In the

     United States Court of Appeals
                 For the Seventh Circuit
                      ____________________


Nos. 19-1380, 19-1387 & 19-1732
JEFFREY ORR, et al.
                                                  Plaintiffs-Appellees,

                                  v.

LOUIS SHICKER, et al.,
                                              Defendants-Appellants.
                      ____________________

         Appeals from the United States District Court for the
                     Central District of Illinois.
             No. 08-cv-2232 — Harold A. Baker, Judge.
                      ____________________

   ARGUED NOVEMBER 4, 2019 — DECIDED MARCH 23, 2020
               ____________________

    Before WOOD, Chief Judge, and BAUER and BRENNAN, Cir-
cuit Judges.
    WOOD, Chief Judge. Plaintiffs are current and former in-
mates of the Illinois Department of Corrections (IDOC) who
have been diagnosed with hepatitis C. They filed this lawsuit
over ten years ago after fruitless efforts to receive treatment
for their disease while incarcerated. Invoking 42 U.S.C. § 1983,
their complaint alleges that the diagnostic and treatment
2                              Nos. 19-1380, 19-1387 & 19-1732

protocols for IDOC inmates with hepatitis C violate the
Eighth and Fourteenth Amendments. After many years, many
motions, and the consolidation of many cases, the district
court granted class certification and preliminary injunctive re-
lief. The defendants—IDOC, Wexford Health Sources, Inc.,
and several doctors—asked us to accept an appeal from that
decision under Federal Rule of Civil Procedure 23(f). We
agreed to do so and now reverse the grant of class certification
and vacate the injunction.
                               I
                        A. Hepatitis C
    Hepatitis C is a disease caused by the hepatitis C virus
(HCV). Those who contract HCV may suffer inflammation of
the liver, which can impair the functioning of that vital organ.
HCV has six genotypes, the first of which predominates in the
United States.
     A hepatitis C infection is categorized as either acute or
chronic. In the acute phase, the infection does not necessarily
result in any noticeable symptoms, and some people naturally
clear the virus from their bodies. Others develop a chronic in-
fection. Persons suffering from a chronic hepatitis C infection
may develop fibrosis—that is, the build-up of scar tissue in
the liver—which can lead to cirrhosis, a severe condition caus-
ing the affected areas of the liver to stop functioning. Cirrhosis
is irreversible. The rate of fibrosis progression is not the same
in all HCV patients.
    Several tests are available to determine the degree of fibro-
sis in a liver. One test, a FibroScan, uses ultrasound technol-
ogy. This test results in several possible scores: F0 (no fibro-
sis), F1 (mild fibrosis), F2 (moderate fibrosis), F3 (advanced
Nos. 19-1380, 19-1387 & 19-1732                               3

fibrosis), and F4 (cirrhosis). Another test is the AST (which
stands for the enzyme aspartate aminotransferase) to Platelet
Ratio Index, or APRI. The APRI is calculated by dividing the
patient’s AST level by her platelet count.
                B. The Lawsuit’s Early Years
    Plaintiffs have all been diagnosed with hepatitis C. They
filed this lawsuit in 2008, alleging that IDOC’s medical direc-
tors and Wexford, a private company that administers
healthcare to inmates in IDOC’s custody, were deliberately
indifferent to their medical conditions when they chose not to
provide necessary treatment. Plaintiffs sought injunctive re-
lief and class certification. In response to a motion from the
defendants, the district court struck these requests.
    In April 2009, Plaintiffs filed amended motions for class
certification and injunctive relief; the district court denied
those motions in November 2009. Two years later, the court
dismissed Wexford from the suit with prejudice. In 2013, the
court consolidated several related cases. The case dragged
along until, in January 2016, Plaintiffs moved for reconsidera-
tion of the denial of class certification. The district court de-
nied their motion, but it indicated a willingness to certify a
class and so granted them leave to renew their request. The
court noted that since the case was filed in 2008, there had
been significant developments in the treatment of hepatitis C.
In the beginning, the disease had virtually no cure free from
serious risks and extensive costs. New treatments that were
both significantly more effective and lower in cost had since
become available.
  In November 2016, Plaintiffs filed an amended complaint,
naming IDOC, Wexford (again), IDOC’s medical director
4                               Nos. 19-1380, 19-1387 & 19-1732

(initially Dr. Michael Puisis, who was later succeeded by Dr.
Louis Shicker and then Dr. Steven Meeks), and several other
doctors as defendants. As before, they alleged that the defend-
ants were deliberately indifferent to their medical needs in vi-
olation of the Eighth and Fourteenth Amendments. They also
filed an amended motion for class certification and a motion
for injunctive relief.
    In April 2017, the district court entered an order summa-
rizing the current status of the case. The order reviewed the
testimony given by witnesses and experts in 2016 and noted
that there was “no dispute that Hep C is a serious medical
condition or that it is present in significant numbers in the
IDOC prison population.” The court nonetheless recognized
a “sharp dispute as to when treatment should begin.” At the
time of the April 2017 order, IDOC’s policy was to begin treat-
ment for inmates once their FibroScan score reached F3. Plain-
tiffs argued that treatment should begin as soon as possible.
    The district court also found that “all inmates should be
tested for Hepatitis C upon admission” to IDOC. It deter-
mined that inmates “who test positively and have at least one
year to serve on their sentence from admission to release date
should be offered treatment with direct acting antiviral drugs
as soon as possible after diagnosis, and, in any event, no later
[than] testing at a fibrosis level of 2.” Nevertheless, the district
court concluded that its findings were “too general to meet
the specific requirements of a preliminary injunction order”
and that it needed additional information about Plaintiffs’ “fi-
brosis levels, contributing conditions, sentence length, etc.”
   In January 2018, Plaintiffs filed still another motion for in-
junctive relief; the district court scheduled an evidentiary
hearing for January 2019. (The court also consolidated this
Nos. 19-1380, 19-1387 & 19-1732                               5

case with three others in which Wexford remained a defend-
ant.) In early January 2019, IDOC implemented an updated
protocol for treating inmates with hepatitis C (“the 2019 Pro-
tocol”).
                    C. The 2019 Protocol
    Under the 2019 Protocol, IDOC contracts with the Univer-
sity of Illinois – Chicago (“UIC”) for treatment of hepatitis C
patients using a telemedicine system. The primary care phy-
sicians at each IDOC facility perform initial testing to ascer-
tain which inmates are eligible for treatment at UIC. The UIC
specialists then “determine the specific regimen for patients
found ready for HCV treatment.”
    When an inmate enters IDOC’s custody, he receives an
HCV antibody test at the receiving and classification prison,
unless he declines testing. If the test shows that the inmate is
HCV positive and the inmate wants to be evaluated for possi-
ble treatment, further testing is performed at the prison to
which the inmate is transferred, in order to “determine if the
patient has chronic HCV disease or [if] the HCV infection has
resolved (which can happen in about 15-25% [of cases]).” If
chronic HCV is present, IDOC physicians evaluate “absolute
exclusion criteria”—in other words, they determine whether
the inmate has less than 12 months remaining on his sentence,
has refused treatment, has unstable medical or psychiatric
conditions, or other contraindications to HCV therapy.
    If no exclusionary factors are present, the inmate under-
goes further tests, including a FibroScan and APRI calcula-
tion, to determine the severity of the liver disease, although
the protocol states that a FibroScan should not be done for in-
mates with certain health conditions. Inmates with a
6                              Nos. 19-1380, 19-1387 & 19-1732

FibroScan of F2 or less are vaccinated against hepatitis A and
B, where appropriate. Inmates with a FibroScan of F2 and
whose APRI score is greater than or equal to .7 are referred to
UIC. Inmates with FibroScan results of F3 or F4 undergo fur-
ther testing before referral to UIC.
    The 2019 Protocol also establishes priorities for the treat-
ment of HCV-positive inmates. IDOC recognized that
“[a]lthough nearly all patients with chronic hepatitis C infec-
tion are candidates for HCV therapy, there may be more ur-
gency to treat certain patients first.” Under the 2019 Protocol,
priority level 1 for HCV treatment includes inmates with any
of the following: cirrhosis; a comorbid medical condition as-
sociated with HCV; a Child Turcotte Pugh score of greater
than or equal to 7; or newly incarcerated inmates who are al-
ready receiving HCV treatment. Priority level 2a includes in-
mates with FibroScan results of F3 or F4, an APRI score of
greater than or equal to 2, Metavir Stage 3 or 4, a hepatitis B
co-infection, or HIV. Priority level 2b includes inmates with a
FibroScan result of F2 and an APRI score of greater than or
equal to .7. Priority level 3 includes inmates with a FibroScan
result of F0 or F1, and an APRI score of less than .7. The level 3
inmates are enrolled in the Hepatitis C Chronic Clinic, where
they are seen every 12 months for (1) a targeted physical and
other lab tests, (2) a FibroScan, (3) the calculation of their
APRI, and (4) HCV education.
    If an inmate is referred to UIC, all lab work and necessary
forms are completed at the prison and forwarded to UIC. Af-
ter the UIC practitioner sees the inmate, prescriptions or or-
ders from UIC are faxed or forwarded to the prison “for med-
ication changes, lab tests to be ordered, studies to be obtained,
or special requests for the IDOC medical team to address.”
Nos. 19-1380, 19-1387 & 19-1732                                  7

    The 2019 Protocol also requires that IDOC’s medical direc-
tor receive a monthly report of inmates determined to be in-
eligible for treatment and the reason why. It requires notifica-
tion of the medical director when there are inmates who are
eligible for treatment but who “are having significant delays
… that could impact their ability to get treated while they are
incarcerated.”
                   D. Lippert Consent Decree
    Also in January 2019, the state defendants (whittled down
to Dr. Shicker, Dr. Meeks, and IDOC) filed a motion to stay
the evidentiary hearing in light of another case, Lippert v.
Ghosh, No. 10-4603 (N.D. Ill.). In Lippert, several IDOC inmates
alleged that they were receiving inadequate medical care for
various medical conditions in violation of the Eighth Amend-
ment. They sought class certification and injunctive and de-
claratory relief. The district court in Lippert certified a class of
“all prisoners in the custody of [IDOC] with serious medical
or dental needs.” In December 2018, the Lippert parties
reached an agreement on a consent decree.
    The consent decree, which was approved in May 2019, ap-
plies to “all prisoners in the custody of [IDOC] with serious
medical or dental needs.” It requires IDOC to implement
changes in several areas, including chronic disease care,
which covers HCV. Under the decree, IDOC will “provide
sufficient nursing staff and clinicians” to complete medical
evaluations within seven business days of an inmate’s arriv-
ing at IDOC’s receiving and classification prisons. If an eval-
uation shows that an inmate needs medical care, IDOC must
ensure follow-up for appropriate care. In addition, the decree
provides for the appointment of a court monitor to “conduct
a staffing analysis and create and implement an
8                              Nos. 19-1380, 19-1387 & 19-1732

Implementation Plan to accomplish the obligations and objec-
tives in this Decree.” Twice each year, the monitor will report
to the parties and the court regarding compliance with the de-
cree.
   The Lippert consent decree includes a dispute resolution
provision. If those plaintiffs believe that the defendants are
not in substantial compliance with any provision of the de-
cree, they are to provide written notice “referencing the spe-
cific provision or provisions” not being complied with to the
defendants, who must then provide a written response within
30 days. If the plaintiffs reject the response, the parties must
meet to try to resolve the dispute. If unsuccessful, the parties
may jointly or individually seek relief from the court.
                  E. February 4, 2019 Order
    We may now return to our case. The district court denied
the state defendants’ motion to stay the January 2019 hearing
in light of the proposed consent decree in Lippert, finding that
there were “no concrete proposals for treatment and handling
of the combined cases.” The court also denied Wexford’s mo-
tion for reconsideration. On January 22 and 23, 2019, the dis-
trict court conducted an evidentiary hearing on Plaintiffs’ mo-
tions for class certification and injunctive relief. Many wit-
nesses testified. Plaintiffs also provided a spreadsheet con-
taining some medical information about the horde of named
plaintiffs (approximately 1,800), including their APRI scores,
which were calculated by a law clerk.
   On February 4, 2019, the district court granted in part and
denied in part Plaintiffs’ motions for class certification and in-
junctive relief. First, the court certified two classes of inmates.
Class 1 includes:
Nos. 19-1380, 19-1387 & 19-1732                                 9

   All current and future prisoners in IDOC custody who
   have been, or will be, diagnosed with chronic hepatitis
   C virus, have at least six months or more remaining [on
   their] sentence, and who have not previously received
   treatment which resulted in a sustained viral response.
Class 2 includes:
   All current and future prisoners in the IDOC custody
   who have been, or will be, diagnosed with chronic hep-
   atitis C virus, have at least one year remaining on their
   sentence, have a fibrosis level of greater than or equal
   to two and an APRI score greater than or equal to .7,
   and have not received direct-acting antiviral drugs.
    The court denied injunctive relief for Class 1, but it granted
an injunction for Class 2, finding that Plaintiffs had demon-
strated “a more than negligible chance that refusing treatment
for inmates with a fibrosis level of at least two could amount
to deliberate indifference.” It acknowledged that the 2019 Pro-
tocol set treatment policies for inmates with a fibrosis level of
two or more, but it concluded that injunctive relief was neces-
sary because otherwise IDOC could abandon the protocol.
Moreover, it said, the 2019 Protocol “does not actually man-
date treatment” because it simply required the referral of in-
mates with a fibrosis level of two or more to UIC for “possible
HCV therapy.” In addition, the protocol set no deadline for
“determining when HCV therapy starts.”
   On the record before it, the court concluded that the bal-
ance of harms “weighs in favor of granting a preliminary in-
junction that adopts, with some reasonable modifications, the
January 2019 protocol.” It ordered Defendants to “commence
immediately the treatment of the class 2 plaintiffs’ Hepatitis
10                             Nos. 19-1380, 19-1387 & 19-1732

C in accordance with” a modified version of the 2019 Protocol.
The new version added three requirements:
     (1) inmates with a fibroscan greater than two and an
     APRI greater than .7 will be referred to UIC and will
     receive HCV therapy unless UIC determines that treat-
     ment is contraindicated; (2) fibroscan and APRI levels
     will be determined within three months of admission
     to the IDOC; (3) HCV therapy will be started within
     three months of a determination that the fibrosis level
     is at least two and the APRI is at least .7.
    Defendants appeal the class certification and injunctive re-
lief. We stayed the district court’s order pending appeal.
                               II
    We review a district court’s decision to grant or deny class
certification for an abuse of discretion. Red Barn Motors, Inc. v.
NextGear Capital, Inc., 

, 1101 (7th Cir. 2019). Alt-
hough this is a deferential standard, it “must also be exacting”
because a decision regarding certification “can have a consid-
erable impact on the playing field of litigation.”

action is “an exception to the usual rule that litigation is con-
ducted by and on behalf of the individual named parties
only.” Califano v. Yamasaki, 

, 700–01 (1979).
    Federal Rule of Civil Procedure 23 requires a plaintiff
seeking class certification to satisfy all four requirements of
Rule 23(a)—numerosity, commonality, typicality, and ade-
quacy of representation—and any one of the general catego-
ries of Rule 23(b). Oshana v. Coca-Cola Co., 

, 513
(7th Cir. 2006). “Failure to meet any of the Rule’s require-
ments precludes class certification.” Arreola v. Godinez, 

, 794 (7th Cir. 2008). The party seeking class
Nos. 19-1380, 19-1387 & 19-1732                               11

certification bears the burden of showing by a preponderance
of the evidence that certification is proper. Bell v. PNC Bank,
Nat. Ass’n, 

, 373 (7th Cir. 2015). The district court
found that Plaintiffs satisfied all the requirements for class
certification. Defendants retort that the proposed class is defi-
cient in all respects.
    We begin our analysis with Rule 23(a)’s numerosity crite-
rion, which requires that the proposed class be “so numerous
that joinder of all members is impracticable.” FED. R. CIV. P.
23(a)(1). A “class can be certified without determination of its
size, so long as it’s reasonable to believe it large enough to
make joinder impracticable and thus justify a class action
suit.” Arnold Chapman & Paldo Sign & Display Co. v. Wagener
Equities Inc., 

, 492 (7th Cir. 2014). “While there is
no magic number that applies to every case, a forty–member
class is often regarded as sufficient to meet the numerosity re-
quirement.” Mulvania v. Sheriff of Rock Island Cnty., 

, 859 (7th Cir. 2017).
   In its February 4 order, the district court found that nu-
merosity was satisfied because “[j]oinder of all IDOC inmates
with Hepatitis C, believed to exceed 2,000, is impractical.”
Plaintiffs add that numerosity is also easily satisfied because
more than 1,800 individual plaintiffs have been added to the
lawsuit. Defendants complain only that the district court did
not cite any evidence showing the number of inmates in-
cluded in each class.
    While the district court ideally would have been more pre-
cise in its order, it did not abuse its discretion in finding that
numerosity was satisfied. Plaintiffs showed that HCV is prev-
alent in prison populations. There are thousands of current
and future IDOC inmates with hepatitis C. The real question
12                            Nos. 19-1380, 19-1387 & 19-1732

is whether each of the two classes is sufficiently numerous for
class treatment. (The identified classes do not sweep in all
IDOC inmates with hepatitis C.)
    For Class 1, we look at the number of current and future
prisoners in IDOC custody with hepatitis C who “have at least
six months or more remaining [on their] sentence, and who
have not previously received treatment which resulted in a
sustained viral response.” While the record does not reflect
how many inmates, if any, have received treatment that re-
sulted in a sustained viral response, part of the problem is the
very lack of data and testing. Given what we do know about
the great number of prisoners with hepatitis C, it was permis-
sible for the district court to infer that there are enough un-
treated inmates to make joinder impracticable.
    Finding numerosity is more difficult with Class 2 because
it is narrower. Nonetheless, there was (barely) enough evi-
dence to support the district court’s finding of numerosity by
a preponderance of the evidence. Recall that Class 2 includes
all current and future prisoners with hepatitis C who “have at
least one year remaining on their sentence, have a fibrosis
level of greater than or equal to two and an APRI score greater
than or equal to .7, and have not received direct-acting antivi-
ral drugs.” At the January 2019 hearing, Plaintiffs provided a
spreadsheet containing medical information for many of the
proposed class members. According to the spreadsheet, hun-
dreds of these inmates have an APRI greater than .7. Despite
significant gaps in the spreadsheet, we are willing to defer to
the district court’s finding that Class 2 is “large enough to
make joinder impracticable.”
    Although others may have seen things differently, the dis-
trict court was also within bounds when it found that the
Nos. 19-1380, 19-1387 & 19-1732                                13

commonality requirement was satisfied for both classes.
Commonality requires a plaintiff to show that “there are ques-
tions of law or fact common to the class.” FED. R. CIV. P.
23(a)(2). A court need find only a single common question of
law or fact, but it needs to identify more than the fact that eve-
ryone suffered as a result of a violation of the same provision
of law. Wal-Mart Stores, Inc. v. Dukes, 

, 350, 359
(2011). Plaintiffs’ claims “must depend upon a common con-
tention … of such a nature that it is capable of classwide res-
olution—which means that determination of its truth or fal-
sity will resolve an issue that is central to the validity of each
one of the claims in one stroke.”

The key to com-
monality is “not the raising of common ‘questions’ ... but, ra-
ther, the capacity of a class-wide proceeding to generate com-
mon answers apt to drive the resolution of the litigation.”

the proposed class are what have the
potential to impede the generation of common answers.”

February 4 order, the district court identified a com-
mon question of law or fact: “whether every inmate with Hep-
atitis C in the IDOC should be treated.” Relying on advances
in medical treatment for hepatitis C, Plaintiffs argue that
every inmate with the disease should be treated as soon as
possible. They urge that IDOC’s current policies, which do
not mandate treatment for all inmates who test positively for
hepatitis C, constitute deliberate indifference to a serious
medical need.
    Defendants counter that there are no questions of law
common to each class because medical care, by its nature, is
individualized. Hepatitis C, they point out, progresses at dif-
ferent rates, and so treatment for each person requires discre-
tion and will differ based on that person’s medical history.
14                             Nos. 19-1380, 19-1387 & 19-1732

This discretion is built into IDOC’s policy. Moreover, inmates
present with different risk factors and different responses to
treatment, Defendants argue, and so certification of any class
involving medical care is tricky and potentially dangerous.
    Although the physical symptoms and progression suf-
fered by each inmate undoubtedly vary, there is still a general
question that can yield a common answer. Plaintiffs assert
Eighth and Fourteenth Amendment challenges to Defend-
ants’ system-wide policies and practices, which allegedly re-
flect deliberate indifference to Plaintiffs’ serious medical
needs. Thus, “[w]hat all members of the [class] have in com-
mon is their alleged exposure, as a result of specified
statewide [IDOC] policies and practices that govern the over-
all conditions of health care services[,] to a substantial risk of
serious future harm to which the defendants are allegedly de-
liberately indifferent.” Parsons v. Ryan, 

, 678 (9th
Cir. 2014). “[A]lthough a presently existing risk may ulti-
mately result in different future harm for different inmates—
ranging from no harm at all to death—every inmate suffers
exactly the same constitutional injury when he is exposed to
a single statewide … policy or practice that creates a substan-
tial risk of serious harm.”

question for each class is thus the following:
whether the specified policies and practices to which all IDOC
inmates are subjected expose them to a substantial risk of
harm. “These policies and practices are the ‘glue’ that holds
together the putative class[;] either each of the policies and
practices is unlawful as to every inmate or it is not. That in-
quiry does not require us to determine the effect of those pol-
icies and practices upon any individual class member (or class
Nos. 19-1380, 19-1387 & 19-1732                                15

members) or to undertake any other kind of individualized
determination.”

not fare as well with Rule 23(a)’s typicality
and representation requirements. Looking first at adequacy of
representation, Rule 23(a)(4), we must decide whether the
Plaintiffs have shown that the “representative parties will
fairly and adequately protect the interests of the class.” FED.
R. CIV. P. 23(a)(4). A class representative must be part of the
class and must “possess the same interest and suffer the same
injury” as the other class members. 

.
    We are stymied at the outset because, despite its certifica-
tion of the two classes, the district court failed to name a rep-
resentative for either class or to explain this omission. We thus
have no way to assess adequacy of representation. On the as-
sumption that the court would have accepted Plaintiffs’ pro-
posed representatives, we have different problems: this rec-
ord does not reveal whether they would be adequate. In their
most recent certification motion, Plaintiffs simply listed nine
potential representatives’ names and their (sometimes for-
mer) places of incarceration. This is not enough, nor does an-
ything else fill the gap.
    With respect to Class 1, we need a representative or repre-
sentatives who are “current and future prisoners in IDOC cus-
tody,” who have been diagnosed with hepatitis C, who “have
at least six months or more remaining [on their] sentence, and
who have not previously received treatment which resulted
in a sustained viral response.” Based on IDOC’s website, at
the time of the district court’s February 4 order, six of the nine
proposed class representatives were no longer in custody, and
so they cannot serve. As for the other three, Plaintiffs
16                            Nos. 19-1380, 19-1387 & 19-1732

provided no evidence about whether any had received treat-
ment that resulted in a sustained viral response.
    For Class 2, the representatives must be “current and fu-
ture prisoners in the IDOC custody” and must “have at least
one year remaining on their sentence, have a fibrosis level of
greater than or equal to two and an APRI score greater than
or equal to .7, and have not received direct-acting antiviral
drugs.” Once again, Plaintiffs’ evidence falls short. The six no
longer in custody are equally inappropriate for Class 2. In ad-
dition, five of those six do not have a fibrosis level of two or
more and an APRI score of .7 or higher, and thus they do not
fit the criteria for Class 2. (There is no information about the
fibrosis level or APRI score of the sixth person.) Of the three
proposed representatives who remain in custody, Plaintiffs’
most recent data shows that two have an APRI score lower
than .7, and there is no evidence about the third’s APRI score.
“Rule 23 does not set forth a mere pleading standard. A party
seeking class certification must affirmatively demonstrate his
compliance with the Rule … .” 

.
Plaintiffs have thus failed to demonstrate that their proposed
class representatives are adequate. The district court erred by
skipping this important step and finding, in the face of these
facts, that Rule 23(a)(4) was satisfied.
    The lack of a named representative also makes it impossi-
ble to find typicality. To satisfy the typicality requirement,
Plaintiffs must show that the “claims or defenses of the repre-
sentative parties are typical of the claims or defenses of the
class.” FED. R. CIV. P. 23(a)(3). Typicality requires “enough
congruence between the named representative’s claim and
that of the unnamed members of the class to justify allowing
the named party to litigate on behalf of the group.” Spano v.
Nos. 19-1380, 19-1387 & 19-1732                              17

Boeing Co., 

, 586 (7th Cir. 2011). As there is no
named representative, there is no way to compare anyone’s
claims with those of the absentees. If we look at the nine prof-
fered representatives, typicality is missing because it appears
that they do not, or no longer, belong to either class.
    As we said at the outset, a class can be certified only if it
meets all four criteria in Rule 23(a). Although we have no
quarrel with the district court’s findings of numerosity and
commonality, Plaintiffs’ showing of typicality and adequacy
of representation fall short. We thus conclude that the district
court abused its discretion in certifying these two classes.
                              III
    We now turn to the preliminary injunction. Before we ad-
dress the merits, we take a moment to explain which plaintiffs
are entitled to seek this remedy. Having rejected both of the
classes the district court certified, the classes are not the
proper parties at this point. But, as we noted, unlike in most
class actions, where only a small number of individual plain-
tiffs are suing on their own behalf and as possible class repre-
sentatives, in this case an extraordinary number of individual
plaintiffs joined the case personally—not as a class repre-
sentative, not as part of a class, but independently. We do not
know as much about these people as would be ideal, but
Plaintiffs proffered a spreadsheet at the hearing held on Janu-
ary 22 and 23, 2019, and that spreadsheet contains just enough
information to assure us that many of the more than 1,800 in-
dividual plaintiffs meet the following criteria:
       •   currently in IDOC custody,
       •   diagnosed with chronic hepatitis C virus,
       •   at least one year remaining on their sentence,
18                             Nos. 19-1380, 19-1387 & 19-1732

       •   fibrosis level of greater than or equal to two
       •   an APRI score greater than or equal to .7,
       •   have not received direct-acting antiviral drugs.
Plaintiffs’ spreadsheet shows, for example, that Brandon
Blasa, Ben McCreadie, Charles Sultan, and Nazim Useni all
have hepatitis C, a fibrosis level of at least two, and an APRI
score greater than .7. IDOC’s website indicates that these in-
mates currently remain in IDOC custody, and they all have at
least one year remaining on their sentences. While we do not
know whether they have received direct-acting antiviral
drugs, given that the 2019 Protocol was implemented only re-
cently, we can assume that at least some have not. In other
words, had Class 2 survived, they likely would have fallen
within its scope. We therefore proceed to consider the merits
of the preliminary injunction.
    In its February 4 order, the district court granted relief to
inmates in Class 2, and thus to individual plaintiffs meeting
those criteria. It ordered Defendants to “commence immedi-
ately the treatment of the class 2 plaintiffs’ Hepatitis C in ac-
cordance with” the 2019 Protocol. As we noted earlier, the
court also slightly revised the 2019 Protocol by adding stricter
referral and treatment requirements.
    We review a district court’s decision to grant a preliminary
injunction for an abuse of discretion. Girl Scouts of Manitou
Council, Inc. v. Girl Scouts of U.S. of Am., Inc., 

,
1086 (7th Cir. 2008). “A preliminary injunction is an extraordi-
nary remedy never awarded as of right.” Winter v. Nat. Res.
Def. Council, Inc., 

, 24 (2008). In fact, a “preliminary
injunction is an exercise of a very far-reaching power, never
Nos. 19-1380, 19-1387 & 19-1732                                 19

to be indulged in except in a case clearly demanding it.” Girl

.
    In Winter, the Supreme Court instructed that “[a] plaintiff
seeking a preliminary injunction must establish that he is
likely to succeed on the merits, that he is likely to suffer irrep-
arable harm in the absence of preliminary relief, that the bal-
ance of equities tips in his favor, and that an injunction is in
the public 

. If the plaintiff fails to meet
these requirements, the court must deny the injunction. The
mere possibility of irreparable injury is not enough.

In each case, “courts must balance the competing claims of
injury and must consider the effect on each party of the grant-
ing or withholding of the requested relief.”

(cleaned
up).
    We start with likelihood of success on the merits. This case
turns on the medical care that the individual plaintiffs still be-
fore us received and whether it was the result of deliberate
indifference—i.e. the knowing disregard of a substantial risk
of serious harm. Petties v. Carter, 

, 728 (7th Cir.
2016). “To determine if the Eighth Amendment has been vio-
lated in the prison medical context, we perform a two-step
analysis, first examining whether a plaintiff suffered from an
objectively serious medical condition, and then determining
whether the individual defendant was deliberately indiffer-
ent to that condition.”

    Defendants contend that the district court erred in finding
that the individual plaintiffs showed the necessary likelihood
of success on the merits. They emphasize that the 2019 Proto-
col provides for treatment of F2, F3, and F4 inmates and, be-
cause the protocol was created by experts at UIC, it is impos-
sible to say that the Defendants were deliberately indifferent.
20                              Nos. 19-1380, 19-1387 & 19-1732

    We can assume for the sake of argument that these indi-
vidual plaintiffs have shown an adequate chance of success.
Hepatitis C is a serious medical condition. Under the 2019
Protocol, inmates in this class are “refer[red] to UIC for possi-
ble HCV therapy,” but treatment is not guaranteed, and it
comes only after the disease has progressed. Inmates with the
described characteristics, we accept for present purposes,
have a chance of succeeding because their disease has pro-
gressed far enough that treatment should no longer be discre-
tionary (as long as there are no counterindications).
    This takes us to irreparable harm, which we have defined
as harm that “cannot be repaired” and for which money com-
pensation is inadequate. Graham v. Med. Mut. of Ohio, 

, 296 (7th Cir. 1997). “The moving party must demonstrate
that he will likely suffer irreparable harm absent obtaining
preliminary injunctive relief.” Whitaker By Whitaker v. Kenosha
Unified Sch. Dist. No. 1 Bd. of Educ., 

, 1044 (7th
Cir. 2017) (emphasis added). This requires “more than a mere
possibility of harm.”

see 

. “Is-
suing a preliminary injunction based only on a possibility of
irreparable harm is inconsistent with our characterization of
injunctive relief as an extraordinary remedy that may only be
awarded upon a clear showing that the plaintiff is entitled to
such relief.” 

.
    The individual plaintiffs have not cleared that hurdle. The
district court justified its contrary finding in its February 4 or-
der, which stated that “the fibrosis test can be a rough indica-
tor—a level two may actually be a level three. Combine this
uncertainty with the inevitable prison delays means that wait-
ing until level three could well create a substantial risk to
Nos. 19-1380, 19-1387 & 19-1732                              21

inmates of liver damage, liver cancer, and painful extrahe-
patic conditions.”
    We conclude, however, that the court’s analysis was
flawed insofar as it found only that a substantial risk “could”
arise, not that irreparable harm was likely. The evidence pre-
sented during the preliminary injunction hearings was like-
wise equivocal. Many of the doctors testified that hepatitis C
is a slow-moving disease and that rates of progression vary
between individuals. Dr. Patel, a physician at UIC, testified
that there was “probably not significant harm without treat-
ment from stage 2 to stage 3.” In 2016, Dr. Batey, a court-re-
cruited expert, testified that sometimes hepatitis C does not
progress for years in patients who do not undergo treatment.
This evidence does not show likely harm.
    To be sure, the fact that a disease may progress slowly
does not mean that IDOC may refuse to treat it. But IDOC is
not refusing to treat inmates with hepatitis C. The 2019 Proto-
col lists very specific guidelines for diagnosing and treating
inmates with hepatitis C. The individual plaintiffs have not
shown that the treatment under the 2019 Protocol will likely
cause them irreparable harm.
    The district court also feared that “[w]ithout an injunctive
order, the IDOC may abandon the current protocol.” It was
concerned that the 2019 Protocol “does not actually mandate
treatment. Instead, inmates with level two or greater are re-
ferred to UIC for ‘possible HCV therapy.’” The existence of
the Lippert consent decree, however, provides the assurance
that the court wanted. The Lippert class includes “all prisoners
in the custody of [IDOC] with serious medical or dental
needs.” The consent decree covers inmates with chronic dis-
eases, specifically including HCV. It includes a dispute
22                             Nos. 19-1380, 19-1387 & 19-1732

resolution provision, which permits inmates to provide notice
to IDOC if they believe that IDOC is not in substantial com-
pliance with the decree. If the inmates and IDOC cannot reach
a resolution, either side may seek relief from the district court
overseeing Lippert. Thus, if IDOC were to “abandon the cur-
rent protocol,” the individual plaintiffs in this case could seek
relief. And if inmates are not receiving treatment, but are
simply being “referred to UIC,” they could seek relief under
the Lippert consent decree. The individual plaintiffs therefore
have not shown that they are likely to suffer irreparable harm
absent the preliminary injunction, and thus it was error to
grant injunctive relief.
                               IV
    The district court abused its discretion in certifying both
classes and in granting preliminary injunctive relief for Class
2 (or, more precisely, for the individual plaintiffs who meet
that class’s criteria). We therefore VACATE its order and re-
mand for further proceedings consistent with this opinion.