Garrett v. Davis

Opinion

United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit

____________

FILED

August 4, 2026

No. 25-40327 ____________ Lyle W. Cayce Clerk

Michael Garrett,

Plaintiff—Appellant,

versus

Director Lorie Davis; Bobby Lumpkin; Eric Guerrero,

Defendants—Appellees.

______________________________

Appeal from the United States District Court for the Southern District of Texas USDC No. 2:13-CV-70

______________________________

Before Willett, Engelhardt, and Douglas, Circuit Judges. Per Curiam: * This is Michael Garrett’s fourth trip to our court. Twice before, we remanded because the district court applied the wrong legal standard to his Eighth Amendment sleep-deprivation claim. Yet the renewed decision again measures Garrett’s claim against requirements our precedents reject. Because those errors infected both the objective and subjective components

_____________________

*

This opinion is not designated for publication. See 5th Cir. R. 47.5.

No. 25-40327

of the analysis, we VACATE and REMAND for further proceedings consistent with this opinion.

I

This case has spanned more than a decade. Garrett, a Texas prisoner, alleges that since at least 2008 the Texas Department of Criminal Justice’s (TDCJ) schedules have afforded him no more than four hours of sleep each night. He sued several TDCJ officials under 42 U.S.C. § 1983, contending that the schedule—first at the McConnell Unit where Garrett then resided— violates the Eighth Amendment. He sought an injunction requiring at least six hours of nightly sleep.

A magistrate judge dismissed the complaint for failure to state a claim.

We reversed, holding that Garrett plausibly alleged both the objective and subjective components of an Eighth Amendment claim. Garrett v. Thaler, 560 F. App’x 375, 379-80 (5th Cir. 2014) (per curiam) (“Garrett I”).

On remand, the district court held a two-day bench trial. Garrett offered unrebutted expert testimony about the schedule permitting at most only four hours of sleep per night and that chronic sleep restriction carries serious health risks, including risks associated with Garrett’s hypertension and kidney disease. The district court nevertheless found no constitutional violation, reasoning principally that Garrett had “not establish[ed] a cause- and-effect relationship between his sleep schedule and any medical complaint he [had] made” and that the prison schedule served “a legitimate penological interest.” See also Garrett v. Lumpkin, 840 F. App’x 807, 808 (5th Cir. 2021) (per curiam) (“Garrett II”). Garrett appealed.

While that appeal was pending, TDCJ transferred Garrett to the Estelle Unit. Id. We remanded so the district court could determine in the first instance whether conditions at the new unit violated the Eighth Amendment. Id.

3

No. 25-40327

The district court found “that Garrett [was] similarly being deprived of sleep at the Estelle Unit,” affording him “only three and one-half hours of sleep, which is further interrupted by hourly cell door operations, bright lights, and a bed-book count at 1:00 a.m.” Yet the court again rejected the claim because Garrett had not “sufficiently linked” the deprivation to his health complaints and because the schedule served “legitimate penological purposes.” Garrett appealed again.

We again vacated. Garrett v. Lumpkin, 96 F.4th 896, 902 (5th Cir.

2024) (“Garrett III”). We explained that the objective component requires proof that the challenged conditions pose “a substantial risk of serious harm”—not proof that they caused a particular injury. Id. at 900–01 (emphasis added). And we explained that the schedule’s penological purpose does not determine whether officials acted with deliberate indifference. Id. at 901. We remanded with instructions to apply those standards. Id. at 902.

On remand, TDCJ reported another schedule change, prompting additional discovery and supplemental briefing. The parties agreed that the revised schedule affords Garrett only three total hours of nighttime sleep— from 11:00 p.m. to midnight and from 1:00 a.m. to 3:00 a.m.—subject to interruptions from cell-door operations, lighting, and bed counts.

The district court nevertheless denied relief. On the objective component, it faulted Garrett’s expert for showing “only an association, not a cause-and-effect relationship, between the amount of continuous sleep and health disorders.” It also required proof of a “specific number of hours of continuous sleep” and credited Garrett’s ability to sleep “in chunks throughout the day and night.” On the subjective component, the court reasoned that, “without any proof as to what threshold of sleep is required to avoid a substantial risk of serious harm,” it “cannot determine that TDCJ

4

No. 25-40327

was aware of that threshold and was deliberately indifferent to it.” The court dismissed Garrett’s claims, and Garrett appealed.

II

“The standard of review for a bench trial is well established: findings of fact are reviewed for clear error and legal issues are reviewed de novo.” Barto v. Shore Constr., L.L.C., 801 F.3d 465, 471 (5th Cir. 2015) (quoting Becker v. Tidewater, Inc., 586 F.3d 358, 365 (5th Cir. 2009)). But clear-error deference does not extend “to [factual] determinations reached by application of an incorrect legal standard.” Fuji Photo Film Co. v. Shinohara Shoji Kabushiki Kaisha, 754 F.2d 591, 595 n.4 (5th Cir. 1985). Generally, when such “a legal error . . . affect[s] . . . factual findings, remand is the proper course unless the record permits only one resolution of the factual issue.” Ball v. LeBlanc, 792 F.3d 584, 596 (5th Cir. 2015) (citation omitted). Remand is likewise proper where application of the wrong legal standard may have influenced the district court’s ultimate conclusion. Harris v. Forklift Sys., Inc., 510 U.S. 17, 23 (1993); June Med. Servs., L.L.C. v. Phillips, 22 F.4th 512, 521–22 (5th Cir. 2022).

III

The Eighth Amendment prohibits “cruel and unusual punishments.”

U.S. Const. amend. VIII. Sleep is not a luxury; it is a basic human necessity. See Harper v. Showers, 174 F.3d 716, 720 (5th Cir. 1999) (“[S]leep undoubtedly counts as one of life’s basic needs.”). Prison conditions therefore violate the Eighth Amendment when they “pose ‘an unreasonable risk of serious damage’ to a prisoner’s health—an objective test—and prison officials must have acted with deliberate indifference to the risk posed—a subjective test.” Ball, 792 F.3d at 592 (quoting Helling v. McKinney, 509 U.S. 25, 33–35 (1993)). We clarified both components in Garrett III, 96 F.4th at

5

No. 25-40327

900–02. The district court’s renewed analysis again departed from that framework.

A

The objective-prong analysis contains three legal errors: the court (1) demanded causation rather than substantial risk, (2) required a quantitative societal benchmark our cases do not demand, and (3) failed to account for duration. Although differently framed, each error reflects the same misconception. The inquiry is whether Garrett’s actual conditions pose a substantial risk of serious harm—not whether science has proved individualized causation, identified a precise physiological threshold, or supplied a universal minimum amount of sleep.

1

First, the district court imposed too demanding a causation requirement. It correctly recited that Garrett “need not prove actual harm, but only a substantial risk of serious harm.” But it then discounted his expert’s opinion because she identified “only an association, not a cause- and-effect relationship, between the amount of continuous sleep and health disorders.” That is the very requirement Garrett III rejected. Garrett need not show that the challenged schedule caused a particular injury—or that sleep restriction invariably causes disease. He must prove a substantial risk of serious harm. See Garrett III, 96 F.4th at 900–01.

Precedent confirms the distinction. The Supreme Court in Helling credited statistical and scientific evidence that secondhand smoke created serious health risks without requiring proof that it caused disease in a particular inmate. 509 U.S. at 33, 36. And our court in McCord v. Maggio imposed no requirement that unsanitary conditions be proved to cause illness before they could violate the Constitution. 927 F.2d 844, 846–47 (5th Cir. 1991). By demanding demonstrated causation rather than substantial risk, the

6

No. 25-40327

district court effectively restored the actual-injury requirement we rejected in Garrett III.

2

Second, the district court required Garrett to identify the precise amount of sleep “a people in a civilized world must actually get before suffering significant health effects.” Our cases impose no such numerical benchmark. 1 Instead, we have explained that “a prisoner must show . . . the conditions were objectively so serious as to deprive prisoners of the minimal civilized measure of life’s necessities.” Alexander v. Texas Dep’t of Crim. Just., 951 F.3d 236, 241 (5th Cir. 2020) (citation omitted); see also Farmer v. Brennan, 511 U.S. 825, 834 (1994) (describing the objective prong as requiring that the deprivation be “sufficiently serious” such that it “must result in the denial of the minimal civilized measure of life’s necessities”) (citations omitted)). That inquiry turns on the totality of the conditions the prisoner actually endured—not on a universally quantified minimum. See Alexander, 951 F.3d at 241.

The district court therefore erred by treating the absence of a scientifically fixed minimum amount of continuous sleep as dispositive. The Eighth Amendment does not require constitutional deprivations to arrive with calibrated thresholds. Garrett’s burden is to prove that the sleep

_____________________

1 See, e.g., Ball, 792 F.3d at 590–93 (no minimum or maximum temperature);

Palmer v. Johnson, 193 F.3d 346, 353 (5th Cir. 1999) (no minimum temperature, number of bathroom facilities, or square footage); Berry v. Brady, 192 F.3d 504, 507 (5th Cir. 1999) (no minimum food); Talib v. Gilley, 138 F.3d 211, 214 n.3 (5th Cir. 1998) (no minimum daily calories, types of food, or frequency of meals); McCord, 927 F.2d at 846–47 (no minimum amount of sewage disposal or quality of water); Ruiz v. Estelle, 679 F.2d 1115, 1152 (5th Cir. 1982), amended in part and vacated in part on other grounds, 688 F.2d 266 (5th Cir. 1982) (no minimum amount of exercise).

7

No. 25-40327

deprivation he experienced, considered in context, posed “a substantial risk of serious harm.” Garrett III, 96 F.4th at 901.

3

Third, the district court overlooked an indispensable consideration:

duration. Whether a condition poses substantial risk of serious harm depends on “the totality of the specific circumstances presented.” Palmer, 193 F.3d at 353. Duration is part of that totality. Berry, 192 F.3d at 507; see also Hutto v. Finney, 437 U.S. 678, 686 (1978) (“[T]he length of confinement cannot be ignored in deciding whether the confinement meets constitutional standards.”), abrogated in part on other grounds as recognized by, Dep’t of Agric. Rural Dev. Rural Hous. Serv. v. Kirtz, 601 U.S. 42, 56 (2024). That is because “[a] facially permissible form of punishment may . . . through continual use inflict cruel and unusual punishment.” Cooper v. Sheriff, Lubbock Cnty., 929 F.2d 1078, 1083 (5th Cir. 1991).

Conditions-of-confinement cases routinely account for both the degree and duration of deprivation. See, e.g., Talib, 138 F.3d at 214 n.3 (“Whether the deprivation . . . falls below this threshold depends on the amount and duration of the deprivation.” (emphasis added)); Ruiz, 679 F.2d at 1152 (considering, among other factors, “the overall duration of his confinement” (emphasis added)); see also Hutto, 437 U.S. at 686–87 (“A filthy, overcrowded cell and a diet of ‘grue’ might be tolerable for a few days and intolerably cruel for weeks or months.”). On remand, the district court must consider how long Garrett has endured challenged conditions, along with whether daytime naps or other sleep opportunities meaningfully mitigate the long-term restriction of nighttime sleep. 2

_____________________

2 Garrett argues that the district court exceeded the scope of remand by considering his opportunities for daytime sleep, contending that Garrett III confined the objective-

8

No. 25-40327

B

The objective-prong errors necessarily affected the subjective inquiry.

To satisfy the subjective prong, “a prisoner must show that prison officials acted with ‘deliberate indifference’ to the risk posed.” Garrett III, 96 F.4th at 901 (quoting Ball, 792 F.3d at 592). Consistent with our remand instructions, the district court correctly abandoned the penological-purpose test. But it declined to assess the subjective component after reasoning that, “without any proof as to what threshold of sleep is required to avoid a substantial risk of serious harm, the Court cannot determine that TDCJ was aware of that threshold and was deliberately indifferent to it.” This conclusion rested on the same mistaken premise that infected the objective analysis.

The district court must therefore reassess deliberate indifference under the correct objective standard. We express no view on the proper result. The inquiry should consider the extent of TDCJ’s knowledge of the risks Garrett identified and the significance of any measures officials took in response to his complaints. See id.

C

Garrett asks us to do more than remand and instead render judgment in his favor. He contends that the “record permit[s] only one resolution— that [he] established his Eighth Amendment claim.” We disagree. The

_____________________

prong inquiry to nighttime sleep only. We disagree. The focus of Garrett’s complaint is “sleep deprivation,” not “continuous nighttime sleep.” And Garrett III did not hold, expressly or impliedly, that daytime sleep is categorically irrelevant to whether a prisoner’s conditions of confinement pose a substantial risk of serious harm. Consistent with the totality-of-the-circumstances approach that governs the objective prong, see Palmer, 193 F.3d at 353, the district court did not impermissibly exceed the scope of remand when it credited Garrett’s daytime sleep opportunities as one relevant consideration bearing on the severity and duration of his sleep deprivation.

9

No. 25-40327

district court’s legal errors leave unresolved factual questions, so the “only one resolution” exception permitting rendition does not apply. See Ball, 792 F.3d at 596. Those questions include (1) whether the Estelle Unit’s schedule—considering duration and all meaningful sleep opportunities— poses a substantial risk of serious harm, and (2) whether TDCJ officials “knew of and disregarded an excessive risk to [Garrett’s] health or safety” based “on the totality of the record evidence.” Alexander, 951 F.3d at 241; Ball, 792 F.3d at 595. We express no view on the answers.

IV

This is Garrett’s fourth appeal and our third remand on the merits.

We do not take that history lightly. The district court’s renewed decision again rests on legal standards our precedents do not permit, and unresolved factual questions remain. We therefore VACATE and REMAND for further proceedings consistent with this opinion.