the facts found in the second step to determine whether they demonstrate a likelihood that the prisoner’s claim will succeed at trial.

The outcome in most cases will depend on whether the court finds the testimony of the prisoner’s expert witnesses credible and persuasive. In Brooks, in addition to establishing that the currént three-drug protocol created “a substantial risk of serious harm,” Baze, 553 U.S. at 50, 128 S.Ct. at 1520, the prisoner had to demonstrate that the substitution of midazolam for pentobarbital as the first drug in the protocol constituted a “substantial change” in the State’s method of execution; otherwise, his claim was time-barred. Whether he carried that burden was a factual determination.

We reviewed the District Court’s finding that he failed to carry that burden for clear error. We consider first the holdings the Brooks Court reached in affirming the District Court’s decision that Brooks failed to establish a substantial likelihood of success on the merits of his Eighth Amendment claim.

After that, we Consider the holdings the panel reached in affirming the District Court’s finding that Brooks failed to proffer any evidence that would establish a substantial likelihood that the substitution of midazolam for pentobarbital affected a “substantial change” in the State’s method of execution and thus reset the statute of limitations clock. i.

The Brooks panel assumed that the District Court, in taking the first step in deciding the likelihood-of-success issue, concluded that Brooks’ statement of his Eighth Amendment claim satisfied the Baze standard: In his intervenor complaint, Brooks ... alleged that midazolam — the first of the three drugs used in Alabama’s execution protocol — will not properly anesthetize him so as to prevent him from feeling an “unconstitutional level of pain” associated with the injection of the other two drugs that will kill him (rocu-ronium bromide and potassium chloride).

He. also claims that midazolam may exhibit a “ceiling effect” — that is, at a certain point, an increase in the dose administered will not have any greater effect on an inmate. Brooks says that there are three alternative methods of execution available to the ADOC that significantly reduce the risk of an unconstitutional level of pain: (1) a single injection of pentobarbital; (2) a single injection of sodium thiopental; or (3) a single injection of midazolam. Id. at 819. Next, the panel looked to the evidence Brooks proffered in support of these allegations. Like the District Court, it found none.

The Court agreed with the District Court that Brooks proffered nothing to suggest he was likely to succeed at trial under the Baze standard. As for the risk of pain the current protocol allegedly created, “[o]n this record, [the Court concluded that] Brooks has not established a substantial likelihood that the State’s lethal injection protocol creates a ‘demonstrated risk of severe pain.’ ” Id, at 818.

Among other things, Brooks proffered no evidence indicating a likelihood that a 500-milligram dose of midazolam would not render the prisoner insensate prior to the administration of the second and third drugs, the paralytic and potassium chloride. As for the single-drug protocol of compounded pentobarbital, sodium thiopental, or midazolam Brooks proposed as an alternative method of execution, the panel concluded that Brooks had not shown a substantial likelihood that any of the three drugs, alone, presented a feasible and readily implementable alternative as defined by Baze. Id. at 819-22.

Moreover, “[o]n this record, Brooks has failed to show a substantial likelihood that a single-drug execution protocol using only midazo-lam is a feasible, readily implementable, and significantly safer method of execution,” Id. at 821.

In concluding that the District Court had not abused its discretion in deciding that Brooks had not established a likelihood of success, the Court made it clear that its decision was based solely on the record presented: [n]othing we say should be read as holding that single-injection drug protocols could not offer valid alternatives.-Rather, on this record, we hold only that Brooks has failed to show that Alabama’s three-drug protocol creates “a demonstrated risk of severe pain” and that “that risk is substantial when compared to the known and available alternatives.” Id. at 822 (emphasis added). ii.

Additionally, the District Court found as a fact that Brooks failed to proffer any evidence demonstrating that it was substantially likely that the substitution of midazolam for pentobarbital affected a substantial change in the State’s method of execution. The finding served as an alternative ground for concluding that Brooks failed to establish the first factor of the stay test, a substantial likelihood of success on the merits. To make the requisite showing, Brooks had to present the District Court with some evidence. Mere allegations were insufficient. But Brooks presented nothing.

We therefore affirmed the Court’s finding: The crux of Brooks’s argument is that the three-drug protocol Alabama implemented on September 11, 2014, constitutes a substantial change because mi-dazolam as the first drug — as opposed to pentobarbital or sodium thiopental — is not an effective analgesic (or pain reliever). But he has provided no evidence to show that pentobarbital or sodium thio-pental would have been any more effective in numbing him against the alleged risk of pain posed by the administration of the second and third drugs, which have remained essentially unchanged since 2002.

Because he has proffered nothing to establish, by a substantial likelihood, that midazolam constituted a “substantial change” from the earlier protocol ], we cannot say that the 2014 switch to midazolam triggered a new statute-of-limitations period, Id. (emphasis added). This language is not a holding that Brooks’ complaint failed to allege that the substitution of.midazolam for pentobarbital in the three-drug protocol did not affect a substantial change in the State’s method of execution.

Rather, it represents an implicit holding that the District Court did clearly err in finding, based on Brooks’ nonexistent proffer, that Brooks had not demonstrated a likelihood that the substitution affected a substantial change. *** In sum, we concluded that the District Court did not abuse its discretion in finding that Brooks failed to demonstrate a substantial likelihood he would succeed on the merits. Our conclusion was based solely on the insufficiency of Brooks’ proffer.

His proffer did not present evidence sufficient to demonstrate he was likely to meet both prongs of the Baze standard, and it failed to present'any evidence whatsoever that the substitution of midazolam for pen-tobarbital constituted a substantial change in protocol and thus reset the statute of limitations clock. For both of those reasons, we held that the District Court did not abuse its discretion-in denying Brooks’ motion for a stay. That is the law of the case. The ADOC agrees that no abuse of discretion occurred, but disagrees -that that holding is the only law of the case.

It contends that given the virtual identity of the allegations in Brooks and Appellants’ respective complaints, the law of the case also includes a holding that -Appellants’ complaints fail to state a claim for relief for:two reasons: First, the substitution of midazolam- for pentobarbital as the initial drug in the three-drug protocol did not create -a “substantial change” in the protocol, one that “significantly alterfed] the method of execution,” Gissendaner v. Comm’r, Ga. Dept. of Corrections, 779 F.3d 1275, 1282 (11th C

[…]