James Howe v. Latoya Hughes ( 2023 )


Menu:
  •                                  In the
    United States Court of Appeals
    For the Seventh Circuit
    ____________________
    No. 22-1368
    JAMES G. HOWE, et al.,
    Plaintiffs-Appellees,
    v.
    LATOYA HUGHES, * in her official capacity as Acting Director
    of the Illinois Department of Corrections, et al.,
    Defendants-Appellants.
    ____________________
    Appeal from the United States District Court for the
    Southern District of Illinois.
    No. 3:14-cv-844 — Staci M. Yandle, Judge.
    ____________________
    ARGUED JANUARY 10, 2023 — DECIDED JULY 24, 2023
    ____________________
    Before SCUDDER, KIRSCH, and JACKSON-AKIWUMI, Circuit
    Judges.
    SCUDDER, Circuit Judge. Very weighty interests are at stake
    when a state institutes a program of civil commitment for sex
    * Latoya Hughes replaced Rob Jeffreys as Acting Director of the Illi-
    nois Department of Corrections on April 1, 2023, so we have substituted
    her as a party to this case. See Fed. R. App. P. 43(c)(2).
    2                                                   No. 22-1368
    offenders who, though never tried for or convicted of a crime,
    are found too dangerous for release. On the one hand is the
    state’s interest in promoting public safety, and on the other is
    an individual’s liberty interest. The Fourteenth Amendment’s
    Due Process Clause permits the balance of these interests to
    tip in the state’s favor—but only if the state adheres to partic-
    ular mandates to ensure the liberty restrictions go no further
    and last no longer than necessary. The necessary balance re-
    quires states to afford sex offenders treatment sufficient to
    permit a realistic opportunity for rehabilitation and, ulti-
    mately, release.
    The broader issues presented in this case are all about this
    constitutionally necessary balancing—all about whether Illi-
    nois, in implementing a civil commitment program under the
    state’s Sexually Dangerous Persons Act, is complying with its
    obligations under the Fourteenth Amendment. The record be-
    fore us leaves us concerned that the state is not holding up its
    end of the balance. Civil detainees under the state’s Sexually
    Dangerous Persons Program receive minimal treatment, rais-
    ing serious questions whether rehabilitation and release are
    realistically available to them.
    Yet we can decide this appeal without fully immersing
    ourselves in the broader issues. The district court, though un-
    derstandably focused on curing the constitutional defects in
    Illinois’s civil commitment program, issued too broad an in-
    junction under the strictures of the Prison Litigation Reform
    Act. We therefore reverse and remand.
    No. 22-1368                                                       3
    I
    A
    The Fourteenth Amendment guarantees that no state shall
    “deprive any person of life, liberty, or property, without due
    process of law.” How this applies in the context of criminal
    law is all too familiar. To find someone guilty of a crime and
    thus deprive them of liberty, the state must abide by a wide
    array of due process protections. See, e.g., In re Winship, 
    397 U.S. 358
    , 364 (1970); Estelle v. Williams, 
    425 U.S. 501
    , 512 (1976).
    And to cement the deprivation of liberty with a criminal sen-
    tence, the state must continue to ensure due process of law.
    See Hicks v. Oklahoma, 
    447 U.S. 343
    , 346 (1980). For all but the
    most serious offenses, this process culminates in a sentence
    that establishes a definite term of imprisonment—one with an
    end date.
    In the context of civil commitment, though, the protections
    of the Fourteenth Amendment operate a little differently. The
    Supreme Court “repeatedly has recognized that civil commit-
    ment for any purpose constitutes a significant deprivation of
    liberty that requires due process protection.” Addington v.
    Texas, 
    441 U.S. 418
    , 425 (1979). At the same time, however, the
    Court has acknowledged that states have “a legitimate inter-
    est … in providing care to [their] citizens who are unable be-
    cause of emotional disorders to care for themselves.” 
    Id. at 426
    . This interest, the Court has explained, parallels states’
    “authority under [their] police power to protect the commu-
    nity from the dangerous tendencies of some who are mentally
    ill.” 
    Id.
     So the Court has upheld state programs that civilly de-
    tain individuals as long as the detainee suffers from a mental
    illness and exhibits some form of violent behavior. See, e.g.,
    Addington, 
    441 U.S. at 426
     (“[T]he State has no interest in
    4                                                     No. 22-1368
    confining individuals involuntarily if they are not mentally ill
    or if they do not pose some danger to themselves or others.”);
    Kansas v. Hendricks, 
    521 U.S. 346
    , 357–58 (1997) (“States have
    in certain narrow circumstances provided for the forcible civil
    detainment of people who are unable to control their behavior
    and who thereby pose a danger to the public health and
    safety.”); Foucha v. Louisiana, 
    504 U.S. 71
    , 78 (1992) (“[K]eeping
    [a detainee] against his will in a mental institution is improper
    absent a determination in civil commitment proceedings of
    current mental illness and dangerousness.”).
    No doubt civil confinement “constitutes a significant dep-
    rivation of liberty” and “can engender adverse social conse-
    quences,” thereby “requir[ing] due process protection.” Ad-
    dington, 
    441 U.S. at
    425–26. The Fourteenth Amendment re-
    quires states to balance their interests—caring for citizens suf-
    fering from mental illness and protecting the community—
    against the liberty interests of those who it seeks to civilly de-
    tain. See 
    id.
     at 425 (citing Mathews v. Eldridge, 
    424 U.S. 319
    , 335
    (1976)). This balance can only be struck where the state’s in-
    terest in civil commitment is not punitive. See 
    id. at 428
    ; Allen
    v. Illinois, 
    478 U.S. 364
    , 373–74 (1986); Foucha, 
    504 U.S. at 80
    ;
    Hendricks, 
    521 U.S. at 361
    .
    Unlike incarceration following a criminal conviction, civil
    commitment often does not have a set end date. That follows
    from the purpose of civil commitment—to provide rehabilita-
    tion and treatment, not retribution and deterrence. See Allen,
    
    478 U.S. at
    369–70; Hendricks, 
    521 U.S. at
    363–64. “[B]ecause it
    is impossible to predict how long it will take for any given
    individual to recover,” the Supreme Court has explained, it is
    permissible “to leave the length of commitment indetermi-
    nate, subject to periodic review.” Jones v. United States, 463
    No. 22-1368 
    5 U.S. 354
    , 368 (1983). That periodic review is important: a civil
    detainee “is constitutionally entitled to ‘immediate release
    upon a showing that [he] is no longer dangerous or mentally
    impaired.’” Hughes v. Dimas, 
    837 F.3d 807
    , 808 (7th Cir. 2016)
    (alteration in original) (quoting Hendricks, 
    521 U.S. at
    368–69).
    As long as the state diligently reevaluates detainees’ mental
    health status and dangerousness, though, it may continue to
    commit them indefinitely. See Hendricks, 
    521 U.S. at 358
    .
    This is where treatment comes into play. Civil detainees
    must actually “receive treatment for the disorders that led to
    their confinement and be released when they’ve improved
    enough no longer to be dangerous.” Hughes, 
    837 F.3d at
    808
    (citing Hendricks, 
    521 U.S. at
    368–69, and Allen, 
    478 U.S. at
    369–
    74); see also Youngberg v. Romeo, 
    457 U.S. 307
    , 317 (1982)
    (“When a person is institutionalized—and wholly dependent
    on the State— … a duty to provide certain services and care
    does exist.”). It is not enough for a state to purport to offer
    treatment but in reality provide next to nothing (or nothing at
    all). See United States ex rel. Stachulak v. Coughlin, 
    520 F.2d 931
    ,
    936 (7th Cir. 1975) (“It is well settled that realities rather than
    benign motives or noncriminal labels determine the relevance
    of constitutional policies.”), abrogated on other grounds by Ad-
    dington, 
    441 U.S. at
    427–31. Indefinite civil commitment with
    no meaningful treatment and no realistic possibility of release
    violates the constitutional command that “the nature of com-
    mitment [must] bear some reasonable relation to the purpose
    for which the individual is committed.” Foucha, 
    504 U.S. at 79
    .
    The essential takeaway is that civil detainees are “entitled
    to non-punitive programs designed using the exercise of pro-
    fessional judgment.” Allison v. Snyder, 
    332 F.3d 1076
    , 1080 (7th
    Cir. 2003) (citing Romeo, 
    457 U.S. at
    321–22). Indeed, states
    6                                                    No. 22-1368
    must invest resources to ensure that their civil commitment
    programs—both in design and in practice—provide detainees
    sufficient treatment to make release a realistic possibility. If a
    state fails to do so and its program becomes “such a substan-
    tial departure from accepted professional judgment, practice,
    or standards as to demonstrate that the person responsible ac-
    tually did not base [their decisions] on such a judgment,” then
    the state has violated the Fourteenth Amendment. Romeo, 
    457 U.S. at 323
    .
    B
    Illinois’s Sexually Dangerous Persons Act authorizes the
    state’s Director of Corrections—currently Latoya Hughes, the
    lead defendant—to involuntarily commit and indefinitely de-
    tain individuals who have been charged with a crime and
    found to “suffer[ ] from a mental disorder … coupled with
    criminal propensities to the commission of sex offenses” and
    “propensities toward acts of sexual assaults or acts of sexual
    molestation of children.” 720 ILCS 205/1.01, /3, /3.01, /8. The
    Act ensures that the Director “shall provide care and treat-
    ment for the person committed to him designed to effect re-
    covery.” 
    Id.
     at 205/8. Once detainees have recovered enough
    to be deemed “no longer dangerous,” the state must dis-
    charge them. 
    Id.
     at 205/9(e). The state may conditionally dis-
    charge detainees under continued supervision, but only if “it
    is impossible to determine with certainty under conditions of
    institutional care that the person has fully recovered.” 
    Id.
    Recognize that the Act, by its terms, aligns with the man-
    dates of the Due Process Clause. Indeed, the Supreme Court
    reviewed the Act in Allen v. Illinois, 
    478 U.S. 364
     (1986), and
    came to that exact conclusion. The Court determined that the
    Act provides for civil—rather than criminal—commitment,
    No. 22-1368                                                    7
    stressing that Illinois “disavowed any interest in punishment,
    provided for the treatment of those it commits, and estab-
    lished a system under which committed persons may be re-
    leased after the briefest time in confinement.” 
    Id. at 370
    . It is
    because of these very characteristics of the Act—the lack of
    punitive intent, the availability of treatment, and the realistic
    possibility of release—that Illinois does not violate the Four-
    teenth Amendment when it detains sexually dangerous per-
    sons. See Hughes, 
    837 F.3d at 808
     (reversing and remanding
    where it was “unresolved” in the record whether Illinois was
    providing detainees with treatment and release consistent
    with the Act); Allison, 
    332 F.3d at 1079
     (explaining that be-
    cause sexually dangerous persons receive treatment under
    the Act, Illinois’s decision to house them in prison facilities
    does not “signify punishment”).
    II
    Our observation that the Act is constitutional in its general
    design does not resolve this appeal, however. Grave ques-
    tions remain about how the Act is applied in practice. Recall
    that “realities rather than benign motives or noncriminal la-
    bels determine the relevance of constitutional policies,”
    Coughlin, 
    520 F.2d at 936
    , so civil detainees must actually “re-
    ceive treatment for the disorders that led to their confinement
    and be released when they’ve improved enough no longer to
    be dangerous.” Hughes, 
    837 F.3d at 808
    . This case implicates
    these practical concerns.
    A
    The plaintiffs—James Howe, Jacob Kallal, and George
    Needs—have been committed under the Act to the Big
    Muddy River Correctional Center Sexually Dangerous
    8                                                 No. 22-1368
    Persons Program. Howe was conditionally discharged on
    May 24, 2023, while Kallal and Needs remain at Big Muddy.
    The treatment program at Big Muddy includes three kinds
    of group therapy: core therapy, offense-specific therapy, and
    didactic (or educational) therapy. Every detainee participates
    in a core therapy group that meets weekly. Only those detain-
    ees who have acknowledged their prior sexual misconduct
    may participate in offense-specific and didactic therapy
    groups. Offense-specific therapy groups cover topics such as
    Victim Empathy, Cycles of Sexual Offending, and Relapse
    Prevention. Similarly, didactic therapy groups address topics
    like Anger Management, Social Skills, Substance Abuse, and
    Expressive Art. Didactic therapy sessions are offered on a ro-
    tating basis. Big Muddy does not provide individual therapy.
    The treatment program consists of four sequential phases.
    To advance to the next phase, a detainee must maintain a
    clean disciplinary record, show progress in therapy, and com-
    plete certain assessments. Therapists evaluate detainees semi-
    annually and provide them with a copy of their evaluations
    as well as six-month treatment plans, but detainees may dis-
    cuss their evaluations with the therapists only in group ther-
    apy, not one-on-one.
    The state contracts with Wexford Health Sources, a
    healthcare services company, to evaluate detainees for re-
    lease. Wexford reviews the risk posed by each detainee seek-
    ing release with reference to the detainee’s semi-annual re-
    ports, treatment plans, therapists’ case notes, and interviews
    with therapists and the detainee. Based on this information,
    Wexford recommends either discharge, conditional dis-
    charge, or continuation in the Big Muddy program.
    No. 22-1368                                                    9
    B
    In July 2014 the plaintiffs filed this lawsuit alleging that
    the treatment program at Big Muddy was being run in a con-
    stitutionally deficient manner. Over four years later, in Octo-
    ber 2018, the district court held a two-day bench trial in which
    the plaintiffs testified as fact witnesses and Dr. Dean
    Cauley—a former clinical team leader and therapist at the
    Florida Civil Commitment Center for Sexually Violent Preda-
    tors—testified as an expert witness. The district court issued
    its findings of fact and conclusions of law three years later, in
    September 2021.
    We follow the lead of both parties in accepting the district
    court’s factual findings, though we note that significant time
    has passed since the trial, which has no doubt led to opera-
    tional changes within the Big Muddy program. Indeed, both
    parties disputed the existence, nature, and legal significance
    of intervening developments in their post-trial papers and at
    oral argument. But it is not our role to pass judgment on such
    assertions for the first time on appeal, nor is any purported
    legal development relevant to our legal analysis. We are con-
    fident that a justiciable controversy remains regardless of any
    operational changes at Big Muddy or the conditional release
    of one of the named plaintiffs. See Friends of the Earth, Inc. v.
    Laidlaw Env’t Servs. (TOC), Inc., 
    528 U.S. 167
    , 189 (2000); Bow-
    sher v. Synar, 
    478 U.S. 714
    , 721 (1986). To the extent that any
    changes to the Big Muddy program have legal significance to
    the nuts and bolts of what prospective relief is appropriate,
    the district court is better situated to explore those issues and
    will have a chance to do so on remand. Our focus, meanwhile,
    is on how things were at Big Muddy in October 2018.
    10                                                  No. 22-1368
    At that time Big Muddy housed 170 civil detainees with
    three therapists on staff. The 170 detainees were split into
    twelve core therapy groups, which met once per week for one
    hour. Many of the offense-specific and didactic therapy
    groups were on hold—including, for example, the Victim Em-
    pathy group, which had not been offered since 2015.
    Testifying for the plaintiffs, Dr. Cauley surveyed data
    from a national network of civil commitment facilities and
    opined that Big Muddy’s program offerings and operation fell
    significantly short of professional norms. Every other facility
    on record, Dr. Cauley explained, provides multiple core
    group therapy sessions per week, whereas Big Muddy stands
    alone in offering just one weekly session. And as a result of
    such infrequent meetings, Big Muddy only provides one hour
    of weekly core therapy, in contrast to the national average (7.5
    hours) and the generally accepted minimum (5 hours). Simi-
    larly, Dr. Cauley testified that certain offense-specific and di-
    dactic groups, such as Substance Abuse and Victim Empathy,
    are “critical” to detainees’ treatment and should always be
    available, even to detainees who have not yet acknowledged
    their prior misconduct. But Big Muddy had indefinitely can-
    celed these treatment offerings. Finally, Dr. Cauley explained
    that evaluators should prioritize indicators such as age and
    the passage of time over past conduct when considering de-
    tainees’ discharge petitions. Wexford, however, did not even
    consider age or passage of time and instead gave meaningful
    weight to detainees’ past conduct.
    The district court accepted Dr. Cauley’s findings and con-
    cluded that the disparity between Big Muddy’s treatment
    program and professional standards amounted to a constitu-
    tional violation. Alongside entering its findings of fact and
    No. 22-1368                                                   11
    conclusions of law, the district court issued a three-pronged
    permanent injunction, requiring that Big Muddy
    •   provide the plaintiffs a minimum of 7.5
    hours of core group therapy per week, with
    each session lasting no less than 90 minutes;
    •   reinstate all inactive offense-specific and di-
    dactic groups, including groups that
    Dr. Cauley identified as foundational (such
    as Substance Abuse) as well as other groups
    on hold (such as Expressive Art); and
    •   use independent evaluators other than Wex-
    ford to perform discharge evaluations.
    The state moved for reconsideration. It claimed the injunc-
    tion was overbroad under the Prison Litigation Reform Act,
    which limits the possible scope of prospective relief in civil
    actions “with respect to prison conditions.” 
    18 U.S.C. § 3626
    (a)(1)(A). The district court saw no overbreadth and
    therefore denied the state’s motion.
    The state now appeals.
    III
    Without disputing that constitutional violations were oc-
    curring at Big Muddy, the state presses a narrower point on
    appeal. It challenges only the scope of the injunction under
    the PLRA. On that limited front, we agree. The PLRA restricts
    the district court’s injunctive power, and the current injunc-
    tion is overbroad.
    A
    Through the Prison Litigation Reform Act, Congress man-
    dated that “[p]rospective relief in any civil action with respect
    12                                                    No. 22-1368
    to prison conditions shall extend no further than necessary to
    correct the violation of the Federal right of a particular plain-
    tiff or plaintiffs.” 
    18 U.S.C. § 3626
    (a)(1)(A). The plaintiffs first
    attempt to sidestep the PLRA by asserting that it does not ap-
    ply to their claims at all.
    Their efforts on this front are forfeited. The district court
    made clear as early as 2014, when it first screened the plain-
    tiffs’ claims, that it saw this case as governed by the PLRA.
    Furthermore, the state expressly argued that the injunction vi-
    olated the PLRA in its October 2021 motion for reconsidera-
    tion. Even then the plaintiffs did not raise this argument until
    seven months later, once the case was on appeal. That is too
    late. See Allen v. City of Chicago, 
    865 F.3d 936
    , 943–44 (7th Cir.
    2017); see also United States v. Knox, 
    624 F.3d 865
    , 873 n.6 (7th
    Cir. 2010) (treating arguments as forfeited, rather than
    waived, when it was “not clear” that the defendant made a
    strategic choice not to make them).
    Because the district court will need to reconsider the scope
    of its injunction on remand, we nevertheless exercise our dis-
    cretion to consider the plaintiffs’ forfeited arguments to make
    clear that they fail as a legal matter. See Hacker v. Dart, 
    62 F.4th 1073
    , 1082 (7th Cir. 2023). The plaintiffs’ first contention—that
    they are not “prisoners” under the PLRA—misreads the rele-
    vant statutory language. While other provisions of the PLRA
    are limited to suits brought “by prisoners,” the provision at
    issue here is not so constrained. Compare 42 U.S.C. § 1997e(a)
    (imposing exhaustion requirements on suits brought “by a
    prisoner”), with 
    18 U.S.C. § 3626
    (a)(1)(A) (restricting federal
    courts’ injunctive authority “in any civil action with respect
    to prison conditions”).
    No. 22-1368                                                   13
    The plaintiffs next contend that because their treatment is
    the very “basis” for their continued confinement, not a condi-
    tion of their confinement, their suit is not one “with respect to
    prison conditions.” But this view runs headlong into the Su-
    preme Court’s conditions-of-confinement jurisprudence,
    which applies when interpreting the PLRA. See Porter v.
    Nussle, 
    534 U.S. 516
    , 527–28 (2002). In Wilkinson v. Dotson, 
    544 U.S. 74
     (2005), the Court held that a prisoner’s challenge to
    parole proceedings was a challenge to the conditions of his
    confinement even though parole proceedings are effectively
    the basis for a prisoner’s continued confinement. See 
    id. at 82
    .
    The Court has since articulated the standard this way: if an
    inmate’s legal victory does not “necessarily imply the invalid-
    ity of his conviction or sentence,” his claim remains a “prison-
    conditions claim” governed by the PLRA. Nance v. Ward, 
    142 S. Ct. 2214
    , 2222 (2022) (quoting Heck v. Humphrey, 
    512 U.S. 477
    , 487 (1994)). Applied to this case, a victory for the plain-
    tiffs (such as an injunction requiring constitutionally ade-
    quate treatment) would not necessarily imply their speedier
    discharge from the Big Muddy program. So the plaintiffs’ ar-
    gument on this front fails as well.
    B
    Returning to the text of the PLRA, we see that Congress
    mandated that any injunction with respect to prison condi-
    tions “extend no further than necessary to correct the viola-
    tion,” be “narrowly drawn,” and be “the least intrusive means
    necessary to correct the violation.” 
    18 U.S.C. § 3626
    (a)(1)(A).
    Interpreting this very limitation, we have explained that fed-
    eral courts must take care not to “conflate[ ] what is constitu-
    tionally adequate … with what is constitutionally required.”
    14                                                    No. 22-1368
    Westefer v. Neal, 
    682 F.3d 679
    , 683–84 (7th Cir. 2012) (emphasis
    in original).
    Our case law has identified two dimensions to the PLRA’s
    mandate against overbreadth. First, an injunction that im-
    poses numeric requirements cannot do so at a level that ex-
    ceeds the constitutional floor. See Rasho v. Jeffreys, 
    22 F.4th 703
    ,
    713–14 (7th Cir. 2022) (reversing a district court injunction
    that “set staffing at levels sufficient to exceed the constitutional
    minimum” (emphasis in original)). That is not to say that fed-
    eral courts may never set numeric benchmarks. See id. at 713.
    To the contrary, the Supreme Court blessed such an injunc-
    tion in Brown v. Plata, 
    563 U.S. 493
    , 538–41 (2011), after inten-
    sive fact-finding established that a specific numeric target
    “was necessary to remedy the constitutional violations.” 
    Id. at 540
    . But such injunctions cannot extend further than neces-
    sary to address the violation at issue.
    The second dimension to the PLRA’s limit on prospective
    relief is that such relief may not be too prescriptive. At bot-
    tom, federal courts must ensure that “substantial discretion
    and flexibility” remain “in the hands of the prison adminis-
    trators.” Westefer, 
    682 F.3d at 685
    ; see also Romeo, 
    457 U.S. at 317
    . That is essential if an injunction is to be narrowly drawn
    and as unintrusive as possible.
    The injunction before us here is overbroad in both re-
    spects. It requires that the Big Muddy Program provide 7.5
    hours of weekly core group therapy despite the district
    court’s express finding that 5 hours of core therapy per week
    would be constitutionally adequate. Indeed, Dr. Cauley—the
    plaintiffs’ expert—favorably referenced other treatment pro-
    grams that provide only 5 hours of core therapy per week.
    Although he explained that those programs had other
    No. 22-1368                                                   15
    therapeutic and educational opportunities available for de-
    tainees, the injunction entered by the district court does not
    give Big Muddy flexibility to implement, say, 7.5 hours of any
    form of weekly group therapy. Rather, the injunction requires
    7.5 hours of core group therapy, which is above the constitu-
    tional floor and therefore extends further than necessary to
    remedy the violation.
    The remaining provisions of the injunction fail for a differ-
    ent reason: they are too prescriptive. As to the reinstatement
    of canceled offense-specific and didactic groups, the injunc-
    tion does not give Big Muddy the option to rotate between
    programs based on actual need. While the evidence may sup-
    port a finding that some groups (like Victim Empathy or Sub-
    stance Abuse) should be offered with more consistency, it
    does not support the same finding with respect to all groups
    (such as Expressive Art, which the injunction includes). And
    the injunction’s prohibition on Wexford providing evaluation
    services to the program is even less grounded in the evidence.
    While it is true that a detainee’s past conduct is only relevant
    to their continued detention to the extent it sheds light on
    their present dangerousness, see O’Connor v. Donaldson, 
    422 U.S. 563
    , 574–75 (1975); Allen, 
    478 U.S. at 371
    , on this record
    we see no constitutional hook to single out a specific contrac-
    tor and prohibit them from providing services to state facili-
    ties. Doing so is not the least intrusive means available to cor-
    rect the constitutional violation at issue.
    IV
    It is sufficient for purposes of resolving this appeal to con-
    clude the injunction is overbroad. But because we remand for
    the district court to craft a new injunction consistent with the
    16                                                   No. 22-1368
    PLRA and our opinion, we offer some parting thoughts to
    help guide this litigation across the finish line.
    First, the parties should ensure that they present thorough
    evidence to the district court as to what kind and level of treat-
    ment they believe constitutes the constitutional floor. See
    Westefer, 
    682 F.3d at
    683–84; Romeo, 
    457 U.S. at
    322–23. If the
    parties fail to do so, or if their evidence is irreconcilable, the
    district court remains free to employ the resources at its dis-
    posal to develop a strong evidentiary foundation for its forth-
    coming injunction. See, e.g., Fed. R. Evid. 706(a) (permitting
    the court to appoint its own expert witness).
    Second, the district court must take care to connect any pro-
    spective relief to a specific constitutional violation. For in-
    stance, if the district court remains concerned about Wexford
    relying too heavily on past conduct when evaluating detain-
    ees for discharge, it needs to explain how its concern maps
    onto the protections of the Fourteenth Amendment. See
    Hughes, 
    837 F.3d at 808
     (explaining that the Fourteenth
    Amendment requires inmates be released if they are no
    longer dangerous); Allen, 
    478 U.S. at 371
     (approving of Illi-
    nois’s civil commitment program because it “primarily” uses
    past conduct “to show the accused’s mental condition and to
    predict future behavior”). And it must tailor its injunction to
    addressing only those constitutional concerns.
    Third, the state may not continue to rely on cost and logis-
    tical difficulties to evade its constitutional obligations. The
    PLRA does not shield states from unwanted expenditures
    necessary to comply with constitutional mandates. See Arm-
    strong v. Schwarzenegger, 
    622 F.3d 1058
    , 1071 (9th Cir. 2010)
    (“[T]he question is not whether the relief the court ordered …
    is expensive, or difficult to achieve, but whether the same
    No. 22-1368                                                    17
    vindication of federal rights could have been achieved with
    less involvement by the court in directing the details of [the
    prison’s] operations.”). If Illinois continues to operate the Big
    Muddy program in violation of the Fourteenth Amendment,
    it may expose itself to far more drastic and costly measures
    than those it has so far avoided in this case. See generally
    Plata, 
    563 U.S. 493
     (affirming an injunction that required
    California to reduce its prison population by roughly 40,000
    inmates).
    Fourth, this case is nearly a decade old, and we urge the
    district court and the parties to work together to bring it to a
    prompt resolution following remand. The district court can
    and should set an aggressive briefing schedule, and the par-
    ties must do everything in their power to comply. The district
    court may suggest the parties try mediation to speed up the
    process. If it does so, we expect they will engage in good faith.
    *      *      *
    No doubt the Constitution permits Illinois to operate a
    civil confinement program for sexually dangerous persons.
    But it does so only as long as there is a careful balancing of the
    state’s interest in the safety of its citizens against those indi-
    viduals’ liberty interest. By providing inadequate treatment
    and, as a result, depriving detainees of a realistic possibility
    of release, Illinois has failed to uphold its end of the balance.
    The state should right these constitutional wrongs without
    delay.
    Still, the district court’s injunction is overbroad under the
    PLRA, so we VACATE the injunction and REMAND for fur-
    ther proceedings consistent with this opinion.