ROVNER, Circuit Judge.
The plaintiffs in this case brought a class action pursuant to 42 U.S.C. § 1983 alleging that the policies and practices of the Marion County Sheriff's Department and the Consolidated City of Indianapolis and Marion County (collectively referred to as the "Sheriff") caused them to be detained in the Marion County Jail awaiting release for an unreasonably long period of time, in violation of the Fourth Amendment. The plaintiffs sought to certify five subclasses in that action, and the district court granted certification as to two of those subclasses, but denied it as to the remaining three. The plaintiffs then filed a petition in this court seeking permission to appeal the denial of two of those class certifications pursuant to Federal Rule of Civil Procedure 23(f). Specifically, the plaintiffs contested the court's denial of two classes, consisting of all individuals who, from December 19, 2012 to the present, were held in confinement by the Sheriff after legal authority for those detentions ceased, due to: (1) the Sheriff's practice of operating under a standard of allowing up to 72 hours to release prisoners who are ordered released; and (2) the Sheriff's practice of employing a computer system inadequate for the purposes intended with respect to the timely release of prisoners.
We granted permission for the interlocutory appeal pursuant to Rule 23(f), and now proceed to the appeal on the merits. We hold that the district court erred in its decision denying class certification and remand the case to the district court for further proceedings.
In order to certify a class, "`a district court must find that each requirement of Rule 23(a) (numerosity, commonality, typicality, and adequacy of representation) is satisfied as well as one subsection of Rule 23(b).'" Harper v. Sheriff of Cook County, 581 F.3d 511, 513 (7th Cir. 2009). The plaintiffs sought to certify a subclass based on the Sheriff's policy, practice or custom of allowing the jail staff to hold inmates for up to 72 hours before releasing them. The district court held that the subclass as so defined would presuppose that some members were detained for less than 48 hours and others for greater than 48 hours. The district court believed that such a range of detention periods was problematic, because the court—relying on County of Riverside v. McLaughlin, 500 U.S. 44 (1991)—believed that the 48-hour line was a critical defining period in establishing the reasonableness of the detention. According to the district court, detentions of less than 48 hours would be presumptively reasonable, and those that extended beyond 48 hours would be presumptively unreasonable, thus subjecting those members within the class to two different burdens of proof.
The district court erred in applying the 48-hour presumption to this context and in relying on it as a basis to deny class certification. The court relied for that denial on the Supreme Court's holding in McLaughlin, which addressed the detention resulting from a warrantless arrest and held that the amount of time between the warrantless arrest and a judicial determination of probable cause was presumptively reasonable if it was 48 hours or less, and presumptively unreasonable if longer. See McLaughlin, 500 U.S. at 56-57; Portis v. City of Chicago, Illinois, 613 F.3d 702, 703-04 (7th Cir. 2010). That time period necessarily would include the time involved in processing and booking the defendant, determining the appropriate charge and preparing charging documents, assigning and transporting to court, and ultimately obtaining a judicial determination of probable cause. See McLaughlin, 500 U.S. at 55.
The class proposed by the plaintiffs involved a markedly different situation. It is composed of persons for whom legal authority for detention has ceased, whether by acquittal after trial, release on recognizance bond, completion of jail time in the sentence, or otherwise. For those persons, all that is left is for the officials to merely process the release. None of the myriad steps required in McLaughlin, between an arrest and a judicial determination of probable cause, are required here; the class members already qualify for release, and all that is left are the ministerial actions to accomplish that release which are within the control of the jail officials. Evidence in the record indicates that the average time period to effect such a release is 2-4 hours in counties in general, and up to 6 hours if problems are encountered, but even if we doubled those times, release still would be accomplished within 12 hours. Because the tasks involved in the situation presented here are significantly less onerous and less time-consuming than the ones involved in McLaughlin, the 48-hour rule makes no sense in this context.
Accordingly, the district court erred in denying the subclass based on its perception that the 48-hour rule in McLaughlin would create different burdens and challenges among the potential subclass members. The only other reason given by the court for denying the subclass was that individual variables could complicate the timing of the release, but those variables were present in McLaughlin as well and they did not preclude class status. The Court in McLaughlin recognized that at some point the State has no legitimate interest in detaining persons for an extended period of time, and if the regular practice exceeds that time period deemed constitutionally-permissible, the State is not immune from systemic challenges such as a class action. 500 U.S. at 55, 58-59. At some point well short of the 24-plus hours alleged here, there is no reason to believe that individual issues would account for that delay.
The defendants nevertheless claim that cases from our circuit, Portis and Harper, foreclose class certification here. They assert that those cases and McLaughlin provide that common questions do not predominate where the core complaint challenges the length of detention rather than the conditions of confinement, and that any extended detention must be evaluated on a case-by-case basis. Those cases do not support that conclusion. In Harper, 581 F.3d at 512, the plaintiff sought to bring a class action alleging that new detainees remanded to the sheriff's custody were unconstitutionally required to undergo certain intake procedures. We held that the claims were not appropriate for class disposition because Harper was not challenging specific intake procedures, but instead was asserting that the Sheriff was unconstitutionally holding detainees after bond was posted. Id. at 514-15. The unconstitutionality depended on the length of the delay, which was an individualized determination because Harper did not allege any overriding policy or practice causing the delay. Id. at 514-15. The only common issue alleged by Harper was whether it was reasonable to assign a jail identification number before releasing a detainee on bond—with the concomitant delay that process entailed—but we held that the issue was not central to his claim because it could not cause the type of injuries asserted, and because it could not be unconstitutional unless it took an unreasonable amount of time in an individual case, thus again not supporting class disposition. Id. at 515-16. Similarly, in Portis, 613 F.3d at 703, we addressed a class challenge by persons subjected to custodial arrests for fine-only offenses, alleging that the failure to release them within two hours from generation of the central booking number was unreasonable in violation of the Fourth Amendment. We held that the Constitution forbids detentions that are unreasonable in length, but that an arbitrary inflexible time period such as two hours for release was not justified. Id. at 704. Accordingly, the claims required an individualized determination as to whether the delay in release was unreasonable, and class certification was improper. We noted, however, that "[t]he Supreme Court suggested in McLaughlin that class treatment might be appropriate if the class sought to establish that a jurisdiction had adopted a policy of deliberate delay." Id. at 705. The Portis court explicitly stated that it did not foreclose the possibility of class-wide relief if the record established such deliberate delay. Id. at 705-06. Neither Portis nor Harper preclude class certification in a case such as this one, in which the plaintiffs assert that the defendants' policy or practice caused them to be detained for an unconstitutionally-unrea-sonable length of time. See also Phillips v. Sheriff of Cook County, 828 F.3d 541, 550 (7th Cir. 2016) (the proper focus in determining commonality is whether the prospective class can "articulate at least one common question that will actually advance all of the class members' claims").
On appeal, the defendants assert that the plaintiffs' case nevertheless cannot be certified because the alleged 72-hour policy and practice upon which they rely does not exist or, if it exists, it is constitutional. The district court refused to address the merits of the allegation that the Sheriff operated under such a 72-hour rule, stating that it was not appropriate to adjudicate the case on its merits at the stage of class certification. As the Supreme Court has noted, "Rule 23 does not set forth a mere pleading standard," and a party seeking class certification "must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011). To certify a class, the trial court must satisfy itself "after a rigorous analysis" that the Rule 23(a) prerequisites are established, and "[f]requently that `rigorous analysis' will entail some overlap with the merits of the plaintiff's underlying claim." Id. at 350-51. That does not mean that the Rule 23 analysis is transformed into a summary judgment motion. The defendants argue the merits of the claim, challenging the existence and constitutionality of the policies, without referencing the Rule 23 factors or establishing how those factors are impacted. The district court properly recognized that its role in assessing class certification did not include a determination of the case on the merits. On remand, the court should, of course, consider factual and legal issues comprising plaintiffs' cause of action insofar as those issues are necessary to a determination of the Rule 23 factors. Id.; Parko v. Shell Oil Co., 739 F.3d 1083, 1085 (7th Cir. 2014); Szabo v. Bridgeport Machines, Inc., 249 F.3d 672, 675-76 (7th Cir. 2001).
The plaintiffs also challenge the denial of certification for a class consisting of all individuals who, from December 12, 2012 to the present, were held in confinement by the Marion County Sheriff after legal authority for those detentions ceased, due to the Sheriff's practice of employing a computer system inadequate for the purposes intended that dramatically delayed the release of prisoners. The plaintiffs presented evidence that the Sheriff acquired a computer system, OMS, from Global Tel Link which offered the system to the Sheriff free of charge in return for the Sheriff's extension of GTL's contract for inmate collect call services. GTL provided the hardware, software, installation and maintenance of the phones and lines, and profited from the high rate charged to inmates for such collect calls. Evidence indicated that the Sheriff profited from the arrangement as well, receiving commissions totaling approximately $800,000 per year.
In addition, the plaintiffs produced some evidence that the Sheriff bypassed the standard review process in choosing and retaining the computer system, disregarding the impact on the release times. At the time that it acquired OMS, the Sheriff knew that the Marion County Courts were using software called Odyssey. The Sheriff was also aware that, prior to acquiring any jail information management software, it was imperative to know whether the software would be compatible with the court's Odyssey software. Evidence nevertheless indicated that the Sheriff acquired OMS outside the standard channels bypassing the Information Services Agency which was the Marion County agency formed to coordinate information technologies among the City and County agencies and through which such acquisitions are supposed to occur in order to assure that the systems integrate with those of other agencies. There is a dispute as to whether the Sheriff vetted OMS; the Sheriff's chief administrative officer testified that prior to acquiring OMS the Sheriff's department received assurances from three IT people that OMS would work with Odyssey, but the plaintiffs introduced evidence from those three persons that they never vetted the system and were never asked to do so.
As the district court noted, OMS was beset with technical issues from the start; it could not interface with DEXTER, the computerized transfer system that allowed agencies such as the Sheriff, the Public Defender, the Prosecutor's Office, Community Corrections, and the Indianapolis Police Department to exchange information with each other on Odyssey. And with OMS the Sheriff could not receive electronic court information, which required the Sheriff's office staff to manually update and process the codes received from the court and to rely on emails, paper records, faxes, and telephone calls to gather information to make release decisions.
Finally, the record contains evidence that the Sheriff chose to remain with OMS even in the face of the significant delays in release times, and did not take efforts to measure the magnitude of the problem. That delay in release often totaling 72 hours was significant both in the pure sense and in proportion to the time that the prisoners could be properly detained. For instance, the plaintiffs produced evidence of individuals who served short sentences for DUI convictions of 5-9 days, but for whom the 2-4 day delay in release increased their incarceration time by 40-50%. Another person was arrested and released by the court on bond two days later, but detained 3 more days awaiting release by the Sheriff, thus more than doubling the period of detention. An expert produced by the plaintiff found that from June to December 2014, 38,000 extra days were spent in jail by inmates pending release compared to the practice before OMS. The delays resulted in complaints to the Sheriff's department by judges, defense counsel and family members, but in the face of those widespread excessive delays the Sheriff chose to continue with OMS rather than implement the compatible system recommended by its IT people.
The class sought by the plaintiffs consisted of all individuals who, from December 19, 2012 to the present, were victims of the computer issues spawned by OMS. The district court denied certification of the subclass because it held that the class was not "identifiable." The court noted that the class description must be sufficiently definite to allow ascertainment of the class members. Dist. Ct. Op. at 12 citing Alliance to End Repression v. Rockford, 565 F.2d 975, 977 (7th Cir. 1977). It held that "[t]he definition of the subclass that Plaintiffs advance is problematic because the technical issues plaguing OMS is one of the overriding reasons for the over-detention of the entire class. As such, the court is not convinced it is a policy or practice from which this subclass can be anchored." [emphasis in original] Dist. Ct. Op. at 12. The court appears to deny certification of the subclass because the policy or practice caused the over-detention of the entire class. That, however, would be a basis to grant certification of the class as a whole rather than as a subclass, not to deny certification because it is not limited to a portion of the class. We have recognized that a class may lack the definitiveness required for class certification if there is no way to know or readily ascertain who is a member of the class, but that issue is not apparent here. Jamie S. v. Milwaukee Pub. Sch., 668 F.3d 481, 495 (7th Cir. 2012). The class is defined by all persons whose excessive detention resulted from the inadequate computer system. Given the evidence of a dramatic increase in detention times in correlation with the implementation of the computer system, the evidence that release information could not be properly transmitted electronically in that system and the court having to resort to alternative channels of communication, and the absence of any evidence that delays of that length could be attributable to individual factors, the class is capable of definition both by the timing and the length of the delay in release. Compare Oshana v. Coca-Cola Co., 472 F.3d 506, 513-14 (7th Cir. 2006) (in consumer fraud action alleging that Coca-Cola was deceptive in failing to inform consumers that its fountain Diet Coke contained saccharine, proposed class was not sufficiently definite because it required only the purchase of fountain Diet Coke and could include millions who were not deceived). But we need not further define the class in this appeal to conclude that the district court's reasoning, in denying class status because the alleged violation applied to the class as a whole rather than a subclass, was an improper basis to deny certification.
With respect to the denial of certification as to this class, the defendants argue that the denial should be upheld because the plaintiffs failed to present evidence sufficient to establish the existence of an unconstitutional policy or practice and also failed to present sufficient evidence of intent. We do not understand the district court's decision to hold that the plaintiffs have not established that their injury resulted from a policy or practice, as the defendants suggest; such an argument would be difficult to make, as the determination to process releases through the OMS system certainly appears to fall within our understanding of what constitutes a policy or practice. See generally Glisson v. Indiana Dept of Corrections, 849 F.3d 372, 378-80 (7th Cir. 2017)(en banc). And as to establishing deliberate indifference, we have recognized in another context that deliberate indifference can be shown by systemic and gross deficiencies in equipment or procedures that result in a deprivation of a constitutional right. See Phillips, 828 F.3d at 554. But as we discussed earlier, the district court did not consider whether the policies or practices existed, nor did it consider whether the evidence indicated deliberate indifference, because the court determined that it was not allowed to engage in an analysis of the merits, and therefore we are presented only with the evidence recited above without any factual findings by the court on those matters. The court on remand should consider all issues related to the Rule 23 factors even if they overlap with the merits, and can make the appropriate fact findings at that time. The issues have never been decided by the district court and are not properly before us at this time.
The decision of the district court is VACATED and the case REMANDED for further proceedings consistent with this opinion.