Whole Woman's Health v. Ken Paxton

Case: 17-51060 Document: 00515984856 Page: 1 Date Filed: 08/18/2021 United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit FILED August 18, 2021 No. 17-51060 Lyle W. Cayce Clerk Whole Woman’s Health, on behalf of itself, its staff, physicians and patients; Planned Parenthood Center for Choice, on behalf of itself, its staff, physicians, and patients; Planned Parenthood of Greater Texas Surgical Health Services, on behalf of itself, its staff, physicians, and patients; Planned Parenthood South Texas Surgical Center, on behalf of itself, its staff, physicians, and patients; Alamo City Surgery Center, P.L.L.C., on behalf of itself, its staff, physicians, and patients, doing business as Alamo Women’s Reproductive Services; Southwestern Women’s Surgery Center, on behalf of itself, its staff, physicians, and patients; Curtis Boyd, M.D., on his own behalf and on behalf of his patients; Jane Doe, M.D., M.A.S., on her own behalf and on behalf of her patients; Bhavik Kumar, M.D., M.P.H., on his own behalf and on behalf of his patients; Alan Braid, M.D., on his own behalf and on behalf of his patients; Robin Wallace, M.D., M.A.S., on her own behalf and on behalf of her patients, Plaintiffs—Appellees, versus Ken Paxton, Attorney General of Texas, in his official capacity; Sharen Wilson, Criminal District Attorney for Tarrant County, in her official capacity; Barry Johnson, Criminal District Attorney for McLennan County, in his official capacity, Defendants—Appellants. Case: 17-51060 Document: 00515984856 Page: 2 Date Filed: 08/18/2021 Appeal from the United States District Court for the Western District of Texas USDC No. 1:17-CV-690 Before Owen, Chief Judge, and Jones, Smith, Stewart, Dennis, Elrod, Haynes, Graves, Higginson, Costa, Willett, Ho, Engelhardt, and Wilson, Circuit Judges. ∗ Jennifer Walker Elrod and Don R. Willett, Circuit Judges, joined by Owen, Chief Judge, and Jones, Smith, Haynes, Ho, Engelhardt, and Wilson, Circuit Judges: ∗∗ We must decide whether the district court erred in permanently enjoining Texas’s Senate Bill 8 (SB8), which prohibits a particular type of dilation and evacuation (D&E) abortion method. SB8 refers to the prohibited method as “live dismemberment” because doctors use forceps to separate, terminate, and remove the fetus. SB8 requires doctors to use alternative fetal-death methods. The district court declared SB8 facially unconstitutional. It held that SB8 imposes an undue burden on a large fraction of women, primarily because it determined that SB8 amounted to a ban on all D&E abortions. But viewing SB8 through a binary framework—that either D&Es can be done only by live dismemberment or else women cannot receive abortions in the second trimester—is to accept a false dichotomy. Instead, the record shows that doctors can safely perform D&Es and comply with SB8 using methods that are already in widespread use. In permanently enjoining SB8, the district ∗ Judges Southwick, Duncan, and Oldham are recused. ** Chief Judge Owen and Judge Haynes concur in the judgment only. 2 Case: 17-51060 Document: 00515984856 Page: 3 Date Filed: 08/18/2021 No. 17-51060 court committed numerous, reversible legal and factual errors: applying the wrong test to assess SB8, disregarding and misreading the Supreme Court’s precedents in Planned Parenthood of Southeastern Pennsylvania v. Casey and Gonzales v. Carhart, and bungling the large-fraction analysis. Accordingly, we VACATE the district court’s permanent injunction. Moreover, remanding to the district court would be futile here because the record permits only one conclusion. The plaintiffs have failed to carry their heavy burden of proving that SB8 would impose an undue burden on a large fraction of women. We REVERSE and RENDER. I. Dilation and evacuation is an abortion method commonly used after the beginning of the 15th week. It begins with the dilation phase, which is lengthy and can take two or even three days to complete. First, the woman is given the option of conscious sedation and then is administered medication for dilation. If medication cannot alone cause sufficient dilation, the doctor injects a local anesthetic directly into the woman’s cervix. After the cervix has been numbed, the doctor inserts osmotic dilators into the cervical canal, which absorb liquid and expand to allow the removal of the fetus and placenta. Starting around 18 weeks gestation, this expansion process normally happens overnight, requiring the woman to come back the next day for the rest of the abortion procedure. Once sufficient dilation has occurred, the second phase begins and the doctor evacuates (removes) the fetus. Doctors use three main evacuation methods: (1) the suction method alone to terminate, separate, and remove the fetus; (2) suction and forceps together to terminate, separate, and remove the fetus; or (3) various fetal-death techniques (e.g., digoxin injections) to terminate the fetus before using forceps (sometimes combined with suction) 3 Case: 17-51060 Document: 00515984856 Page: 4 Date Filed: 08/18/2021 No. 17-51060 to separate and remove the fetus. Unlike the dilation phase, evacuation is relatively brief and can be done in “a few minutes.” In 2017, the Texas legislature enacted SB8, which allows any abortion accomplished by dilation and suction alone (the first method) or accomplished by fetal death caused without forceps followed by evacuation with forceps (the third method), but regulates the second method by prohibiting a doctor from using forceps to separate the fetal tissue and thereby terminate the fetus via live dismemberment. 1 SB8 states: A person may not intentionally perform a dismemberment abortion unless the dismemberment abortion is necessary in a medical emergency. 2 A “dismemberment abortion” is defined by the legislature as: an abortion in which a person, with the purpose of causing the death of an unborn child, dismembers the living unborn child and extracts the unborn child one piece at a time from the uterus through the use of clamps, grasping forceps, tongs, scissors, or a similar instrument that, through the convergence of two rigid levers, slices, crushes, or grasps, or performs any combination of those actions on, a piece of a the unborn child’s body to cut or rip the piece from the body. 3 A “medical emergency” is defined as a: life-threatening physical condition aggravated by, caused by, or arising from a pregnancy that, as certified by a physician, places the woman in danger of death or a serious risk of substantial 1 See Act of May 26, 2017, 85th Leg. R.S., ch. 441, § 6, 2017 Tex. Gen. Laws 1164, 1165–67 (eff. Sept. 1, 2017) (codified as Tex. Health & Safety Code §§ 171.151–.154). 2 Id. § 171.152. 3 Id. § 171.151. 4 Case: 17-51060 Document: 00515984856 Page: 5 Date Filed: 08/18/2021 No. 17-51060 impairment of a major bodily function unless an abortion is performed. 4 When a medical emergency arises, the doctor may proceed straight to live dismemberment with forceps. 5 SB8 does not regulate the dilation phase of the abortion or any other evacuation method. SB8 does not ban the use of suction during any abortion procedure. SB8 does not prohibit a doctor from having forceps “on hand” to use after fetal death has occurred or to use if a medical emergency arises. 6 What SB8 does do is prohibit one particular evacuation method in one particular set of circumstances—live dismemberment by forceps when a medical emergency does not exist. Thus, doctors may comply with SB8 by using only suction to achieve fetal death and remove the fetus—or, at later gestational ages, using either suction or a digoxin injection to cause fetal death before forcep-dismemberment and removal. 7 The plaintiffs here, six abortion clinics and five individual doctors who provide abortions, brought this facial challenge against SB8 in federal court. They allege that SB8 imposes an undue burden on women seeking abortions in the second trimester of pregnancy. The defendants are various Texas law 4 Id. § 171.002. 5 Id. § 171.152. 6 Although SB8 prohibits using “clamps, grasping forceps, tongs, scissors, or . . . similar instrument[s]” to cause fetal death, id. § 171.151, we will refer to those items collectively as “forceps.” 7 A potassium-chloride injection and umbilical-cord transection are additional alternatives to live dismemberment, and the State presented testimony about them at the trial. As far back as Stenberg v. Carhart, 530 U.S. 914, 925 (2000), the Supreme Court has recognized potassium chloride, in particular, as an established method of causing fetal death. We need not discuss these additional alternatives, however, because digoxin and suction are already widely used and are alone sufficient for our holding in this case that the plaintiffs failed to prove an undue burden on a large fraction of women in the relevant circumstances. 5 Case: 17-51060 Document: 00515984856 Page: 6 Date Filed: 08/18/2021 No. 17-51060 enforcement officials. Texas argues that SB8 does not impose an undue burden on a large fraction of women in the relevant circumstances because there are safe and available alternatives for causing fetal death without forceps. The district court granted a temporary restraining order preventing SB8’s enforcement, followed by a five-day bench trial. The district court subsequently ruled that SB8 is facially unconstitutional and entered a permanent injunction. Texas appealed. A panel of our court held the case in abeyance pending the Supreme Court’s decision in June Medical Services L.L.C. v. Russo, 140 S. Ct. 2103 (2020). Once the June Medical opinion was issued, we ordered supplemental briefing from the parties on the effect, if any, of June Medical on this appeal. Texas moved for a stay of the district court’s injunction pending appeal. A two-member majority of the panel denied the motion with Judge Willett in dissent. See Whole Woman’s Health v. Paxton, 972 F.3d 649 (5th Cir. 2020). The panel subsequently issued its opinion on the merits, ruling that SB8 is unconstitutional under Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292, 2309 (2016), with Judge Willett in dissent again. See Whole Woman’s Health v. Paxton, 978 F.3d 896 (5th Cir.), vacated and reh’g en banc granted, 978 F.3d 974 (5th Cir. 2020). A majority of the members of our court voted to take the case en banc. II. A. We review the district court’s permanent injunction for abuse of discretion. Scott v. Schedler, 826 F.3d 207, 211 (5th Cir. 2016). The district court abuses its discretion if it “(1) relies on clearly erroneous factual findings when deciding to grant or deny the permanent injunction, (2) relies on erroneous conclusions of law when deciding to grant or deny the permanent injunction, or (3) misapplies the factual or legal conclusions when fashioning 6 Case: 17-51060 Document: 00515984856 Page: 7 Date Filed: 08/18/2021 No. 17-51060 its injunctive relief.” Alcatel USA, Inc. v. DGI Techs., Inc., 166 F.3d 772, 790 (5th Cir. 1999) (alteration omitted) (quoting Peaches Ent. Corp. v. Ent. Repertoire Assocs., 62 F.3d 690, 693 (5th Cir. 1995)). We review questions of fact for clear error and legal conclusions de novo. Scott, 826 F.3d at 211. A clear error has occurred when we are “left with the definite and firm conviction that a mistake has been committed.” June Medical, 140 S. Ct. at 2141 (Roberts, C.J., concurring) (quoting United States v. U.S. Gypsum Co., 333 U.S. 364, 395 (1948)). If “a district court’s findings rest on an erroneous view of the law, they may be set aside on that basis.” Pullman-Standard v. Swint, 456 U.S. 273, 287 (1982); see also Aransas Project v. Shaw, 775 F.3d 641, 658 (5th Cir. 2014) (“When, as here, a court’s factual finding ‘rest[s] on an erroneous view of the law’, its factual finding does not bind the appellate court.” (quoting Swint, 456 U.S. at 287)); Thornburg v. Gingles, 478 U.S. 30, 79 (1986) (holding that appellate courts’ power to correct extends to “finding[s] of fact that [are] predicated on a misunderstanding of the governing rule of law” (quoting Bose Corp. v. Consumers Union of U.S., 466 U.S. 485, 501 (1984))). And “when the record permits only one resolution of the factual issue after the correct law is applied, remand is unnecessary.” Aransas Project, 775 F.3d at 658 (citing Swint, 456 U.S. at 292); see also Swint, 456 U.S. at 292 (“[W]here findings are infirm because of an erroneous view of the law, a remand is the proper course unless the record permits only one resolution of the factual issue.”). B. In Planned Parenthood of Southeastern Pennsylvania v. Casey, the Supreme Court repudiated lower courts’ post-Roe v. Wade practice of invalidating abortion regulations that “in no real sense deprived women of the ultimate decision” to have an abortion. 505 U.S. 833, 875 (1992). Casey established three principles: (1) the woman has a “right . . . to choose to have 7 Case: 17-51060 Document: 00515984856 Page: 8 Date Filed: 08/18/2021 No. 17-51060 an abortion before viability . . . without undue interference from the State”; (2) the State has the “power to restrict abortions after fetal viability”; and (3) the State has “legitimate interests from the outset of the pregnancy in protecting the health of the woman and the life of the fetus.” Id. at 846. In Casey, the Court set out the familiar undue-burden test, stating that “[o]nly where state regulation imposes an undue burden on a woman’s ability to make” the decision to have an abortion does the State violate the Due Process Clause. Id. at 874. “A finding of an undue burden is a shorthand for the conclusion that a state regulation has the purpose or effect of placing a substantial obstacle in the path of a woman seeking an abortion of a nonviable fetus.” Id. at 877 (emphasis added). The Casey Court further explained that “[t]he fact that a law which serves a valid purpose, one not designed to strike at the right itself, has the incidental effect of making it more difficult or more expensive to procure an abortion cannot be enough to invalidate it.” Id. at 874. When a plaintiff claims that an abortion law is facially invalid—as opposed to unconstitutional as applied to her—we ask whether the law would impose a substantial obstacle on a “large fraction” of women in the relevant circumstances. Id. at 895. 8 We first determine the denominator of the fraction by identifying the number of women “for whom the law is a restriction, not the [number of women] for whom the law is irrelevant.” Id. at 894. After determining that proper denominator, courts should deduce the numerator—the number of women for whom the abortion regulation 8 The large-fraction test is a generous exception to the normal burden that litigants bear in facial challenges. In non-abortion cases, a plaintiff must establish that no set of circumstances exists under which the law would be constitutional. See Women’s Med. Pro. Corp. v. Voinovich, 130 F.3d 187, 193–95 (6th Cir. 1997) (noting the Supreme Court’s “inconsistent” rules in facial challenges between abortion cases and non-abortion cases); see also Planned Parenthood of Cent. N.J. v. Farmer, 220 F.3d 127, 142–43 (3d Cir. 2000) (same). 8 Case: 17-51060 Document: 00515984856 Page: 9 Date Filed: 08/18/2021 No. 17-51060 would impose an “undue burden.” Id. at 895. The plaintiff bears the burden of proving a large fraction—and that burden is “heavy.” Gonzales v. Carhart, 550 U.S. 124, 167–68 (2007). III. The district court concluded that SB8 amounts to a complete ban on “standard D&E” abortions. This conclusion rested on four errors—each of which independently compels reversal. First, the district court applied an incorrect legal test to assess SB8. Second, the district court disregarded Casey, Gonzales, and Hellerstedt by dismissing the State’s interests and committing myriad other legal errors. Third, the district court failed to properly evaluate SB8’s burdens under Casey and, in doing so, improperly concluded that the only safe second-trimester abortion procedure is live dismemberment by forceps. Finally, the district court misapplied the large- fraction test by incorrectly determining the number of women upon whom SB8 would place an undue burden (the numerator) and incorrectly determining the number of women to whom SB8 would apply (the denominator). In sum, the district court’s opinion rested on bad law, bad facts, and bad math. We address each error in turn. A. 1. For decades, Casey’s undue-burden test was the governing standard for assessing abortion regulations. Five years ago, in Whole Woman’s Health v. Hellerstedt, the Supreme Court stated that Casey’s undue-burden test “requires that courts consider the burdens a law imposes on abortion access together with the benefits those laws confer.” Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292, 2309 (2016). This language in Hellerstedt came 9 Case: 17-51060 Document: 00515984856 Page: 10 Date Filed: 08/18/2021 No. 17-51060 to be recognized by some as a “balancing test.” Id. at 2324 (Thomas, J., dissenting). 9 Last summer in June Medical—issued after the district court enjoined SB8—the Supreme Court again tackled the meaning of “undue burden.” 140 S. Ct. at 2112 (plurality opinion). The four-Justice plurality considered the law’s benefits together with its burdens. Id. Chief Justice Roberts wrote separately, concurring in the judgment but disavowing any balancing test. Id. at 2135–37 (Roberts, C.J., concurring). The Chief Justice explained that the proper standard is the straightforward undue-burden test and that neither Casey nor Hellerstedt established a balancing test. “In neither [Hellerstedt nor Casey] was there [a] call for consideration of a regulation’s benefits.” Id. at 2139. The Chief Justice noted that the Court in Hellerstedt explicitly stated that it “appl[ied] the undue burden standard of Casey” and that it needed “[n]othing more” than the burdens analysis to hold the challenged law unconstitutional. Id. at 2138–39. As the Chief Justice put it, Hellerstedt, properly understood, was simply an iteration of Casey’s undue-burden standard, which “require[s] a substantial obstacle before striking down an abortion regulation.” Id. at 2139. “Laws that do not pose a substantial obstacle to abortion access are permissible, so long as they are ‘reasonably related’ to a legitimate state interest.” Id. at 2135 (quoting Casey, 505 U.S. at 878). The only relevance of an abortion regulation’s asserted “benefits” is “in considering the threshold requirement that the State have a ‘legitimate 9 Previously, our circuit explicitly eschewed a benefits-versus-burdens balancing test. “In our circuit, we do not balance the wisdom or effectiveness of a law against the burdens the law imposes.” Whole Woman’s Health v. Cole, 790 F.3d 563, 587 n.33 (5th Cir.) (quoting Whole Woman’s Health v. Lakey, 769 F.3d 285, 297 (5th Cir.) (citing Planned Parenthood of Greater Tex. Surgical Health Servs. v. Abbott, 748 F.3d 583, 593–94 (5th Cir. 2014) (Abbott II)), vacated in part, 574 U.S. 931 (2014))), modified, 790 F.3d 598 (5th Cir. 2015), rev’d and remanded sub nom. Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292 (2016). 10 Case: 17-51060 Document: 00515984856 Page: 11 Date Filed: 08/18/2021 No. 17-51060 purpose’ and that the law be ‘reasonably related to that goal.’” Id. at 2138 (first quoting Casey, 505 U.S. at 878 (plurality opinion); and then quoting id. at 882 (joint opinion)). The Chief Justice opined in June Medical that trying to weigh the State’s interest in protecting fetal life is impossible—and therefore a balancing test is impossible—because how do you “assign weight to such imponderable values?” Id. at 2136. Agreeing with all but two pages of the plurality’s opinion, Chief Justice Roberts said that the inquiry should have ended after the plurality analyzed the law’s burdens on abortion access. 2. Under the Marks rule, the Chief Justice’s concurrence is June Medical’s controlling opinion. In Marks v. United States, the Supreme Court instructed that “[w]hen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.” 430 U.S. 188, 193 (1977) (internal quotation marks omitted) (quoting Gregg v. Georgia, 428 U.S. 153, 169 n.15 (1976) (opinion of Stewart, Powell, and Stevens, JJ.)). We have clarified that this principle “is only workable where there is some common denominator upon which all of the justices of the majority can agree.” United States v. Duron-Caldera, 737 F.3d 988, 994 n.4 (5th Cir. 2013) (internal quotation marks omitted) (quoting United States v. Eckford, 910 F.2d 216, 219 n.8 (5th Cir. 1990)). In June Medical, the common denominator is the undue-burden (substantial-obstacle) analysis, which took up more than 80% of the plurality’s reasoning. Indeed, the Chief Justice concluded that, “for the reasons the plurality explain[ed],” the law “imposed a substantial obstacle” to abortion access. June Medical, 140 S. Ct. at 2139, 2141 (Roberts, C.J., concurring). The only part the Chief Justice disagreed with was the 11 Case: 17-51060 Document: 00515984856 Page: 12 Date Filed: 08/18/2021 No. 17-51060 plurality’s two-page benefits analysis. So the Chief Justice’s test is a narrower version (only burdens) of the plurality’s test (benefits and burdens). Accordingly, the Chief Justice’s concurrence controls and we do not balance the benefits and burdens in assessing an abortion regulation. We agree with the Eighth and Sixth Circuits in holding that the Chief Justice’s concurrence controls. See Hopkins v. Jegley, 968 F.3d 912, 915 (8th Cir.) (“Chief Justice Robert[s]’s vote was necessary in holding unconstitutional Louisiana’s admitting-privileges law, so his separate opinion is controlling.”), reh’g and reh’g en banc denied, No. 4985329 (2020); 10 EMW Women’s Surgical Ctr., P.S.C. v. Friedlander, 978 F.3d 418, 437 (6th Cir.) (“The Chief Justice’s opinion in June Medical Services concurs in the judgment on the narrowest grounds, so it is the ‘controlling opinion’ from that decision.” (quoting Marks, 430 U.S. at 193)), reh’g en banc denied, No. 104-1 (6th Cir. Dec. 31, 2020). 11 These circuits held that the Chief 10 See also Little Rock Fam. Plan. Servs. v. Rutledge, 984 F.3d 682, 687 n.2 (8th Cir. 2021) (“Chief Justice Roberts’s concurring opinion [in June Medical] is controlling.”), petition for cert. filed, No. 20-1434 (Apr. 13, 2021). 11 While noting that the Chief Justice’s concurrence offered the narrowest basis for June Medical’s judgment, the Seventh Circuit has taken a somewhat different approach to Marks’s application to June Medical. Planned Parenthood of Ind. & Ky., Inc. v. Box, 991 F.3d 740, 741 (7th Cir. 2021), petition for cert. filed, No. 20-1375 (Mar. 29, 2021). The Seventh Circuit views only one part of the Chief Justice’s concurrence as binding—the part where the Chief Justice agreed with the plurality that Hellerstedt “was entitled to stare decisis effect on essentially identical facts.” Id. at 748. The Seventh Circuit also stated that “the Marks rule tells us that June Medical did not overrule [Hellerstedt]” and that “[Hellerstedt] remains precedent binding on lower courts.” Id. On this point, we agree with the Seventh Circuit. Where we diverge from the Seventh Circuit is our respect for the full weight of the Chief Justice’s controlling concurrence, which observed that neither Casey nor Hellerstedt established a balancing test. “As middle-management circuit judges, we cannot overrule the Supreme Court. But neither should we ‘underrule’ it.” Whole Woman’s Health, 978 F.3d at 920 (Willett, J., 12 Case: 17-51060 Document: 00515984856 Page: 13 Date Filed: 08/18/2021 No. 17-51060 Justice’s concurrence “clarified that the undue burden standard is not a balancing test.” EMW Women’s Surgical Ctr., P.S.C., 978 F.3d at 437; 12 see also Hopkins, 968 F.3d at 915 (“According to Chief Justice Roberts, the appropriate inquiry under Casey is . . . ‘not whether benefits outweighed burdens’. . . . [Benefits are] ‘consider[ed] [only in] the threshold requirement that the State [has] a “legitimate purpose” and that the law be “reasonably related to that goal.”’” (first quoting June Medical, 140 S. Ct. at 2137–38; then quoting Casey, 505 U.S. at 878)). dissenting). “Our duty is to harmonize its decisions as well as possible.” Nelson v. Quarterman, 472 F.3d 287, 339 (5th Cir. 2006) (Jones, C.J., dissenting on other grounds). Like the Seventh Circuit, the Eleventh Circuit has chosen to underrule the Chief Justice’s controlling concurrence. In Reproductive Health Services v. Strange, the court noted that the June Medical plurality opinion applied a benefits-versus-burdens balancing test. 3 F.4th 1240, 1259 (11th Cir. 2021). The court also explained that the plurality opinion and the Chief Justice’s concurrence shared a “common ground,” which is the “conclusion that the . . . statute constituted an undue burden.” Id. Despite acknowledging the shared analysis and conclusion of the plurality opinion and the Chief Justice’s concurrence, the Eleventh Circuit confusingly held that the Chief Justice’s concurrence was not “narrower” than the plurality opinion and thus not controlling under Marks. Id. 12 Even though it acknowledged that the EMW panel had held that the Chief Justice’s opinion in June Medical was controlling under the Marks rule, a subsequent panel of the Sixth Circuit decided to disregard the EMW panel’s holding when it denied a state’s motion to stay pending appeal in an abortion case involving waiting periods. Bristol Reg’l Women’s Ctr., P.C. v. Slatery, 988 F.3d 329, 337–38 (6th Cir.), opinion vacated, 994 F.3d 774 (6th Cir. 2021). Judge Thapar dissented and pointed out that the panel majority erred because “the holding of a published panel opinion [EMW] binds all later panels unless overruled or abrogated en banc or by the Supreme Court.” Id. at 346 (Thapar, J., dissenting) (quoting Wright v. Spaulding, 939 F.3d 695, 700 (6th Cir. 2019)). The Sixth Circuit decided to take the case straight to en banc review. See Bristol Reg’l Women’s Ctr., P.C. v. Slatery, 993 F.3d 489 (6th Cir. 2021). Moreover, a more recent Sixth Circuit opinion confirms that that circuit views the Chief Justice’s concurrence as controlling. See Preterm-Cleveland v. McCloud, 994 F.3d 512, 524 (6th Cir. 2021) (en banc) (explaining that the EMW decision applied Marks “to determine that the June Medical concurrence was the narrowest opinion and, therefore, the governing law”). 13 Case: 17-51060 Document: 00515984856 Page: 14 Date Filed: 08/18/2021 No. 17-51060 Under the Chief Justice’s controlling concurrence in June Medical, the district court erred by balancing SB8’s benefits against its burdens. That is reason alone to reject the district court’s findings. See Swint, 456 U.S. at 287 (“[A] district court’s findings [that] rest on an erroneous view of the law . . . may be set aside on that basis.”). But, as explained below, the district court erred under all of the Supreme Court’s relevant precedents—Casey, Hellerstedt, Gonzales, and June Medical. B. 1. Despite Casey’s clear language, repeated in Gonzales, that the State has legitimate and substantial interests in fetal life throughout pregnancy, the district court dismissed the State’s interests as deserving “only marginal consideration” and “having [] primary application once the fetus is capable of living outside the womb.” What is more, the State asserted several interests in passing SB8 in addition to respect for fetal life—benefits to patients both physically and psychologically, medical and societal ethics, and informed consent for women seeking abortions. The Supreme Court accepted all of these interests in Gonzales. 550 U.S. at 158, 163. Yet the district court disregarded all of them here, contravening both Casey and Gonzales. See Casey, 505 U.S. at 846; Gonzales, 550 U.S. at 145–46. First, the State asserted its interest in the health and safety of women seeking abortions. The State presented evidence showing that women seeking abortions benefit physically and psychologically when fetal death occurs before dismemberment. For example, the Planned Parenthood Federation of America Manual of Medical Standards and Guidelines tells patients that a study showed that “more than 90 percent of women . . . prefer[] knowing that fetal death occurred before the abortion surgery began.” The American Institute of Ultrasound and Medicine agrees and has also found that doctors have a similar preference and believe that 14 Case: 17-51060 Document: 00515984856 Page: 15 Date Filed: 08/18/2021 No. 17-51060 inducing fetal death can help with emotional difficulties for the patient. Casey noted that “most women considering an abortion would deem the impact on the fetus relevant, if not dispositive, to the decision.” 505 U.S. at 882. Beyond psychological benefits, terminating the fetus before dismembering it makes the abortion physically easier for the mother. As the Supreme Court noted in Gonzales: “Fetal demise may cause contractions and make greater dilation possible. Once dead, moreover, the fetus’s body will soften and its removal will be easier.” 550 U.S. at 136. Second, the State asserted its interest in providing a greater degree of dignity in a soon-to-be-aborted fetus’s death. The State argues that, by requiring doctors to choose alternatives to a brutal abortion procedure, SB8 evinces the State’s “profound respect for the life within the woman.” Id. at 157. Dismemberment D&Es are self-evidently gruesome. It has long been illegal to kill capital prisoners by dismemberment. See In re Kemmler, 136 U.S. 436, 447 (1890). It is also illegal to dismember living animals. Tex. Penal Code § 42.092. The State urges that SB8 would simply extend the same protection to fetuses. 13 In its opinion, the district court dismissed the State’s interest in respecting fetal life with the comment that “[a]n abortion always results in 13 The State also argues that SB8 may protect fetuses from feeling the pain of being dismembered alive. The Supreme Court “has given state and federal legislatures wide discretion to pass legislation in areas where there is medical and scientific uncertainty.” Gonzales, 550 U.S. at 163. The record here reveals that scientists are unsure at what gestational age a fetus begins to feel pain. The plaintiffs and the State presented conflicting expert testimony and there appears to be a wide range of views. Faced with this uncertainty, the State is permitted to exercise its “wide discretion” and err on the side of caution—especially in light of the numerous benefits provided by killing the fetus before it is dismembered and evacuated. June Medical, 140 S. Ct. at 2136 (Roberts, C.J., concurring) (quoting Gonzales, 550 U.S. at 163). “Medical uncertainty does not foreclose the exercise of legislative power in the abortion context any more than it does in other contexts.” Gonzales, 550 U.S. at 164. 15 Case: 17-51060 Document: 00515984856 Page: 16 Date Filed: 08/18/2021 No. 17-51060 the death of the fetus.” The district court also noted that the State’s interest “does not add weight to tip the balance in the State’s favor.” The district court’s analysis cannot be reconciled with the Supreme Court’s instruction in Gonzales: The government may use its voice and its regulatory authority to show its profound respect for the life within the woman. . . . Where it has a rational basis to act, and it does not impose an undue burden, the State may use its regulatory power to bar certain procedures and substitute others, all in furtherance of its legitimate interests in regulating the medical profession in order to promote respect for life, including life of the unborn. 550 U.S. at 157–58 (emphases added). Third, the State asserted its interest in promoting societal and medical ethics. “There can be no doubt the government ‘has an interest in protecting the integrity and ethics of the medical profession.’” Id. at 157 (quoting Washington v. Glucksberg, 521 U.S. 702, 731 (1997)). SB8’s provisions are supported by general principles of medical ethics, which require accounting for the harms to and dignity of both the mother and the fetus. Finally, the State asserted its interest in ensuring that women give informed consent to abortions. The State contends that SB8 by its very nature furthers this important interest. Although SB8’s constitutionality does not depend on whether it has an informed-consent requirement, the law nevertheless promotes informed consent even without technically requiring that abortion providers use more detailed consent forms. In Gonzales, the Supreme Court upheld the Partial-Birth Abortion Ban Act despite the fact that the law did not include an informed-consent requirement because: It is self-evident that a mother who comes to regret her choice to abort must struggle with grief more anguished and sorrow more profound when she learns, only after the event, what she once did not know: that she allowed a doctor to pierce the skull 16 Case: 17-51060 Document: 00515984856 Page: 17 Date Filed: 08/18/2021 No. 17-51060 and vacuum the fast-developing brain of her unborn child, a child assuming the human form. Gonzales, 550 U.S. at 159–60 (emphasis added). “It is . . . precisely [a] lack of information concerning the way in which the fetus will be killed that is of legitimate concern to the State.” Id. at 159. “The State has an interest in ensuring so grave a choice is well informed.” Id. What was true in Gonzales is true here. Women who receive live- dismemberment D&Es are not being told what is going to happen to the fetus. In this case, the plaintiffs’ consent forms do not explain in “clear and precise terms” what a live-dismemberment abortion entails. Id. (quoting Nat’l Abortion Fed’n v. Ashcroft, 330 F. Supp. 2d 436, 466 n.22 (S.D.N.Y. 2004)). For example, Plaintiff Southwestern’s form tells the patient that “the pregnancy tissue will be removed during the procedure” and does not explain that the fetus’s body parts—arms, legs, ribs, skull, and everything else—will be ripped apart and pulled out piece by piece. Plaintiff Alamo’s consent form states that the doctor will “empt[y] the uterus either by vacuum aspiration or evacuation (manual removal of the fetus by forceps).” Plaintiff Whole Woman’s Health’s form states: “The physician will use . . . instruments such as forceps to remove the pregnancy from the uterus . . . in multiple fragments.” The district court cast aside all of these interests—even though each was recognized as legitimate and substantial in Gonzales and even though a “central premise of [Casey] was that the Court’s precedents after Roe had ‘undervalue[d] the State’s interest in potential life.’” Gonzales, 550 U.S. at 157 (quoting Casey, 505 U.S. at 873). 2. In addition to dismissing all of the State’s interests, the district court contravened the Supreme Court’s precedents in several other ways. First, the district court disregarded Roe by deeming the abortion right “absolute.” 17 Case: 17-51060 Document: 00515984856 Page: 18 Date Filed: 08/18/2021 No. 17-51060 “[W]e do not agree” that “the woman’s right is absolute.” Roe v. Wade, 410 U.S. 113, 153 (1973). Of course, no constitutional rights, even those expressly enshrined in the Bill of Rights, are absolute. Second, the district court’s faulty framework led it to place the burden of proof on the wrong party and turn the State’s legislative power on its head. It did so by holding that SB8 was unconstitutional because live dismemberment is a common abortion method in the second trimester. This was exactly backwards. Since Casey, we have recognized that abortion doctors do not get to set their own rules. They are not permitted to self- legislate or self-regulate simply by making an abortion method “common.” Abortion doctors do not have “unfettered choice[s].” Gonzales, 550 U.S. at 163. Indeed, not even the woman—the patient—gets “to terminate her pregnancy at whatever time, in whatever way, and for whatever reason she alone chooses.” Roe, 410 U.S. at 153. To the contrary, when the State enacts laws reasonably related to a legitimate interest, abortion doctors must find “different and less shocking methods to abort the fetus . . . thereby accommodating legislative demand.” Gonzales, 550 U.S. at 160. Third, the district court incorrectly defined “substantial obstacle.” Casey, 505 U.S. at 877. “Substantial” is defined as “of considerable importance, size, or worth.” Substantial, New Oxford Am. Dictionary 1736 (3d ed. 2010); see also Toyota Motor Mfg., Ky., Inc. v. Williams, 534 U.S. 184, 196 (2002) (“‘[S]ubstantially’ in the phrase ‘substantially limits’ suggests ‘considerable’ or ‘to a large degree.’”). The definition of “substantial” is consistent with the purpose of Casey’s substantial-obstacle test: to establish a relatively high bar for striking down laws—especially in facial challenges— that regulate abortions. See also Gonzales, 550 U.S. at 156, 167 (explaining that a facial-challenge plaintiff bears a “heavy burden” of proving that a law would impose a “substantial obstacle”). And yet the district court construed “substantial” to mean “no more and no less than ‘of substance.’” This 18 Case: 17-51060 Document: 00515984856 Page: 19 Date Filed: 08/18/2021 No. 17-51060 construction would yield essentially all abortion regulations unconstitutional and cannot be harmonized with the Supreme Court’s precedent. “[N]ot every law which makes a right more difficult to exercise is, ipso facto, an infringement of that right.” Casey, 505 U.S. at 873. In sum, the district court committed numerous legal errors and contravened Casey, Gonzales, and Hellerstedt by balancing SB8’s benefits against its burdens; diminishing the State’s compelling, numerous, and evidence-supported interests in preventing live-dismemberment abortions; granting the right to abortion an “absolute” status; placing the burden of proof on the wrong party; and erroneously defining “substantial” in “substantial obstacle.” These legal errors undermine the deference that we would normally owe the district court’s factual findings. See Thornburg, 478 U.S. at 79 (holding that appellate courts can correct errors, “including those that may infect a so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law” (quotation and citations omitted)). C. We now turn to the district court’s analysis of SB8’s burdens and its attendant factual findings. Because the district court’s myriad and fundamental legal errors evinced “a misunderstanding of the governing rule of law,” Bose Corp., 466 U.S. at 501, its factual “findings may be set aside on that basis,” Swint, 456 U.S. at 287. See also Aransas Project, 775 F.3d at 658 (“When, as here, a court’s factual finding ‘rest[s] on an erroneous view of the law’, its factual finding does not bind the appellate court.” (quoting Swint, 456 U.S. at 287)); Women’s Med. Ctr. of Nw. Hous. v. Bell, 248 F.3d 411, 419 (5th Cir. 2001) (“Although the ultimate decision whether to grant or deny a preliminary injunction is reviewed only for abuse of discretion, a decision grounded in erroneous legal principles is reviewed de novo.”). We therefore owe no deference to the district court’s factual findings. But, as 19 Case: 17-51060 Document: 00515984856 Page: 20 Date Filed: 08/18/2021 No. 17-51060 demonstrated below, even if we were to consider the district court’s factual findings under a clear-error standard, they fail to demonstrate an “undue burden” on the protected right. The district court disregarded and distorted the record to hold that SB8 would result in a complete ban on D&E abortions, in large part due to its erroneous definition of “substantial obstacle.” The district court first assumed, as a matter of law before even alluding to anything in the record, that requiring fetal death before live dismemberment by forceps would be “banning the standard D&E procedure.” The district court read Gonzales to describe the “standard D&E” as the “procedure performed before fetal demise.” This was error. In Gonzales, the Supreme Court described the typical D&E, and within that description noted that “[s]ome doctors, especially later in the second trimester, may kill the fetus a day or two before performing the surgical evacuation. They inject digoxin or potassium chloride into the fetus, the umbilical cord, or the amniotic fluid.” 550 U.S. at 136. The Court also pointed out that “[o]ther doctors refrain” from causing fetal death because they believe it provides no medical benefit. Id. After making these statements, the Court proceeded to describe partial-birth abortions—“a variation of this standard D & E.” Id. In other words, the Court’s description of the “standard D&E” included the option of fetal death before live dismemberment. The district court here misread Gonzales and thereby incorrectly concluded that there was only one kind of “standard D&E.” More broadly, the district court failed to sufficiently appreciate the direct applicability of Gonzales to the facts and many of the legal issues in this case. Gonzales’s facts are extremely similar to the situation presented here. In Gonzales, the Supreme Court upheld the federal Partial-Birth Abortion Ban Act and vacated two permanent injunctions of it. 550 U.S. at 133, 168. The Act proscribed “intact” dilation and extraction abortions, during which 20 Case: 17-51060 Document: 00515984856 Page: 21 Date Filed: 08/18/2021 No. 17-51060 the fetus is removed in one—as opposed to, as here, multiple—piece(s). 14 Id. at 136–37. The Court concluded that “the medical uncertainty over whether the Act’s prohibition create[d] significant health risks provide[d] a sufficient basis to conclude . . . that the Act d[id] not impose an undue burden.” Id. at 164. The Court noted that Congress was legitimately concerned “with the effects on the medical community and on its reputation caused by the practice of partial-birth abortion” and that the Act furthered the State’s legitimate and substantial interests in promoting ethics in the medical profession. Id. at 157. Moreover, the Act furthered the State’s interest in promoting “respect for life” by prohibiting procedures that are “laden with the power to devalue human life.” Id. at 158. Another “consideration[]” that supported the Court’s conclusion that the Act did not impose an undue burden was that “alternatives” to the prohibited procedure were available. Id. at 164. The district court was not at liberty to deviate from the teachings of Gonzales, and neither are we. Errors also pervaded the district court’s analysis of the alternatives to live dismemberment. The district court found that requiring fetal death before live dismemberment was an undue burden for “all women seeking a second-trimester abortion at 15 weeks” and beyond. Its bases for this sweeping conclusion were that the alternative methods would delay a woman’s abortion, which, according to the court, was sufficient by itself to create an undue burden, and that the alternative methods were unsafe and ineffective. 15 So, according to the district court, even if SB8 is not an explicit 14 Dismemberment abortions are “brutal.” Gonzales, 550 U.S. at 182 (Ginsburg, J., dissenting). 15 The district court apparently copied and pasted into its opinion facts from other district and circuit court opinions. That was inappropriate. The analysis is case-specific, litigation-specific, and fact-specific, and the district court erred by relying on other cases’ factual descriptions as bases for its ruling. For example, the district court borrowed facts from West Alabama Women’s Center v. Miller, 217 F. Supp. 3d 1313, 1339 (M.D. Ala. 2016). 21 Case: 17-51060 Document: 00515984856 Page: 22 Date Filed: 08/18/2021 No. 17-51060 ban, it would operate as a functional ban on second-trimester abortions. Contrary to the district court’s holding, the record shows that performing a D&E that complies with SB8, using either suction or digoxin, is safe, effective, and commonplace. 16 1. Suction Suction is a relatively simple technique. The woman is dilated enough to allow the placement of the “suction catheter” into the woman’s uterus. The suction then removes the amniotic fluid and fetus. Relevant to this case, the record describes three different ways suction can be used: (1) as a stand- alone method to cause fetal death and remove the entire fetus; (2) as a fetal- death technique to be followed by forceps for complete removal of the fetus; and (3) as a complement to forceps during live-dismemberment abortions to ensure that all amniotic fluid and pieces of the fetus have been removed. The But Miller involved a truncated preliminary-injunction record that included just one state- called witness. Id. Here, the district court held a five-day bench trial with dozens of witnesses and hundreds of exhibits. The district court should have relied on the voluminous and comprehensive record before it, not other courts’ opinions with materially different records. 16 We contrast this case with the Eleventh Circuit’s decision in West Alabama Women’s Center v. Williamson, 900 F.3d 1310 (11th Cir. 2018), cert. denied sub nom. Harris v. W. Ala. Women’s Ctr., 139 S. Ct. 2606 (2019). The most significant difference is that the Alabama district court found the fetal-demise law unconstitutional “as applied to the plaintiffs” whereas the plaintiffs here argue that SB8 is facially unconstitutional. W. Ala. Women’s Ctr. v. Miller, 299 F. Supp. 3d 1244, 1289 (M.D. Ala. 2017). Also, SB8 differs from the Alabama statute in meaningful ways, as do the cases’ records. As explained in footnote 14 supra, the record evidence in this case is markedly more developed and, moreover, flatly contradicts the Alabama record in critical respects. Even so, the smaller record in the Alabama case quantified the number of women impacted by the law. Id. at 1278. And the district court noted that not all doctors in Alabama are trained to perform D&Es, so finding any doctors willing to provide abortions in Alabama is difficult. The district court there found that requiring doctors to learn not only D&E but also the alternative fetal-death techniques would result in a substantial obstacle. Id. at 1284–85. As we explain above the line, the plaintiffs here did not even attempt to quantify the number of women who would face a substantial obstacle under SB8. 22 Case: 17-51060 Document: 00515984856 Page: 23 Date Filed: 08/18/2021 No. 17-51060 first two of these methods are allowed by SB8 because SB8 is violated only when a fetus is killed by dismemberment with forceps. See Tex. Health & Safety Code § 171.151. There was substantial trial testimony about suction. According to the record, some doctors use suction as a stand-alone method up to 17 weeks, while others begin using alternative methods, like digoxin or live dismemberment, at an earlier point. 17 Plaintiffs’ expert, Dr. Amna Dermish, the Regional Medical Director for Plaintiff Planned Parenthood of Greater Texas, testified that she could “guarantee” compliance with SB8 in the “vast majority of cases” through 16 weeks, 6 days using suction alone to cause fetal death and complete the procedure. The State’s expert, Dr. Chireau, reviewed over 100 studies to offer her opinion that suction alone is sufficient to complete abortions through 16 weeks, 6 days. Another plaintiffs’ expert, Dr. Mark Nichols, a Medical Director for Planned Parenthood, testified that he has used suction to cause fetal death and complete an abortion through 15 weeks, 6 days. Plaintiff Dr. Robin Wallace, a Family Physician for Plaintiff Southwestern Women’s Surgery Center, testified that some doctors rely on suction through 16 weeks, 6 days. Dr. Edward Aquino, who provides abortions at Plaintiff Alamo’s San Antonio location, testified that the increased size of suction cannulas in recent years has allowed doctors to more commonly use suction as a stand-alone method. 17 Judge Dennis’s dissent quibbles at some length on a perceived distinction between “alternative” methods and “additional” procedures. In Gonzales, the Supreme Court upheld the Partial-Birth Abortion Ban Act and described the same procedures proposed by Texas in this case as “alternatives”—despite the fact that more steps had to be taken to complete an abortion under the Act. 550 U.S. at 136, 164, 166–67. Even if Texas’s proposed alternatives to live dismemberment could be construed as “additional” procedures, that would not render SB8 unconstitutional. See Casey, 505 U.S. at 873 (“[N]ot every law which makes a right more difficult to exercise is, ipso facto, an infringement of that right.”). 23 Case: 17-51060 Document: 00515984856 Page: 24 Date Filed: 08/18/2021 No. 17-51060 The district court’s only reference to suction was in a footnote, which stated that “before 15 weeks,” doctors do not usually use forceps because “the fetus and all other in utero materials will pass through a dilated cervix using only suction.” That suction is commonly used “before 15 weeks” says nothing about whether it can also be used safely and effectively after 15 weeks. Indeed, according to the plaintiffs and their own witnesses, this safe and common abortion procedure can be used to comply with SB8 up to almost 17 weeks. In 2015, 1,520 of the 3,150 abortions (48%) performed in Texas during weeks 15–22 occurred in weeks 15 and 16. The testimony of the plaintiffs themselves, their experts, and their doctors that suction can be used to comply with SB8 in many abortions during weeks 15 and 16 casts serious doubt on the plaintiffs’ efforts to carry their heavy burden of proving an undue burden on a large fraction of women. As we show below, adding the second alternative to live dismemberment—digoxin—removes any doubt that plaintiffs have failed to carry their burden. 2. Digoxin The district court found that using digoxin to cause fetal death is unsafe, ineffective, and would delay a woman’s abortion procedure. The district court found that digoxin use before 18 weeks would be experimental. Many of the district court’s digoxin findings are contradicted by the plaintiffs’ own evidence and practices; others are simply unsupported by the record. a. Safety and Risk Digoxin has long been recognized as a common method of causing fetal death during an abortion. Two decades ago, in Stenberg, the Supreme Court noted that “[s]ome physicians use . . . digoxin to induce fetal demise prior to a late D & E (after 20 weeks), to facilitate evacuation.” Stenberg v. Carhart, 530 U.S. 914, 925 (2000) (quoting Carhart v. Stenberg, 11 F. Supp. 24 Case: 17-51060 Document: 00515984856 Page: 25 Date Filed: 08/18/2021 No. 17-51060 2d 1099, 1104 (D. Neb. 1998)). 18 Seven years later, in Gonzales, the Supreme Court again acknowledged that “[s]ome doctors, especially later in the second trimester, may kill the fetus [using digoxin] a day or two before 18 As discussed below, three of the plaintiffs in this case require the use of digoxin to achieve fetal demise prior to D&Es performed after 18 or 20 weeks. SB8 follows the lead of these plaintiffs and requires fetal demise prior to all D&E procedures, whether by digoxin or suction. In no way then is SB8 a ban on D&Es; rather it is a regulation of the method of performing a D&E. Judge Dennis’s dissent nevertheless mischaracterizes SB8 as a ban and contends that “the Supreme Court has already decided this exact case, holding that a Nebraska law was unconstitutional because it could be interpreted to be the sort of ban that the Texas statute openly embodies.” Post at 67 (citing Stenberg, 530 U.S. at 945). Not so. The Supreme Court struck down the partial-birth-abortion ban in Stenberg primarily because it lacked a necessary health exception for the mother. See Stenberg, 530 U.S. at 930–31. SB8 has a health exception. The Stenberg Court then noted that, although the law targeted “D&X”—dilation and extraction—abortions in which the fetus is pulled into a breech position in the vaginal cavity before dismemberment, the law could also be read to cover the more common D&E method. Id. at 926–27. And the Court noted as well that, at least in 2000, “[t]he D & E procedure carries certain risks. The use of instruments within the uterus creates a danger of accidental perforation and damage to neighboring organs. Sharp fetal bone fragments create similar dangers. And fetal tissue accidentally left behind can cause infection and various other complications.” Id. at 926. Fast forward twenty-one years and some, including Judge Dennis’s dissent, consider D&Es “very safe.” Post at 69. We see no principled reason to decline to analyze Texas’s SB8 on its own terms, cognizant of the current medical realities. Indeed, we glean from Supreme Court precedent a duty to test the statute before us given the facts before us—that is what the Supreme Court did in Gonzales when it considered (and upheld) the federal Partial-Birth Abortion Ban Act in its own right rather than simply invalidating it on loose analogy to the Nebraska statute at issue in Stenberg. See Gonzales, 550 U.S. at 140–41, 161–62, 168. Under Judge Dennis’s dissent’s approach, any regulation affecting abortion procedures in any way can be deemed unconstitutional simply because another regulation has been so deemed— despite any differences between the regulations or the facts. But cf. Hellerstedt, 136 S. Ct. at 2306 (“A statute valid as to one set of facts may be invalid as to another.” (quoting Nashville, C. & St. L. Ry. Co. v. Walters, 294 U.S. 405, 415 (1935))). That itself would fly in the face of Roe and its progeny, which recognize that states can impose regulations affecting abortion in some circumstances. Roe, 410 U.S. at 164–65; Casey, 505 U.S. at 846. Moreover, it would disavow any possibility of progress in medicine and science and instead shackle the states’ regulatory power to abortion standards from the last century. 25 Case: 17-51060 Document: 00515984856 Page: 26 Date Filed: 08/18/2021 No. 17-51060 performing” the D&E. 550 U.S. at 136. In fact, the Court in Gonzales found that “an injection that kills the fetus” allows a doctor to perform the D&E without violating the Partial-Birth Abortion Ban Act. Id. at 164. The use of digoxin to cause fetal death before a D&E is hardly a novel phenomenon. The plaintiffs here know this because they have used and continue to use digoxin. In 2007, one month after the Supreme Court described digoxin as a “safe alternative” fetal-death method in Gonzales, the nation’s largest abortion provider, Planned Parenthood Federation of America, mandated that all of its affiliates use digoxin to cause fetal death before most surgical abortions at or above 18 weeks. Plaintiff Alamo is so sure that digoxin is safe that it requires digoxin’s use to cause fetal death in abortions after 18 weeks. Plaintiff Southwestern requires digoxin beginning at 20 weeks. Plaintiff Planned Parenthood of Greater Texas required the use of digoxin starting at 18 weeks. During the district court’s five-day bench trial, every doctor who testified had used digoxin to cause fetal death except one, and he works with other doctors who have used it. Plaintiff Planned Parenthood of Greater Texas’s consent form lists some of the “risks and side effects” of digoxin (like extramural delivery and pain), then tells patients that “there are no published reports of serious problems from using digoxin before abortion.” The form also assures patients that “[s]ome clinicians also believe that using digoxin makes it easier to do the abortion. Studies have shown that it is safe to use digoxin for this purpose.” Despite arguing in this case that digoxin provides no health benefits to the woman, Alamo’s consent form also assures patients that “the injection [of digoxin] . . . help[s] the woman’s body prepare for the abortion process” and that “the abortion process is made easier and safer by injecting the fetus” with digoxin. Even with all of this evidence in the plaintiffs’ own 26 Case: 17-51060 Document: 00515984856 Page: 27 Date Filed: 08/18/2021 No. 17-51060 documents, the district court found that digoxin presents “significant health risks.” 19 This holding contradicted the State’s and the plaintiffs’ evidence. The district court also found that the “pain and invasiveness” of digoxin was one reason that its use was “a substantial obstacle” to a woman’s abortion right. This finding ignored the record evidence that patients undergoing D&E are given the option of sedation even when digoxin is not administered. And Plaintiff Dr. Wallace admitted that, when she performs an abortion involving digoxin, she injects a local numbing anesthetic before injecting the digoxin. Plaintiff Planned Parenthood of Greater Texas tells its patients that any pain from the digoxin injection will “go away quickly.” b. Efficacy When digoxin is used, its success rate is between 90 and 100%. Plaintiffs Southwestern and Alamo describe digoxin failures as “unusual.” Plaintiff Dr. Wallace testified that digoxin is 98% successful. Plaintiffs’ expert Dr. Dermish testified that digoxin is 95% successful. Another expert testified that several studies have shown either 0% failure rates or 99% effectiveness rates. Plaintiff Southwestern’s “Consent for Digoxin Injection” form states in unequivocal terms that digoxin failing to cause fetal death “is uncommon and may or may not delay the expected completion time of your abortion procedure.” Dr. Chireau testified extensively about myriad studies that found digoxin safe and effective. One study found digoxin “safe and effective” with a 100% success rate for intra-fetal injections in a study 19 The plaintiffs also argue that digoxin presents significant risks and is contraindicated for women with certain heart conditions. And the plaintiffs state that for obese women or women with fibroids, administering digoxin is “difficult or impossible.” Plaintiffs ignore that their own documents state that obese women and women with fibroids are considered to have “special conditions requiring special evaluation and management” for the D&E itself. In other words, according to the plaintiffs’ arguments, if digoxin is unsafe, then D&E itself is unsafe. Indeed, it is unclear whether certain women with these conditions are able to receive a D&E abortion at all. 27 Case: 17-51060 Document: 00515984856 Page: 28 Date Filed: 08/18/2021 No. 17-51060 with 107 abortions performed in weeks 17–24. Another study showed a 98% success rate with digoxin for 1,600 abortions performed in weeks 18–22. What is more, the plaintiffs admit that if digoxin does not cause fetal death after one attempt, it can be injected again. 20 Plaintiff Alamo’s digoxin consent form tells patients: “If fetal death has not been induced [on the first attempt], a second injection of Digoxin can be administered at the physician’s discretion.” Plaintiff Southwestern tells its patients that digoxin failure is “unusual” and that a “second injection may be administered” if the first fails. In ruling for the plaintiffs on digoxin’s efficacy, the district court ignored the plaintiffs’ own extensive documentation that digoxin is highly effective. c. Delay The district court’s holding as to the delay digoxin would cause was both factually and legally incorrect. The record does not support the district court’s factual finding that digoxin’s use would cause a delay for all women seeking what “otherwise is a one-day standard D&E procedure.” The district court found that a woman “undergoing a digoxin injection would be required to make an additional trip to the clinic 24 hours before her appointment for the standard D&E procedure.” The district court assumed that for women receiving a digoxin injection, that injection would happen a day after the State’s mandatory 24-hour waiting period and a day before the one-day D&E. This finding is refuted in several ways by the plaintiffs’ documents. First, many D&E abortions are not one-day procedures. The plaintiffs admit that starting at 18 weeks, doctors use laminaria to achieve the necessary 20 Judge Dennis’s contention that a second digoxin injection is “wholly experimental” and “too dangerous to administer” stands at odds with what the plaintiffs have been telling their patients for years. Post at 98–99. 28 Case: 17-51060 Document: 00515984856 Page: 29 Date Filed: 08/18/2021 No. 17-51060 dilation. “Because laminaria expand gradually, patients usually have them inserted and return the next day to complete the procedure.” So, according to the plaintiffs, the “standard D&E” is a two-day procedure starting at 18 weeks. Their documents also say that the D&E can become a two-day procedure as early as 16 weeks, 6 days. At trial, one of the plaintiffs’ doctors also noted that later in the second trimester, the dilation process can take up to two days such that the D&E procedure is not completed until the third day. Second, digoxin works within several hours, and it can be administered at the same time or close to the beginning of the dilation process. Thus, it is not true that using digoxin would add another day to every woman’s one-day D&E abortion. Even so, the district court also legally erred by concluding that a one- day delay is sufficient, by itself, to create an undue burden. The Supreme Court has approved regulations embodied in 24-hour waiting periods for all women and parental-consent and judicial-bypass laws covering minors that, by their nature, may entail many days (and even weeks) before an abortion is finally approved. See Casey, 505 U.S. at 885–86 (holding that Pennsylvania’s 24-hour waiting period, even if it caused “a delay of much more than a day,” was not an undue burden); June Medical, 140 S. Ct. at 2136–37 (Roberts, C.J., concurring) (explaining that Casey held that Pennsylvania’s 24-hour waiting period and parental-consent and doctor-notification requirements did not create substantial obstacles even though they risked delays, increased costs, and “had little if any benefit”); see also Tex. Fam. Code Ann. §§ 33.003, 33.004 (requiring trial and appellate courts to rule on a minor’s application for judicial bypass within five business days of the initial request or notice of appeal). If these procedures are not constitutionally infirm because of the delays involved, then—even assuming that the district court was right on the facts of delay—adding a one-day delay to assure a less brutal pregnancy 29 Case: 17-51060 Document: 00515984856 Page: 30 Date Filed: 08/18/2021 No. 17-51060 termination and vindicate the State’s interest in human dignity is not an undue burden. d. Before 18 Weeks As for the beginning of week 15 up through 17 weeks, 6 days, the district court held that administering a digoxin injection during that period would be “arguably experimental” and weighed that against the State. Under Gonzales, this was yet another legal error made by the district court. In Gonzales, the Supreme Court confronted a record with conflicting testimony about the safety of intact D&E abortions and the alternatives. See 550 U.S. at 161–62. The question became, then, whether the Partial-Birth Abortion Ban Act was constitutional in light of that medical uncertainty. Id. at 163. The answer was a resounding yes: “The Court’s precedents instruct that the Act can survive this facial attack. . . . [S]tate and federal legislatures [have] wide discretion to pass legislation in areas where there is medical and scientific uncertainty.” Id. (emphasis added). Indeed, “medical uncertainty” about whether the Act’s requirements “create[d] significant health risks provide[d] a sufficient basis to conclude” that the law there did not impose an undue burden. Id. at 164 (emphasis added). Gonzales was not the first time that the Court emphasized legislatures’ right to regulate in areas “fraught with medical and scientific uncertainties.” Kansas v. Hendricks, 521 U.S. 346, 360 n.3 (1997) (quoting Johnson v. United States, 463 U.S. 354, 370 (1983)). “Legislative options must be especially broad” in this context and “courts should be cautious not to rewrite legislation.” Id. (emphasis added) (quoting Johnson, 463 U.S. at 370); see also Marshall v. United States, 414 U.S. 417, 427 (1974) (“[L]egislative options must be especially broad” in areas “fraught with medical and scientific uncertainties.”). In his June Medical concurrence, the Chief Justice reaffirmed courts’ obligation to give legislatures broad deference in the context of scientific or 30 Case: 17-51060 Document: 00515984856 Page: 31 Date Filed: 08/18/2021 No. 17-51060 medical uncertainty—as taught by both Casey and Gonzales. June Medical, 140 S. Ct. at 2136 (Roberts, C.J., concurring) (“[W]e have explained that the ‘traditional rule’ that ‘state and federal legislatures [have] wide discretion to pass legislation in areas where there is medical and scientific uncertainty’ is ‘consistent with Casey.’” (quoting Gonzales, 550 U.S. at 163)). Judges are simply ill-suited to make such decisions. “Attempting to do so would be like ‘judging whether a particular line is longer than a particular rock is heavy.’” Id. (quoting Bendix Autolite Corp. v. Midwesco Enters., Inc., 486 U.S. 888, 897 (1988) (Scalia, J., concurring in the judgment)). Staying in our judicial lane accords with our broader duty to recognize and respect the institutional competency of legislatures. 21 Medical uncertainty does not foreclose the exercise of legislative power in the abortion context any more than it does in other contexts. Cf. Hendricks, 521 U.S. at 360 n.3. The Court specifically addressed this in Gonzales: A zero tolerance policy would strike down legitimate abortion regulations, like the present one, if some part of the medical community were disinclined to follow the proscription. This is too exacting a standard to impose on the legislative power, exercised in this instance under the Commerce Clause, to regulate the medical profession. Considerations of marginal safety, including the balance of risks, are within the legislative competence when the regulation is rational and in pursuit of legitimate ends. When standard medical options are available, mere convenience does not suffice to displace them; and if some procedures have different risks than others, it does not follow that the State is altogether barred from imposing reasonable regulations. The Act is not invalid on its face where there is uncertainty over whether the barred procedure is ever 21 Judge Dennis’s dissent chides us for deferring to the legislature. Post at 99–100. But that is precisely what the Supreme Court has directed us to do in situations like this. 31 Case: 17-51060 Document: 00515984856 Page: 32 Date Filed: 08/18/2021 No. 17-51060 necessary to preserve a woman’s health, given the availability of other abortion procedures that are considered to be safe alternatives. Gonzales, 550 U.S. at 166–67. This case is even easier than Gonzales because the plaintiffs here did not contradict the State’s evidence about digoxin’s use before 18 weeks, much less show that its use presents “significant health risks.” Id. at 164. Dr. Chireau testified about a study that noted Planned Parenthood of Los Angeles’s mandatory use of digoxin for all second-trimester abortions (weeks 13–26). 22 Dr. David Berry, a maternal-fetal medicine specialist in Austin, testified that he knows of doctors who have administered digoxin before 18 weeks. The plaintiffs do not refute this evidence; they just blame the State (the party without the burden of proof) for not producing more evidence. In sum, in making its findings about digoxin, the district court failed to apprehend that the plaintiffs’ own extensive use of digoxin, notices and consent forms, and written minimization of risks not only conflict with their testimony in this case, but also certainly raise serious questions about the debatability of the actual risk of using digoxin to cause fetal death. The plaintiffs have long used digoxin to ensure that they do not violate the Partial- 22 In their en banc brief and at oral argument, the plaintiffs argued that the sentence in the study referencing the policy was a “typo.” The disputed sentence says: “Although PPLA . . . protocols dictate use of digoxin for all second-trimester abortion. . . .” The study was published in 2009 and concluded that intra-fetal or intra-amniotic “injection of digoxin is safe and effective for inducing fetal death prior to second-trimester surgical abortion.” The abstract is available here: https://www.contraceptionjournal.org/article/S0010-7824 (09)00409-0/fulltext. Ten years later, while this case was pending before this court, in 2019, the author of the study, Dr. Deborah Nucatola, sent a letter to the editor and explained that the original statement was “not correct” because, she says, Planned Parenthood of Los Angeles’s description of its digoxin policy in other years shows that the original statement could not have been true: https://www.contraception journal.org/article/S0010-7824(19)30386-5/fulltext. 32 Case: 17-51060 Document: 00515984856 Page: 33 Date Filed: 08/18/2021 No. 17-51060 Birth Abortion Ban Act. Surely, no reasonable abortion provider would subject women to “significant” health risks from digoxin just to avoid their own federal liability. See Gonzales, 550 U.S. at 164. Because there are safe, medically recognized alternatives to live- dismemberment-by-forceps D&E (suction and digoxin), and because women seeking a D&E abortion are not significantly affected by a non-forceps fetal- death requirement, the district court’s undue-burden analysis is incorrect as a matter of law. SB8 falls comfortably within the orbit of Casey/Gonzales as a regulation that respects the important state and societal interests involved in proscribing a brutal procedure, yet does not pose a substantial obstacle to women seeking abortions in the relevant circumstances. D. The district court’s final flaw was its large-fraction analysis. In this facial challenge, it is the plaintiffs who bear the “heavy burden” of showing that SB8 would be unconstitutional in a “large fraction of relevant cases.” Gonzales, 550 U.S. at 167–68. The numerator is the number of women for whom the law is an undue burden. And the denominator is the number of women in the relevant circumstances—i.e., the women for whom the law “is an actual rather than an irrelevant restriction.” Hellerstedt, 136 S. Ct. at 2320. Because the district court concluded that SB8 was a complete ban on 1 the standard D&E, it found that the fraction of burdened women was . The 1 district court botched both numbers in this fraction. First, the district court erred by finding that the denominator included only women with fetuses at the gestational age of 15–20 weeks. 23 In fact, the 23 The second trimester spans from 13 to 26 weeks gestation. Texas law bans abortions after 22 weeks unless the abortion is necessary to protect the woman’s health or the fetus has a severe abnormality. Tex. Health & Safety Code §§ 171.044, .046. After 15 weeks, the D&E procedure is a common abortion method. So SB8 affects only abortions between 15 and 22 weeks, which make up about 5% of total abortions in Texas. See Induced 33 Case: 17-51060 Document: 00515984856 Page: 34 Date Filed: 08/18/2021 No. 17-51060 correct denominator, as all parties to the case acknowledge, is all women with fetuses in the gestational age of 15–22 weeks. The relevant denominator is therefore larger by two weeks gestation than the district court stated. The extra duration is important because, as discussed above, the plaintiffs already use (and even require) digoxin after 18 weeks. 24 Second, the numerator is not equal to the denominator, which is what the district court implicitly found by holding that SB8 constituted a “ban.” There are safe and widely used alternatives to live-dismemberment D&E for the entire second trimester. Regarding suction, the record shows that doctors can sometimes use this method to complete abortions up through 16 weeks, 6 days. As for digoxin, and as explained by Chief Judge Owen, “[t]here is no basis in the record for concluding that the use of digoxin, standing alone, constitutes a substantial obstacle . . . at or after 15 weeks gestation.” Post at 40. Indeed, digoxin is used ubiquitously, including by the plaintiffs themselves, beginning at the first day of the 18th week. The plaintiffs bear the heavy burden here. If there are actual cases in which neither suction nor digoxin is medically indicated and only live- dismemberment D&E by forceps is medically approved, the plaintiffs did not describe them. The plaintiffs made no effort to quantify the number of Termination of Pregnancy Statistics, Tex. Health & Human Servs., https://hhs.texas.gov/ about-hhs/records-statistics/data-statistics/itop-statistics (last visited June 21, 2021). Texas presented evidence that 92% of countries ban almost all abortions after 12 weeks gestation. Only three countries’ abortion laws are roughly as permissive as Texas (Singapore, the Netherlands, and the United Kingdom), whereas only six countries are more permissive than Texas (China, North Korea, Vietnam, Canada, Cuba, and Bahrain). 24 Judge Dennis’s dissent contends that “the appropriate denominator is the class of women actually affected by SB8, which is composed of only those women who would undergo a forceps-assisted D&E in Texas without their doctors’ first inducing fetal demise in the absence of SB8.” Post at 103. This is not how the district court characterized the denominator, nor is it what the parties agree is the correct denominator: women seeking abortions in the gestational age of 15–22 weeks. 34 Case: 17-51060 Document: 00515984856 Page: 35 Date Filed: 08/18/2021 No. 17-51060 women who might be subjected to an additional burden if a digoxin injection is used and the injection prolongs the entire procedure by an additional day. The sum total of the plaintiffs’ efforts in this area is one expert’s testimony about the negative effects an “additional trip to an abortion clinic” would have on low-income women in Texas. Even if this were true, some (unspecified number of) women does not constitute a large fraction. Plus, under Casey, that would not constitute an undue burden, without more, anyway. See 505 U.S. at 886 (rejecting the argument that a waiting period imposed on women with the “fewest financial resources” would constitute an undue burden). Similarly, the plaintiffs made no effort to quantify the “unreliability” of digoxin beyond stressing a 90–100% success rate for a single injection and conceding heightened effectiveness with a second injection. The district court acknowledged that digoxin’s failure rate is only “between 5% and 10%.” This high efficacy rate made the plaintiffs’ “burden” even heavier to show that digoxin’s high success rate is not enough. Some or all of this data should have enabled the district court to determine whether in fact a “large fraction” of the women seeking second-trimester abortions in Texas would suffer a substantial obstacle through the operation of SB8. Instead, the district court accepted plaintiffs’ all-or-nothing “ban” argument. The district court did not just err by accepting the plaintiffs’ false dichotomy; it also turned facial validity on its head and placed the burden of proof on the wrong party. The district court concluded that prohibiting only one method of D&E (live dismemberment by forceps) is unconstitutional all of the time because the other methods that achieve fetal death (like digoxin) do not work some of the time. This distorts the State’s burden. The State need not prove that every alternative works every time for every woman. As Gonzales instructs, a prohibition of a particular method is “permissible” when “a woman [can] still obtain an abortion through an acceptable 35 Case: 17-51060 Document: 00515984856 Page: 36 Date Filed: 08/18/2021 No. 17-51060 alternative method.” Preterm-Cleveland, 994 F.3d at 534. The plaintiffs’ own practices show that such alternatives are available and widely used for the vast majority of abortions in most of the relevant weeks of gestation. Thus, the plaintiffs cannot show that SB8 poses a substantial obstacle in a large fraction of cases. See Gonzales, 550 U.S. at 164; Stenberg, 530 U.S. at 931–36. Finally, because the plaintiffs rested only on their argument that SB8 is a ban on all D&E abortions, they did not develop any evidence related to SB8’s specific impact on abortion access. During oral argument, the plaintiffs’ attorney said that there was record evidence that “at least three providers would stop providing abortions if SB8 took effect.” En Banc Oral Argument at 39:31–39:48. Actually, the record shows that one doctor testified that she would alter her practice only to stop providing abortions after 17 weeks. This same doctor testified that another doctor at her clinic told her that he would also stop providing abortions after 17 weeks; this testimony was struck as hearsay. One other abortion doctor, who has been practicing for over 40 years, said that he would retire. Contrast this to Hellerstedt where these same plaintiffs argued to the Supreme Court that 50% of Texas’s abortion clinics (20 out of 40 clinics) would close if the challenged law had taken effect. 136 S. Ct. at 2301. That argument was crucial to the Supreme Court’s determination that the admitting-privileges law was facially invalid. See id. at 2312. No reading of this record supports anything remotely similar here. Indeed, at en banc oral argument, the plaintiffs conceded that they were not arguing that clinics would close because of SB8. IV. SB8 was signed into law four years ago—four years in which federal courts have halted Texas’s duly enacted and modest legislation from taking effect. The parties produced mountains of evidence and presented that 36 Case: 17-51060 Document: 00515984856 Page: 37 Date Filed: 08/18/2021 No. 17-51060 evidence to the district court during a week-long trial. The district court abused its discretion by applying the wrong legal test to assess SB8, dismissing and ignoring the State’s important and substantial interests, placing the burden of proof on the wrong party, explicitly and erroneously stating that the abortion right is “absolute” and evaluating SB8 under that view, erroneously defining “substantial obstacle,” incorrectly determining that SB8 constitutes a “ban” on D&E abortions, ignoring vast swaths of testimony about suction, making findings about digoxin that contradict the plaintiffs’ own digoxin use and practices, weighing medical uncertainty against the State, and incorrectly determining both the numerator and denominator in the large-fraction analysis. As it was in Gonzales, remanding to the district court would be futile here because the voluminous record permits only one conclusion. 25 The safety, efficacy, and availability of suction to achieve fetal death during abortions in weeks 15 and 16 combined with the safety, efficacy, and availability of digoxin to do the same in weeks 18–22 mean that the plaintiffs have utterly failed to carry their heavy burden of showing that SB8 imposes an undue burden on a large fraction of women in the relevant circumstances. * * * The district court’s permanent injunction is VACATED. We REVERSE the judgment of the district court and RENDER judgment in the State’s favor. SB8 is constitutional. 25 Our effort to apply Supreme Court precedent to SB8 very well may be called “Sisyphean,” but that does not dissuade us from the task. See post at 67. Nor should it, as we intermediate court judges must always roll the stones of Supreme Court precedent up the hills before us. As it is with Sisyphus, so it is with us: “The struggle itself . . . is enough to fill a man’s heart. One must imagine Sisyphus happy.” Albert Camus, The Myth of Sisyphus 123 (Justin O’Brien, trans. 1955). 37 Case: 17-51060 Document: 00515984856 Page: 38 Date Filed: 08/18/2021 No. 17-51060 Priscilla R. Owen, Chief Judge, concurring in the judgment: I concur in reversing the district court’s judgment, which held Texas’s prohibition of dismemberment abortions 1 facially unconstitutional and permanently enjoined its enforcement. 2 Reversal is required because prohibiting dismemberment of a living fetus with the purpose of causing the death of an unborn child by a means described in Texas Health and Safety Code § 171.151 3 does not “operate as a substantial obstacle to a woman’s choice to undergo an abortion” “in a large fraction of the cases in which [it] is relevant.” 4 All agree that the relevant focus is on abortions occurring from 15 to 22 weeks of gestation. In order to avoid the risk of violating Texas law and incurring criminal penalties, abortion providers can cause fetal demise before proceeding to use forceps or the other devices described in § 171.151 to perform an abortion. The record developed in this case clearly reflects that fetal demise prior to 1 See Tex. Health and Safety Code Ann. §§ 171.151-154 (West 2017). 2 ROA.1615-17. 3 See Tex. Health & Safety Code Ann. § 171.151 (West 2017): In this subchapter, “dismemberment abortion” means an abortion in which a person, with the purpose of causing the death of an unborn child, dismembers the living unborn child and extracts the unborn child one piece at a time from the uterus through the use of clamps, grasping forceps, tongs, scissors, or a similar instrument that, through the convergence of two rigid levers, slices, crushes, or grasps, or performs any combination of those actions on, a piece of the unborn child’s body to cut or rip the piece from the body. The term does not include an abortion that uses suction to dismember the body of an unborn child by sucking pieces of the unborn child into a collection container. The term includes a dismemberment abortion that is used to cause the death of an unborn child and in which suction is subsequently used to extract pieces of the unborn child after the unborn child’s death. 4 Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 895 (1992). 38 Case: 17-51060 Document: 00515984856 Page: 39 Date Filed: 08/18/2021 No. 17-51060 extraction with forceps or similar devices can be accomplished with little or no risk to the mother’s health by using digoxin. The remaining question is whether prohibiting dismemberment abortion creates a delay that amounts to a substantial obstacle to obtaining an abortion. The record reflects that ensuring fetal demise after 17 weeks and six days of gestation can be accomplished without any delay at all in the abortion process. In abortions performed at 15 weeks to 17 weeks and six days, there may be an additional delay for some women of approximately 24 hours beyond Texas’s statutory 24-hour waiting period. This additional 24- hour delay does not constitute a substantial obstacle and does not render the Texas statutes at issue unconstitutional. 5 Even were an additional 24-hour delay a substantial obstacle, there would not be such a delay for a large fraction of women seeking an abortion at 15 to 22 weeks of gestation. I As a preliminary matter, it is unnecessary to decide whether Chief Justice Roberts’s concurring opinion in June Medical Services L.L.C. v. Russo 6 governs and therefore supersedes the balancing test set forth in Whole Woman’s Health v. Hellerstedt,7 as Judge Elrod and Judge Willett’s 5 See ante at 29-30 (first citing Casey, 505 U.S. at 885-86, and then citing June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103, 2136-37 (Roberts, C.J., concurring)). 6 See 140 S. Ct. at 2135-39 (Roberts, C.J., concurring) (rejecting a balancing test and concluding that, so long as the state has a “legitimate purpose” and the statute is “reasonably related to that goal,” “the only question for a court is whether a law has the ‘effect of placing a substantial obstacle in the path of a woman seeking an abortion of a nonviable fetus’” (quotation at 2138) (quoting Casey, 505 U.S. at 877, 878, 882)). 7 See 136 S. Ct. 2292, 2309 (2016) (“The rule announced in Casey . . . requires that courts consider the burdens a law imposes on abortion access together with the benefits those laws confer.”). 39 Case: 17-51060 Document: 00515984856 Page: 40 Date Filed: 08/18/2021 No. 17-51060 plurality opinion posits that it does. 8 Under either view of the governing parameters, the Texas laws are constitutional because they do not place a substantial obstacle in the path of a woman seeking to abort a nonviable fetus. II As discussed in Judge Elrod and Judge Willett’s opinion, the record reflects that digoxin is a means of causing fetal demise that has been widely and successfully used by many of the abortion providers who are parties to this litigation. The use of digoxin rarely causes injury to or complications for the mother. There is no basis in the record for concluding that the use of digoxin, standing alone, constitutes a substantial obstacle to obtaining an abortion at or after 15 weeks of gestation. As the plurality opinion explains, the record is clear regarding digoxin’s safety and efficacy. 9 The remaining potential obstacle digoxin poses to women seeking an abortion is delay. The district court concluded that administering digoxin would create an additional 24-hour delay for “all women” seeking an abortion past 15 weeks. 10 This conclusion was clearly erroneous. Based on the record evidence, administering digoxin would create approximately an additional 24-hour delay (beyond Texas’s 24-hour waiting period) for some, but not all, or even most, abortions occurring from 15 weeks to 17 weeks and six days. It would create no additional delay for abortions performed after 17 weeks and six days. 8 See ante at 9-13 (discussing the two standards of review and concluding that, under Marks v. United States, 430 U.S. 188, 193 (1977), Chief Justice Roberts’s formulation controls). 9 See ante at 24-28. 10 ROA.1610. 40 Case: 17-51060 Document: 00515984856 Page: 41 Date Filed: 08/18/2021 No. 17-51060 In Texas, a physician cannot begin an abortion, including the dilation process, until the 24-hour waiting period has concluded. 11 Patients must first attend an initial office visit, during which the physician performs an ultrasound and provides state-mandated information. 12 The patients may then return 24 hours later for the actual abortion procedure, beginning with dilation. 13 The duration of dilation varies depending on the method used, which in turn depends on the fetus’s “gestational age.” 14 For pregnancies from 15 weeks to 17 weeks and six days of gestation, dilation is often achieved with medication. 15 The physician administers the medication, which is effective within several hours. 16 Once the medication takes effect, the physician may begin the evacuation portion of the abortion procedure. 17 Importantly, however, dilation and evacuation are not always performed on the same day in abortions occurring between 15 weeks and 17 weeks and six days. A potential delay arises for logistical reasons: a physician’s ability to perform the evacuation on the same day as the dilation depends on the timing of the patient’s initial appointment, which sets the 24- hour waiting period. 18 As one of the physicians who is also a plaintiff in this case explained, if the patient’s initial visit occurs early enough in the 11 See ROA.2012-15, 2111-12. 12 See ROA.2012-13, 2111. 13 See ROA.2014-15, 2111-12. 14 ROA.2111. 15 See ROA.1918, 2014-15, 2111-12. 16 See ROA.1923, 1924, 2014-15. 17 See ROA.1924, 2111-12. 18 See ROA.2111-13. 41 Case: 17-51060 Document: 00515984856 Page: 42 Date Filed: 08/18/2021 No. 17-51060 morning, the patient is able to return and receive dilation medication the following morning, and several hours later, the physician may perform the evacuation. 19 But if the initial appointment—and the patient’s subsequent return—occur “beyond a certain time of day, usually 10:00 or 11:00 in the morning,” the physician may not be able to “assure enough time for adequate dilation with the remainder of the clinic day.” 20 In such cases, the physician “place[s] . . . dilators and allow[s] them to work overnight and ask[s] the patient to return on a third day to have her D&E procedure completed.” 21 This plaintiff estimated that “[a]bout half” of patients between 15 weeks’ and 17 weeks and six days’ gestation were able to undergo dilation and evacuation on the same day, while the other half had to undergo dilation and evacuation over a two-day period. 22 For pregnancies at and beyond 18 weeks, dilation is achieved using osmotic dilators, or laminaria. 23 The physician places the laminaria inside the patient’s cervix on one day, and the patient generally returns the following day for the evacuation. 24 In some cases, however, a second set of laminaria is required, such that the first set of laminaria is placed one day, the patient returns the following day for removal of the first set and placement of a second set, then the patient returns once more on the third day for removal 19 See ROA.2111-12. 20 ROA.2112, 2113. 21 ROA.2113. 22 ROA.2113. 23 See ROA.1918, 1923. 24 See ROA.1923, 2015. 42 Case: 17-51060 Document: 00515984856 Page: 43 Date Filed: 08/18/2021 No. 17-51060 of the second set of laminaria and the actual evacuation procedure. 25 The record does not clearly indicate how often two sets of laminaria are needed. The evidence reflects that even without the use of digoxin to ensure fetal demise, many second-trimester dilation and evacuation abortions are multi-day procedures. Including the waiting period, roughly one-half of abortions performed between 15 weeks and 17 weeks and six days take two days, while the remaining half take three days. Most abortions performed during and after week 18 take three days, but some may take up to four. The use of digoxin to ensure fetal demise would have no effect on approximately two-thirds or more of abortions occurring from 15 to 22 weeks of gestation. In theory, digoxin becomes effective over a period of 30 minutes to 24 hours, depending on its method of administration—intra-cardiac, intra- fetal, or intra-amniotic. 26 In practice, however, the record reflects that physicians choose to administer it and wait 24 hours to ensure fetal death before performing the evacuation. 27 Digoxin can be administered on the day that dilation begins. 28 Accordingly, for abortions in which physicians would otherwise be able to perform dilation and evacuation on the same day— roughly one-half of abortions performed between 15 weeks and 17 weeks and six days—digoxin might add an additional day to the procedure. But for the other half of abortions performed between 15 weeks and 17 weeks and six days, and all abortions performed past 17 weeks and six days, using digoxin to cause fetal demise would not result in any delay in the abortion process. 25 See ROA.1923-24. 26 See ROA.4433, 4582-83, 4653; see also ROA.2101-04, 2659. 27 See ROA.1937, 1941, 2029, 2041, 2101-04, 2113, 2150. 28 See ROA.2091, 4312. 43 Case: 17-51060 Document: 00515984856 Page: 44 Date Filed: 08/18/2021 No. 17-51060 The record in the present case contains data about abortions in Texas from 2011 to 2015. The number of late-term abortions (15 weeks to 22 weeks) pales in comparison to the number of abortions performed up to 15 weeks of gestation: 29 Abortions at Less than Abortions at 15 Weeks 15 Weeks to 22 Weeks 2011 69,913 2,287 2012 65,642 2,434 2013 60,915 2,147 2014 50,979 3,135 2015 50,746 3,175 If, as the record evidence reflects, about one-half of abortions performed between 15 weeks and 17 weeks and 6 days of gestation will not be delayed at all by using digoxin to cause fetal demise prior to proceeding with evacuation, the data also reflects that only about one-third of all abortions performed from 15 weeks up to the 22nd week of gestation would be delayed by approximately an additional 24 hours (that is, delayed another 24 hours beyond the initial waiting period): 30 Abortions at One-Half of Abortions 15 weeks to Abortions at from 15 17 weeks and 15 weeks to weeks to 22 % 6 days 17 weeks and weeks 6 days 2011 1503 752 2287 32.88% 2012 1639 820 2434 33.69% 2013 1425 713 2147 33.21% 2014 2315 1158 3135 36.94% 29 See ROA.4242-4259. 30 See ROA.4242-4259. 44 Case: 17-51060 Document: 00515984856 Page: 45 Date Filed: 08/18/2021 No. 17-51060 2015 2088 1044 3175 32.88% But even were a delay to occur in all abortions from 15 to 22 weeks, as already noted above, a 24-hour delay, in addition to a 24-hour waiting period, does not constitute a substantial obstacle. 31 Requiring someone seeking to abort a fetus at or beyond 15 weeks of gestation to wait 24 hours to reflect upon the decision, 32 and to wait an additional 24 hours to ensure the demise of the fetus in utero before proceeding with an abortion that may or will involve the use of forceps or similar devices to dismember the fetus does not present a substantial obstacle to a woman seeking an abortion. Judicial bypass proceedings for minors that can delay an abortion well beyond 48 hours have been upheld by the Supreme Court. The Court’s opinion in Ohio v. Akron Center for Reproductive Health 33 is instructive. The Court explained that “the Bellotti principal opinion indicated that courts must conduct a bypass procedure with expedition to allow the minor an effective opportunity to obtain the abortion.” 34 The judicial bypass procedure under consideration in Akron Center required the trial court to render its decision no more than five business days after the minor filed a complaint, required the state court of appeals to docket an appeal no more than four days after the minor filed a notice of appeal, and required the court 31 See ante at 29-30 (framing the issue in terms of an “undue burden” rather than a “substantial obstacle”). 32 See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 885 (1992) (upholding a “24–hour waiting period between the provision of the information deemed necessary to informed consent and the performance of an abortion,” reasoning in part that “[t]he idea that important decisions will be more informed and deliberate if they follow some period of reflection does not strike us as unreasonable, particularly where the statute directs that important information become part of the background of the decision.”). 33 497 U.S. 502 (1990). 34 Id. at 513 (citing Bellotti v. Baird, 443 U.S. 622, 644 (1979)). 45 Case: 17-51060 Document: 00515984856 Page: 46 Date Filed: 08/18/2021 No. 17-51060 of appeals to render its decision no more than five days after docketing the appeal. 35 The Supreme Court held that these bypass procedures withstood a facial challenge. 36 It did so even though the Sixth Circuit had construed “days” to mean business days and had calculated that the statute permitted a delay of up to 22 days, and even though the record included an affidavit averring that “a 3–week delay could increase by a substantial measure both the costs and the medical risks of an abortion.” 37 Though the Supreme Court questioned the soundness of construing “day” to mean “business day,” it proceeded to hold that “the mere possibility that the procedure may require up to 22 days in a rare case is plainly insufficient to invalidate the statute on its face.” 38 The Supreme Court pointed out that in Planned Parenthood of Kansas City, Missouri, Inc. v. Ashcroft, 39 “for example, [it had] upheld a Missouri statute that contained a bypass procedure that could require 17 calendar days plus a sufficient time for deliberation and decisionmaking at both the trial and appellate levels.” 40 In Ashcroft, the Eighth Circuit had rejected Planned Parenthood’s argument that “the statute does not assure that the procedure will be . . . expeditious.” 41 The Eighth Circuit concluded that Missouri’s bypass “statute sets forth reasonable time requirements for court action on the petition” and explained that “[a]lthough the statute does no more than direct 35 See id. 36 See id. at 514. 37 Id. at 513. 38 Id. at 514. 39 462 U.S. 476 (1983). 40 Akron Ctr., 497 U.S. at 514 (citing Ashcroft, 462 U.S. at 477 n.4, 491 n.16). 41 Planned Parenthood Ass’n of Kansas City, Mo., Inc. v. Ashcroft, 655 F.2d 848, 860 (8th Cir. 1981), aff’d in part & rev’d in part, 462 U.S. 476 (1983). 46 Case: 17-51060 Document: 00515984856 Page: 47 Date Filed: 08/18/2021 No. 17-51060 the Missouri Supreme Court to promulgate rules for expedited appellate review, we are confident the Missouri Supreme Court will exercise its jurisdiction in a manner that recognizes the serious dangers caused by delay.” 42 The Supreme Court affirmed the Eighth Circuit’s judgment “insofar as it . . . upheld the State's parental and judicial consent provision.” 43 An additional delay of 24 hours caused by the use of a medical procedure to ensure that a living fetus in not dismembered or disemboweled in utero does not constitute a substantial obstacle for women seeking an abortion. At least some of the Supreme Court’s rationale in upholding a State’s imposition of a 24-hour waiting period before proceeding with an abortion provides support for this conclusion. In Casey, the Supreme Court reasoned that the statute at issue “permit[ted] avoidance of the waiting period in the event of a medical emergency and the record evidence show[ed] that in the vast majority of cases, a 24–hour delay does not create any appreciable health risk.” 44 Texas’s prohibition of dismemberment abortions does not apply in cases of medical emergency. 45 There is no indication in the record that when an additional day to perform an abortion would be necessary to induce fetal demise, such a delay would create an appreciable risk to the woman seeking an abortion. The Supreme Court reasoned in Casey that, “[i]n theory, at least, the waiting period is a reasonable measure to implement the State's interest in protecting the life of the unborn, a measure that does 42 Id. 43 Ashcroft, 462 U.S. at 494 (holding that bypass provisions in the version of MO. REV. STAT. § 188.028 (2019) in effect in 1981, see Act of June 29, 1979, No. 523, § 188.028, 1979 Mo. Laws 375, 376-78, were constitutional). 44 Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 885 (1992). 45 See Tex. Health & Safety Code Ann. § 171.152(a) (West 2017). 47 Case: 17-51060 Document: 00515984856 Page: 48 Date Filed: 08/18/2021 No. 17-51060 not amount to an undue burden.” 46 Prohibiting dismemberment of a living fetus by the use of forceps or similar devices in the manner described in Texas Health and Safety Code § 171.151 (including removing the living “unborn child” from the uterus “one piece at a time” and “cut[ting]” or “rip[ping]” “piece[s]” from the living unborn child’s body with forceps or other similar devices) 47 is a reasonable measure to protect the unborn and does not amount to an undue burden. The Texas laws at issue in the present appeal should not have been struck down by the district court. III Though I conclude that any delay caused by the Texas laws at issue is not a substantial obstacle in the path of a woman seeking to abort a nonviable fetus because of the brevity of any additional delay beyond the waiting period, even assuming that a 24-hour delay were a substantial obstacle, the prohibition of dismemberment abortions would not “be unconstitutional in a large fraction of relevant cases.” 48 The record reflects that there would be an additional delay of 24 hours for something less than approximately one- third, at most, of those obtaining an abortion from 15 to 22 weeks of gestation. That is because not all physicians would use digoxin from 15 weeks to 16 weeks and 6 days of gestation to cause fetal demise. They would use suction to cause the death of the fetus. Physicians and experts disagreed as to when digoxin would be used during 15 weeks to 17 weeks six days of gestation, which is the only time frame 46 Casey, 505 U.S. at 885. 47 Tex. Health and Safety Code Ann. § 171.151 (West 2017). 48 Gonzales v. Carhart, 550 U.S. 124, 167-168 (2007). 48 Case: 17-51060 Document: 00515984856 Page: 49 Date Filed: 08/18/2021 No. 17-51060 during which digoxin might cause a 24-hour delay, since the record reflects that all abortions performed at and beyond 18 weeks take longer than 48 hours. The fraction of women for whom the use of digoxin might cause an additional 24-hour delay is therefore less than all abortions from 15 to 22 weeks. The number does not exceed or even reach 37% of all abortions from 15 to 22 weeks based on the record, and there is considerable evidentiary support for the conclusion that the fraction is much smaller, ranging from 4.63% to 9.57% in a given year. Some of Plaintiffs’ own experts testified that digoxin would not be used until 17 weeks of gestation because suction or vacuum abortions would cause fetal demise up through 16 weeks and six days of gestation. That testimony supports the much smaller fractions ranging from 4.63% to 9.57 % and certainly a fraction of less than 36.9%. In determining the “fraction of relevant cases,” the denominator consists of the cases in which the statute is “relevant,” 49 encompassing “those [women] for whom [the provision] is an actual rather than an irrelevant restriction.” 50 The Supreme Court explained in Gonzales v. Carhart that “relevant cases” means “all instances in which the doctor proposes to use the prohibited procedure.” 51 The denominator in the present case is comprised of women who seek an abortion from 15 to 22 weeks of gestation, the period in which the record reflects that physicians might perform a dismemberment abortion. As noted, there is conflicting evidence as to the numerator, again, assuming for the sake of argument that a delay of 24 hours is a substantial 49 Id. at 168; see Casey, 505 U.S. at 895. 50 Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292, 2320 (2016) (alterations in original) (quoting Casey, 505 U.S. at 895). 51 Gonzales, 550 U.S. at 168. 49 Case: 17-51060 Document: 00515984856 Page: 50 Date Filed: 08/18/2021 No. 17-51060 obstacle. The number of women who might experience a delay depends on the number of women whose abortion procedure would be extended 24 hours due to the injection of digoxin. One of the plaintiffs’ witnesses testified that “in the vast majority of cases . . . in a situation of normal anatomy, I would be able to absolutely complete [a suction abortion causing fetal demise] through 16.6 [sixteen weeks and six days of gestation].” 52 This witness stated that “there may be a few select cases” 53 in which the “uterine anatomy might make the use of a suction cannula difficult or impossible,” 54 but even if SB8 went into effect, this physician would continue to perform suction abortions through 16 weeks and six days of gestation. 55 She would not use digoxin during abortions at 15 weeks to 16 weeks and 6 days. 56 So, if the numerator were based on this witness’s testimony, potential delay due to the use of digoxin might only occur at week 17 through week 17 and 6 days. Up to the 17th week, digoxin would not be used to cause fetal demise; only suction would be used, so there would be no additional delay from 15 weeks to 16 weeks and 6 days of gestation. A witness for the State similarly testified that suction could be used to cause fetal demise through 16 weeks and 6 days of gestation, and digoxin would not be necessary. 57 If we considered only the testimony of these witnesses, the percentage of abortions delayed due to use of digoxin would range from 4.63% to 9.57% from 2011 to 2015: 58 52 ROA.2227-2228. 53 ROA.2227. 54 ROA.2223. 55 ROA.2221. 56 ROA.2221. 57 See ROA.2587-2590 (Chireau, a witness for the State). 58 See ROA.4242-59. 50 Case: 17-51060 Document: 00515984856 Page: 51 Date Filed: 08/18/2021 No. 17-51060 Abortions One-Half of Abortions during Week Abortions from 15 17 During weeks to 22 Week 17 weeks % 2011 211 106 2287 4.63% 2012 249 125 2434 5.14% 2013 217 109 2147 5.08% 2014 599 300 3135 9.57% 2015 568 284 3175 8.94% But there is testimony from other witnesses indicating that suction alone does not or cannot always cause fetal demise and therefore, some other means of ensuring the death of the fetus prior to use of forceps or a similar device would be employed. Some physicians would use digoxin instead of suction alone at varying stages of gestation from 15 weeks up to the end of the 16th week. 59 There was testimony that, from 15 weeks up to 18 weeks, some abortion providers have forceps at hand in case suction aspiration cannot fully evacuate the fetus. 60 Some providers said they were unwilling to risk violating Texas law if they began, but were unable to complete, an abortion 59 See, e.g., ROA.1921 (one physician stating that he generally stops completing abortions without the use of forceps, i.e., with suction alone, at 15 weeks); ROA.1972 (the same physician explaining that he does not use suction alone during week 16, although he sometimes uses suction alone during week 15); ROA.2012 (another physician stating that he “commonly prepare[s] for the use of forceps [and not suction alone] around 15 ½ weeks”); ROA.2176-77 (another physician stating that she “switch[es]” from suction to forceps at 15 weeks (quotation at 2177)); ROA.2205 (another physician explaining that she began keeping forceps on hand at 15 weeks, and that she could not generally know before beginning a procedure whether she would need to use forceps); ROA.2223-24, 2226-28 (that same physician explaining that there were some cases before 16.6 weeks in which she would not be able to use suction to comply with the statute); ROA.2689 (another physician stating that he could not offer the opinion that the use of suction would be possible in every case up to 16 weeks); ROA.2807 (another physician describing the ability to use suction alone at 15 to 16 weeks as “unpredictable”). 60 See, e.g., ROA.2012, 2205. 51 Case: 17-51060 Document: 00515984856 Page: 52 Date Filed: 08/18/2021 No. 17-51060 using suction, then switched to forceps to complete the procedure without having first caused fetal demise. 61 (The record reflects that virtually all physicians cause fetal demise before performing an abortion at and after 18 weeks of gestation in order to avoid the risk of violating the federal ban on partial-birth abortion. However, as discussed above, using digoxin causes no additional delay when aborting a fetus at 18 to 22 weeks of gestation because the abortion process in all of those cases already extends more than 24 hours beyond Texas’s initial statutory 24-hour waiting period.) Different physicians employ differing practices. The Plaintiffs did not quantify how many women across Texas would experience a delay of an additional 24 hours due to causing fetal demise by the use of digoxin. It was their burden to do so in this facial challenge. IV To the extent that consideration of the benefits of Texas’s prohibition on fetal demise by dismemberment is a relevant inquiry, the record is silent as to how a means of bringing about fetal demise prior to dismemberment, such as digoxin, actually affects the fetus. Other than reflecting that digoxin causes fetal death in a large percentage of cases in which it is administered within 24 hours before an abortion, there is no evidence as to how digoxin brings about fetal death. There is no evidence as to potential pain or suffering while the fetus succumbs after introduction of digoxin into the womb, and if there is such a potential, the nature and duration of any pain or suffering. The State has expressed its interest in prohibiting the dismemberment of a living fetus. This is congruent with the widely accepted principle that dismemberment of living mammals should be prohibited. For example, 61 See ROA.2223-28. 52 Case: 17-51060 Document: 00515984856 Page: 53 Date Filed: 08/18/2021 No. 17-51060 unwanted dogs, cats, puppies and kittens in shelters must be humanely euthanized under Texas law. 62 The plaintiffs have not demonstrated that causing fetal demise by the use of digoxin is morally or even factually equivalent to fetal demise by dismemberment. Both procedures are abhorrent. But it cannot be said on this record that Texas has no legitimate interest in requiring fetal demise by a means other than dismemberment during an abortion. * * * I concur in reversing the district court’s judgment and rendering judgment that the facial challenges asserted in this case to the constitutionality of Texas’s prohibition of dismemberment abortion fail. 62 See 25 Tex. Admin. Code § 169.84(a), (c) (2013) (for dogs and cats in the custody of an animal shelter, requiring the animal be euthanized only by sodium pentobarbital, and for any animal other than a dog or cat in the custody of an animal shelter, requiring the animal “be humanely euthanized only in accordance with the methods, recommendations, and procedures of the American Veterinary Medical Association” in their latest guidelines for the euthanasia of animals “applicable to that species of animal”); Tex. Health & Safety Code Ann. § 821.052 (West 2015) (same). 53 Case: 17-51060 Document: 00515984856 Page: 54 Date Filed: 08/18/2021 No. 17-51060 James C. Ho, Circuit Judge, concurring: The plurality opinion chronicles the numerous errors committed by the district court. I write separately to focus on one particular error. Constitutional challenges to abortion laws are governed, not by the text or original meaning of the Constitution, but by decisions of the Supreme Court. “[W]hat distinguishes abortion from other matters of health care policy in America—and uniquely removes abortion policy from the democratic process established by our Founders—is Supreme Court precedent.” Jackson Women’s Health Org. v. Dobbs, 945 F.3d 265, 277 & n.1 (5th Cir. 2019) (Ho, J., concurring in the judgment), cert. granted, _ S. Ct. _. Compare, e.g., Jacobson v. Massachusetts, 197 U.S. 11, 26 (1905) (rejecting substantive due process claim that “a compulsory vaccination law is . . . hostile to the inherent right of every freeman to care for his own body” and “nothing short of an assault upon his person”). So we focus on Supreme Court precedent. That precedent recognizes that scientists and medical experts disagree over a number of issues affecting abortion policy in states across the country. And when experts disagree, legislators decide—and judges defer. See, e.g., Gonzales v. Carhart, 550 U.S. 124, 163 (2007) (“The Court has given state and federal legislatures wide discretion to pass legislation in areas where there is medical and scientific uncertainty.”) (collecting cases); id. at 164 (“Medical uncertainty does not foreclose the exercise of legislative power in the abortion context any more than it does in other contexts.”); see also June Medical Servs. v. Russo, 140 S. Ct. 2103, 2136 (2020) (Roberts, C.J., concurring in the judgment) (same). This is not only the “‘traditional rule,’” but the only sensible one. June Medical, 140 S. Ct. at 2136 (Roberts, C.J., concurring in the judgment) (quoting Gonzales, 550 U.S. at 163). As the Chief Justice has observed, courts 54 Case: 17-51060 Document: 00515984856 Page: 55 Date Filed: 08/18/2021 No. 17-51060 should focus on the “sort of inquiry familiar to judges”—namely, the resolution of legal disputes—and not the heady medical and scientific controversies for which judges lack the proper qualifications to decide. Id. The district court here repeatedly violated these principles and failed to defer, as the plurality details. See ante, at 15–16 & n.13. So did the 2–1 panel majority, which chastised state officials for relying on experts that the panel deemed “less mainstream” on such hotly debated matters as the gestational age at which an unborn child begins to feel pain. Whole Woman’s Health v. Paxton, 978 F.3d 896, 910 (5th Cir. 2020), vacated and reh’g en banc granted, 978 F.3d 974 (5th Cir. 2020). And three of our dissenting colleagues today make the same move. They acknowledge that scientists disagree on these issues. Yet they insist that legislatures must take one particular side of that debate over the other. Post, at 100 n.8 (Dennis, J., dissenting). “Follow the science,” it’s often said. And rightly so. But what do we do when scientists disagree? The Supreme Court’s abortion precedents are unequivocal: Judges have no business deciding which scientists are right and which ones are wrong. Moreover, this principle is especially vital because, as it turns out, scientists don’t always follow the science themselves. I write separately to explore this concern. I. We take for granted today the overwhelming medical and scientific consensus that germs cause disease, and that handwashing is therefore essential to basic human hygiene. But it was not always so. To the contrary, germ theory and handwashing were once the subject of severe scorn and ridicule among “mainstream” scientists. In fact, it took the outspoken efforts of a few 55 Case: 17-51060 Document: 00515984856 Page: 56 Date Filed: 08/18/2021 No. 17-51060 dissenters within the medical community who were willing to withstand years of ridicule and peer pressure in order to challenge—and eventually, change—the reigning consensus. A. Ignác Semmelweis was a Hungarian physician who practiced obstetrics in the maternity clinic of the Vienna General Hospital during the late 1840s. At the time, a disease known as “childbed fever” was killing many of the women who gave birth there. Sherwin B. Nuland, The Doctors’ Plague: Germs, Childbed Fever, and the Strange Story of Ignác Semmelweis 79–85 (2004). The maternity clinic had two wards: one attended only by midwives, and the other attended only by physicians. And significantly, the physicians not only delivered babies—they also performed autopsies on women who succumbed to childbed fever. Id. at 97. Semmelweis observed that women who gave birth in the ward attended by midwives died at significantly lower rates than women who gave birth in the ward attended by physicians. Id. He hypothesized that the physicians who were also examining the bodies of women dying of childbed fever were transmitting contaminated particles from the infected patients to healthy women during childbirth. Id. at 100–01. At the time, those physicians saw no reason to wash their hands between conducting autopsies and treating healthy women in the delivery ward. So they didn’t. Id. at 100. To test his hypothesis, Semmelweis advised physicians to wash their hands in chlorine solution after performing autopsies and before treating healthy women. Id. at 101. As a result, “something remarkable [began] to happen.” Id. Childbed fever deaths in the physicians’ ward plummeted. The death rate fell to “virtually equal” in the two wards. Id. Semmelweis 56 Case: 17-51060 Document: 00515984856 Page: 57 Date Filed: 08/18/2021 No. 17-51060 proved that the hospital could virtually eliminate the spread of infection simply by insisting that physicians wash their hands. Id. at 104–05. Semmelweis’s discovery saved lives. But instead of being praised or even accepted, he was ridiculed as an “agitator” and marginalized within the scientific community for his “unorthodox and highly irregular ways of doing things.” Id. at 147, 157. More senior colleagues expressed “alarm [at] the increasing influence of younger physicians” like Semmelweis. Id. at 120. So, to use modern parlance, they cancelled him. Semmelweis was denied another term as an instructor at the medical school because of “the way he kept demanding that students and staff wash in the chloride solution.” Id. at 125. And even when he was later accepted for another teaching position, he was restricted in what courses he could teach and what materials he could access. Id. at 128. A European medical publication advised readers: “We thought that this theory of chlorine disinfection had died out long ago . . . . [O]ur readers should not allow themselves to be misled by this theory.” Id. at 144–45. Why did the scientific community “turn[] its collective back on” Semmelweis, even when it turned out that he was so obviously right—and on a matter so critical to patients’ lives? Id. at 158. Why couldn’t he “change their fatalistic attitude about the inevitability of recurrent epidemics”? Id. at 157. Those who have studied the events observe that it would “prove to be intolerable” for respected (“mainstream,” if you will) doctors to admit that they were horribly, brutally wrong—for they could not accept “the possibility that they had been killing their patients for years or decades.” Id. at 118. 57 Case: 17-51060 Document: 00515984856 Page: 58 Date Filed: 08/18/2021 No. 17-51060 B. A similar fate befell Joseph Lister. A British surgeon nearly twenty years after Semmelweis’s discovery, Lister sought to explain the causes of infection in surgical wounds that led to post-operation deaths. Lindsey Fitzharris, The Butchering Art: Joseph Lister’s Quest to Transform the Grisly World of Victorian Medicine 155–60 (2017). Lister developed a “germ theory of disease”—that certain diseases are caused by the invasion of the body by microscopic organisms. Id. at 159. And he sought out to find a “means of destroying microorganisms within the wound itself before infection could set in.” Id. He began the practice of treating wounds with a carbolic-acid antiseptic to disinfect the skin, “prevent germs from entering wounds, [and] destroy[] those that had already entered the body.” Id. at 168–70. And he advanced the technique of sterilizing surgical instruments with his antiseptic solutions before using them on patients. Id. at 177. By the time of his death, Lister would be acclaimed as “the greatest modern Englishman” and “the world’s greatest surgeon.” Laurence Farmer, Master Surgeon: A Biography of Joseph Lister 129 (1962). But throughout his career, he encountered fierce opposition, even mockery. His contemporaries could not accept his suggestion that invisible germs floating in the air could somehow cause disease. So they dismissed him as “crazy, rash, and blinded by enthusiasm.” Id. at 76. Others denigrated him as “mentally unhinged” and possessed by “a ‘grasshopper in the head.’” Fitzharris, supra, at 220. They disparaged his work as “the latest toy in medical science,” “unnecessary and overly complicated distractions,” “quackery,” and “medical hocus-pocus.” Id. at 203, 215, 218. One renowned English surgeon, in an address to the British Medical 58 Case: 17-51060 Document: 00515984856 Page: 59 Date Filed: 08/18/2021 No. 17-51060 Association, ridiculed Lister’s work as worse than “an innocent fallacy”— as nothing more than “unsupported fancies, which have little other existence than what is found in the imagination of those who believe in them.” Id. at 193. The editor of the magazine Medical Record captured the dominant mood this way: “We are likely to be as much ridiculed in the next century for our blind belief in the power of unseen germs, as our forefathers were for their faith in the influence of spirits, of certain planets and the like, inducing certain maladies.” Candice Millard, Destiny of the Republic: A Tale of Madness, Medicine and the Murder of a President 184 (2012). As with Semmelweis, Lister’s colleagues resisted his methods for the simple reason that they “direct[ly] conflict[ed] with [their own] technique[s].” Fitzharris, supra, at 180. “It was difficult for many surgeons at the height of their careers to face the fact that for the past fifteen or twenty years they might have been inadvertently killing patients by allowing wounds to become infected.” Id. at 185. II. The reaction of the “mainstream” scientific community to Semmelweis and Lister may seem outrageous to us today. But it is surprisingly typical, as explained by academics in a field known as the philosophy of science. Scientific progress is often arduous work. For science is at bottom “a conservative activity.” Samir Okasha, Philosophy of Science: A Very Short Introduction 71, 75 (2nd ed. 2016). That is, scientists typically “accept the [prevailing] paradigm unquestioningly,” and devote their research primarily to “develop[ing] and extend[ing] the existing paradigm.” Id. at 75. Scientists generally assume that any “experimental result which conflicts with the paradigm . . . is faulty, not that the paradigm 59 Case: 17-51060 Document: 00515984856 Page: 60 Date Filed: 08/18/2021 No. 17-51060 is wrong.” Id. at 76. So challenges to prevailing scientific wisdom are often dismissed. And the more entrenched the existing paradigm, the greater the upheaval, and the more vigorous the resistance will be to any challenge to the governing paradigm—as “a burgeoning sense of crisis envelops the scientific community.” Id. at 76. It may not be enough, then, that an existing paradigm deserves to be supplanted, and that a new paradigm proves to be superior. The most scientifically sound and intellectually rigorous viewpoint does not necessarily prevail. Scientists may be subject to “peer pressure” and even “mob psychology.” Id. at 77. So which view ultimately prevails may depend more on personality than merit. “If a given paradigm has very forceful advocates, it is more likely to win widespread acceptance.” Id. As a result, some academics have even begun to wonder whether “[s]cience . . . can no longer be construed simply as the ideal of the quest for truth (i.e., pure science).” Fabrice Jotterand, The Politicization of Science and Technology: Its Implications for Nanotechnology, 34 J.L. Med. & Ethics 658, 658 (2006). After all, “[s]cience, through its technological applications, has become the source of economic power and, by extension, political power.” Id. As a result, “[s]cience, with its political implications, has entered what [one scholar] calls the era of ‘post-academic’ science.” Id. And “[t]he role played by cultural-political factors in scientific research lies at the basis of a shift in how scientific inquiry is conducted.” Id. at 661. Indeed, scientific resistance to novel ideas is so pervasive that medical historians have coined a term for it: “the term ‘Semmelweis reflex’ is used to refer to the knee-jerk tendency to reject new evidence because it contradicts established norms.” Lindsey Fitzharris, The Unsung Pioneer of Handwashing, Wall St. J. (Mar. 19, 2020), available at 60 Case: 17-51060 Document: 00515984856 Page: 61 Date Filed: 08/18/2021 No. 17-51060 https://www.wsj.com/articles/the-unsung-pioneer-of-handwashing- 11584627614. The bottom line is this: Of course we should “follow the science.” But that doesn’t mean we should always blindly follow the scientists. Because, like the rest of us, scientists are, first and foremost, human beings. They’re susceptible to peer pressure, careerism, ambition, and fear of cancel culture, just like the rest of us—as courts have recognized. See, e.g., Ott v. St. Luke Hosp. of Campbell Cnty., Inc., 522 F. Supp. 706, 711 (E.D. Ky. 1981) (a “Lister or Semmelweis” might well discover the need for “salutary changes in [medical or scientific] procedures,” yet his views “may be excluded simply because he ‘makes waves’”); Kosilek v. Spencer, 774 F.3d 63, 78 (1st Cir. 2014) (en banc) (noting concern that medical debate over sex reassignment surgery may be “politically” driven); Gibson v. Collier, 920 F.3d 212, 222 (5th Cir. 2019) (same). 1 1 Similar concerns about intimidation and politicization within the scientific community have been expressed in a number of recent press accounts. See, e.g., Adam O’Neal, A Scientist Who Said No to Covid Groupthink, Wall St. J. (June 11, 2021), available at https://www.wsj.com/articles/a-scientist-who-said-no-to-covid-groupthink- 11623430659 (profiling Filippa Lentzos, a scientist and expert on biological threats who was “wary” about voicing her theory on the origins of COVID-19 because it “challenged the enforced consensus,” noting that “there are power plays,” “agendas,” and “strong vested interests” in the scientific community that cause dissenters to “fear[] for their careers [and] for their grants”); Katherine Eban, The Lab-Leak Theory: Inside the Fight to Uncover COVID-19’s Origins, Vanity Fair (June 3, 2021), available at https://www.vanityfair.com/news/2021/06/the-lab-leak-theory-inside-the-fight-to- uncover-covid-19s-origins (“[F]ormer Centers for Disease Control director Robert Redfield received death threats from fellow scientists after telling CNN that he believed COVID-19 had originated in a lab. ‘I was threatened and ostracized because I proposed another hypothesis,’ Redfield told Vanity Fair. ‘I expected it from politicians. I didn’t expect it from science.’”); John Tierney, The Panic Pandemic: Fearmongering from journalists, scientists, and politicians did more harm than the virus, City Journal (Summer 2021), available at https://www.city-journal.org/panic-pandemic (“There’s always a certain amount of herd thinking in science, but I’ve never seen it reach this level.”) 61 Case: 17-51060 Document: 00515984856 Page: 62 Date Filed: 08/18/2021 No. 17-51060 III. Doctors and scientists deserve enormous respect. We ignore their advice at our peril. But we also follow them blindly at our peril. Consider the story of Baby Richard. Born at just 21 weeks, he weighed less than a pound, small enough to fit in the palm of a hand. He had small air sacs instead of developed lungs. Oxygen was not flowing to his brain. He needed IV fluid, a breathing tube, and blood pressure support to sustain himself. He was immediately rushed to a neonatal intensive care unit. Tommy Brooksbank, ‘Miracle baby’ born at 21 weeks heads home from hospital just in time for Christmas, Good Morning America (Dec. 25, 2020), available at https://www.goodmorningamerica.com/family/story/miracle- baby-born-21-weeks-heads-home-hospital-74848084. Richard’s doctors gave him a “0% chance of survival.” Id. As his neonatologist, Dr. Stacy Kern, later noted, “many NICUs around the world are not even resuscitating babies born at 22 weeks.” Id. See also id. (noting that, “[a]ccording to the Department of Health and Human Services, babies born before 22 weeks are typically not resuscitated because their bodies are simply too immature to be treated with intensive care”). (quoting Harvard epidemiologist Martin Kulldorff); see also, e.g., Lesley Stahl, State Bills Would Curtail Health Care for Transgender Youth, 60 Minutes (May 23, 2021), available at https://www.cbsnews.com/news/transgender-health-care-60-minutes-2021-05-23/ (quoting Dr. Laura Edwards-Leeper, a psychologist at a major youth gender clinic in Boston who has “helped hundreds of teens and young adults transition successfully after a comprehensive assessment”: “It greatly concerns me where the field has been going. I feel like what is happening is unethical and irresponsible in some places. . . . Everyone is very scared to speak up because we’re afraid of not being seen as affirming or being supportive of these young people or doing something to hurt the trans community. But even some of the providers are trans themselves and share these concerns.”). 62 Case: 17-51060 Document: 00515984856 Page: 63 Date Filed: 08/18/2021 No. 17-51060 For weeks, Richard was hooked up to two ventilators to keep him breathing. But his oxygen levels continued to decline. So doctors invited his mother to Baby Richard’s bedside—to say goodbye. But then they touched. “She puts her hand on him and all the sudden his oxygen saturation goes up to the 80s then 90s, and I look at her and go, ‘I guess he just needed his mom,’” Dr. Kern later said. Id. “It was one of the most incredible things I’ve ever seen. He just continued to surprise us day after day.” Id. After six months in the hospital, Baby Richard came home—just in time for Christmas. Id. He recently celebrated his first birthday. See Sydney Page, A newborn weighed less than a pound and was given a zero percent chance of survival. He just had his first birthday., Wash. Post (June 23, 2021), available at https://www.washingtonpost.com/lifestyle/2021/06/23/ premature-baby-survive-birthday-record/. IV. States have a profound interest in respecting unborn life. See, e.g., Gonzales, 550 U.S. at 157 (“The government may use its voice and its regulatory authority to show its profound respect for the life within the woman.”). Surely that interest includes protecting the unborn from unnecessary pain and suffering. See, e.g., Jackson Women’s, 945 F.3d at 280 (Ho, J., concurring in the judgment) (“A State has an unquestionably legitimate (if not compelling) interest in preventing gratuitous pain to the unborn.”). Indeed, if states must avoid unnecessary pain to convicted murderers on death row as a matter of constitutional mandate, then surely states may avoid unnecessary pain to innocent unborn babies as a matter of constitutional discretion. “It would be surprising if the Constitution requires States to use execution methods that avoid causing unnecessary pain to 63 Case: 17-51060 Document: 00515984856 Page: 64 Date Filed: 08/18/2021 No. 17-51060 convicted murderers, but does not even permit them from preventing abortions that cause unnecessary pain to unborn babies.” Id. at 280 (citing Baze v. Rees, 553 U.S. 35, 49 (2008) (plurality opinion)). “Not surprisingly, then, members of the Supreme Court have acknowledged that avoidance of pain is indeed a valid state interest in the abortion context.” Id. (citing Webster v. Reprod. Health Servs., 492 U.S. 490, 552 (1989) (Blackmun, J., concurring in part and dissenting in part) (“I should think it obvious that the State’s interest in the protection of an embryo . . . increases progressively and dramatically as the organism’s capacity to feel pain, to experience pleasure, to survive, and to react to its surroundings increases day by day.”) (quoting Thornburgh v. Am. Coll. of Obstetricians & Gynecologists, 476 U.S. 747, 778 (1986) (Stevens, J., concurring)); Webster, 492 U.S. at 569 (Stevens, J., concurring in part and dissenting in part) (“There can be no interest in protecting the newly fertilized egg from physical pain or mental anguish, because the capacity for such suffering does not yet exist; respecting a developed fetus, however, that interest is valid.”). The record of this case demonstrates that scientists disagree about what gestational phase an unborn child begins to feel pain. See ante, at 15 n.13; see also Jackson Women’s, 945 F.3d at 274–75; id. at 279–80 (Ho, J., concurring in the judgment). Accordingly, the Supreme Court’s abortion precedents require courts to defer to legislators to resolve those debates. See, e.g., Gonzales, 550 U.S. at 163–64. But rather than defer to Texas legislators to make that judgment call, the 2–1 panel scolded them for relying on doctors the panel majority deemed outside the “mainstream.” Whole Woman’s, 978 F.3d at 910. If society takes seriously its obligation to protect the most innocent among us from unnecessary pain, it’s hard to imagine a more important issue 64 Case: 17-51060 Document: 00515984856 Page: 65 Date Filed: 08/18/2021 No. 17-51060 on which to defer to legislative judgments than the medical debate over an unborn child’s capacity to feel pain. *** Someday, scientists may look back on today’s abortion debates as shocking and barbaric—just as we look back in disbelief at those who ridiculed and ostracized proponents of handwashing and sterilizing surgical instruments to prevent disease and infection. Indeed, many have that view today. According to Carter Snead, one of the nation’s leading scholars on public bioethics and an expert witness in this case, “132 countries out of 194 that I looked at ban abortion outright, at all gestational stages, with certain exceptions defined by law,” while 178 countries generally ban abortion after a gestational age of 12 weeks. So “92 percent of all countries presumptively ban abortions at 12 weeks or less.” Texas does not ban abortion until 22 weeks. So Texas law is not only valid under the Constitution and Supreme Court precedent—it’s also more permissive than the overwhelming majority of laws around the world. Yet federal courts have blocked it for four years. This in spite of the fact that, when it comes to medical disputes surrounding abortion, Supreme Court precedent requires judges to defer to—not overturn—the will of the voters and the judgment of the legislators they elected to office. “The right to vote means nothing if we abandon our constitutional commitments and allow the real work of lawmaking to be exercised by [federal judges], rather than by elected officials accountable to the American voter.” Texas v. Rettig, 993 F.3d 408, 410–11 (5th Cir. 2021) (Ho, J., dissenting from denial of rehearing en banc). After four years, the court today finally allows the law to take effect. I concur. 65 Case: 17-51060 Document: 00515984856 Page: 66 Date Filed: 08/18/2021 No. 17-51060 James L. Dennis, Circuit Judge, joined by Stewart and Graves, Cir- cuit Judges, dissenting: In Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833, 846 (1992) (plurality opinion), three Supreme Court Justices set forth the core principles that have come to guide the modern jurisprudence of abor- tion. The foremost among these was that women have a constitutional right “to choose to have an abortion before [fetal] viability and to obtain it without undue interference from the State.” Id. In other words, “[b]efore viability, the State’s interests are not strong enough to support a prohibition of abor- tion or the imposition of a substantial obstacle to the woman’s effective right to elect the procedure.” Id. The corollary to this principle is known as the undue burden standard, under which state regulations that have “the pur- pose or effect ” of “plac[ing] a substantial obstacle in the path of a woman seeking an abortion before the fetus attains viability” are unconstitutional. Id. at 877. Notwithstanding Casey’s clear statement that “[u]nnecessary health regulations that have the purpose or effect of presenting a substantial obstacle to a woman seeking an abortion impose an undue burden on the right,” id. at 878, our court has frequently failed to identify and strike down laws that tar- get abortion rights under the semblance of regulating the procedure. Five years ago, the Supreme Court reversed our upholding of a Texas law that, although ostensibly a medical regulation, provided very few if any actual medical benefits and instead mainly served to hinder a woman’s right to a previability abortion. See Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292, 2318 (2016). Only two years later, our court declined to heed the Hel- lerstedt decision and approved a virtually identical Louisiana law, substituting our own strained reading of the evidence for the findings of the district court in order to conclude that the burdens the law placed on women’s abortion 66 Case: 17-51060 Document: 00515984856 Page: 67 Date Filed: 08/18/2021 No. 17-51060 choice did not outweigh its benefits. June Med. Servs., L.L.C. v. Gee, 905 F.3d 787, 815 (5th Cir. 2018). As one might expect, the Supreme Court again reversed our decision, reprimanding us for defying on-point binding prece- dent and failing to defer to the district court’s factual findings that were plau- sible in light of the full record, as an appeals court must on clear error review. June Med. Servs., L.L.C., v. Russo, 140 S. Ct. 2103, 2121, 2124-25 (2020) (plu- rality opinion); id. at 2133-34, 2141 (Roberts, C.J., concurring). Today, in a Sisyphean return to form, our court upholds a Texas law that, under the guise of regulation, makes it a felony to perform the most common and safe abortion procedure employed during the second trimester. In an opinion that fortunately lacks fully binding precedential effect, the en banc plurality disregards the two major lessons of June Medical. First, it ig- nores on-point Supreme Court precedent in multiple ways. For one, the plu- rality wrongly declares a single Justice’s concurrence to be precedential in order to impose a variation of the undue burden standard that the Court has explicitly rejected. See Hellerstedt, 136 S. Ct. at 2309. And, even under the plurality’s preferred standard, the Supreme Court has already decided this exact case, holding that a Nebraska law was unconstitutional because it could be interpreted to be the sort of ban that the Texas statute openly embodies. Stenberg v. Carhart, 530 U.S. 914, 945(2000). Second, just as in June Medical, the en banc plurality fails to defer to the district court’s well-reasoned and well-supported factual findings regarding the burdens and benefits associated with the Texas law, instead substituting its own reading of the evidence to make findings of fact in the first instance. This would be bad enough on its own, but the actual findings that the plurality makes are contrary to the great weight of the evidence in the record and place us at odds with virtually every other court to have considered the matter. In a final, entirely new sort of error, the plurality faults the district court for “botch[ing]” the large fraction analysis, Plurality at 33, which asks 67 Case: 17-51060 Document: 00515984856 Page: 68 Date Filed: 08/18/2021 No. 17-51060 whether the challenged restriction is an undue burden for a large portion of the women affected by it. But the plurality “bungl[es]” the analysis itself, Plurality at 3, incorrectly minimizing the statute’s impact by wrongly includ- ing in its evaluation a large number of women whose lives will be wholly un- affected by SB8. The court’s decision today will, in the name of “medical ethics,” force many women to unnecessarily undergo what the en banc plurality wrongfully characterizes as “alternatives” to the very common and safe pro- cedure that Texas has banned—painful, invasive, expensive, and in some cases experimental additional treatments that carry with them significantly elevated risks to the women’ health and well-being. Further burdening abor- tion access, many abortion providers will likely decline to perform later-term abortions rather than face the dilemma today’s ruling foists upon them: be- come a felon or do a risky procedure that is contrary to the doctor’s medical judgment regarding the patient’s best interests. This outcome is neither cor- rect as a logical matter nor consistent with our duties as a lower federal ap- pellate court, and I respectfully but emphatically dissent. I. A. As courts have long recognized, dilation and evacuation (“D&E”) is “the most commonly used method for performing previability second tri- mester abortions.” Stenberg, 530 U.S. at 945. The procedure is generally performed as a two-step process. The first step remains the same throughout all stages of the pregnancy: doctors induce dilation through medication alone or in combination with small sticks made from an expanding organic or syn- thetic material called laminaria. But the technique employed at the second step—evacuation—varies depending on how advanced the woman’s preg- nancy is at the time of the procedure. 68 Case: 17-51060 Document: 00515984856 Page: 69 Date Filed: 08/18/2021 No. 17-51060 The en banc plurality claims that there are three seemingly equally ac- ceptable “main” options that a doctor may elect to employ during the evac- uation phase of the D&E. Plurality at 3. According to the plurality, once a woman’s cervix has been dilated, a doctor may evacuate the contents of her uterus using either suction alone, a combination of suction and forceps, or various “fetal-death” techniques in conjunction with suction and forceps. Plurality at 3. This characterization of the procedure is inaccurate. The rec- ord and the district court’s findings make clear that there are only two ways to perform the second step of a D&E: suction alone or in conjunction with forceps or similar implements, with the gestational age of the fetus the pri- mary factor dictating which technique the doctor can safely and effectively employ. Generally, during the first trimester, 1 the contents of the uterus can be evacuated via suction with a plastic tube called a “cannula” in a process termed “suction aspiration.” The suction causes the fetal tissue to separate, resulting in fetal demise, and it removes the residual contents of the womb. But beginning during the second trimester at around fifteen weeks of preg- nancy, the most common method of abortion both in Texas and nationally involves the additional use of forceps or similar handheld medical imple- ments. Performed in an outpatient setting, this very safe, approximately ten- minute procedure differs from the early-stage procedure in that, rather than relying solely on suction during the step-two evacuation phase, the physician uses forceps to reach into the uterine cavity and manually remove the fetal tissue through the cervix. Because of its size and position, doctors use the 1 The gestational age of a fetus is measured by the time elapsed since the woman’s last menstrual period. Pregnancy is commonly separated into three trimesters. The first trimester runs from the first through twelfth week and the second trimester runs from the thirteenth through twenty-sixth week. See Stenberg, 530 U.S. at 923-25. The third tri- mester begins the twenty-seventh week and continues through the end of the pregnancy. 69 Case: 17-51060 Document: 00515984856 Page: 70 Date Filed: 08/18/2021 No. 17-51060 forceps to “disarticulate” or separate the fetal tissue into pieces small enough to be removed through the dilated opening. Once the removal is com- plete, the doctor uses suction to remove any residual material remaining in the uterus. As other courts have recognized and as will be discussed, there are some additional measures that doctors can perform during a D&E wherein various techniques are used to independently produce fetal demise prior to evacuation, but these are not an alternative method of evacuation as the en banc plurality seems to claim. See EMW Women’s Surgical Ctr., P.S.C. v. Friedlander, 960 F.3d 785, 798 (6th Cir. 2020) (“Fetal-demise procedures are not, by definition, alternative procedures. A patient who undergoes a fetal- demise procedure must still undergo the entirety of a standard D&E. Instead, fetal-demise procedures are additional procedures.”), cert. granted in part on other grounds sub nom. Cameron v. EMW Women’s Surgical Ctr., P.S.C., 141 S. Ct. 1734 (2021). The tissue separation that occurs during a forceps-assisted D&E results in fetal demise, and the procedure does not require an addi- tional, antecedent step of producing fetal demise through other methods. As is discussed in more detail below, performing such an extra step significantly increases the health risks and physical, emotional, and financial costs associ- ated with the procedure. B. In 2017, Texas enacted Senate Bill 8 (“SB8”). Along with a number of other provisions exhibiting hostility to a woman’s constitutional right to obtain a previability abortion, the law prohibits so-called “dismemberment abortions.” 2 Act of May 26, 2017, 85th Leg. R.S., ch. 441, § 6, 2017 Tex. 2 SB8 does not contain any legislative findings, and the district court did not make any factual findings regarding the Texas legislature’s intent in enacting it. But when con- sidering similar bans, well-respected jurists have posited that the abortion method was 70 Case: 17-51060 Document: 00515984856 Page: 71 Date Filed: 08/18/2021 No. 17-51060 Gen. Laws 1164, 1165–67 (eff. Sept. 1, 2017) (codified as TEX. HEALTH & SAFETY CODE §§ 171.151–.154). Obviously, this pejorative label, which the en banc plurality largely adopts, is not found in any medical texts. But the statute defines the procedure as one in which the physician, “with the pur- pose of causing the death of an unborn child, dismembers the living unborn child and extracts the unborn child one piece at a time from the uterus through the use of clamps, grasping forceps, tongs, scissors, or a similar in- strument.” 3 Id. Violation of the statute is a felony offense punishable by a minimum of 180 days to a maximum of two years in jail and a fine of up to $10,000. Texas asserts in the present litigation that SB8 proscribes the use of forceps or similar instruments to produce fetal demise during the second step of the D&E procedure. The State concedes that SB8 does not prohibit a suc- tion-aspiration abortion, and it likewise asserts that an abortion in which fetal demise occurs prior to the evacuation of the uterus with forceps is outside the statute’s ambit. In other words, a physician performing a D&E in which forceps are needed could typically avoid criminal liability only by taking the targeted “not because the procedure kills the fetus, not because it risks worse complica- tions for the woman than alternative procedures would do, not because it is a crueler or more painful or more disgusting method of terminating a pregnancy.” Stenberg, 530 U.S. at 951-52 (Ginsburg, J., concurring) (quoting Hope Clinic v. Ryan, 195 F.3d 857, 881 (7th Cir. 1999) (Posner, J., dissenting)). “Rather . . . the law prohibits the procedure because the state legislators seek to chip away at the private choice shielded by Roe v. Wade, 410 U.S. 113 (1973), even as modified by Planned Parenthood of Southeastern Pa. v. Casey, 505 U.S. 833 (1992).” Id. at 952. “[I]f a statute burdens constitutional rights and all that can be said on its behalf is that it is the vehicle that legislators have chosen for expressing their hostility to those rights, the burden is undue.” Hope Clinic, 195 F.3d. at 881 (Posner, J., dissenting); see also, Casey, 505 U.S. at 877 (stating that a law imposes an undue burden if it has “the purpose or effect ” of “plac[ing] a substantial obstacle in the path of a woman seeking an abortion before the fetus attains viability” (emphasis added)). 3 The statute includes an exception for medical emergencies. TEX. HEALTH & SAFETY CODE § 171.152. 71 Case: 17-51060 Document: 00515984856 Page: 72 Date Filed: 08/18/2021 No. 17-51060 additional, medically unnecessary step of inducing fetal demise in utero be- fore performing the evacuation phase, regardless of the doctor’s professional medical judgment whether such action is safe or appropriate. The Plaintiffs in this case, who are six licensed abortion clinics and five abortion providers that operate in Texas, filed the present lawsuit against the defendants, who are various Texas law enforcement officers acting in their official capacity. Plaintiffs contended that SB8 places an unconstitu- tional undue burden on a woman’s ability to obtain a previability abortion. Following an extensive five-day bench trial and consideration of testimony from numerous medical experts and a multitude of professional literature, the district court agreed that the statute is unconstitutional. The court issued a thorough memorandum opinion that meticulously reviewed and parsed the complex evidence the parties had introduced, weighed the competing narratives, made credibility determinations, and oth- erwise resolved complicated factual disputes in the manner that district courts are uniquely situated to do within our judicial system. See Whole Woman’s Health v. Paxton, 280 F. Supp. 3d 938, 941 n.5 (W.D. Tex. 2017) (“In making these findings and conclusions, the court has considered the rec- ord as a whole. The court has observed the demeanor of the witnesses and has carefully weighed that demeanor and the witnesses’ credibility in deter- mining the facts of this case and drawing conclusions from those facts. Fur- ther, the court has thoroughly considered the testimony of both sides’ expert witnesses and has given appropriate weight to their testimony in selecting which opinions to credit and upon which not to rely.”). The court evaluated each of the State’s proposed methods by which a doctor could comply with SB8, and, “[a]fter considering all of the medical expert testimony, the court conclude[d] that pre-evacuation fetal demise provides no additional medical benefit to a woman undergoing a standard D & E abortion.” Id. at 949. In- stead, the court found, each of the proposed techniques significantly 72 Case: 17-51060 Document: 00515984856 Page: 73 Date Filed: 08/18/2021 No. 17-51060 increases the risk and physical, emotional, and financial cost associated with the D&E procedure. Id. at 953. The court acknowledged that Texas has a legitimate interest in promoting “respect for the life of the unborn,” but it explained that this interest did not outweigh the considerable burden SB8 imposes on a woman’s ability to obtain the previability abortion to which she is constitutionally entitled. Id. The court thus concluded that “requiring a woman to undergo an unwanted, risky, invasive, and experimental procedure in exchange for exercising her right to choose an abortion, substantially bur- dens that right.” Id. And the district court accordingly declared SB8 facially unconstitutional and permanently enjoined its enforcement. Id. at 954. The State appealed, and we held this case in abeyance while the Su- preme Court decided June Medical Services L.L.C. v. Russo, 140 S. Ct. 2103 (2020), a case much like this one in which a majority of this court defied on- point Supreme Court precedent and substituted its own stilted interpretation of the evidence for the district court’s first-hand findings. See June Med. Servs., L.L.C. v. Gee, 913 F.3d 573, 574, 579-84 (5th Cir. 2019) (Dennis, J., dissenting from denial of en banc rehearing). The Supreme Court in June Medical—including the Chief Justice in his separate concurrence—rebuked this court’s temerity, chastising us about the importance of stare decisis and the deference that appeals courts owe to a district court’s factual findings. See 140 S. Ct. at 2121, 2124-25 (plurality opinion); id. at 2133-34, 2141 (Roberts, C.J., concurring). But after the Supreme Court issued June Medical, Texas filed a motion for a stay of the district court’s injunction in this case in light of that decision, somehow interpreting the Supreme Court’s admonishment that our court should heed controlling precedent and defer to a district court’s findings of fact as an invitation for our court to depart from Supreme Court jurisprudence and overturn the district court’s factual findings. Rec- ognizing the absurdity of this proposition, a majority of a panel of this court denied Texas its requested stay, 972 F.3d 649 (5th Cir. 2020), and then 73 Case: 17-51060 Document: 00515984856 Page: 74 Date Filed: 08/18/2021 No. 17-51060 affirmed the district court’s decision on the merits, 978 F.3d 896 (5th Cir. 2020). But an en banc majority of this court vacated that decision, 978 F.3d 974 (5th Cir. 2020), and it now reverses with only a plurality agreeing upon a rationale. II. We review the district court’s decision to permanently enjoin enforce- ment of SB8 for abuse of discretion. See Jackson Women’s Health Org. v. Dobbs, 945 F.3d 265, 270 (5th Cir. 2019), cert. granted in part, 209 L. Ed. 2d 748 (May 17, 2021). The court’s underlying conclusions of law are reviewed de novo. Guzman v. Hacienda Records & Recording Studio, Inc., 808 F.3d 1031, 1036 (5th Cir. 2015). Its findings of fact, on the other hand, are reviewed for clear error. Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 573 (1985). “If the district court’s account of the evidence is plausible in light of the rec- ord viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evi- dence, the factfinder’s choice between them cannot be clearly erroneous.” Id. at 573-74. And “[w]hen findings are based on determinations regarding the credibility of witnesses, [Federal] Rule [of Civil Procedure] 52(a) de- mands even greater deference to the trial court’s findings; for only the trial judge can be aware of the variations in demeanor and tone of voice that bear so heavily on the listener’s understanding of and belief in what is said.” Id. at 575. The en banc plurality relies on statements in Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 501 (1984), and Pullman-Standard v. Swint, 456 U.S. 273, 287 (1982), to boldly state that, because the district court em- ployed the wrong legal standard, “[w]e therefore owe no deference to the district court’s factual findings.” Plurality at 19. As discussed below, the 74 Case: 17-51060 Document: 00515984856 Page: 75 Date Filed: 08/18/2021 No. 17-51060 district court employed the correct legal standard, so this contention fails from the start. But even were that not the case, the plurality gives no clear reason for its holding that a district court’s mistake regarding the rule for de- termining whether an abortion restriction is constitutional relieves us of our duty to defer to the underlying factual findings that the district court applied that standard to. “Clear error review follows from a candid appraisal of the compara- tive advantages of trial courts and appellate courts. While we review tran- scripts for a living, they listen to witnesses for a living. While we largely read briefs for a living, they largely assess the credibility of parties and witnesses for a living.” June Med. Servs., L.L.C., 140 S. Ct. at 2141 (Roberts, C.J., con- curring) (internal quotes and citation omitted). To be sure, the Supreme Court has stated that “[a] finding of fact in some cases is inseparable from the principles through which it was deduced,” and there may be times when an error of law makes it appropriate to set aside a “so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law.” Bose, 466 U.S. at 501 & n.17. But when factual questions are not intertwined with questions of law, district courts remain in a far better position than appellate courts to evaluate credibility and parse conflicting evidence in order to resolve them. And while a misunderstanding of the governing law might affect which factual disputes a district court chooses to resolve, see Swint,456 U.S. at 287 n.17 (“The presence of . . . legal errors may justify a remand by the Court of Appeals to the District Court for additional factfinding under the correct legal standard.”), it generally has lit- tle bearing on whether the purely factual findings that a district court does make are accurate. The en banc plurality does not explain why the district court’s appli- cation of what the plurality believes was an incorrect legal standard—weigh- ing SB8’s burdens against its benefits to determine its constitutionality— 75 Case: 17-51060 Document: 00515984856 Page: 76 Date Filed: 08/18/2021 No. 17-51060 would in any way undermine the district court’s first-hand reading of the ev- idence of what those burdens and benefit are. It thus provides no reason for our withholding the clear-error deference mandated for district courts’ fac- tual determinations. See id. at 287 (“[Federal] Rule [of Civil Procedure] 52(a) broadly requires that findings of fact not be set aside unless clearly er- roneous. It does not make exceptions or purport to exclude certain categories of factual findings from the obligation of a court of appeals to accept a district court’s findings unless clearly erroneous.”). The Supreme Court has had to remind our court in recent years that, even in abortion cases, we are an ap- pellate court that should not second guess a district court’s reading of con- flicting evidence. See June Med. Servs., L.L.C., 140 S. Ct. at 2121, 2124-25 (plurality opinion); id. at 2133-34, 2141 (Roberts, C.J., concurring). I would take that lesson to heart and hold that the clear error standard of review ap- plies to the district court’s factual findings in the present case. III. On the merits, the en banc plurality claims that the district court com- mitted a range of legal errors by employing the wrong legal standard, failing to heed binding Supreme Court precedent, not sufficiently crediting the State’s legitimate interests in enacting SB8, and making several other miscel- laneous mistakes. It also asserts that the district court’s factual findings re- garding the burdens SB8 imposes on abortion access are unsupported or con- tradicted by the record. And the plurality contends that the district court misapplied the “large fraction” analysis when determining what proportion of women seeking previability abortions would be unduly burdened by SB8. Each of the plurality’s claims of error is wrong and provides no grounds for reversal, and each will be considered and rejected in turn. 76 Case: 17-51060 Document: 00515984856 Page: 77 Date Filed: 08/18/2021 No. 17-51060 A. It has been clear since the Supreme Court’s landmark decision in Roe v. Wade, 410 U.S. 113 (1973), that the Fourteenth Amendment guarantees a woman’s right to choose to undergo a previability abortion. Two decades later, in Planned Parenthood of Southeastern Pennsylvania v. Casey, the Court reaffirmed Roe’s “essential holding” and set forth a three-part legal frame- work for assessing the constitutionality of abortion restrictions: First is a recognition of the right of the woman to choose to have an abortion before viability and to obtain it without un- due interference from the State. Before viability, the State’s interests are not strong enough to support a prohibition of abor- tion or the imposition of a substantial obstacle to the woman’s effective right to elect the procedure. Second is a confirmation of the State’s power to restrict abortions after fetal viability, if the law contains exceptions for pregnancies which endanger the woman’s life or health. And third is the principle that the State has legitimate interests from the outset of the pregnancy in protecting the health of the woman and the life of the fetus that may become a child. 505 U.S. at 846. “Casey, in short, struck a balance.” Gonzales v. Carhart, 550 U.S. 124, 146 (2007). On the one hand, it protected women’s fundamental rights by mandating that “a State may not prohibit any woman from making the ulti- mate decision to terminate her pregnancy before viability.” Casey, 505 U.S. at 879. On the other, it recognized that a state may enact previability regula- tions designed “to further the health or safety of a woman seeking an abor- tion” or “to express profound respect for the life of the unborn.” Id. at 877- 78. But “a statute which, while furthering the interest in potential life or some other valid state interest, has the effect of placing a substantial obstacle in the path of a woman’s choice cannot be considered a permissible means of 77 Case: 17-51060 Document: 00515984856 Page: 78 Date Filed: 08/18/2021 No. 17-51060 serving its legitimate ends.” Id. at 877. Thus, state regulations may have neither “the purpose [n]or [the] effect of placing a substantial obstacle in the path of a woman seeking an abortion of a nonviable fetus.” Id. The “short- hand” for a substantial obstacle is an undue burden. Id. 1. Five years ago, in Whole Woman’s Health v. Hellerstedt, the Supreme Court confirmed that the undue burden “rule announced in Casey . . . re- quires that courts consider the burdens a law imposes on abortion access to- gether with the benefits those laws confer.” 136 S. Ct. at 2309 (citing the Casey Court’s balancing of a law’s benefits against its burdens). That is to say, in order to determine if a burden on a woman’s right to choose is “un- due,” courts must assess the benefits of the state’s regulation relative to the obstacles it erects to women obtaining a previability abortion. Id. A majority of the Court expressly rejected an approach that considers only the burdens imposed by an abortion restriction, stating that this “articulation of the rele- vant standard is incorrect.” Id. And, applying the correct balancing test, the Court reversed this court’s decision upholding a Texas law that, among other things, required abortion providers to obtain admitting privileges at a local hospital. Id. at 2313-14. In light of the district court’s findings that the law had little if any medical benefit and imposed significant obstacles to many women obtaining a previability abortion, the Court held that the law uncon- stitutionally erected a substantial barrier to a large fraction of women exercis- ing their constitutional right to choose. Id. at 2313-15. As noted, the Supreme Court issued its most recent ruling explaining and applying the undue burden test during the pendency of this appeal in June Medical, in which it once again reversed this court’s ruling upholding an abortion restriction. 140 S. Ct. at 2114. June Medical concerned a Louisiana admitting-privileges statute that was virtually identical to the one considered 78 Case: 17-51060 Document: 00515984856 Page: 79 Date Filed: 08/18/2021 No. 17-51060 in Hellerstedt, and the Court unsurprisingly came to the same conclusion, in- validating the law because it imposed an undue burden on a woman’s right to obtain a previability abortion. Id. at 2112-13. A four-Justice plurality applied the balancing approach elucidated in Hellerstedt, weighing the statute’s as- serted benefits against its burdens. See id. at 2121-32. In a solo opinion con- curring in the judgment, Chief Justice Roberts rejected the balancing test, stating that, other than with respect to the preliminary inquiry as to whether the challenged law is rationally related to a legitimate state interest, the undue burden test requires looking only to the burdens of an abortion regulation. See id. at 2136-37 (Roberts, C.J., concurring in the judgment). Citing Marks v. United States, 430 U.S. 188, 193 (1977), the en banc plurality declares today that Chief Justice Roberts’s solo concurrence consti- tutes June Medical’s holding and is accordingly binding on this court. It therefore holds that the district court erred by employing the legal standard set forth in Hellerstedt and balancing the benefits of SB8 relative to the bur- dens it places on a woman’s constitutional right to choose. Plurality at 11-14. For reasons that were discussed at length in the previous panel opinions, the plurality is wrong. See Whole Woman’s Health, 972 F.3d at 652-53; Whole Woman’s Health, 978 F.3d at 904-05. To recapitulate, “[o]rdinarily, ‘[w]hen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Jus- tices, the holding of the Court may be viewed as the position taken by those Members who concurred in the judgment[ ] on the narrowest grounds.’” United States v. Duron-Caldera, 737 F.3d 988, 994 n.4 (5th Cir. 2013) (second alteration in original) (quoting Marks, 430 U.S. at 193)). But we have long held that the Marks “principle . . . is only workable where there is some ‘com- mon denominator upon which all of the justices of the majority can agree.’” Id. (quoting United States v. Eckford, 910 F.2d 216, 219 n.8 (5th Cir. 1990)). When a concurrence does not share a “common denominator” with, or 79 Case: 17-51060 Document: 00515984856 Page: 80 Date Filed: 08/18/2021 No. 17-51060 cannot “be viewed as a logical subset of,” a plurality’s opinion, it “does not provide a controlling rule” that establishes or overrules precedent. Id. In June Medical, the only common denominator between the plurality and the concurrence is their shared conclusion that the challenged Louisiana law constituted an undue burden. Compare 140 S. Ct. at 2132 (plurality opin- ion), with id. at 2141-42 (Roberts, C.J., concurring in the judgment). What they obviously disagreed on is the proper test for conducting the undue-bur- den analysis: the June Medical plurality applied Hellerstedt’s balancing of the law’s burdens against its benefits, while the concurrence analyzed only the burdens. In fact, the Chief Justice expressly disavowed the plurality’s test. See id. at 2136. Our precedents make clear that a concurrence is not a logical subset of a plurality opinion or vice versa in these circumstances. See Duron- Caldera, 737 F.3d at 994 n.4 (holding that, in the Supreme Court’s decision in “Williams[ v. Illinois, 132 S. Ct. 2221 (2012)], there is no such common denominator between the plurality opinion and Justice Thomas’s concurring opinion. Neither of these opinions can be viewed as a logical subset of the other. Rather, Justice Thomas expressly disavows what he views as ‘the plu- rality’s flawed analysis,’ including the plurality’s ‘new primary purpose test.’” (quoting Williams, 132 S. Ct. at 2255, 2262 (Thomas, J., concurring) (emphasis added))). Basic logic reaffirms that a rule that asks simply whether a given factor is present in sufficient quantities is not a logical subset of a rule that calls for that factor to be weighed against another variable. Consider this counterfac- tual: If the June Medical plurality’s rule were “Unconstitutional if A or B is present” and the concurrence’s were “Unconstitutional if A is present,” then the concurrence would be a logical subset of the plurality’s opinion. 4 All 4 This appears to be how the en banc plurality conceptualizes the matter, as it stresses that the Chief Justice agreed with the portion of the June Medical plurality’s 80 Case: 17-51060 Document: 00515984856 Page: 81 Date Filed: 08/18/2021 No. 17-51060 possible unconstitutional outcomes produced by the concurrence’s test would also be unconstitutional outcomes under the plurality’s, and the Venn diagram of results would show the circle representing the concurrence fully contained within the circle representing the plurality. Contrast this to the situation we are now presented with: The June Medical plurality’s rule is “Unconstitutional if A (burdens) is greater than B (benefits)” and the Chief Justice’s concurrence’s standard is “Unconstitutional if A (burdens) is greater than X (an acceptable level).” In situations in which an abortion re- striction has virtually no benefits but imposes only modest burdens, it would be unconstitutional under the June Medical plurality’s test but not the Chief Justice’s. And in situations in which a law has tremendous benefits and im- poses a lesser but nonetheless significant burden, the law would be unconsti- tutional under the Chief Justice’s test but not the plurality’s. The Venn dia- gram is divergent, with neither set of outcomes entirely contained within the other. Both the Seventh and Eleventh Circuits have arrived at the same con- clusion, recognizing that the Chief Justice’s single-justice concurrence is not a logical subset of the June Medical plurality’s opinion. See Reprod. Health Servs. v. Strange, 3 F.4th 1240, 1259 (11th Cir. 2021) (“The Chief Justice’s concurrence cannot fairly be considered narrower than the plurality opinion because, although they came to the same result, the Chief Justice and the plurality diverged on the reasoning supporting that result. As a result, the only common ground between the plurality and Chief Justice Roberts is in the shared conclusion that the Louisiana statute constituted an undue opinion analyzing the burdens imposed by the challenged law. Plurality at 11-12. But the June Medical plurality did not reason that a previability abortion restriction is unconstitu- tional if it has burdens or benefits. It concluded that a previability abortion restriction is unconstitutional if the law’s burdens are greater than its benefits. Simply identifying that a law imposes burdens on the right to abortion is not sufficient to resolve the case under the Hellerstedt formulation that the June Medical plurality applied. 81 Case: 17-51060 Document: 00515984856 Page: 82 Date Filed: 08/18/2021 No. 17-51060 burden. The benefits-burdens approach to the undue burden analysis from Whole Woman’s Health therefore continues to bind us.”); Planned Parenthood of Indiana & Kentucky, Inc. v. Box, 991 F.3d 740, 748 (7th Cir. 2021) (“In June Medical, there is one critical sliver of common ground between the plurality and the concurrence: Whole Woman’s Health was entitled to stare decisis ef- fect on essentially identical facts. The Marks rule therefore applies to that common ground, but it applies only to that common ground.”). The en banc plurality’s approach to applying the Marks rule would have far-reaching consequences, as it would allow “a single Justice writing only for himself . . . the authority to bind th[e] Court to propositions it has already rejected.” Ramos v. Louisiana, 140 S. Ct. 1390, 1402 (2020) (Gor- such, J., plurality opinion). Anytime a fractured opinion arose, any Justice on the court could seize the opportunity to rewrite precedent, regardless of the disagreement of the rest of the Court. Indeed, in Hellerstedt, a majority of the Court explicitly declined to adopt the approach later favored by the Chief Justice in June Medical. Hellerstedt, 136 S. Ct. at 2309. The plurality allows this binding ruling to be disregarded based on the will of a single Jus- tice, which is far from what the Marks court intended when it said that a case’s holding can be ascertained when there is agreement on a dispositive point by a majority of Justices concurring in the judgment. Marks, 430 U.S. at 193. Thus, under our precedents, June Medical did not serve to displace the balancing test called for by Hellerstedt, which remains controlling law. The district court cited and applied the correct legal standard, and the en banc plurality errs by concluding otherwise. 2. Before I proceed to an in-depth discussion of the en banc plurality’s further errors, it is worth noting that this should be an exceptionally easy case 82 Case: 17-51060 Document: 00515984856 Page: 83 Date Filed: 08/18/2021 No. 17-51060 even under the plurality’s preferred legal standard because the Supreme Court has already decided it. In Stenberg v. Carhart, the Supreme Court con- sidered a Nebraska statute that prohibited “deliberately and intentionally de- livering into the vagina a living unborn child, or a substantial portion thereof, for the purpose of performing a procedure that the person performing such procedure knows will kill the unborn child.” 530 U.S. 914, 938 (2000) (quot- ing NEB. REV. STAT. ANN. § 28–326(9)). Nebraska contended that the stat- ute was constitutional because it merely prohibited “dilation and extraction” (“D&X”) abortions, an alternative abortion method that is not implicated in the present case. But the Supreme Court struck down the Nebraska law spe- cifically because the text of the prohibition could reach the same common D&E procedure that SB8 bans. Stenberg, 530 U.S. at 938. Evidence before the trial court makes clear that D & E will often involve a physician pulling a “substantial portion” of a still living fetus, say, an arm or leg, into the vagina prior to the death of the fetus. Indeed D & E involves dismemberment that commonly occurs only when the fetus meets resistance that re- stricts the motion of the fetus: The dismemberment occurs be- tween the traction of the instrument and the counter-traction of the internal os of the cervix. And these events often do not occur until after a portion of a living fetus has been pulled into the vagina. . . . Even if the statute’s basic aim is to ban D & X, its lan- guage makes clear that it also covers a much broader category of procedures. . . . Both procedures can involve the introduc- tion of a “substantial portion” of a still living fetus, through the cervix, into the vagina[.] .... In sum, using this law some present prosecutors and fu- ture Attorneys General may choose to pursue physicians who use D & E procedures, the most commonly used method for 83 Case: 17-51060 Document: 00515984856 Page: 84 Date Filed: 08/18/2021 No. 17-51060 performing previability second trimester abortions. All those who perform abortion procedures using that method must fear prosecution, conviction, and imprisonment. The result is an undue burden upon a woman’s right to make an abortion deci- sion. We must consequently find the statute unconstitutional. Id. at 938-39, 945 (internal citations and alterations omitted). The Supreme Court explicitly stated that banning the performance of a standard D&E, which it repeatedly emphasized involved the evacuation of a living fetus in which fetal demise has not yet been induced, resulted in an undue burden and was therefore constitutionally impermissible. The Court declared the Nebraska law unconstitutional because it could be interpreted to include such a ban. See id. at 945. What the Nebraska statute could be read to extend to, SB8 does directly, targeting and prohibiting the standard D&E procedure. And it is no answer that SB8’s prohibition may be evaded through the various fetal demise techniques the State advocates, for the Ne- braska law, which only applied to procedures involving a “living unborn child,” id. at 922, could have been avoided through the same means. The Supreme Court specifically noted that “[s]ome physicians . . . induce fetal demise prior to a late D & E (after 20 weeks),” id. at 925, but the possibility was immaterial to the Supreme Court’s decision, which should dictate ours. If the Nebraska law was unconstitutional, it necessarily follows that SB8 is as well. Sixteen years prior to Hellerstedt, employing the legal standard that the en banc plurality contends June Medical restored, the Supreme Court held that prohibiting a standard D&E imposed an undue burden on a woman’s 84 Case: 17-51060 Document: 00515984856 Page: 85 Date Filed: 08/18/2021 No. 17-51060 constitutional right to abortion. This is precisely what SB8 does, and that should be the end of this case. 5 3. Nonetheless, the plurality ignores the binding Stenberg precedent that should mandate the resolution of this case under any legal standard, and in- stead contends that the district court committed a host of additional legal er- rors unrelated to the Marks question. None of these assertions withstand even a cursory examination. The plurality argues that the district court com- mitted what it characterizes as legal errors by failing to credit a number of the State’s valid interests in enacting SB8. The district court did not fail to do so, but merely determined either that SB8 failed to advance those interests or that they were not sufficient to outweigh the burdens that SB8 imposes on access to previability abortion. 5 The plurality mischaracterizes Stenberg’s holding as resting “primarily” on the Nebraska law’s lack of a health exception. Plurality at 25 n.18. However, the Court said explicitly in Stenberg that the Nebraska law was unconstitutional “for at least two independent reasons.” 530 U.S. at 930 (emphasis added). The first independent reason was the lack of a health exception. Id. But the Court’s second independent reason was that the law “‘impos[ed] an undue burden on a woman’s ability’ to choose a D & E abortion, thereby unduly burdening the right to choose abortion itself.” Id. (quoting Casey, 505 U.S. at 874). This was so, the Court explained, because D&E was “the most commonly used method for performing previability second trimester abortions.” Id. at 945. Thus, Stenberg is clear that if a state law unduly burdens “the most commonly used method for performing previability second trimester abortions” than “[t]he result is an undue burden upon a woman’s right to make an abortion decision.” Id. at 945-46. That is the precise situation that we are presented with in evaluating SB8, because, as a factual matter, D&E remains the most commonly used method for performing previability second trimester abortions. The conclusion that SB8 is unconstitutional is not based on a comparison to “abortion standards from the last century” and does not require one to “disavow” medical progress. See Plurality at 25 n.18. Rather, this conclusion is based on a straight-forward application of the Supreme Court’s precedential Stenberg holding to the facts of the present case. 530 U.S. at 945-46. 85 Case: 17-51060 Document: 00515984856 Page: 86 Date Filed: 08/18/2021 No. 17-51060 The plurality claims that the district court erred by treating the State’s interest in preserving fetal life as “only [a] marginal consideration” that has “its primary application once the fetus is capable of living outside the womb.” Plurality at 14; see Whole Woman’s Health, 280 F. Supp. 3d at 953. Similarly, it faults the district court for stating that a woman’s right to a previ- ability abortion is “absolute.” Plurality at 17-18; see Whole Woman’s Health, 280 F. Supp. 3d at 953. As a threshold matter, these offhand and isolated statements are gleaned from the conclusion of the district court’s memoran- dum opinion, and there is no sign that the district court materially relied upon them in its substantive reasoning. Moreover, the district court said little more than the Supreme Court stated in Casey, in a passage that the plurality conveniently omits from its description of that case’s holding: “Before via- bility, the State’s interests are not strong enough to support a prohibition of abortion or the imposition of a substantial obstacle to the woman’s effective right to elect the procedure.” 505 U.S. at 846; see Plurality at 8. In other words, a woman’s right to a previability abortion is absolute in the sense that a state’s interests are never enough to justify its placing an undue burden on her exercise of that right. Casey, 505 U.S. at 846. Thus, the district court was correct that, prior to viability, the State’s interest in protecting fetal life is necessarily outweighed by a woman’s right to obtain an abortion free from any substantial obstacle imposed by state regulation. As the district court stated, “The State’s valid interest in promoting respect for the life of the un- born, although legitimate, is not sufficient to justify such a substantial obsta- cle to the constitutionally protected right of a woman to terminate a preg- nancy before fetal viability.” 280 F. Supp. 3d at 953. The plurality also contends that the district court failed to credit sev- eral additional interests the State asserted SB8 serves, including the physical and psychological benefits to a woman’s health that result from inducing fetal demise prior to evacuation, the provision of dignity in death to fetuses 86 Case: 17-51060 Document: 00515984856 Page: 87 Date Filed: 08/18/2021 No. 17-51060 immediately prior to the second phase of the D&E procedure, the promotion of societal and medical ethics, and ensuring women give informed consent to abortions. Plurality at 14-17. But that the State contended that SB8 pro- moted these interests does not mean the district court was required to find that it was so. As I have stated, the district court conducted an extensive hearing and reviewed voluminous evidence to make its factual findings. The plurality cherry picks evidence in the record stating that some women feel better knowing that fetal demise occurred prior to the evacuation phase of a D&E, but the district court also heard evidence of the painful, invasive, and risky techniques that must be used to induce fetal demise. The district court ulti- mately concluded that, on balance, “pre-evacuation fetal demise provides no additional medical benefit to a woman undergoing a standard D & E abor- tion.” 6 Whole Woman’s Health, 280 F. Supp. 3d at 948. As has been stated 6 The plurality chides the district court at length for relying on the decisions of other courts considering similar laws, including the well-reasoned opinion in West Alabama Women’s Center v. Miller, in which a district court struck down virtually identical legislation to SB8. 299 F. Supp. 3d 1244, 1268 (M.D. Ala. 2017), aff’d sub nom. W. Alabama Women's Ctr. v. Williamson, 900 F.3d 1310 (11th Cir. 2018). The plurality states the district court “should have relied on the voluminous and comprehensive record before it, not other courts’ opinions with materially different record.” Plurality at 21 n.15. But the plurality fails to heed its own advice and extensively relies on off-hand statements in Stenberg v. Car- hart, 530 U.S. 914, 925 (2000), and Gonzales v. Carhart, 550 U.S. 124, 136 (2007), to claim that independently inducing fetal demise is both widely practiced and can potentially make a D&E easier in various ways. Plurality at 24-26. Needless to say, the prevalence and rel- ative advantages of various methods of conducting a complex modern medical procedure are not the type of widely known and uncontroversial facts of which we may take judicial notice, let alone from sources more than a decade-and-a-half old. See Fed. R. Evid. 201(b). Based on a wealth of scientific literature and expert testimony, the district court in this case found that “pre-evacuation fetal demise provides no additional medical benefit to a woman undergoing a standard D & E abortion,” Whole Woman’s Health v. Paxton, 280 F. Supp. 3d 938, 948 (W.D. Tex. 2017), and the plurality points to no evidence in the record of this case compelling enough to make that finding clearly erroneous. 87 Case: 17-51060 Document: 00515984856 Page: 88 Date Filed: 08/18/2021 No. 17-51060 repeatedly, weighing conflicting evidence, judging credibility, and making factual determinations about the effects of a medical procedure are the prov- ince of a district court, and the evidence on this point contained in the record is far from so one-sided as to render the district court’s determination im- plausible. Similarly, the district court found that, unlike any other medical regu- lation, SB8 “requires a doctor—in contravention of the doctor’s medical judgment and the best interest of the patient—to conduct a medical proce- dure that delivers no benefit to the woman.” Id. at 953. It further found that, in some cases, the techniques for inducing fetal demise advocated by the State were experimental and without clear evidence as to their safety or effi- cacy. Id. at 949. As will be discussed in more detail below, these findings are supported by the record, and thus the district court committed no error in finding that, on balance, SB8 is inconsistent with principles of medical ethics and did not further any state interest in protecting the integrity of the profes- sion. The plurality also mischaracterizes the district court’s consideration of the State’s interest in protecting the dignity of fetuses. It contends that the district court stated that “the State’s interest ‘does not add weight to tip the balance in the State’s favor.’” Plurality at 16. But the district court’s memorandum opinion says just the opposite: “The evidence before the court is graphic and distasteful. But this evidence is germane only to the State’s interest in the dignity of fetal life and is weighed on the State’s side of the scale.” Whole Woman’s Health, 280 F. Supp. 3d at 947. The district court merely reasoned that “[a]n abortion always results in the death of a fetus,” “[t]he extraction of the fetus from the womb occurs in every abortion,” and “[d]ismemberment of the fetus is the inevitable result.” Id. Thus, the court found that any increase in fetal dignity afforded by SB8 over the standard 88 Case: 17-51060 Document: 00515984856 Page: 89 Date Filed: 08/18/2021 No. 17-51060 D&E procedure was marginal, and thus it was not enough to “tip the balance in the State’s favor.” Id. The district court did not err by so finding. Finally, the plurality’s reasoning as to how SB8 promotes informed consent is circuitous and puzzling. As the plurality concedes, the law says nothing regarding what information about the procedure abortion providers are required to convey to patients. Plurality at 16. Instead, if I understand the en banc plurality correctly, it is arguing that SB8 furthers the interest of informed consent by bringing the D&E procedure more into line with women’s expectations. In support of this, the plurality posits that “[w]omen who receive live-dismemberment D&Es are not being told what is going to happen to the fetus.” Plurality at 17. First, this is clearly the type of factual finding that appellate courts are ill-suited to make, and it is based on little more than the plurality’s supposi- tion. The plurality cites various abortion-provider consent forms that were introduced into evidence, and it seems to contend that the forms are mislead- ing because they describe the procedure in accurate, clinical terms rather than containing a graphic and disparaging description that condemns the pro- cedure as barbaric. Plurality at 17. And the plurality points to no evidence in the record regarding what further details are conveyed to patients orally and through other materials, and the district court made findings on neither this point nor what women generally believe occurs during an abortion prior to receiving information on the procedure. Moreover, it is unclear that SB8 would in fact bring the abortion pro- cedure more in line with patients’ expectations even accepting the plurality’s contentions. The district court found that the women whom SB8 affected would be “in a unique position” because no other “medical context” re- quires a doctor to perform an unnecessary procedure that the doctor believes is contrary to the patient’s best interests. Whole Woman’s Health, 280 F. 89 Case: 17-51060 Document: 00515984856 Page: 90 Date Filed: 08/18/2021 No. 17-51060 Supp. 3d at 953. A reasonable inference from this finding is that, even if women do not expect fetal demise to specifically occur as a result of the evac- uation procedure, they likewise do not expect a doctor to perform an extra step that the doctor considers unnecessary and liable to expose the patient to additional risk without any reciprocal medical benefit. Further, SB8 imposes certain procedures regardless of the choice reached by a woman through dis- cussion and consultation with her physician—hardly a situation that respects a patient’s informed consent. Thus, the district court did not err in finding that SB8 does not on balance promote informed consent, and therefore this interest does not add any weight to the benefit side of the equation. 4. The en banc plurality next misrepresents the district court’s analysis to claim that the district court placed the burden of proof on the wrong party. Plurality at 18. The plurality’s entire basis for this contention is that, in hold- ing SB8 unconstitutional, the district court relied in part on the fact that a standard D&E without the separate step of inducing fetal demise is the most commonly used method of surgical abortion in Texas and nationally. Accord- ing to the plurality, the district court was permitting abortion providers to “set their own rules” and “self-legislate or self-regulate simply by making an abortion method ‘common.’” Plurality at 18. But this totally misconstrues the district court’s reasoning, which merely considered what proportion of abortions would be affected by SB8 in evaluating the burden the legislation places on a woman’s right to choose. As the district court explicitly noted, the Supreme Court has em- ployed this exact analysis in landmark abortion-rights cases. See Whole Woman’s Health, 280 F. Supp. 3d at 945 (citing Stenberg, 530 U.S. at 939). In Stenberg, the Court struck down a Nebraska abortion restriction specifically because it could be interpreted to prohibit “the most commonly used method 90 Case: 17-51060 Document: 00515984856 Page: 91 Date Filed: 08/18/2021 No. 17-51060 for performing previability second trimester abortions.” 530 U.S. at 945. Similarly, in Planned Parenthood of Central Missouri v. Danforth, the Court de- clared unconstitutional a Missouri ban on saline amniocentesis because, at the time, it was the “most commonly used” method of abortion “nationally by physicians after the first trimester.” 428 U.S. 52, 78 (1976); see also Gon- zales , 550 U.S. at 153, 165 (holding that the federal “Partial-Birth Abortion Act,” 18 U.S.C. § 1531, which banned the D&X procedure, did “not con- struct a substantial obstacle to the abortion right,” because the D&E proce- dure—the “most commonly used and generally accepted method” of second trimester abortions—remained available). The district court did not err by considering the ubiquity and general acceptance of D&E within the medical community in determining the degree of burden SB8 imposes on women’s constitutional right to obtain a previability abortion. The plurality also contends that the district court committed legal er- ror by incorrectly defining “substantial obstacle,” focusing on the district court’s statement that the term means “no more and no less than ‘of sub- stance.’” Plurality at 18-19; see Whole Woman’s Health, 280 F. Supp. 3d at 944. But the district court’s incidental statement was part of its larger dis- cussion of the undue burden standard set forth in Hellerstedt. The sentences immediately preceding the excerpt on which the plurality wrongly focuses stated the correct standard in no uncertain terms: “Whether an obstacle is substantial—and a burden is therefore undue—must be judged in relation to the benefits that the law provides. Where a law’s burdens exceed its benefits, those burdens are by definition undue, and the obstacles they embody are by definition substantial.” Whole Woman’s Health, 280 F. Supp. 3d at 944 (cit- ing Hellerstedt, 136 S. Ct. at 2300, 2309-10, 2312, 2318). The district court determined that SB8 erected a substantial obstacle because any benefits from the law were significantly outweighed by the burdens it places on the consti- tutional right to a previability abortion, and there is no indication that the 91 Case: 17-51060 Document: 00515984856 Page: 92 Date Filed: 08/18/2021 No. 17-51060 incidental statement the plurality disfavors affected that analysis in the least. The plurality is thus also wrong to claim this was legal error. B. I now turn to the true gravamen of the plurality’s dispute with the dis- trict court: The plurality disagrees with the district court’s reading of the ev- idence regarding the burdens attendant to the various fetal-demise tech- niques that the State claims can be used to evade SB8’s prohibition. The plurality focuses on two potential fetal-demise methods, suction aspiration and digoxin injection, 7 and concludes that the district court clearly erred by finding that the techniques are not “safe, effective, and common- place.” Plurality at 22. Notably, we—an appellate court that generally should not make factual findings—seem to be the only federal court that has ever found that safe and effective means of complying with this sort of fetal- demise mandate exist, and at least two of our sister circuits have affirmed district courts that found that the methods being considered here are not safe or effective. See EMW Women’s Surgical Ctr. P.S.C., 960 F.3d at 807-08; W. Ala. Women’s Ctr., 900 F.3d at 1324-28; see also Glossip v. Gross, 576 U.S. 863, 882 (2015) (“Our review is even more deferential where, as here, multiple trial courts have reached the same finding, and multiple appellate courts have affirmed those findings.”). At the risk of belaboring the point, the plurality repeats the errors of the past and does what a majority of the Supreme Court in June Medical, including Chief Justice Roberts, clearly told us not to do: Substitute our view of conflicting evidence for that of the district court and 7 The State also offered potassium-chloride injections and umbilical-cord transec- tion as possible methods of complying with SB8. Because the plurality does not rely on these possibilities, this dissent will not examine them at length. Suffice it to say, for the reasons found by the district court and discussed in the panel opinion, these options are even riskier and less feasible than the procedures the plurality contends are viable here. 92 Case: 17-51060 Document: 00515984856 Page: 93 Date Filed: 08/18/2021 No. 17-51060 displace its well-considered factual findings simply because we do not like the outcome. See June Med. Servs., L.L.C., 140 S. Ct. at 2121, 2124-25 (plurality opinion); id. at 2133-34, 2141 (Roberts, C.J., concurring). As I said at the outset, there is a “fundamental flaw” in the plurality’s description of these fetal-demise procedures as “alternatives.” EMW Women’s Surgical Ctr. P.S.C., 960 F.3d at 798. Instead, they are, “by defini- tion, . . . additional procedures,” and “[a]dditional procedures, by nature, expose patients to additional risks and burdens.” Id.; see also, e.g., W. Ala. Women’s Ctr., 900 F.3d at 1326 (noting the State’s concession that fetal de- mise procedures “would always impose some increased health risks on women”). This fact alone—that an abortion restriction would require a woman to undergo a riskier procedure in order to procure an abortion—has been sufficient in other cases for the Supreme Court to conclude that the law was unconstitutional. See Danforth, 428 U.S. at 78-79 (invalidating an abor- tion restriction that “force[d] a woman and her physician to terminate her pregnancy by methods more dangerous to her health than the method out- lawed”); Gonzales, 550 U.S. at 161 (“The prohibition in the Act would be unconstitutional, under precedents we here assume to be controlling, if it ‘subject[ed] [women] to significant health risks.’” (alterations in original) (quoting Ayotte v. Planned Parenthood of N. New England, 546 U.S. 320, 327 (2006))); see also Planned Parenthood of Cent. N.J. v. Verniero, 41 F. Supp. 2d 478, 500 (D.N.J. 1998) (“By relegating physicians to the performance of more risk-laden abortion procedures, the Act imposes an undue burden on the woman’s constitutional right to terminate her pregnancy.”), aff’d sub nom. Planned Parenthood of Cent. N.J. v. Farmer, 220 F.3d 127 (3d Cir. 2000). Moreover, even if the mere existence of increased risk without any re- ciprocal medical benefit were not sufficient to invalidate SB8, the plurality errs here by doing its own fact finding and second-guessing the district 93 Case: 17-51060 Document: 00515984856 Page: 94 Date Filed: 08/18/2021 No. 17-51060 court’s assessment of the efficacy and degree of risk associated with these two techniques. I will consider each of them in turn. 1. The plurality first contends that inducing fetal demise through suction aspiration is a viable method of complying with SB8. Plurality at 22-24. The plurality acknowledges that the district court did not make any factual find- ings on the feasibility of using suction to induce fetal demise after fifteen weeks’ gestation. Plurality at 24. But the plurality takes it upon itself to make the factual findings on this point that the district court did not, boldly declar- ing that the evidence in the voluminous record is so one-sided as to permit only one conclusion. Plurality at 37. As a threshold matter, neither the State nor the plurality contends that fetal demise can be induced by suction alone at or after seventeen weeks’ ges- tation. See Plurality at 24. Texas bans most abortions outright after twenty- two weeks, see TEX. HEALTH & SAFETY CODE § 171.044, and even under the plurality’s overly generous view of the evidence, suction would be a feasible method of inducing fetal demise for only two of the seven weeks during which most Texas D&Es are performed using forceps. Less than half of the fifteen- to-twenty-two-week abortions conducted in Texas in 2015 fell into this two- week period, and thus even the plurality admits that using only suction aspi- ration to induce fetal demise would not be feasible in the majority of abortions under consideration here. Plurality at 24. Moreover, as even one of the judges concurring in the judgment rec- ognizes, the actual evidence that suction alone can be used safely and effec- tively after fifteen week’s gestation is equivocal at best. The en banc plurality makes much of the fact that several witnesses stated that they had at points performed suction aspiration abortions after the fifteen-week point and could perhaps do so regularly if they were forced to under penalty of law. Plurality 94 Case: 17-51060 Document: 00515984856 Page: 95 Date Filed: 08/18/2021 No. 17-51060 at 23. But those same experts testified that it is unpredictable whether suc- tion will in fact work at this later stage and “suction alone is often not suffi- cient to complete the procedure.” One doctor explained that he generally does not know prior to beginning the D&E whether forceps will be required, and he has needed to resort to them in some instances as early as ten weeks’ gestation. Another doctor explained that using suction becomes difficult as a practical matter around the fifteen-week point because the larger suction cannula that is needed to perform the procedure at that stage is unwieldy, which likely leads to an increased risk of injury to the patient undergoing the procedure. Additionally, several doctors testified that some women’s uter- ine anatomy may make the use of a suction cannula difficult or impossible, exposing those women to a heightened risk of injury if the doctor is forced to use only suction aspiration. In sum, the record at most suggests that, if forced, doctors might be able to employ suction aspiration alone to induce fetal demise in some cases after the fifteen-week mark, though doing so would often be contrary to their preferences, medical judgment, and the patients’ best interests. This evi- dence is hardly so compelling as to allow only one possible reading regarding the feasibility of using suction alone to induce fetal demise after fifteen weeks’ gestation, and the plurality errs by abandoning its proper role as an appellate court in order to make a factual finding on the matter. See Swint, 456 U.S. at 292 (“[F]actfinding is the basic responsibility of district courts, rather than appellate courts, and . . . the Court of Appeals should not have resolved in the first instance this factual dispute which had not been consid- ered by the District Court.” (quoting DeMarco v. United States, 415 U.S. 449, 450 n.* (1974))). 95 Case: 17-51060 Document: 00515984856 Page: 96 Date Filed: 08/18/2021 No. 17-51060 2. The en banc plurality also argues that abortion providers can safely and effectively cause in utero fetal demise prior to the evacuation phase of a D&E—and thereby avoid criminal sanctions for violating SB8—by injecting the chemical digoxin into the fetus or amniotic fluid. The plurality relies heavily on the fact that some abortion providers have a policy of using digoxin to induce fetal demise when performing later- term abortions, typically after eighteen or twenty weeks’ gestation at the ear- liest. Plurality at 26. In its medical wisdom, the plurality seems to say that what is good for the goose is good for the gander; what is fit for some later- term D&Es must be suitable for all forceps-assisted D&Es at all gestational stages. But this point is not the coup de grâce the plurality believes. First, the plurality greatly overstates the prevalence of the technique; only two out of the twenty-one clinics in Texas have a policy of using digoxin for their later-term abortions, and in 2015, the injections were employed in less than 200 of the 3,150 Texas abortions that were performed after fifteen weeks’ gestation. Moreover, as the plurality fully acknowledges, the abortion providers that do use digoxin injections when performing later-term abor- tions frequently do so in order to fully ensure compliance with 18 U.S.C. § 1531, the federal law that prohibits performing a D&X abortion, which is a procedure that can be done inadvertently during a standard D&E if a woman’s cervix dilates more than anticipated. Plurality at 32-33; see Gonza- les, 550 U.S. at 154. The plurality posits that, “[s]urely, no reasonable abor- tion provider would subject women to ‘significant’ health risks from digoxin just to avoid their own federal liability,” Plurality at 33, but this underscores precisely why SB8 would impose such a large burden on women’s abortion access. As doctors testified to repeatedly in this case, many abortion provid- ers do not perform these later-term abortions, and more would stop providing 96 Case: 17-51060 Document: 00515984856 Page: 97 Date Filed: 08/18/2021 No. 17-51060 earlier term abortions if they were forced to include an additional risky, inva- sive, painful, and medically unnecessary step in the procedure that is contrary to their medical judgment regarding the patient’s best interest. And digoxin injections are all those things. The method requires a physician to insert a 3- to 4-inch surgical “spinal” needle either transabdom- inally (through the woman’s abdomen) or transvaginally (through the vaginal wall or the cervix). Obviously, the injection is invasive and painful, and it often requires the patient to receive an additional numbing injection or to un- dergo intravenous sedation. The plurality makes little effort to claim other- wise, stating only that anesthetic is available and, according to the reassur- ances one abortion provider gives in an effort to settle patients’ fears, the pain will fade quickly. Plurality at 27. As should be apparent, the administering of additional anesthetic and sedation can itself be quite painful and invasive, and it inherently imposes additional health risks. But the burdens imposed by digoxin injections are not limited to those associated with the immediate discomfort of the procedure. There was mul- titudinous evidence and expert testimony at trial that digoxin injections carry significant health risks as compared to a D&E procedure performed without the injections, including a heightened risk of infection, bleeding, tachycardia, nausea, vomiting, dizziness, fainting, and even extramural delivery—the un- expected and spontaneous expulsion of the fetus from the uterus while the woman is outside of a clinical setting and without the aid of a medical profes- sional. One study from the record found that the risk of hospitalization is six times greater when digoxin injections are administered than when a standard D&E is performed without the injections. One doctor testified that he had discontinued a policy of administering digoxin like the ones the en banc plu- rality cites specifically because of concern over the health risks associated with the unnecessary and nonbeneficial procedure. Digoxin injections are also contraindicated or outright impossible to administer to patients with 97 Case: 17-51060 Document: 00515984856 Page: 98 Date Filed: 08/18/2021 No. 17-51060 certain conditions who collectively account for a large portion of the popula- tion, including those with heart conditions, certain fetal or uterine anatomy, and even obesity and fibroids. The plurality dismisses these latter concerns by stating that a D&E itself might be unsafe for women with these conditions, citing a passage from an abortion provider’s documentation that states these are “special conditions requiring special evaluation and management.” Plu- rality at 27 n.19. But that special measures may need to be employed to per- form a D&E safely on women with these conditions has no bearing on whether this procedure—digoxin injections—can ever be safely administered to these women. The plurality reverses the district court’s factual finding that the use of digoxin carries “significant health risk” not simply on the basis of conflicting evidence, which we are not supposed to do, but against the overwhelming weight of the evidence in the record. Digoxin, moreover, fails to actually induce fetal demise about 5-10% of the time, with its effectiveness dependent on variables such as uterine anat- omy and fetal positioning. The plurality hand waves away this fact—that as many as one-in-ten digoxin injections expose a woman to a painful, invasive, and risky technique without even accomplishing the central goal of the pro- cedure. Plurality at 27-28. Doctors can simply try again with a second injec- tion, the Plurality states. But the plurality fails to acknowledge that, in addi- tion to the pain and health risks normally associated with an initial digoxin injection, there is no documented testing regarding the efficacy and safety of administering a second injection; in fact, every abortion provider who testi- fied in this case stated that, because of this unknown risk, they do not employ a second digoxin injection if the first one fails. Instead, they currently simply proceed with a standard D&E when a digoxin injection does not induce fetal demise, a fallback measure that SB8 criminalizes. In a sizable number of cases, then, SB8 would require doctors to undertake a wholly experimental 98 Case: 17-51060 Document: 00515984856 Page: 99 Date Filed: 08/18/2021 No. 17-51060 second digoxin injection that they presently deem too dangerous to adminis- ter. Similarly, as the district court observed, nearly every study in the rec- ord concerning the safety and efficacy of digoxin injections included only pregnancies at or after eighteen weeks’ gestation, with only a few studies in- cluding cases at seventeen weeks. No study considered the efficacy, dosage, or safety of injecting digoxin into women before seventeen weeks’ gestation. Indeed, that no abortion providers administer digoxin prior to eighteen weeks’ gestation is a testament to how risky and untested the procedure is at these earlier stages, for many of the same § 1531-compliance concerns that exist after eighteen weeks’ gestation also exist before that point. In light of the lack of evidence regarding its safety, the district court found that requir- ing digoxin injections before eighteen weeks of pregnancy would subject women to an arguably experimental procedure without any counterbalancing medical benefit. The plurality calls these well-supported findings error, relying on statements in Gonzales that state legislatures enjoy substantial latitude to reg- ulate abortion where there is scientific uncertainty. Plurality at 30-32 (citing 550 U.S. at 161-62, 166-67). First, the plurality once again repeats the mis- takes of the past. In Hellerstedt, the Supreme Court rebuked our court for relying on these same passages from Gonzales to declare that “medical un- certainty underlying a statute is for resolution by legislatures, not the courts.” 136 S. Ct. at 2309 (quoting Whole Woman’s Health v. Cole, 790 F.3d 563, 587 (5th Cir. 2015)). A majority of the Supreme Court held this “artic- ulation of the relevant standard is incorrect,” and stated clearly that courts should not defer to legislatures or refrain from making findings based on con- flicting scientific evidence when Constitutional rights are at issue: “The statement that legislatures, and not courts, must resolve questions of medical uncertainty is also inconsistent with this Court’s case law,” the Court 99 Case: 17-51060 Document: 00515984856 Page: 100 Date Filed: 08/18/2021 No. 17-51060 explained. Id. at 2310. Such deference wrongly “equate[s] the judicial re- view applicable to the regulation of a constitutionally protected personal lib- erty with the less strict review applicable where, for example, economic leg- islation is at issue.” Id. at 2309. Instead, “[c]ourt[s] retain[] an independent constitutional duty to [make and] review factual findings where constitu- tional rights are at stake.” 8 Id. at 2310 (quoting Gonzales, 550 U.S. at 165). That is exactly what the district court did here, finding that the wholly un- known risks associated with the digoxin protocols the State advocates would be a burden to the health and well-being of women who seek abortions. Moreover, the situation we are presented with is materially different from Gonzales. As the plurality states, the record in Gonzales contained con- flicting expert testimony and other evidence about the safety of D&X abor- tions and available alternatives. Plurality at 30 (citing 550 U.S. at 161). Here, the record contains no evidence about the safety of employing a second di- goxin injection or administering digoxin prior to seventeen weeks’ gestation, and a plethora of evidence stating that doctors currently refrain from doing so because of the risks associated with performing an untested procedure. Indeed, the only evidence the plurality can point to is one expert’s mention of a study that the author has since publicly acknowledged contained incor- rect information regarding the gestation age at which digoxin was used. Plu- rality at 32 & n. 22. Aside from this information that the plurality concedes is erroneous, the plurality offers only one expert’s comment that he knows of doctors who have employed digoxin before eighteen weeks, which is not re- motely probative of whether it is in fact safe and effective to do so. 8 For this same reason, the plurality and concurrence’s contentions regarding the possibility that SB8 guards against fetal pain are misplaced. The majority of the scientific literature and expert testimony in the record indicates that fetal pain is not possible at these early stages of development, and the district court was well within its rights to so find. 100 Case: 17-51060 Document: 00515984856 Page: 101 Date Filed: 08/18/2021 No. 17-51060 The plurality further faults the district court’s well-supported finding that employing a digoxin injection generally adds a day to what is typically a one-day procedure, again doing its own fact finding to imagine various sce- narios where this might not be the case. Plurality at 28-29. It further states that it was error for the district court to even consider this matter as a burden on abortion access, citing the Supreme Court’s upholding of mandatory wait- ing periods and other regulations that by their nature cause a delay in a woman receiving an abortion. Plurality at 29 (citing Casey, 505 U.S. at 885- 86). But this misses the point. The delay occasioned by requiring digoxin injections is only one of the burdens on abortion access that would result from requiring the technique. The district court did not err by holding that, when this delay is considered together with the increased travel and other financial costs it leads to, and in conjunction with the painful and invasive nature of the procedure, its significant health risks, and its lack of consistent efficacy, the burdens add up to a substantial obstacle—particularly when viewed in light of the total absence of medical benefit associated with the technique. Based on the pain and invasiveness of the procedure, the delay in care and logistical difficulties it necessitates, its unreliability, the unknown risks for women before eighteen weeks’ gestation, and the known heightened risk of complication in all instances, the district court found that digoxin is not a safe and viable method of inducing fetal demise before the evacuation phase of a D&E abortion. These findings, along with those regarding the unfeasi- bility of other methods of inducing fetal demise, are all very well supported by the record, and the plurality errs by substituting its own findings for those of the district court. See Anderson, 470 U.S. at 574 (stating that, as an appel- late court, even if we disagree with the findings below, we cannot reverse them so long as they are based on a “permissible view[ ] of the evidence”). 101 Case: 17-51060 Document: 00515984856 Page: 102 Date Filed: 08/18/2021 No. 17-51060 C. Lastly, the plurality contests the district court’s large-fraction analy- sis, variously claiming that the court “bungl[ed]” and “botched” the evalu- ation. Plurality at 3, 33. But the plurality’s analysis is itself riddled with er- rors and predicated on its unsupported assumptions and faulty factual find- ings. The Supreme Court held in Casey that an abortion regulation is fa- cially unconstitutional if “it will operate as a substantial obstacle to a woman’s choice to undergo an abortion” in “a large fraction of the cases in which [it] is relevant.” 505 U.S. at 895. The Court reaffirmed and clarified that standard in Hellerstedt, in which it held that the phrase refers to a large fraction of “those women for whom the provision is an actual rather than an irrelevant restriction.” 136 S. Ct. at 2320 (alterations omitted) (quoting Ca- sey, 505 U.S. at 895). That category is narrower “than all women, pregnant women, or even women seeking abortions[.]” Id. (internal quotation marks omitted). Thus, the appropriate denominator—the number of women for whom SB8 is an actual rather than irrelevant restriction—is not “all women with fetuses in the gestational age of 15-22 weeks” as the plurality claims. Plural- ity at 34. For a great many of those women, SB8 is a totally “irrelevant re- striction.” Id. The vast majority are not seeking abortions, 9 and of that small 9 The plurality seems to acknowledge that these women are unaffected by SB8, stating in a footnote that “SB8 affects only abortions between 15 and 22 weeks[.]” Plurality at 33 n.23 (emphasis added). The plurality also misrepresents the district court’s finding on this matter, which was not that SB8 affected “only women with fetuses at the gestational age of 15-20 weeks.” Plurality at 33. Rather, the district court found that “the class of women here consists of all women in Texas who are 15 to 20 weeks pregnant and seek an outpatient second-trimester D & E abortion.” Whole Woman’s Health, 280 F. Supp. 3d at 952 (emphasis added). At most, the district court’s formulation was slightly underinclusive in that some D&E abortions performed prior to fifteen weeks’ and from twenty to twenty-two 102 Case: 17-51060 Document: 00515984856 Page: 103 Date Filed: 08/18/2021 No. 17-51060 portion who are, some would undergo a procedure that would comply with SB8 in any event, even were the law not in effect. That is because, as the plurality relies on elsewhere in its opinion, some D&Es are currently per- formed through suction aspiration alone after fifteen weeks’ gestation and some D&Es are currently performed with digoxin injections after eighteen weeks’ gestation. Instead, the appropriate denominator is the class of women actually affected by SB8, which is composed of only those women who would undergo a forceps-assisted D&E in Texas without their doctors’ first inducing fetal demise in the absence of SB8. These are the only women “for whom the provision is an actual . . . restriction,” because these are the only women for whom SB8 mandates a change in the procedure they would otherwise un- dergo. Cf. Jackson Women’s Health Org, 945 F.3d at 276 (“The only women to whom the Act is an actual restriction, then, are those who seek abortions before 20 weeks; the Act is redundant of existing Mississippi law as to all abortions after that point.”). The plurality errs by defining the class far more broadly to include many women whose lives will never be the least bit im- pacted by SB8, regardless of whether the law goes into effect. The question, then, becomes what portion of the women who would otherwise receive SB8-noncompliant abortions are unduly burdened by the statute. As I have stated, under controlling precedent, this is the number of those women for whom the burdens SB8 imposes outweigh any benefits re- sulting from the law. Hellerstedt, 136 S. Ct. at 2300, 2309-10, 2312, 2318. In light of the district court’s well-supported findings that the fetal demise weeks’ gestation do not already comply with SB8 and slightly overinclusive in that some D&E abortions performed from fifteen to twenty weeks’ gestation are already SB8 compli- ant. But the district court’s analysis was far closer to the mark than the plurality’s, which includes a majority of women for whom SB8 is a wholly irrelevant restriction. 103 Case: 17-51060 Document: 00515984856 Page: 104 Date Filed: 08/18/2021 No. 17-51060 measures that these women must undergo to comply with SB8 expose the woman to risky, painful, invasive, and untested procedures; have no medical benefit for the patient; cause needless delay and increased financial costs; and only marginally advance the State’s interest in protecting and providing dig- nity to potential human life, it is clear that SB8’s burdens far outweigh its benefits in every such case. Cf. Jackson Women’s Health Org., 945 F.3d at 276 (“Here, the Act is invalid as applied to every Mississippi woman seeking an abortion for whom the Act is an actual restriction, never mind a large frac- tion of them.”). The plurality declares that the district court erred by concluding the fraction of women for whom SB8 is both an actual restriction and an undue burden was 1/1. Plurality at 33. But when the matter is properly framed and the district court’s findings are given the appropriate deference, no other conclusion is possible. *** The plurality concludes its opinion by relisting the litany of mistakes it wrongfully attributes to the district court, none of which in fact occurred. Plurality at 37. And, underscoring its abdication of the role of an appellate court, it declines to order the remedy that would be proper if the district court had in fact misstepped in its analysis, which would be a remand for reconsid- eration in light of our clarification of the pertinent legal principles. The plu- rality instead renders judgment, stating that its view of the evidence is the only possible logical conclusion. For the above reasons, the plurality is wrong. The tendency of our court to eschew settled legal principles when abortion is involved has been documented and discussed elsewhere, see, e.g., June Medical Services, L.L.C., 905 F.3d at 834-35 (Higginbotham, J. dissent- ing) (“[W]hen abortion shows up, application of the rules of law grows 104 Case: 17-51060 Document: 00515984856 Page: 105 Date Filed: 08/18/2021 No. 17-51060 opaque, a phenomenon not unique to this court.” (footnote omitted)), and there seems little need to reprise that debate now, for the matter speaks for itself. The district court’s legal rulings were correct, and its factual findings were not clearly erroneous. And while the plurality laments the amount of time SB8 has been enjoined, that is time in which women in Texas were shielded from the ill effects of a law that is clearly unconstitutional in light of Stenberg, 530 U.S. at 938-39, and the great burdens the statute places on abor- tion access with exceedingly few reciprocal benefits. That the shield is with- drawn today and that women in Texas will be forced to undergo invasive and unsafe techniques to exercise their constitutional right to an abortion—if it does not prevent their exercising that right altogether—is a devastating blow to their self-determination. I hope only that this opinion gives voice to a mod- icum of their frustration, anger, and pain. Once again, I respectfully but em- phatically dissent. 105 Case: 17-51060 Document: 00515984856 Page: 106 Date Filed: 08/18/2021 No. 17-51060 Stephen A. Higginson, Circuit Judge, joined by Costa, Circuit Judge, dissenting: I write separately to make two observations. First, given our court’s plurality opinion’s conclusion that the district court erred because it assessed SB8 under a “balancing test,” without the benefit of Chief Justice Roberts’ “narrower version (only burdens) of the plurality’s test (benefits and burdens),” 1 we should do no more than remand to the district court, confident that it will perform its role finding and applying facts to rules of law we clarify. This leaves us in our lane, not arrogating to ourselves the job district judges perform: above all weighing witness testimony, especially expert witness testimony, elicited by talented, opposing counsel during a week-long trial. Indeed, a circuit decision the plurality cites favorably, Hopkins v. Jegley, 968 F.3d 912, 916 (8th Cir. 2020), did just that—and we should do no more. Second, our court’s plurality opinion’s separate conclusion that the district court erred “under all of the Supreme Court’s precedents”—a contention I think is wrong—should be further reason for us to stay in our lane—i.e., error correction. 2 We should explain our distinguishing 1 This issue currently divides courts. Compare EMW Women’s Surgical Ctr., P.S.C. v. Friedlander, 978 F.3d 418, 433 (6th Cir. 2020) (“[T]he Chief Justice’s position is the narrowest under Marks. His concurrence therefore constitutes June Medical Services’ holding and provides the governing standard here.” (cleaned up)), with Planned Parenthood of Ind. & Ky., Inc. v. Box, 991 F.3d 740, 748 (7th Cir.), petition for cert. filed, No. 20-1375 (Mar. 29, 2021) (“[T]he Marks rule tells us that June Medical did not overrule Whole Woman’s Health.”). 2 I regret our court’s plurality opinion’s characterization of the district judge’s efforts—including that our colleague “disregarded and distorted the record,” “copied and pasted” the facts section of the court’s opinion, “failed to apprehend” the evidence before it, and “botched” a portion of its legal analysis. We weaken what we exaggerate. See Jean-Francois de La Harpe, Mélanie act 1, sc. 1 (1778). 106 Case: 17-51060 Document: 00515984856 Page: 107 Date Filed: 08/18/2021 No. 17-51060 interpretation of Supreme Court doctrine and, with rules of law clarified, return the case to the district court for it to perform its work applying facts assessed at trial to law we have clarified. Stated otherwise, what we should not do is what we recently were admonished not to do in June Medical Services v. Russo, 140 S. Ct. 2103, 2121, (2020): reweigh facts and witness credibility ourselves, here relying selectively on unspecified portions of transcripts from a five-day bench trial where plaintiffs’ expert testimony was heard, and credited or discredited, by a district judge present to observe witness demeanor. 3 See generally Aransas Project v. Shaw, 774 F.3d 324, 325-26, 331 (5th Cir. 2014) (Prado, J., dissenting from denial of rehearing en banc). Regardless of the sensitivity and consequence of any issue that comes to us, our commitment must be to layered judicial responsibilities, where co-equal judges in courts of original jurisdiction adjudicate facts and we do our best not to do so but just to discern error under existing Supreme Court law. It goes without saying that our layer of responsibility starts and finishes with the primacy of the only Court the Framers contemplated, whose rulings we must unerringly follow. Especially when presented with facts we dislike, it can be tempting to arrogate to ourselves the task of constitutional revision, fractionally stepping ahead of, or nudging, the Supreme Court. 4 The imperative, however, is the one followed by this district judge 5 and every 3 Notably, whereas our court’s plurality opinion gives repeated assurances about SB8’s limited impact, as well as assurances about permissible abortion alternatives, the Texas legislature set forth no such findings, indeed, it set forth no legislative findings at all. 4 Cf. Ruth Bader Ginsburg, Four Louisiana Giants in the Law, 48 Loy. L. Rev. 253, 264 (2002) (applauding Judge Rubin’s vision and confidence that “courts need not follow . . . outgrown dogma,” yet acknowledging that only the Supreme Court may revise its interpretation of the Constitution). 5 Cf. Alvin B. Rubin, Views from the Lower Court, 23 UCLA L. Rev. 448, 452 (1976) (“[A]n understandable desire to decide today’s case in accordance with the proclivities of the panel now sitting seems to lead to opinions that fail to accord to prior 107 Case: 17-51060 Document: 00515984856 Page: 108 Date Filed: 08/18/2021 No. 17-51060 court to have considered legislation like SB8. 6 He, they, and we are bound by existing Supreme Court law. Our court’s plurality opinion goes to pains to perceive nuance in Supreme Court precedent twice confirming that bans like SB8 are invalid, 7 as well as to subordinate to one final footnote acknowledgment of contrary circuit law. Suffice it to say, I agree with the consensus of courts that apply Supreme Court law to invalidate similar bans, leaving several judges on those courts free to regret the state of that law. 8 Indeed, Justices themselves confirm and apply, as to this issue, that law even as they critique it 9—and we cannot do more. decisions the willing acceptance and wholehearted enforcement that trial judges are expected to accord appellate decisions.”). 6 EMW Women’s Surgical Ctr., P.S.C. v. Friedlander, 960 F.3d 785, 793 (6th Cir. 2020) (affirming injunction and noting that “in every challenge brought to date [as to ten states’ similar laws], the court has enjoined the law, finding that it indeed unduly burdens”); W. Ala. Women’s Ctr. v. Williamson, 900 F.3d 1310, 1327 (11th Cir. 2018) (affirming injunction, also noting that “every court to consider the issue has ruled that laws banning dismemberment abortions are invalid and that fetal demise methods are not a suitable workaround”), cert. denied sub nom. Harris v. W. Ala. Women’s Ctr., 139 S. Ct. 2606 (2019). 7 See Gonzales v. Carhart, 550 U.S. 124, 147, 150 (2007); Stenberg v. Carhart, 530 U.S. 914 (2000). 8 E.g., W. Ala. Women’s Ctr. v. Williamson, 900 F.3d 1310, 1314 (11th Cir. 2018), cert. denied sub nom. Harris v. W. Ala. Women’s Ctr., 139 S. Ct. 2606 (2019) (Carnes, C.J., writing for the majority) (“Some Supreme Court Justices have been of the view that there is constitutional law and then there is the aberration of constitutional law relating to abortion. If so, what we must apply here is the aberration.”); id. at 1329-30 (“In our judicial system, there is only one Supreme Court, and we are not it. . . . The primary factfinder is the district court, and we are not it. Our role is to apply the law the Supreme Court has laid down to the facts the district court found.”); id. at 1330 (Dubina, J., concurring) (“I am not on the Supreme Court, and as a federal appellate judge, I am bound by my oath to follow all of the Supreme Court’s precedents, whether I agree with them or not.”). 9 Harris v. W. Ala. Women’s Ctr., 139 S. Ct. 2606, 2607 (2019) (Thomas, J., concurring in certiorari denial as to the Eleventh Circuit’s adherence to Stenberg and the 108 Case: 17-51060 Document: 00515984856 Page: 109 Date Filed: 08/18/2021 No. 17-51060 undue burden test to invalidate similar legislation while still offering sharp critique: “[W]e cannot continue blinking the reality of what this Court has wrought.”). 109