Skip navigation
× You have 2 more free articles available this month. Subscribe today.

25 Years of the Prison Litigation Reform Act

At the time the PLRA was enacted over 40 state prison systems and hundreds of jails around the country were under some form of court supervision, either through injunctions or consent decrees, due to unconstitutional conditions. After the Attica rebellion of 1971 federal courts gradually ended two centuries of the “hands off” doctrine and began applying the constitution to prisoners.

Within a quarter-century, prisoners had gained legal rights to adequate medical care, to due process, to be free from physical torture, religious rights and more. Judicial intervention was due to centuries of neglect and abuse by the legislative and executive branches at every level of government which created and perpetuated these abuses.

The United States started the 1990s with a million people locked in prisons and jails. As the prison population ramped up under President Clinton’s 1994 Crime Bill various state and federal legislators and attorney generals saw a need to curtail prisoners’ ability to enforce their constitutional rights, especially those pertaining to overcrowding. Thus began the political push for the PLRA which was ultimately successful in largely curtailing the ability of prisoners to challenge their conditions of confinement and to retain counsel to do so.

The PLRA is 25 years old this year, and it’s worth taking stock of what it has wrought.

The Prospective Relief Provisions

The PLRA limits “prospective relief”—mainly meaning injunctions—by restricting the relief that courts can grant and that parties can agree to, making settlement more difficult, restricting the duration of injunctions, and placing especially rigid restrictions on injunctions requiring release of prisoners. Parties have been able to reach agreements mitigating some of these restrictions, but these arrangements have not been seriously tested in appellate courts and may be precarious.

The PLRA bars courts from entering injunctions in prison cases unless they first find that the relief granted “is narrowly drawn, extends no further than necessary to correct the violation of the Federal right, and is the least intrusive means necessary to correct the violation of the Federal right.” (18 U.S.C. § 3626(a)(1)). By itself that was not a big change from prior law, but the PLRA requires these “need-narrowness-intrusiveness” findings when parties seek to settle cases with court-enforceable consent judgments. (Before, and still in non-prison cases, parties can generally agree to whatever relief they prefer when they settle.) This requirement was made retroactive, so injunctions previously entered without the findings (almost all of them, since no one knew they would be required) could be immediately terminated. The result in the first years of the statute was to sweep away the legal protections gained by hundreds of thousands of prisoners over the preceding decades. In the years immediately after the PLRA’s enactment, almost every consent decree or injunction entered prior to its enactment was terminated. PLN reported extensively on these developments at the time. Not surprisingly, prison and jail conditions did not improve with the termination of judicial relief.

This findings requirement weakened a big incentive for defendants to settle—avoiding a court finding or admission that they did anything wrong. Some litigants have navigated this problem by agreeing to findings, to be approved by the court, that say only that the elements of the PLRA have been satisfied, without identifying the defendants’ transgressions.

In other cases, though, defendants have refused to settle even with such general findings. In those, prisoners and their lawyers have had to choose between the large expense and uncertainty of trial in big cases, or accepting a private settlement agreement (PSA). PSAs need not conform to PLRA requirements or include the need-narrowness-intrusiveness findings, but also cannot be enforced in federal court. Plaintiffs must either enforce the agreement in state court (which no advocates seem to have tried yet), or abandon the agreement and reinstate the litigation after a lapse of months or years. In my experience, compliance with PSAs has ranged from satisfactory to mediocre. Few litigators in recent years seem to opt for the PSA.

If a PLRA-compliant injunction is entered, by consent or court decision, defendants may move to terminate it two years later, and if the motion is denied, move again in a year, and every year thereafter. To avoid termination, the plaintiffs must show a current and ongoing violation of Federal law—not just non-compliance with the order or consent judgment. While few defendants have made all the motions they could under this provision—it’s work for them too—this termination regime poses the problem that many Federal law violations in prison litigation simply cannot be remedied within two years, and allowing termination motions in the middle of the process serves only to disrupt implementation of the remedy.

Some litigants have addressed this problem by agreeing to longer time frames than the PLRA’s two-year/one-year scheme. Litigants have also agreed to terms for terminating the judgment that differ from the “current and ongoing violation of the Federal right” standard provided in the PLRA’s termination provisions—for example, by agreeing that termination will be conditioned on complying with the terms of the remedial order, rather than requiring relitigation of the underlying federal law violation. District courts have approved these arrangements, consistently with the general principle that statutory rights can be waived by litigants unless there is “some affirmative indication of Congress’ intent to preclude waiver” (U.S. v. Mezzanatto, 513 U.S. 196, 201 (1995)), which there is not. However, there is as yet no appellate law addressing the validity of such agreed departures from the PLRA’s literal terms.

Ironically, even defendants who refuse to negotiate consent judgments are often interested in keeping the federal judge involved in the case after entry of a PSA. A number of private settlement agreements name the federal judge as a mediator (with or without using that word), and the parties agree to be bound by the judge’s recommendation with respect to extending the duration of the PSA if there are compliance problems.

Other litigants have accomplished the same end by entering interim agreements, supervised and enforceable by the court, while postponing any actual judgment in the case until, one hopes, compliance is achieved and the problem solved. (See, e.g., Disability Law Center v. Massachusetts Dept. of Correction, 960 F.Supp.2d 271, 275-80 (D.Mass. 2012)). These real-world agreements with prison officials and their lawyers contrast sharply with the rhetoric of PLRA proponents about the evils of federal court involvement in prison matters.

The other major PLRA restrictions on prison injunctions are the provisions governing “prisoner release orders,” defined as “any order, including a temporary restraining order or preliminary injunctive relief, that has the purpose or effect of reducing or limiting the prison population, or that directs the release from or nonadmission of prisoners to a prison.” (18 U.S.C. § 3626(g)(4)). Prompted by political opposition to prisoner releases, the PLRA provides that no such order can be entered without the prior entry of a “less intrusive” order that after a reasonable time for compliance has failed to remedy the Federal law violation at issue, and even then, only after the convening of a special three-judge court. That court must find by clear and convincing evidence that crowding is the primary cause of a violation of a Federal right, and that the problem cannot be remedied without a prisoner release order. (18 U.S.C. § 3626(a)(3)).

The Supreme Court made clear that these provisions are not the death knell of prisoner release orders, upholding an order that the California prison population be reduced because it was impossible to provide adequate medical and mental health care at the existing level of crowding. (Brown v. Plata, 563 U.S. 493 (2011)). However, such orders have been few and far between. Further, the PLRA provisions are grossly ill-designed to respond to emergencies, as shown during the COVID-19 pandemic. Since it is generally physically impossible to achieve social distancing at normal prison population levels, safety required substantial population reductions very quickly. Quick reductions could not be ordered because of the requirement of a prior “less intrusive” order with time for it to work, and the subsequent convening of a three-judge court. (Smith v. DeWine, 2020 WL 4436362, *16 (S.D.Ohio, Aug. 3, 2020); Maney v. Brown, 464 F.Supp.3d 1191, 1207-08 (D.Or. 2020)).

The Prisoner Litigation Provisions

A number of PLRA provisions are directed at suits brought by prisoners, regardless of the kind of relief sought. Some of these have made little difference, such as the provisions for early screening of prisoner complaints and dismissal of those that are meritless on their face; these simply expedite decisions that would otherwise come in response to motions from the defendants. (28 U.S.C. § 1915(e)(2); 28 U.S.C. § 1915A; 42 U.S.C. § 1997e(c)). A section requiring pre-trial proceedings in prisoner cases to be conducted remotely “to the extent practicable,” and encouraging the conduct of hearings at the prison where the relevant unit of government agrees (18 U.S.C. § 1997e(f)), mostly codified existing practice.

A provision for the revocation of federal prisoners’ earned release credit (good time) by a federal court that believes the plaintiff filed a case for a malicious purpose or to harass the adverse party, or presented false testimony or other evidence to the court (28 U.S.C. § 1932) has almost never been used, probably because of its failure to address the obvious due process problems raised by courts’ ordering more prison time based on their view of plaintiffs’ conduct in litigation. A provision that can be read as barring default judgments for prisoners has been read otherwise by most courts. (See, e.g., Henricks v. Pickaway Corr. Inst., 2016 WL 4705647, *3 (S.D.Ohio, Sept. 8, 2016); Adams v. Smith, 2014 WL 811992, *2 (S.D.Ill., Mar. 3, 2014), interpreting 42 U.S.C. § 1997e(g)).

Other prisoner litigation provisions have had massive consequences for prisoners and for the courts.

The exhaustion requirement. The most consequential prisoner litigation provision is the requirement that prisoners exhaust all “available” administrative remedies before bringing suit based on federal law over prison conditions. (42 U.S.C. § 1997e(a)). The Supreme Court has interpreted that requirement broadly and severely, holding that prisoners with damages claims must still exhaust administrative remedies even if they do not provide for damages (Booth v. Churner, 532 U.S. 731 (2001)); that the phrase “prison conditions” encompasses virtually anything that happens in prison (Porter v. Nussle, 534 U.S. 516 (2002)); and that prisoners are held to a standard of “proper exhaustion,” i.e., strict compliance with the rules and deadlines set by the prison, on pain of dismissal of their actions. (Woodford v. Ngo, 548 U.S. 81, 90 (2006)).

The Court has made clear that the primacy of the prison’s rules cuts both ways; prisoners who have complied with them cannot be subjected to additional requirements by the courts (Jones v. Bock, 549 U.S. 199 (2007)), or—as the lower courts have held—by prison officials. (Williams v. Wilkinson, 659 Fed.Appx. 512, 520-21 (10th Cir. 2016) (unpublished) (holding “the prison’s regulations, not ‘facility practice,’ define proper exhaustion”); Miller v. Tanner, 196 F.3d 1190, 1194 (11th Cir. 1999) (holding that failing to sign and date a grievance was not a failure to exhaust since no rule required it)).

The Supreme Court has also held that whether a remedy is “available” under the statute is a practical determination; remedies that are “officially on the books” but not actually capable of use are not available. Remedies may be unavailable, for example, where they present no actual prospect for relief, where they are too “opaque” for prisoners to navigate, and where prison personnel thwart exhaustion through “machination, misrepresentation, or intimidation.” (Ross v. Blake, 136 S.Ct. 1850 (2016)).

This legal scheme, colliding with the realities of prison and prisoners, has yielded an avalanche of satellite litigation over whether particular prisoners have properly exhausted, and whether the remedy was actually available to them. Prison populations include many people who are least able to comply with systems of bureaucratic rules because of psychiatric, cognitive, and developmental disabilities, low educational attainment and literacy levels, language barriers, youth, and in many cases, general unsophistication about legal and bureaucratic processes.

This problem is compounded by prison grievance policies, which are frequently badly designed or poorly drafted, or both. They are often explained to prisoners—if at all—in incomplete or incomprehensible fashion. They are often administered in ways contrary to the rules provided to prisoners—violating their own rules, or invoking supposed rules that do not appear in grievance policies or instructions, or simply failing to function.

The Supreme Court’s decision in Ross v. Blake, cited above, illustrates some of these problems. The plaintiff, complaining of violent assault by staff, did not pursue a grievance because he understood that in practice, matters under investigation by the system’s “Internal Investigative Unit” could not be grieved, though the rules did not say that. While prison officials disputed his argument, the plaintiff showed that in other cases, those officials had commonly argued that same position that they now denied. The Supreme Court, figuratively shaking its head, remanded the case for the lower courts to determine what the system’s grievance policy actually was, and whether the remedy was genuinely available under the circumstances.

Ross is one of many cases in which grievance systems’ actual operation has departed sharply from the rules provided to prisoners. The Sixth Circuit, reviewing Michigan’s system for hearing sexual abuse complaints, described the disparity between the rules and the actual practices as “a classic case of Orwellian doublethink” rendering the remedy unavailable. (Does 8-10 v. Snyder, 945 F.3d 951, 963 (6th Cir. 2019)).

In a West Virginia case, the prison-issued handbook instructed prisoners to request a paper grievance form, fill it out, and give it to a jail employee to transmit it to an “Administrator.” In reality, there was no “Administrator”; grievances were required to be completed on electronic kiosks; and the processing of those grievances sharply deviated from the procedures described in the handbook. The court held that the system was unavailable because ordinary prisoners could not discern or navigate it. (Baxley v. Jividen, 2020 WL 7489760, *12 & n.16, *14-15 (S.D.W.Va., Dec. 21, 2020)).

In other cases, instructions necessary to complete the grievance process are simply absent. A court reviewing the Alabama grievance process for mental health complaints noted that the materials provided prisoners did not explain “how to file such a grievance . . . what form the grievance should be composed on, to whom it should be given and by what means, what information should be included, who will review it and how quickly, and whether there is any process of appeal”; the formal policy, which was not shown to be provided to prisoners, referenced “the relevant form” without identifying it; the policy forbade giving these forms directly to staff and required sending them to the “designated medical department” without advising potential grievants which department was meant. (Dunn v. Dunn, 219 F.Supp.3d 1100, 1116 (M.D.Ala. 2016)). In some prisons, grievances alleging serious staff misconduct are referred to internal affairs or inspector generals’ offices, leaving the grievance in limbo with no guidance to the prisoner or grievance officials as to how the prisoner can then complete the grievance process. (Reid v. Balota, 962 F.3d 325, 330-31 (7th Cir. 2020)).

In still other instances, officials hold prisoners to rules that do not exist or are not disclosed to prisoners. For example, Pennsylvania has a general grievance system, supplemented by a special procedure for complaints about physical and sexual abuse. Officials have argued that prisoners who complete that special procedure must then start over by filing in the general grievance system, a requirement that does not appear in the rules, and which if credited would mean “the opacity prohibited in Ross [v. Blake] would render the process unavailable.” (Moore v. Lamas, 2017 WL 4180378, *11 n. 15 (M.D.Pa., Sept. 2, 2017)). The federal Bureau of Prisons has a “sensitive” grievance procedure for matters that might present risks to the grievant’s safety or well-being, but in at least one federal prison, an unwritten rule limited that procedure to complaints against members of the prison’s executive or grievance staff; the court held the scheme “opaque” and the remedy therefore unavailable. (Taylor v. Gilbert, 2017 WL 3839390, *4-5 (S.D.Ind., Aug. 31, 2017)).

The Supreme Court in Ross v. Blake also acknowledged that remedies may be made unavailable by “machination, misrepresentation, or intimidation.” Examples are legion, including the practice at the Cook County Jail of reclassifying prisoners’ grievances as “requests” which could not be appealed, preventing the prisoner from exhausting the grievance process (Coleman v. Dart, 2019 WL 670248, *7-8 (N.D.Ill., Feb. 19, 2019)), and the grievance repeatedly rejected by the General Counsel of the Federal Bureau of Prisons for failure to attach a document, despite the prisoner’s showing that he couldn’t get the document (DeBrew v. Atwood, 792 F.3d 118, 126-29 (D.C.Cir. 2015); see also Risher v. Lappin, 639 F.3d 236, 240 (6th Cir. 2011), with nearly identical facts). There is a steady drumbeat of cases of intimidation of prisoners from using the grievance system (Gipson v. Renninger, 750 Fed.Appx. 948, 953 (11th Cir. 2018) (per curiam) (unpublished) (noting unrebutted testimony that the person controlling the grievance process threatened to break the plaintiff’s jaw if he filed a grievance or lawsuit); King v. Zamiara, 680 F.3d 686, 697-709 (6th Cir. 2012) (directing entry of judgment against officials who increased prisoner’s security classification based on his complaints and grievances); Armstrong v. Newsom, 2021 WL 933106, *7-9, 26 (N.D.Cal., Mar. 11, 2021) (finding a pattern of threats, intimidation, and coercion of disabled prisoners to prevent their filing ADA grievances or making other complaints), appeal filed, No. 21-15614 (9th Cir., Apr. 6, 2021)).

In short, the PLRA exhaustion requirement often does not foster an orderly process of dispute resolution and problem-solving, but instead sets up what amounts to a game of “gotcha,” with prisoners unfairly prevented from completing the grievance process through various means within and without the grievance system. In many of the above-cited cases, courts—especially appellate courts—have rejected such claims of non-exhaustion, but there are many others in which courts have allowed prisoners to lose their claims as a result of unclear or unfair grievance proceedings or staff misconduct, such as the prisoner who submitted his request for a Halal meat diet to the chaplain as directed by the Religious Diet Policy, but had his case dismissed for not pursuing a grievance in addition, despite the lack of any instruction in the rules to do so, see: Muhammad v. Mayfield, 933 F.3d 993, 1001 (8th Cir. 2019)). We have no way of knowing how many prisoners have been deterred from filing grievances or appealing them, or from filing suit, by such misdirection and misconduct.

Even when the rules are clear and the grievance officials follow them, prisoners can lose their claims for trivial transgressions, like the prisoner whose grievance was rejected and his lawsuit dismissed because he attached handwritten copies of prior appeals rather than photocopies, explaining that the photocopier was broken (Mack v. Klopotoski, 540 Fed.Appx. 108, 112-13 (3d Cir. 2013) (unpublished)); the prisoner whose case was dismissed for non-exhaustion because he failed to sign and date the back of the grievance form where instructed, despite his providing a signature later and the presence of the date elsewhere on the form, Chamblis v. Bland, 2019 WL 1373677, *4 (E.D.Ark., Mar. 4, 2019), report and recommendation adopted, 2019 WL 1372743 (E.D.Ark., Mar. 26, 2019)); the prisoner whose grievance was rejected and his lawsuit dismissed because he placed more than one appeal in a single envelope, then filed a second grievance rather than re-sending the appeal in its own envelope, Hatcher v. Rubenstein, 2018 WL 6036433, *10 (S.D.W.Va., Aug. 8, 2018), report and recommendation adopted, 2018 WL 4628321 (S.D.W.Va., Sept. 27, 2018); and the prisoner who wrote outside the lines on the grievance form (Bracero v. Sec’y, Fla. Dep’t of Corr., 748 Fed.Appx. 200, 302 (11th Cir. 2018) (per curiam) (unpublished) (noting that “the grievances were still legible and just a few lines were outside the boundaries of the space provided. Yet the PLRA demands that prisoners complete the administrative process in accordance with the applicable grievance procedure set by the prison.”)).

Many years ago, in an employment discrimination case, the Supreme Court wisely observed that “technicalities are particularly inappropriate in a statutory scheme in which laymen, unassisted by trained lawyers, initiate the process.” (Love v. Pullman, 404 U.S. 522, 527 (1972)). The failure to extend that principle to prisoners is a major reason PLRA exhaustion has served as such an instrument of injustice.

The physical injury requirement. The PLRA provides: “No Federal civil action may be brought by a prisoner confined in a jail, prison, or other correctional facility, for mental or emotional injury suffered while in custody without a prior showing of physical injury or the commission of a sexual act (as defined in section 2246 of title 18).” (42 U.S.C. § 1997e(e)). (The “sexual act” provision was added long after the PLRA’s enactment.) This provision has denied many prisoners any meaningful remedy for abusive conduct that does not draw blood or leave marks.

The courts have expended much effort in trying to figure out what this badly drafted statute means. For starters, its literal language indicates that prisoners can’t bring lawsuits at all to remedy constitutional violations that don’t cause physical injury. Denying any judicial remedy for constitutional violations would raise significant constitutional problems (Zehner v. Trigg, 133 F.3d 459, 461-62 (7th Cir. 1997)), so courts have consistently held that § 1997e(e) is a limitation on damages, not on the right to bring suit, often citing the title of the statute, “Limitation on Recovery,” in support.

Other problems interpreting this statute stem from Congress’s failure to define its terms. The biggest dispute over § 1997e(e) involves the meaning of “mental or emotional injury.” Some courts hold it means what it says—“such things as stress, fear, and depression, and other psychological impacts” (Amaker v. Haponik, 1999 WL 76798, *7 (S.D.N.Y., Feb. 17, 1999)). Others have said it means anything that is not physical injury (Turner v. Pollard, 564 Fed.Appx. 234, 238 (7th Cir. 2014) (unpublished)).

Courts taking the second view hold that violations of intangible rights like religious freedom or procedural due process are only mental or emotional injuries, so the plaintiff cannot recover compensatory damages for them, but is limited to nominal damages (usually $1.00), and punitive damages in those few cases where the plaintiff can meet the difficult standard for receiving them. (Allah v. Al-Hafeez, 226 F.3d 247 (3d Cir. 2000) (barring compensatory damages for a religious rights violation); Pearson v. Welborn, 471 F.3d 732, 744-45 (7th Cir. 2006) (upholding jury verdict for $1.00 for a prisoner who spent a year in solitary confinement in retaliation for constitutionally protected speech)).

The statutory phrase “physical injury,” which is also undefined, is also the subject of disagreement, though the disagreement is less conceptual, and is found more in the application of law to facts, than the dispute over “mental or emotional injury.” Courts are agreed that § 1997e(e) is satisfied only by physical injury that is “more than de minimis” (Siglar v. Hightower, 112 F.3d 191, 193 (5th Cir. 1997)), but that phrase’s meaning is very much in the eye of the beholder.

One influential district court decision held: “A [more than de minimis] physical injury is an observable or diagnosable medical condition requiring treatment by a medical care professional. It is not a sore muscle, an aching back, a scratch, an abrasion, a bruise, etc., which lasts even up to two or three weeks .... Injuries treatable at home and with over-the-counter drugs, heating pads, rest, etc., do not fall within the parameters of 1997e(e).” (Luong v. Hatt, 979 F.Supp. 481, 485-86 (N.D.Tex. 1997)). By contrast, the Ninth Circuit has explicitly rejected the Luong holding (Oliver v. Keller, 289 F.3d 623, 628 (9th Cir. 2002)) and has held, for example, that a complaint of constant cell illumination which caused migraine headaches, disorientation, and inability to sleep was not about mental or emotional injury but about “various forms of physical injury and discomfort,” and thus satisfied § 1997(e). (Grenning v. Miller-Stout, 739 F.3d 1235, 1238 (9th Cir. 2014)). Not surprisingly, the vagueness of the “more than de minimis” standard has often led to different outcomes for similar injuries.

Even the phrase “Federal civil action” is disputed, with the only federal appellate authority holding that it “means all federal claims, including constitutional claims” (Napier v. Preslicka, 314 F.3d 528, 532 (11th Cir. 2002)). But a number of district courts have held it means claims brought in federal court, regardless of whether they are based on federal law or state law. (See, e.g., Jacobs v. Pennsylvania Dept. of Corrections, 2011 WL 2295095, *23 (W.D.Pa., June 7, 2011)). The second view is more consistent with the statute’s language: “No Federal civil action may be brought,” etc. The dispute makes a difference, since a plaintiff may or may not be able to recover damages for claims based on state law, or for federal claims filed in state court and then removed to federal court by the defendants, depending on the court’s interpretation of that phrase.

As mentioned above, the phrase “or the commission of a sexual act (as defined in section 2246 of title 18)” was added to the statute by amendment, as part of the 2013 reauthorization of the Violence Against Women Act, since there was some dispute whether sexual assault, including forcible rape, constituted physical injury. The amendment made it clear that at least for more serious sexual assaults, plaintiffs may seek compensatory damages regardless of the existence or extent of physical injury.

In forma pauperis (IFP) amendments. The PLRA amended the provisions allowing indigents to sue in federal court without prepayment of filing fees, so prisoners must pay the filing fees and litigation costs in installments of 20% of each month’s funds received. (28 U.S.C. § 1915(a-b)). Courts have upheld this requirement as constitutional, relying heavily on the “safety-valve” provision of 28 U.S.C. § 1915(b)(4), which says that prisoners shall not be prevented from filing or appealing because of lack of funds; that means that the fee provisions do not actually deny access to the courts. (Taylor v. Delatoore, 281 F.3d 844, 848-49 (9th Cir. 2002)). But the fee requirement is a strong practical deterrent even to prisoners with meritorious claims, a deterrent that has been strongly enhanced by the Supreme Court’s holding that multiple fees must be paid off simultaneously, so a prisoner who has filed multiple actions and appeals may be subject to monthly payments of 100%. (Bruce v. Samuels, 577 U.S. 82, 89-90 (2016)).

The three strikes provision. The most morally offensive part of the PLRA is probably the “three strikes” provision, 28 U.S.C. § 1915(g), which bars indigent prisoners entirely fromIFP status if they have had three previous actions or appeals dismissed as frivolous, malicious, or as failing to state a claim on which relief can be granted, unless they can show they are under “imminent danger of serious physical injury” at the time of filing.

That means if prisoners who have had those dismissals cannot pony up the entire $350 filing fee at once, as well as the separate “administrative fee” of $52 (and $505 if they wish to appeal), they are barred from court, no matter how meritorious their claim or how badly they were abused. This prohibition has no time limit; prisoners serving life sentences will be barred from court for a lifetime if they cannot pay the fees up front.

Federal courts have always been able to dismiss cases and to impose sanctions and restrictions on persons who abuse the court system. Before the PLRA this power was exercised with restraint. The courts acknowledged that the constitutional right of access to courts required restrictions to be no broader than warranted by the litigant’s conduct. Permanent exclusion from IFP status based on three dismissed cases is a far harsher penalty than the courts ever imposed. However, that body of constitutional law has been disregarded under the PLRA.

Actions and appeals raising serious claims are routinely dismissed after “three strikers” are denied IFP status. (See, e.g., Brown v. Edinger, 2018 WL 527421, *2-3 (M.D.Pa., Jan. 24, 2018) (barring claim alleging prisoner assault permitted by staff causing eye and facial injuries including fractures), appeal dismissed, No. 18-1198 (3d Cir., Feb. 4, 2020) (unpublished); Robinson-Bey v. Calloway, 2017 WL 6813678, *1 (C.D.Ill., Oct. 13, 2017) (barring claim that correctional staff tore up plaintiff’s permits for an extra mattress, low galley, and low bunk intended to accommodate his severely painful degenerative osteoarthritis, and he was denied specialty care on ground of cost), appeal dismissed, No. 17-3374 (7th Cir., Mar. 1, 2018) (unpublished)).

The statute’s exemption for prisoners who are under “imminent danger of serious physical injury” is at best a half measure and at worse a false promise. Violation of intangible constitutional rights, such as religious freedom or communication with court or counsel, does not fall within the exemption no matter how egregious the facts. Claims of serious physical abuse or medical neglect, no matter how serious, are not exempt from the three strikes bar unless they are ongoing at the time the case is filed; otherwise, they are dismissed as “past harm.” Some courts have even held that when the risk of harm was not imminent at the time of filing, but became so after the case was filed, the three strikes bar should be enforced. (Pinson v. Samuels, 761 F.3d 1, 4-5 (D.C.Cir. 2014)). Where prisoners have alleged an ongoing sequence of violent or threatening events, some courts have displayed extraordinary resistance to acknowledging an ongoing imminent danger. (Turner v. Gibson, 2013 WL 5587391, *3 (E.D.Cal., Oct. 10, 2013) (holding allegations of repeated assaults and other punitive misconduct done in retaliation for plaintiff’s complaints, occurring over a two-year period continuing to the month before the complaint was filed, did not show imminent danger because the possibility of recurrence was speculative)).

The imminent danger exception is supposed to be assessed based on the plausible allegations in the prisoner’s complaint. (Andrews v. Cervantes, 493 F.3d 1047, 1055 (9th Cir. 2007). However, courts have allowed defendants to relitigate claims of imminent danger, bringing in whatever evidence they choose to dispute the plaintiff’s allegations and to obtain the revocation of already-granted IFP status. Often, dismissal follows because the plaintiff cannot pay the entire filing fee at once. In these cases, courts are sometimes required to determine medical issues or other difficult questions at an early stage in the case where appointing counsel for the plaintiff has not been addressed and the pro se plaintiff is unassisted, which is not conducive to informed decision-making. (See Gresham v. Meden, 938 F.3d 847, 850 (6th Cir. 2019) (holding forced use of psychiatric medication Prolixin did not constitute imminent danger, but without considering whether it presented a risk of tardive dyskinesia, a known side effect of the drug)).

The three strikes provision, like the exhaustion requirement, has burdened the courts with a large volume of side litigation about its interpretation and application. As one court observed, “It is no small irony that a rule presumably intended to preserve scarce resources has consumed so much attention.” (Bontemps v. Sotak, 2015 WL 812360, *3 n.2 (E.D.Cal., Feb. 25, 2015), report and recommendation adopted, 2015 WL 1469870 (E.D.Cal., Mar. 30, 2015), order vacated, 699 Fed.Appx. 653 (9th Cir. 2017) (unpublished)).

The Supreme Court has had to resolve two such questions. (Lomax v. Ortiz-Marquez, --- U.S. ----, 140 S. Ct. 1721, 1724 (2020) (holding a dismissal on the specified grounds without prejudice is a strike); Coleman v. Tollefson, 575 U.S. 532, 534 (2015) (holding a dismissal takes effect as a strike immediately, and not after completion or waiver of any appellate proceedings)).

Many other questions have had to be resolved—or remain unresolved—in the lower courts, including but not limited to whether a dismissal for suing an immune defendant is a strike; whether dismissals for miscellaneous misconduct in the course of litigation—rather than for the inadequacies of the complaint—can be dismissals as “malicious” for purposes of three strikes; whether voluntary dismissals, including those taken after an adverse recommendation from a magistrate judge, can be strikes; whether a prisoner whose third strike dismissal occurs in the district court is precluded from IFP status for an appeal from that dismissal, or whether the prohibition only applies to the next separate action; and many others.

This litigation—like litigation under other PLRA provisions—represents time and thought taken away from the courts’ consideration of the merits of prisoners’ and others’ claims, and allows serious questions about the treatment of prisoners to be swept under the rug with technical dismissals.

Attorney’s fees. The PLRA attacks prisoners’ access to courts indirectly as well as directly, by making it harder to obtain lawyers. In civil rights suits under 42 U.S.C. § 1983, plaintiffs who win are entitled to an award of attorney’s fees for the reasonable value of the attorney’s services at market rates. (42 U.S.C. § 1988(b)). The purpose of the attorney’s fees statute was to allow plaintiffs to retain lawyers even if they did not have money to pay them, and if their claims were not likely to generate large enough damage awards to make contingency fee arrangements viable.

The PLRA does not exclude prisoners from recovering attorney’s fees, but it restricts them in multiple ways. First, it limits fees to 150% of the rates paid to criminal defense lawyers appointed for federal criminal defendants. (42 U.S.C. § 1997e(d)(3)). That might not sound so bad, except the defense lawyers get paid regardless of outcome, while civil rights lawyers get paid only if they prevail, and the CJA rates are far below market rates in most jurisdictions.

Even these reduced fees can only be recovered to the extent that the fees were “directly and reasonably incurred” in “proving an actual violation” of federal law (42 U.S.C. § 1997e(d)(1)(A)), which means that prisoners, unlike other litigants, cannot recover fees for cases that are settled rather than proved unless the defendants agree to pay them as part of the settlement. Fees can also be recovered for enforcing the relief ordered for a violation (42 U.S.C. § 1997e(d)(1)(B)(ii)), but most courts have held this means a federal law violation must have previously been found, or must be found in the enforcement proceeding, to support a fee award. “Merely” showing that the order has not been followed is not compensable.

Prisoners who recover damages are also required to contribute “a portion of the judgment (not to exceed 25%)” to help satisfy the attorney’s fee award—which the Supreme Court has said means 25 percent in all cases unless the fee award is satisfied with less (Murphy v. Smith, --- U.S. ----, 138 S.Ct. 784, 786 (2018), interpreting 42 U.S.C. § 1997e(d)(2)).

An even more restrictive provision proscribes that in damages cases, court-ordered fees are limited to 150% of the judgment. Many violations of intangible constitutional rights yield awards of nominal damages of $1.00, because courts and juries can’t figure out how to value them, and in any case the physical injury requirement discussed above has been interpreted to bar compensatory damages entirely for such claims. So, for example, a plaintiff who proved to a jury that he had been sent to solitary confinement for a year in retaliation for exercising First Amendment rights received $1.00 in damages and his attorney was awarded $1.50 in fees. (Pearson v. Welborn, 471 F.3d 732, 742-43 (7th Cir. 2006), applying 42 U.S.C. § 1997e(2)). There is also a requirement that fees be “proportionately related” to the relief granted, but the other restrictions ensure that fees are kept low enough that this provision is rarely invoked.

These provisions together make taking a prisoner case a much less attractive proposition than before the PLRA for lawyers in private practice who have to make a living. Some private lawyers do continue to take prisoner cases (often as a sideline to a police misconduct practice), but others decline to do so, and some very capable and experienced lawyers will only take prisoner cases involving death or catastrophic injury which promise a large enough recovery to make the case economically viable despite the PLRA.

Where things stand. Beginning as soon as the PLRA was enacted, prisoner advocates challenged most of the PLRA’s provisions as unconstitutional, and lost across the board. The PLRA will stay as it is unless Congress changes it. Concerted efforts began early in the Obama administration to amend or repeal its most damaging provisions. Rep. Bobby Davis, D-Va., introduced a reform bill in the House of Representatives, which went nowhere; nothing happened in the Senate. The only successful amendment came a few years later: the above described amendment to the physical injury requirement which allows damage recovery for specified types of sexual assault without proof of physical injury, enacted as part of the 2013 renewal of the Violence Against Women Act. Whether the present Congress or administration will take up amendment or repeal of part or all of the PLRA remains to be seen.

[Editor’s Note: Since 1996 the Human Rights Defense Center has both opposed the PLRA and called for its repeal. Early in the Obama administration there was brief talk about repealing it which quickly died. Despite all the talk about criminal justice reform, anything involving improving conditions of confinement for prisoners, repealing the PLRA or giving prisoners enforceable rights is unheard of and appears to be off limits for virtually all criminal justice reformers, except for HRDC.] 


John Boston is the former director of the prisoner rights unit at the Legal Aid Society in New York City, until his retirement. He is one of the foremost prisoner rights attorneys in the United States and the coauthor of the Prisoners’ Self-Help Litigation Manual. Mr. Boston is also the author of the forthcoming book The PLRA Handbook: Law and Practice under the Prison Litigation Reform Act, which will be published by HRDC later this year.

As a digital subscriber to Prison Legal News, you can access full text and downloads for this and other premium content.

Subscribe today

Already a subscriber? Login