Kosilek v. Spencer

United States Court of Appeals
For the First Circuit

No. 12-2194
MICHELLE KOSILEK,

Plaintiff, Appellee,

v.

LUIS S. SPENCER, Commissioner of the
Massachusetts Department of Correction,

Defendant, Appellant.

APPEAL FROM THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF MASSACHUSETTS

[Hon. Mark L. Wolf, U.S. District Judge]

Before

Lynch, Chief Judge,
Torruella, Howard, Thompson and Kayatta, Circuit Judges.

Richard C. McFarland, Legal Division, Department of
Correction, with whom Nancy Ankers White, Special Assistant
Attorney General, was on brief for appellant.
Joseph L. Sulman, with whom David Brody, Law Office of Joseph
L. Sulman, Frances S. Cohen, Jeff Goldman, Christina Chan, and
Bingham McCutchen LLP, were on brief for appellee.
Andrew D. Beckwith, on brief for the Massachusetts Family
Institute, amicus curiae in support of appellant.
Daniel V. McCaughey, Cori A. Lable, Kristin G. Ali and Ropes
& Gray LLP, on brief for World Professional Association for
Transgender Health, Mental Health America, Callen-Lorde Community
Health Center, Whitman-Walker Health, GLMA: Health Professionals
Advancing LGBT Equality, and Mazzoni Center, amici curiae in
support of appellee.
Matthew R. Segal, Joshua Block, LGBT Project, and David C.
Fathi, National Prison Project, on brief for American Civil
Liberties Union, American Civil Liberties Union of Massachusetts,
Legal Aid Society, Harvard Prison Legal Assistance Project,
Prisoners' Legal Services of New York, and Prisoners' Legal
Services of Massachusetts, amici curiae in support of appellee.
Jennifer Levi and Bennett H. Klein, on brief for Gay & Lesbian
Advocates & Defenders, EqualityMaine, Human Rights Campaign,
MassEquality, Massachusetts Transgender Political Coalition,
National Center for Transgender Equality, National Gay & Lesbian
Task Force, and Transgender New Hampshire, amici curiae in support
of appellee.

Opinion En Banc

December 16, 2014

-2-
 TORRUELLA, Circuit Judge. This case involves important

issues that arise under the Eighth Amendment to the U.S.

Constitution. We are asked to determine whether the district court

erred in concluding that the Massachusetts Department of Correction

("DOC") has violated the Cruel and Unusual Punishment Clause of the

Eighth Amendment by providing allegedly inadequate medical care to

prisoner Michelle Kosilek ("Kosilek"). More precisely, we are

faced with the question whether the DOC's choice of a particular

medical treatment is constitutionally inadequate, such that the

district court acts within its power to issue an injunction

requiring provision of an alternative treatment -- a treatment

which would give rise to new concerns related to safety and prison

security.

After carefully considering the community standard of

medical care, the adequacy of the provided treatment, and the valid

security concerns articulated by the DOC, we conclude that the

district court erred and that the care provided to Kosilek by the

DOC does not violate the Eighth Amendment. We therefore reverse

the district court's grant of injunctive relief, and we remand with

instructions to dismiss the case.

I. Background

This litigation has now spanned more than twenty years

and produced several opinions of significant length. See Kosilek

v. Spencer, 889 F. Supp. 2d 190 (D. Mass. 2012) ("Kosilek II");

-3-
Kosilek v. Maloney, 221 F. Supp. 2d 156 (D. Mass. 2002) ("Kosilek

I"). In light of the expansive record, we recite here only the

facts necessary to clarify the issues on appeal.

A. Michelle Kosilek

Michelle Kosilek -- born in 1949 as Robert Kosilek -- is

an anatomically male prisoner in her mid-sixties who suffers from

gender identity disorder ("GID")1 and self-identifies as a female.

In 1992 Kosilek was convicted of first-degree murder and sentenced

to a term of life imprisonment without parole for the 1990

strangulation of her then-wife, Cheryl McCaul, whose body was found

abandoned in the backseat of a vehicle at a local shopping mall.

See Commonwealth v. Kosilek, 423 Mass. 449, 668 N.E.2d 808 (1996).

While awaiting trial for McCaul's murder, Kosilek twice attempted

to commit suicide. She also once tied a string around her

testicles in an attempt at self-castration, but removed the string

when it became painful. Since 1994, Kosilek has been housed at

MCI-Norfolk, a medium security male prison in Massachusetts.

Throughout the twenty-year duration of her incarceration at MCI-

Norfolk, Kosilek has not attempted to harm herself.

1
The term "gender identity disorder" has recently been replaced
with the term "gender dysphoria" in the medical community. See Am.
Psychiatric Ass'n, Gender Dysphoria,
http://www.dsm5.org/documents/gender%20dysphoria%20fact%20sheet.pdf
(last visited June 3, 2014). To maintain consistency with prior
related litigation and evidence in the record, we continue to use
the term "GID" in this opinion.

-4-
B. Kosilek I

Kosilek first sued the DOC in 1992, alleging that its

failure to provide direct treatment for her GID was a violation of

the Eighth Amendment. At that time, Kosilek was receiving only

"supportive therapy" to cope with the distress caused by her GID.

Kosilek initially sought both damages and injunctive relief

requiring the DOC to provide her with sex reassignment surgery

("SRS"), although only her claim for injunctive relief survived to

trial.

The district court issued a decision in 2002, in which it

concluded that Kosilek had proven the existence of a serious

medical need and had shown that her then-current treatment plan was

inadequate. The court concluded, however, that the DOC was unaware

that a failure to provide additional treatment to Kosilek might

result in serious harm. Moreover, it held that the DOC's failure

to provide treatment was rooted, at least in part, in "sincere

security concerns." As a result, the court ruled that the DOC was

not in violation of the Eighth Amendment.

Despite finding for the DOC, the district court's opinion

made clear that Kosilek required additional treatment for her GID,

and that the DOC would need to develop and implement an improved

treatment plan. The court warned that a failure to provide

treatment in the future, now that the DOC was on notice of the

-5-
potential for harm if only "supportive therapy" was provided, could

amount to an Eighth Amendment violation.

C. The DOC offers treatment

The DOC responded to Kosilek I by revamping its policy

for GID treatment. In the past, the DOC had adopted a policy of

"freezing" a prisoner's treatment at whatever level that prisoner

had attained prior to incarceration. Hormonal treatment, for

example, would be available only to prisoners who had been

prescribed hormones prior to incarceration. In place of this

"freeze-frame" policy, after Kosilek I the DOC adopted a plan that

allowed prisoners to receive additional treatment beyond the level

of that received before entering prison, when such care was

medically required. Under this new plan, medical recommendations

would be made by the University of Massachusetts Correctional

Health Program ("UMass"), a health-services provider contracted by

the DOC. The DOC Commissioner and the DOC Director of Health

Services were responsible for assessing whether any change in

treatment would create increased security concerns.

Kosilek was evaluated by Dr. David Seil, a gender-

identity specialist, who prescribed a course of treatment to

alleviate the mental distress -- often referred to as "dysphoria"

-- associated with her GID. In line with Dr. Seil's

recommendations, in 2003 the DOC began providing Kosilek with

significant ameliorative treatment aimed at directly addressing the

-6-
mental distress caused by GID. In addition to continued mental

health treatment, she was provided female, gender-appropriate

clothing and personal effects, and electrolysis was performed to

permanently remove her facial hair.2 Kosilek also began a course

of hormonal treatments recommended by an endocrinologist. These

treatments resulted in "breast development and shrinkage of her

testicles." All of the treatments described continue to be offered

to Kosilek to the present day.

D. Consideration of SRS

In line with the Harry Benjamin Standards of Care ("the

"Standards of Care" or "the Standards"),3 Dr. Seil recommended that

Kosilek be considered for SRS after one year of hormonal

2
Facial hair removal was delayed because of difficulty finding a
provider that was willing to perform these services on Kosilek.
The minutes of the DOC's Executive Staff Meetings show that they
proactively sought out service providers throughout this period of
delay, and electrolysis was completed in November 2004.
3
The Standards of Care are a set of treatment recommendations
issued by the Harry Benjamin International Gender Dysphoria
Association that provide guidance on the treatment of individuals
with GID. Relevant to Kosilek II is the sixth version of the
Standards of Care. See Harry Benjamin Int'l Gender Dysphoria
Ass'n, Standards of Care for Gender Identity Disorders, Sixth
Version (2001) ("Standards of Care"). A seventh version of the
Standards of Care was published in 2011, and adopts the Harry
Benjamin Association's new name. See World Professional Ass'n for
Transgender Health ("WPATH"), Standards of Care for the Health of
Transsexual, Transgender, and Gender–Nonconforming People, Version
7 (2011). The Standards of Care "are intended to provide flexible
directions for the treatment" of GID, and state that "[i]ndividual
professionals and organized programs may modify" the Standards'
requirements in response to "a patient's unique . . . situation" or
"an experienced professional's evolving [treatment methodology]."
Standards of Care at 1-2 (emphasis added).

-7-
treatment.4 Accordingly, in 2004 the DOC began the process of

finding an appropriate professional to evaluate Kosilek's

eligibility for, and the necessity of, SRS. At the DOC's Executive

Staff Meetings there was some debate regarding who should be hired

to conduct this evaluation. The UMass Mental Health Program

Director, Dr. Kenneth Appelbaum, suggested that the DOC consult

with the Fenway Community Health Center (the "Fenway Center"). The

Fenway Center is a Boston-based facility focused on serving the

lesbian, gay, bisexual, and transgender community. In contrast,

the DOC's Director of Mental Health and Substance Abuse Services,

Gregory Hughes ("Hughes"), suggested consulting with Cynthia

Osborne ("Osborne"), a gender identity specialist employed at the

Johns Hopkins School of Medicine who had experience working with

other departments of correction regarding GID treatment.

Hughes expressed concern with using the Fenway Center

because of "the perception that their approach was to come out with

recommendations that globally endorsed a full panoply of

treatments." It was thought that Osborne, in contrast, "may do

more objective evaluations." Dr. Appelbaum noted, however, that

the Fenway Center's approach was, to his knowledge, probably "more

4
This treatment plan aligns with the Standards of Care's triadic
sequence for GID treatment. This sequence begins with diagnosis
and the provision of therapy, progresses to endocrine treatments,
and culminates with consideration of SRS after at least one full
year living a "real life experience" in the preferred gender role.
Many individuals with GID choose not to complete the full sequence.

-8-
the norm than the exception." The DOC also recognized that having

a Boston-based treatment provider might more easily facilitate the

process of Kosilek's evaluation.

The Fenway Center was retained by the DOC, and Kosilek

was evaluated by Kevin Kapila, M.D., and Randi Kaufman, Psy.D., in

a ninety-minute interview. Drs. Kapila and Kaufman also reviewed

Kosilek's medical records. On February 24, 2005, they issued a

report recommending that Kosilek receive SRS (the "Fenway Report").

The Fenway Report acknowledged Kosilek's positive response to the

treatment provided by the DOC.

Her joy around being feminized through hormone
therapy, facial and body hair removal, and her
ability to have access, and to dress in,
feminine attire and make-up is palpable.
These responses further suggest that being
able to express herself as female has been
helpful in alleviating her gender
dysphoria. . . . [I]t is clear that her
increasingly feminine presentation has been
beneficial to her psychologically.

Nonetheless, it also emphasized that Kosilek remained significantly

distressed by "having male genitalia, as well as not having female

genitalia." In light of this continuing distress, the Fenway

Center doctors stated that "it is quite likely that Michelle will

attempt suicide again if she is not able to change her anatomy."

The report also concluded that Kosilek had fully progressed through

the Standards of Care's triadic sequence, and that she "appear[ed]

to be ready" for SRS. SRS, the doctors believed, would most likely

"allow Michelle to have full relief from the symptoms of gender

-9-
dysphoria" and would quite possibly "increase her chance for

survival" by greatly decreasing the potential for future suicidal

ideation.

The Fenway Report was received by the DOC and reviewed by

Dr. Appelbaum and his UMass colleague, Dr. Arthur Brewer. The

UMass doctors informed the DOC that they found no clear

contraindications to SRS, but noted that they were "unaware of any

other case in which an inmate has undergone sex reassignment

surgery while incarcerated."

After considering the information from UMass, the DOC

decided to have Osborne conduct a peer review of the Fenway Report.

In a letter to Osborne, the DOC stated that it was requesting her

services because "[t]he treatment of Gender Identity Disorder

within a correctional environment is a complicated issue and one

that the Department takes very seriously. We are aware of the

substantial expertise you possess in this area and hope that you

can provide us with assistance in determining appropriate

-10-
treatment."5 On April 12, 2005, the DOC sent Osborne copies of all

previous medical evaluations of Kosilek.

On April 28, 2005, the DOC Director of Health Services,

Susan Martin ("Martin"), wrote UMass, stating her concern that

UMass had not "address[ed] the lack of detail, clarity and specific

recommendations in the evaluation done by the Fenway Clinic," and

had failed to provide an independent recommendation as to the

appropriateness of surgery. She also asked for specific logistical

information, including a list of doctors who might provide the

surgery, what procedures would be performed, and what recovery time

could be expected.

On May 10, 2005, Drs. Appelbaum and Brewer replied,

indicating that they deferred to the Fenway Center's recommendation

of surgery, as they were not experts in the area of SRS -- a

medical procedure specifically excluded from their contract to

provide services to the DOC. They provided a preliminary list of

surgeons to consider, none of whom were licensed to practice

medicine in Massachusetts.

5
Osborne previously worked with the Virginia and Wisconsin
Departments of Correction regarding their treatment of prisoners
with GID. It is unclear from the record whether the Fenway Center
had previously developed treatment plans for GID within a
penological setting. When the DOC asked what consideration the
Fenway Center gave to issues such as "criminal history [and]
violence against women," the center responded that "independent of
other psychological disorders [Fenway experts] don't consider
criminal history, homicide, [or] brutality." On January 5, 2005 --
before the Fenway Center released its report -- the DOC's Director
of Health Services "expressed concern" about these omissions.

-11-
 On May 20, 2005, Osborne finished her peer review of the

Fenway Report. She began by making clear that her review was

limited to reading and evaluating the reports of others. As a

result, she could not independently diagnose Kosilek, but she

agreed with the conclusion that Kosilek suffered from GID. Still,

she disagreed with what she believed to be a lack of

comprehensiveness in the report and an inclination to minimize the

possibility of comorbid conditions. Namely, Osborne highlighted

that Kosilek had previously been diagnosed with Antisocial

Personality Disorder, a diagnosis neither confirmed nor denied by

the Fenway Report, and that the report included no indication that

Kosilek had been assessed for other pathologies likely to lead to

self-harming behavior. Osborne expressed belief that threats of

self-harm or suicide should serve as a contraindication to surgery,

and that such threats were not a valid or clinically acceptable

justification for surgery. In consequence, she disagreed with the

Fenway Center's statements that surgery was medically necessary as

a means to diminish the likelihood that Kosilek would attempt

suicide in the future.

Osborne's report also highlighted that the Standards of

Care admit of flexible application, and noted that the Standards

state that "the diagnosis of GID invites the consideration of a

variety of therapeutic options, only one of which is the complete

therapeutic triad." She emphasized that "[t]here is currently no

-12-
universal professional consensus regarding what constitutes medical

necessity in GID."

In reference to the Standards of Care's application in a

penological setting, Osborne noted that the Standards of Care

include a criterion that candidates for SRS exhibit "satisfactory

control of problems such as sociopathy, substance abuse, psychosis

and suicidality." She believed that this requirement was

inherently in conflict with the Standard's application to

incarcerated persons, as she felt incarceration indicated a lack of

mastery over such antisocial leanings. Moreover, Osborne noted

that non-incarcerated individuals often face external constraints

in their choice of treatments or determine, as a result of their

"real life experience," that other, non-invasive treatments are

personally preferable to SRS. In consequence, she felt that it was

unrealistic for inmates to expect "that prison life [would] provide

no constraints or obstacles to cross gender preferences" and that

it was "outside[] the bounds of good clinical practice" for care

providers to try to meet this expectation. Given the isolation

attendant to incarceration, Osborne also emphasized that prisoners

might often lack awareness of the frequency with which individuals

choose alternative treatments over SRS.

After considering Osborne's peer review, Martin again

reached out to the doctors at UMass. On May 25, 2005, she

expressed continuing concern with the Fenway Report, highlighting

-13-
that Osborne's peer review had raised at least three questions

regarding the report's thoroughness: (1) why the report omitted

consideration of potential comorbidities; (2) why the report did

not rely on formal psychological testing, but only an in-person

interview; and (3) why Kosilek's claims that she would likely seek

to end her life if not provided with SRS were seen to justify,

rather than serve as a contraindication to, surgery. Martin also

expressed dissatisfaction that "the February 24, 2005 evaluation by

the Fenway Clinic does not indicate whether sex reassignment

surgery is a medical necessity for Michele [sic] Kosilek" and

"fails to adequately address the issue of whether the current

treatment provided to Kosilek provides sufficient relief of the

symptoms of gender dysphoria."

A response from Drs. Appelbaum and Brewer came on

June 14, 2005. The doctors made clear that they were not experts

in the treatment of GID, and that they deferred to the Fenway

Center's treatment recommendation. Referring to the differences

between the preferred treatment plans of the Fenway Center and

Osborne, the doctors reminded Martin that Osborne's report had

emphasized the "dearth of empirical research upon which to base

treatment decisions" for GID and had highlighted the lack of

"professional consensus" regarding the "medical necessity" of SRS.

The Fenway Center issued a follow-up report aimed at

answering Osborne's critique of its initial recommendation. In

-14-
this report, Drs. Kapila and Kaufmann noted that suicidal ideation

was common among individuals suffering from GID, and that it often

decreased with the provision of care. Therefore, the likelihood

that Kosilek would become suicidal if denied surgery was, to the

doctors, not a contraindication to her eligibility, but instead was

a symptom that could be alleviated by provision of SRS. The

doctors also disagreed with Osborne's belief that incarceration was

a significant contraindication to surgery, noting that the

Standards of Care specifically state that "[p]ersons who are

receiving treatment for [GID] should continue to receive

appropriate treatment . . . after incarceration. For example,

those who are receiving psychotherapy and/or cross-sex hormonal

treatments should be allowed to continue this medically necessary

treatment . . . ."

The Fenway Center doctors further discussed their belief

that a key step of the triadic sequence, the "real-life

experience," could occur in prison. This treatment prerequisite

requires that an individual live full-time in their preferred

cross-gender role for at least one year prior to being deemed

eligible for SRS. The purpose of this requirement is ensure that

GID patients have an opportunity to experience a full measure of

life in a cross-gender role, including the social scrutiny that may

arise among professional counterparts and peers. Prison, the

Fenway Center's doctors surmised, might be considered a "more

-15-
stringent" real-life experience, because a prisoner's gender

presentation would be subject to full-time monitoring by prison

personnel and other inmates. The report concluded by reiterating

the Fenway Center's recommendation that Kosilek receive SRS. The

doctors recognized that performing "such a procedure would . . .

bring up issues of housing and safety," but emphasized that

"hormone therapy and [SRS] are the only clinical treatments found

to be effective for GID."

E. The DOC's Security Report

On April 25, 2005, the district court issued an order

requiring that the DOC conduct a review of any potential safety and

security concerns arising from the provision of SRS. In the next

month, the DOC worked to formalize its security concerns into a

report, which it eventually submitted to the court on June 10,

2005. As made clear by the minutes of the DOC's staff meetings,

however, these security issues were a topic of discussion prior to

the court's order. Previously, on January 5, 2005, the meeting

attendees had discussed how and if Kosilek's prior violent acts

against her wife should impact their evaluation and treatment plan.

On April 20, 2005, the parties discussed potential security

concerns that would arise should Kosilek be housed, post-surgery,

in MCI-Framingham, Massachusetts' only female prison. During that

meeting, DOC personnel noted that they were prepared to provide an

-16-
evaluation of general climate and security concerns implicated by

the provision of surgery.

On May 19, 2005, DOC Commissioner Kathleen Dennehy

("Dennehy") convened a meeting with the Superintendent of MCI-

Norfolk, Luis Spencer ("Spencer"), and the Superintendent of MCI-

Framingham, Lynne Bissonnette ("Bissonnette"), as well as the DOC's

legal counsel. The purpose of this meeting was to formally discuss

the security concerns previously expressed by both superintendents

in phone conversations with Dennehy. It was also an opportunity to

begin preparation of the report requested by the district court.

The report focused mainly on issues of safety and

security surrounding Kosilek's post-operative housing. Dennehy

conveyed concern regarding housing Kosilek at MCI-Norfolk, noting

that approximately twenty-five percent of male offenders in the

Massachusetts prison system are classified as sex offenders and

concluding that "Kosilek would clearly be a target for assault and

victimization in a male prison." The report also expressed

concerns with housing Kosilek at MCI-Framingham, including the

absence of single-bed cells, such that all inmates had to share

cells, and the possibility that Kosilek's presence might exacerbate

mental distress among the significant portion of MCI-Framingham's

population that had previously experienced domestic abuse and

trauma at the hands of male partners.

-17-
 Given the stated infeasibility of housing Kosilek in the

general population of either MCI-Framingham or MCI-Norfolk, the

report considered segregated housing in a protected ward. It

expressed concern, however, about the possible deleterious impact

on Kosilek's mental health caused by any housing solution that

required long-term isolation. The report also noted that it was

not within the DOC's ability to create a special ward for prisoners

with GID, given that these prisoners present a significant range of

criminal histories, security ratings, and treatment needs that are

antithetical to co-housing.

On June 10, 2005, citing both its internal review of

safety and security and Osborne's reported concerns regarding the

appropriateness of SRS, the DOC informed the district court that it

had chosen to continue Kosilek's current ameliorative treatment,

but not to provide her with SRS.

F. Kosilek II

Trial commenced on May 30, 2006, with what would be the

first of three rounds of testimony. For the sake of clarity and

concision, we summarize this testimony topically, rather than

temporally. We begin with evidence regarding the standard of care

for treatment of GID.

-18-
 1. Testimony related to medical necessity

a. Initial testimony

First to testify in 2006 was an expert witness for

Kosilek, Dr. George Brown, who had previously evaluated Kosilek in

2001 and was an author of the Standards of Care. Prior to

testifying, Dr. Brown issued a written report assessing Kosilek’s

readiness for surgery and evaluating her current mental and

physical presentation, as compared with 2001. Dr. Brown noted that

Kosilek consistently presented as female and that "[a]ccess to

makeup and female undergarments, laser hair removal, along with

hormonal treatments . . . have all seemed to significantly

reinforce and consolidate the outward expression of [Kosilek's]

gender identity as female." Other positive effects of treatment

were also described:

Hormonal treatments have resulted in obvious
breast growth since my last assessment,
decrease in upper body strength, increase in
hip size, changes in amount and texture of
body hair, skin texture changes, testicular
volume decrease, and a large reduction in
spontaneous erections . . . .
Psychologically, the effects of these combined
treatments have [included] . . . resolution of
depression, resolution of suicidality and
suicide gestures and attempts, improved mood
with reduction in irritability, anxiety, and
depression . . . .

Based on his observations, Dr. Brown concluded that Kosilek was

eligible for SRS, having met all of the readiness criteria.

-19-
 Before the court, Dr. Brown's testimony emphasized that

the provision of female clothing and effects, hair removal, and

hormones had resulted in a lessening of "the severity of

[Kosilek's] dysphoria." According to Dr. Brown "[s]he was clearly

less depressed, less anxious, less irritable . . . . She was not

suicidal . . . ." Despite these significant improvements, Dr.

Brown testified that he believed SRS to be an appropriate and

"medical[ly] necessary component" of Kosilek's treatment. He

related instances in which incarcerated persons who could not

complete the triadic sequence6 exhibited an increase in negative

symptoms, including a resurgence of self-harming behavior.

Dr. Brown further testified that, if not granted surgery,

he believed Kosilek’s feelings of "hopelessness will intensify,"

and that she would likely attempt suicide. In reaching this

conclusion, Dr. Brown emphasized that "other parts of the treatment

plan [e.g., hormones, hair removal, and the provision of female

clothing] . . . all contribute in their own way to a level of

improvement." Nonetheless, he felt that, if Kosilek lost hope of

receiving SRS, her current treatment plan would not stop a

deterioration of her mental state and the possible reemergence of

suicidal ideation.

6
The steps of this sequence, if fully completed, progress from
GID diagnosis and therapeutic treatment, through endocrine
treatment, and culminate -- after at least a one-year-long real-
life experience -- with the consideration of SRS.

-20-
 Dr. Kaufman from the Fenway Center also testified,

reiterating that the Fenway Center believed SRS to be an

appropriate and medically necessary step in Kosilek's treatment.

She further stated her belief that, if not given surgery, Kosilek

would present a significant risk of suicide: "if she's not able to

have surgery, I think that she'll be hopeless and feel helpless and

at that point really will have nothing else to live for."

Next to testify was Mark Burrows ("Burrows"), who had

been Kosilek’s treating psychiatrist for approximately five years.

Burrows testified to Kosilek's strong desire for SRS, and to her

feelings of hope associated with completing the formalization of

her gender presentation. Burrows also stated that denying surgery

would likely have a negative impact on Kosilek's mental health. He

believed that it was "slightly" "more probable than not" that a

denial of the surgery would result in Kosilek attempting to commit

suicide. Burrows also spoke about his belief that, if given SRS,

Kosilek should not continue to reside at MCI-Norfolk, as "the risks

involved in her possibly being assaulted are obvious."

Dr. Appelbaum of UMass was also called as a witness for

Kosilek. He testified as to UMass's trust in the Fenway Center's

recommendations, and to his belief that the DOC need not have

sought out a peer review of the Fenway Report, given the Fenway

Center's expertise in the treatment of GID.

-21-
 Kosilek testified next. She expressed the depth of her

desire for SRS, and she stated that she would continue to

experience mental anguish regarding her gender identity so long as

she had male genitalia. If not provided with SRS, Kosilek said

that she "would not want to continue existing [as an anatomical

male]" and might instead attempt to commit suicide. She disagreed

with the suggestion that treatment short of SRS could adequately

relieve her mental distress, stating that "[t]he problem is my

genitals. That's what needs to be fixed." Kosilek also testified

as to feeling discomfort in the all-male environment of MCI-Norfolk

and having a strong desire to be transferred to MCI-Framingham.

She felt that the inmates at MCI-Framingham would be more accepting

and welcoming of her than those at MCI-Norfolk.

The DOC offered testimony from Dr. Chester Schmidt, a

licensed psychiatrist and Associate Director of the Johns Hopkins

School of Medicine. Dr. Schmidt expressed his belief that Kosilek

had undergone an "excellent adaptation" through treatment with

hormones, hair removal, psychotherapy, and the provision of female

garb. These treatments had alleviated the severity of her mental

distress and allowed Kosilek to significantly consolidate her

gender identity. Dr. Schmidt acknowledged that, if not provided

SRS, Kosilek's level of mental distress would likely increase, with

depression or attempts at self-harm possible. On the whole,

however, he believed that her positive adaptation and the

-22-
consolidation of her gender identity indicated that the current

course of treatment provided by the DOC was medically adequate.

Dr. Schmidt explained that the severity of dysphoria associated

with GID may "wax and wane," with patients feeling depressed or

hopeless at times, but generally being able to alleviate these

depressive symptoms with appropriate psychotherapy and medical

interventions. He felt that these measures, in combination with

Kosilek's current course of treatment, would allow her to live

safely and maintain a level of contentment.

On cross-examination, Dr. Schmidt was questioned

regarding his alleged rejection of the Standards of Care.

Dr. Schmidt responded that he found the Standards of Care "very

useful for patients" and that he commonly requested that patients

familiarize themselves with these Standards when they began to seek

care for SRS. Asked if he had stricter requirements for SRS

eligibility than those in the Standards of Care, Dr. Schmidt

emphasized that he neither "advocate[s] for nor . . . speak[s]

against the decisions for the cross-gender hormones or eventually

for surgery." Rather, he believes such decisions are best made by

the patient, based on their personal needs and desires. In line

with this belief, Dr. Schmidt stated that he does not specifically

recommend SRS, but at a patient's request he will release medical

files and send a letter indicating that a patient is ready for

surgery to their chosen SRS provider.

-23-
 Dr. Schmidt further testified that he viewed the

Standards of Care as "guidelines." He explained, however, that

"[t]here are many people in the country who disagree with those

standards who are involved in the [GID] field." Because of this

disagreement, Dr. Schmidt expressed hesitation to refer to the

Standards of Care, or the recommendation for SRS, as medically

necessary. He emphasized the existence of alternative methods and

treatment plans accepted within the medical community. He also

questioned whether the Standards of Care's requirement of a real-

life experience could occur in prison, opining that the real-life

experience required a range of social and vocational experiences

unavailable within a penological setting.

Osborne testified next, reiterating her agreement with

Kosilek's GID diagnosis, but disagreeing that SRS was a medically

necessary treatment. In reference to the Standards of Care,

Osborne testified that she fully agreed that SRS was an effective

and appropriate treatment for GID. She emphasized, however, that

she did not view SRS as medically necessary in light of "the whole

continuum from noninvasive to invasive" treatment options available

to individuals with GID. Regarding Kosilek personally, Osborne

indicated that she believed Kosilek's current treatment plan had

been highly effective in allowing Kosilek to feel "hopeful,

euphoric, and not depressed" about her gender identity. Osborne,

like Dr. Schmidt before her, again expressed skepticism as to

-24-
whether a real-life experience could occur in jail, given that a

single-sex environment necessarily limited the sorts of social and

human interactions available. Osborne agreed that not providing

Kosilek with SRS might give rise to possible suicidal ideation, but

noted that the DOC had significant expertise in treating prisoners

exhibiting self-harming behavior. She felt that Kosilek's current

treatment plan, in conjunction with protective measures aimed at

ensuring her personal safety, was an appropriate and medically

acceptable response to Kosilek's GID.

b. The UMass report

Following the close of initial testimony, the district

court ordered UMass to review the testimony of all medical experts

and to issue a report regarding whether the treatment proposed by

Dr. Schmidt was an adequate method of treating Kosilek's GID. In

this report -- submitted to the court on September 18, 2006 -- Drs.

Appelbaum and Brewer made clear that they "worked with and relied

upon Dr. Kapila and Dr. Kaufman" who "assist[ed] to prepare this

response." The report stated that the UMass doctors "have been

informed by Dr. Kaufman and Dr. Kapila that . . . trial testimony

. . . confirms their opinion that Michelle Kosilek has a 'serious

medical need' because there is a 'substantial risk of serious harm

if it is not adequately treated.'" In conclusion, the report

reiterated that the Fenway Center believed Dr. Schmidt's proposed

-25-
treatment plan would not provide adequate care, and UMass endorsed

that conclusion.

c. The court-appointed expert

At the conclusion of the first round of testimony, the

district court decided to appoint an independent expert to assist

in determining what constituted the medical standard of treatment

for GID. On October 31, 2006, with the parties' input, the

district court selected Dr. Stephen Levine, a practitioner at the

Center for Marital and Sexual Health in Ohio and a clinical

professor of psychiatry at Case Western Reserve University School

of Medicine. Dr. Levine had helped to author the fifth version of

the Standards of Care, and served as Chairman of the Harry Benjamin

International Gender Dysphoria Association's Standards of Care

Committee.

A month after his appointment, Dr. Levine issued a

written report. The report began by explaining the dual roles that

WPATH -- formerly the Harry Benjamin Association and the

organization that wrote the Standards of Care -- plays in its

provision of care to individuals with GID:

WPATH is supportive to those who want sex
reassignment surgery (SRS). . . . Skepticism
and strong alternate views are not well
tolerated. Such views have been known to be
greeted with antipathy from the large numbers
of nonprofessional adults who attend each [of]
the organization's biennial meetings. . . .

The [Standards of Care are] the product
of an enormous effort to be balanced, but it

-26-
 is not a politically neutral document. WPATH
aspires to be both a scientific organization
and an advocacy group for the transgendered.
These aspirations sometimes conflict. The
limitations of the [Standards of Care],
however, are not primarily political. They
are caused by the lack of rigorous research in
the field.

Dr. Levine further emphasized that "large gaps" exist in the

medical community's knowledge regarding the long-term effects of

SRS and other GID treatments in relation to its positive or

negative correlation to suicidal ideation.

Dr. Levine next discussed the possibility of Kosilek

having a real-life experience in prison. He explained that the

Fenway Center, in stating that a real-life experience could be had

in prison, "failed to offer a mild caveat that the real life test

was designed to test the patients' capacity to function as a female

in the community by mastering the demands of . . . family, social

relationships, educational accomplishment, [and] vocational

performance." Such experiences and relationships, Dr. Levine

noted, are not a part of Kosilek's daily life in prison. Dr.

Levine's final conclusion was that:

Dr. Schmidt's view, however unpopular and
uncompassionate in the eyes of some experts in
GID, is within prudent professional community
standards. Treatment stopping short of SRS
would be considered adequate by many
psychiatrists, gender team members, and gender
patients themselves, if Kosilek were a citizen
in the community. . . . [T]here are a number
of acceptable community standards which derive
from differing assumptions about disorders,

-27-
 their causes, and the possible effective
interventions.

He recognized that the different treatment plans advocated by Dr.

Schmidt and the Fenway Center "each . . . [had] merit," as well as

limitations. Dr. Levine further wrote that doctors generally "do

not recommend treatment to GID patients. . . . The decision is

[the patient's], when and if they still want it."

Dr. Levine testified on December 16, 2006. He first

reiterated his belief that Dr. Schmidt's view, although not

preferred by some GID specialists, was within "prudent professional

standards." He noted that Kosilek had received significant relief

on her current treatment plan, and that many patients with GID live

comfortably without completing the triadic sequence. He believed

that Kosilek had already successfully consolidated her gender

identity, such that the removal of her male genitalia might relieve

dysphoria, but it was not necessary to complete that consolidation.

He also indicated variability and difficulty in forecasting

depressive symptoms and self-harming behavior in GID patients. He

explained that he believed Kosilek would certainly express deep

disappointment if denied SRS -- described as the sole current focus

of her life -- but that coping mechanisms might well change her

outlook in months and years to come, allowing her to live happily

without the provision of SRS.

The district court then asked Dr. Levine to narrow the

lens of his inquiry by presuming that there were absolutely no

-28-
external contraindications to surgery and that Kosilek had indeed

had a real-life experience in prison. Given these presumptions,

the court asked Dr. Levine to testify as to whether it would still

be prudent to not provide Kosilek with SRS. Dr. Levine

acknowledged his belief that prudent professionals would generally

not deny surgery to a fully eligible individual. Still, he

hesitated to declare Dr. Schmidt's approach medically unacceptable.

He answered that the provision of SRS would surely be a prudent

course of treatment, but then stated that "I also believe it's

prudent not to give her Sex Reassignment Surgery for lots of

reasons." He again emphasized for the court that the treatment of

GID was an evolving field, in which practitioners could reasonably

differ in their preferred treatment methods. Dr. Levine explained

that in many instances patients cannot or do not want to receive

SRS, and prudent physicians commonly employ a range of treatments

to ameliorate these patients' dysphoria.

d. Additional rounds of testimony

Several witnesses were recalled for additional testimony.

Drs. Kapila and Kaufman appeared again on behalf of Kosilek. Both

reiterated their belief that Kosilek had a serious medical need and

that, given Kosilek's high risk of suicide if denied the surgery,

SRS was the only adequate treatment plan. Dr. Appelbaum also

testified again, as did the UMass Medical Director. Both UMass

-29-
doctors reaffirmed their endorsement of the Fenway Center's

treatment recommendations.

Kosilek also presented additional witness testimony from

Dr. Marshall Forstein, Associate Professor of Psychiatry at Harvard

Medical School, who had previously evaluated Kosilek during

Kosilek I. Dr. Forstein issued a written report, in which he noted

that "the question of the most prudent form of treatment is

complicated by the diagnosis of GID being on the margins of typical

medical practice." Despite this recognition, he testified that he

believed SRS was necessary for Kosilek. He felt that, if she was

not given SRS, there was a significant risk that Kosilek would

attempt suicide or self-mutilation. Although Dr. Forstein believed

that psychotherapy might "help with frustration, with harassment,

and with depression," he was uncertain whether Kosilek could ever

fully "reconcile with being incompletely transitioned."

2. Testimony regarding safety and security concerns

a. Initial testimony

In line with the June 10, 2005, security report prepared

by Commissioner Dennehy, multiple DOC officials testified regarding

the safety and security concerns that were likely to arise if

Kosilek was provided SRS.

First to testify was Spencer, who at that time served as

Superintendent of MCI-Norfolk. Spencer began by explaining the

general layout and security measures at MCI-Norfolk. He also

-30-
explained that the prison had, so far, successfully been able to

accommodate Kosilek's receipt of care without incident. Spencer

was unaware of any issues or incidents of harassment related to

Kosilek's breast growth and increasingly feminine appearance. He

stated, however, that he would have significant concerns housing an

anatomically female prisoner in MCI-Norfolk, an all-male prison.

Despite the lack of historical incidents specific to Kosilek, he

emphasized that "inmates do get assaulted, inmates have been raped

. . . [a]nd putting a female in a correctional environment like

MCI-Norfolk would be of high concern to me." If Kosilek remained

at MCI-Norfolk, Spencer testified that he believed she would only

be safe if housed in the Special Management Unit, a highly

restricted secure building separated from the general population.

Bissonnette, Superintendent of MCI-Framingham, also

testified about the security concerns she believed would arise if

Kosilek was transferred to the all-female prison after receiving

SRS. She explained that MCI-Framingham does not have private

cells, save for the segregation and medical units. All women in

the general population are required to cohabitate, and that prison

would be unable to provide a single-occupancy cell for Kosilek.

She also explained that Kosilek's presence could create significant

disruption in MCI-Framingham's population, given that Kosilek had

been convicted for violently murdering her wife, and that a

-31-
significant portion of women at MCI-Framingham were victims of

domestic abuse.

Bissonnette acknowledged that there were procedures in

place designed to help women cope with exposure to upsetting or

traumatic experiences with other prisoners, but maintained that

these security concerns would require that Kosilek, if transferred

to MCI-Framingham, be housed in the segregated Close Custody unit.

Bissonnette explained that she had significant hesitation about

incarcerating anyone long-term in the Close Custody unit, given the

potential negative effects of such long-term segregation.

Commissioner Dennehy also testified. She described the

security concerns arising from cross-gender housing as "obvious" to

any experienced corrections officer. In line with her belief that

the safety and security concerns about post-operative housing were

clear, Dennehy stated that she would not feel comfortable allowing

SRS -- even if mandated by the court -- if she could not identify

an adequate method of safely housing Kosilek after her operation.

Dennehy reiterated Spencer's and Bissonnette's concerns, stating

that she deeply trusted both Superintendents' professional

judgments regarding the security of housing Kosilek at their

respective facilities. Dennehy also explained why reliance on an

interstate compact to transfer Kosilek would be problematic. She

emphasized that other states take prisoners on a fully voluntary

-32-
basis, and that no state may be willing or able to accommodate a

transfer request for Kosilek.

Commissioner Dennehy was also questioned about negative

press surrounding the DOC's possible provision of SRS to Kosilek.

Specifically, she was asked about her professional relationship

with a state senator who had vocally opposed surgery and sponsored

legislation to deny its provision. She was also asked about any

contact with the then-lieutenant governor, who was another strong

opponent of providing SRS to prisoners. Dennehy stated that she

was aware of negative press reports and political opposition

surrounding Kosilek's request, but that her decision not to provide

SRS was based only on security concerns and had not been influenced

by this public pressure.

The district court recalled Dennehy on October 18, 2006,

to ask additional questions regarding a growing amount of press

coverage surrounding the case. Dennehy acknowledged that she was

aware of significant news coverage of Kosilek's case, but denied

personally following the story in the media. She explained that

there were staff members within the DOC trained to deal with press

inquiries and that she generally received only summaries of news

coverage from her staff. Again, Dennehy strongly denied forming

any opinion about correctional safety procedures based on media

reports or public opinion.

-33-
 b. Commissioner Clarke

Dennehy ended her tenure as DOC Commissioner on April 30,

2007, and in November 2007 the position was filled by Harold

Clarke. After Clarke took over, the district court requested that

he familiarize himself with a selected number of trial transcripts.

Clarke was ordered to file a report, on the basis of those

transcripts, indicating whether he believed that the DOC had

legitimate reasons to refuse Kosilek's request for SRS.

Clarke's report, filed approximately a month after the

district court's order, stated that his conclusions were based on

more than three decades of correctional experience and were not

influenced by political or media pressure. He expressed concern

regarding threats of suicide being used as a means for prisoners to

receive wanted benefits or concessions from staff. Finding it to

be bad practice for prison administrators to give in to demands

accompanied by the threat of suicide, Clarke stated that he

believed the Massachusetts prison system had taken significant

measures to ensure it was prepared to deal with suicidal ideation

among its prison population. In addition to considering the issue

of suicide, Clarke's report reemphasized the significant post-

operative security concerns expressed by his predecessor. He

stated that housing Kosilek at MCI-Norfolk created clear security

concerns related to mixed-gender prison populations, while housing

Kosilek at MCI-Framingham would pose a significant risk of

-34-
destabilizing that environment, given the number of women prisoners

who were victims of domestic violence. Clarke also stated his

belief that a separate unit to house GID prisoners was not

feasible, given that prisoners with GID might have a wide range of

security classifications and security needs, making cohabitation

unsafe. In reference to the possibility of an interstate transfer,

Clarke reiterated the concern that any interstate transfer would be

completely voluntary and that a receiving state might later decide

to return Kosilek, at which time the housing concerns would

reemerge.

Testifying before the court, Clarke acknowledged that he

had received several letters from outraged state politicians

claiming that provision of the surgery would be an "affront to the

taxpayers" and citing state budget concerns as a reason to deny

Kosilek surgery. The letters argued that a strained state budget

should not be used to accommodate what the legislators believed to

be an "elective" procedure and that the DOC would be "unwise" to

provide it. Clarke, however, explained that he had not answered

these letters, as he believed providing an answer would be

inappropriate given his role as DOC Commissioner. He also denied

being in any way influenced by cost concerns in reaching his

conclusion regarding safety and security concerns. Clarke similarly

testified that he was aware of media coverage regarding Kosilek's

-35-
request, but he had not personally viewed the news or heard the

radio stories.

G. Kosilek II

The district court issued an extensive opinion on

September 4, 2012. This opinion concluded that Kosilek had a

serious medical need and that -- based on the court's belief that

Dr. Schmidt was not a prudent professional -- the only adequate way

to treat this need was through SRS. Moreover, the court determined

that the DOC's stated security concerns were merely pretextual and

concluded that the DOC had in fact made its decision based on

public and political pressure. This, the court concluded, amounted

to deliberate indifference under the Eighth Amendment. Stating its

belief that the DOC would continue to deny Kosilek adequate

treatment in the future, the district court granted an injunction

requiring that the DOC provide Kosilek with SRS.

II. Discussion

A. The Eighth Amendment and Medical Care in Prison

"Excessive bail shall not be required, nor excessive

fines imposed, nor cruel and unusual punishments inflicted." U.S.

Const. amend. VIII. From this brief amendment, courts have derived

the principles that govern the permissible conditions under which

prisoners are held and that establish the medical treatment those

prisoners must be afforded. See Farmer v. Brennan, 511 U.S. 825,

832 (1994). Where "society takes from prisoners the means to

-36-
provide for their own needs," the failure to provide such care "may

actually produce physical torture or a lingering death." Brown v.

Plata, 131 S. Ct. 1910, 1928 (2011) (internal quotation marks

omitted). Undue suffering, unrelated to any legitimate penological

purpose, is considered a form of punishment proscribed by the

Eighth Amendment. Estelle v. Gamble, 429 U.S. 97, 103 (1976). The

Eighth Amendment is meant to prohibit "unnecessary and wanton

infliction of pain," which is "repugnant to the conscience of

mankind." Id. at 105-06 (internal quotation marks omitted).

The Amendment's focus on punishment means that not all

shortages or failures in care exhibit the intent and harmfulness

required to fall within its ambit. See Farmer, 511 U.S. at 837

(reasoning that the Eighth Amendment's prohibition of punishment

implies an act done with intentionality). Therefore, to prove an

Eighth Amendment violation, a prisoner must satisfy both of two

prongs: (1) an objective prong that requires proof of a serious

medical need, and (2) a subjective prong that mandates a showing of

prison administrators' deliberate indifference to that need. See

Estelle, 429 U.S. at 106 (holding that inadequate treatment must be

"sufficiently harmful to evidence deliberate indifference to

serious medical needs"); Sires v. Berman, 834 F.2d 9, 12 (1st Cir.

1987) ("A plaintiff must satisfy two elements to present a viable

[Eighth Amendment] claim: he must show a serious medical need, and

he must prove the defendant's purposeful indifference thereto.").

-37-
 First, a medical need must be "serious." Id. This

objective prong requires that the need be "one that has been

diagnosed by a physician as mandating treatment, or one that is so

obvious that even a lay person would easily recognize the necessity

for a doctor's attention." Gaudreault v. Municipality of Salem,

Mass., 923 F.2d 203, 208 (1st Cir. 1990). This prong does not

impose upon prison administrators a duty to provide care that is

ideal, or of the prisoner's choosing. See United States v. Derbes,

369 F.3d 579, 583 (1st Cir. 2004) (stating that prison

administrators are "by no means required to tailor a perfect plan

for every inmate; while [they are] constitutionally obligated to

provide medical services to inmates, these services need only be on

a level reasonably commensurate with modern medical science and of

a quality acceptable within prudent professional standards"

(internal quotation marks and citations omitted)); United States v.

DeCologero, 821 F.2d 39, 43 (1st Cir. 1987) (same); Ferranti v.

Moran, 618 F.2d 888, 891 (1st Cir. 1980) ("[A]llegations [that]

simply reflect a disagreement on the appropriate course of

treatment . . . fall[] short of alleging a constitutional

violation."). Rather, the Constitution proscribes care that is

"'so inadequate as to shock the conscience.'" Torraco v. Maloney,

-38-
923 F.2d 231, 235 (1st Cir. 1991) (quoting Sires, 834 F.2d at

13)).7

Second, even if medical care is so inadequate as to

satisfy the objective prong, the Eighth Amendment is not violated

unless prison administrators also exhibit deliberate indifference

to the prisoner's needs. Estelle, 429 U.S. at 105-06. For

purposes of this subjective prong, deliberate indifference "defines

a narrow band of conduct," Feeney v. Corr. Med. Servs. Inc., 464

F.3d 158, 162 (1st Cir. 2006), and requires evidence that the

failure in treatment was purposeful. See Estelle, 429 U.S. at 105

(holding that "an inadvertent failure to provide adequate medical

care" is not a constitutional violation);8 id. at 106 ("Medical

malpractice does not become a constitutional violation merely

because the victim is a prisoner."); Watson v. Caton, 984 F.2d 537,

540 (1st Cir. 1993) ("The courts have consistently refused to

7
Although these cases address the second, subjective prong of the
Eighth Amendment analysis, we have recognized that "the subjective
deliberate indifference inquiry may overlap with the objective
serious medical need determination" and that "similar evidence
. . . may be relevant to both components." Leavitt v. Corr. Med.
Servs., Inc., 645 F.3d 484, 498 (1st Cir. 2011) (internal quotation
marks and brackets omitted); see also DesRosiers v. Moran, 949 F.2d
15, 18-19 (1st Cir. 1991). As the adequacy of care is germane both
to Kosilek's objective need for surgery and to the DOC's alleged
deliberate indifference to that need, the principles of these cases
are relevant to both steps of our analysis.
8
Although this case does not involve "an inadvertent failure to
provide adequate medical care," see Estelle, 429 U.S. at 106, that
fact alone does not elevate the DOC's choice among alternative
treatments to "deliberate indifference" for purposes of the Eighth
Amendment analysis.

-39-
create constitutional claims out of disagreements between prisoners

and doctors about the proper course of a prisoner's medical

treatment, or to conclude that simple medical malpractice rises to

the level of cruel and unusual punishment."). "The obvious case

would be a denial of needed medical treatment in order to punish

the inmate." Watson, 984 F.2d at 540. While deliberate

indifference may also be exhibited by a "wanton disregard" to a

prisoner's needs, Battista v. Clarke, 645 F.3d 449, 453 (1st Cir.

2011), such disregard must be akin to criminal recklessness,

requiring consciousness of "'impending harm, easily preventable.'"

Watson, 984 F.2d at 540.

When evaluating medical care and deliberate indifference,

security considerations inherent in the functioning of a

penological institution must be given significant weight.

Battista, 645 F.3d at 454 ("[S]ecurity considerations also matter

at prisons . . . and administrators have to balance conflicting

demands."). "[W]ide-ranging deference" is accorded to prison

administrators "in the adoption and execution of policies and

practices that in their judgement are needed to . . . maintain

institutional security." Whitley v. Albers, 475 U.S. 312, 321-22

(1986) (quoting Bell v. Wolfish, 441 U.S. 520, 547 (1979))

(internal quotation marks omitted). In consequence, even a denial

of care may not amount to an Eighth Amendment violation if that

decision is based in legitimate concerns regarding prisoner safety

-40-
and institutional security. Cameron v. Tomes, 990 F.2d 14, 20 (1st

Cir. 1993) (requiring courts to "embrace security and

administration, . . . not merely medical judgments" in assessing

claims of deliberate indifference); Sires, 834 F.2d at 13

("[S]afety factors are properly included in the evaluation of the

medical needs of an inmate."). Importantly, prison administrators

need only have "'responded reasonably to the risk.'" Giroux v.

Somerset Cnty., 178 F.3d 28, 33 (1st Cir. 1999) (quoting Farmer,

511 U.S. at 844).

B. Standard of Review

The test for establishing an Eighth Amendment claim of

inadequate medical care encompasses a multitude of questions that

present elements both factual and legal. Review of such "mixed

questions" is of a variable exactitude; the more law-based a

question, the less deferentially we assess the district court's

conclusion. In Re Extradition of Howard, 996 F.2d 1320, 1328 (1st

Cir. 1993) ("The standard of review applicable to mixed questions

usually depends upon where they fall along the degree-of-deference

continuum . . . .").

The ultimate legal conclusion of whether prison

administrators have violated the Eighth Amendment is reviewed de

novo. See, e.g., Thomas v. Bryant, 614 F.3d 1288, 1307 (11th Cir.

2010) ("Whether the record demonstrates that [the prisoner] was

sprayed with chemical agents . . . and that he suffered

-41-
psychological injuries from such sprayings are questions of fact.

Whether these deprivations are objectively 'sufficiently serious'

to satisfy the objective prong, is a question of law . . . ."

(internal citations omitted)); Hallett v. Morgan, 296 F.3d 732, 744

(9th Cir. 2002) ("The district court's factual findings regarding

conditions at the Prison are reviewed for clear error. However,

its conclusion that the facts do not demonstrate an Eighth

Amendment violation is a question of law that we review de novo."

(citing Campbell v. Wood, 18 F.3d 662, 681 (9th Cir. 1994) (en

banc))); Hickey v. Reeder, 12 F.3d 754, 756 (8th Cir. 1993)

("Whether conduct, if done with the required culpability, is

sufficiently harmful to establish an Eighth Amendment violation is

an objective or legal determination which we decide de novo.");

Alberti v. Klevenhagen, 790 F.2d 1220, 1225 (5th Cir. 1986)

("[O]nce the facts are established, the issue of whether these

facts constitute a violation of constitutional rights is a question

of law that may be assayed anew upon appeal."). Subsidiary legal

questions, such as whether an actor's conduct amounted to

deliberate indifference for purposes of the Eighth Amendment, are

likewise reviewed de novo. Cf. Ornelas v. United States, 517 U.S.

690, 699 (1996) (holding that, for Fourth Amendment purposes,

reasonable suspicion and probable cause determinations should

receive de novo appellate review); United States v. Camacho, 661

F.3d 718, 724 (1st Cir. 2011) (we review de novo a district court's

-42-
subsidiary reasonable suspicion and probable cause determinations

in evaluating a motion to suppress); United States v. Bucci, 582

F.3d 108, 115-17 (1st Cir. 2009).

Our court awards deference to the district court's

resolution of questions of pure fact and issues of credibility.

See, e.g., DesRosiers v. Moran, 949 F.2d 15, 19 (1st Cir. 1991)

(reviewing factual findings regarding the adequacy of care

deferentially); Torraco, 923 F.2d at 234 (finding that issues of

culpability in a deliberate indifference inquiry are usually

questions for a jury). We will reverse the district court's

findings on such factual questions only for clear error.

DesRosiers, 949 F.2d at 19 ("[W]e assay findings of fact in a bench

trial only for clear error."). We find clear error when we are

left with "'a strong, unyielding belief, based on the whole of the

record,' that the judge made a mistake." In re O'Donnell, 728 F.3d

41, 45 (1st Cir. 2013) (quoting Islamic Inv. Co. of the Gulf (Bah.)

Ltd. v. Harper (In re Grand Jury Investigation), 545 F.3d 21, 24

(1st Cir. 2008)). We may also find clear error when the district

court commits an error of law that affects its fact-finding

analysis. See Uno v. City of Holyoke, 72 F.3d 973, 978 (1st Cir.

1995) ("[T]he jurisprudence of clear error 'does not inhibit an

appellate court's power to correct errors of law, including those

that may infect a so-called mixed finding of law and fact, or a

finding of fact that is predicated on a misunderstanding of the

-43-
governing rule of law.'" (quoting Thornburg v. Gingles, 478 U.S.

30, 106 (1986))).

This standard of review tracks the Supreme Court's

framework for appellate review of claims of excessive punishment or

fines under the Eighth Amendment. United States v. Bajakajian, 524

U.S. 321, 336-37 & n.10 (1998). In Bajakajian, the Supreme Court

concluded that the excessiveness of a fine was a question properly

considered de novo by appellate courts, applying "the standard of

gross disproportionality articulated in [its] Cruel and Unusual

Punishments Clause precedents." Id. at 336. "[T]he application of

a constitutional standard to the facts of a particular case," the

Supreme Court reasoned, may appropriately require de novo appellate

review to ensure consistency in the law's development. Id. at 336

n.10; see also Cooper Indus. v. Leatherman Tool Grp., Inc., 532

U.S. 424, 435-36 (2001) (extending de novo review of the

excessiveness inquiry associated with the Excessive Fines Clause of

the Eighth Amendment to punitive damages awards); Ornelas, 517 U.S.

at 699 (holding that "as a general matter determinations of

reasonable suspicion and probable cause should be reviewed de novo

on appeal").

The considerations set forth in Ornelas, and applied in

Bajakajian and Leatherman Tool, are equally relevant here.

"Medical 'need' in real life is an elastic term," Battista, 645

F.3d at 454, "that take[s its] substantive content from the

-44-
particular context[] in which the standards are being assessed."

Ornelas, 517 U.S. at 696. Similarly, the "legal rules" for what

constitutes care in violation of the Eighth Amendment "acquire

content only through application" -- a fact which favors de novo

appellate review "to maintain control of, and to clarify, the legal

principles." See id. at 697.

C. The Objective Prong: Serious Medical Need

To sustain a claim under the objective prong of the

Eighth Amendment, Kosilek must show that she has a serious medical

need for which she has received inadequate treatment. See Estelle,

429 U.S. at 106; Sires, 834 F.2d at 13 (finding no Eighth Amendment

violation where the prisoner failed to "present[] any evidence of

a serious medical need that has gone unmet"); see also Derbes, 369

F.3d at 583 (a prison's constitutional obligation to provide

medical services does not require "a perfect plan for every

inmate"); DeCologero, 821 F.2d at 42 ("[T]hough it is plain that an

inmate deserves adequate medical care, he cannot insist that his

institutional host provide him with the most sophisticated care

that money can buy."). A significant risk of future harm that

prison administrators fail to mitigate may suffice under the

objective prong. Helling v. McKinney, 509 U.S. 25, 35 (1993); see

also Baze v. Rees, 553 U.S. 35, 50 (2008) ("[S]ubjecting

individuals to a risk of future harm . . . can qualify as cruel and

unusual punishment."); Roe v. Elyea, 631 F.3d 843, 858 (7th Cir.

-45-
2011) ("[T]he Eighth Amendment 'protects [an inmate] not only from

deliberate indifference to his or her current serious health

problems, but also from deliberate indifference to conditions

posing an unreasonable risk of serious damage to future health.'"

(quoting Board v. Farnham, 394 F.3d 469, 479 (7th Cir. 2005))).

That GID is a serious medical need, and one which

mandates treatment, is not in dispute in this case. The parties do

not spar over the fact that Kosilek requires medical care aimed at

alleviating the harms associated with GID -- to the contrary, the

DOC has provided such care since 2003. Rather, the parties

disagree over whether SRS is a medically necessary component of

Kosilek's care, such that any course of treatment not including

surgery is constitutionally inadequate. The parties' disparate

positions on this issue are fit for succinct summary.

Kosilek argues that the only constitutionally sufficient

treatment regimen is to adhere to the Standards of Care's triadic

sequence in full, including the provision of SRS. Kosilek

emphasizes that doctors at both UMass and Fenway Clinic -- doctors

hired by the DOC -- confirmed at trial that SRS was "medically

necessary." The failure to provide treatment, these doctors

testified, would almost certainly lead to a deterioration in

Kosilek's mental state and a high likelihood of self-harming

behaviors. In light of this risk, and given that they believed

Kosilek had successfully met all eligibility criteria for SRS,

-46-
these doctors believed that any course of treatment excluding SRS

is insufficient to treat Kosilek's GID.

In contrast, the DOC argues that full progression through

the Standards of Care's triadic sequence is not the only adequate

treatment option, as Kosilek's GID may be appropriately managed

with treatment short of SRS. The DOC maintains that the evidence

does not meet the standards for negligent treatment of a medical

condition, much less the higher Eighth Amendment standard. See

Estelle, 429 U.S. at 106 ("Thus, a complaint that a physician has

been negligent in diagnosing or treating a medical condition does

not state a valid claim of medical mistreatment under the Eighth

Amendment."); Watson, 984 F.2d at 540 (stating that "simple medical

malpractice" does not "rise[] to the level of cruel and unusual

punishment").

Relying on the advice of accredited medical

professionals, the DOC argues that its alternative course of

treatment -- which provides Kosilek such alleviative measures as

psychotherapy, hormones, electrolysis, and the provision of female

garb and accessories -- is sufficient to treat Kosilek's GID and

far exceeds a level of care that would be "so inadequate as to

shock the conscience." See Torraco, 923 F.2d at 235 (quoting

Sires, 834 F.2d at 13). Moreover, this course of treatment has, in

practice, greatly diminished Kosilek's mental distress and allowed

her a fair measure of contentment. Should suicidal ideation arise

-47-
in the future, the DOC contends that -- based on the advice of its

medical experts and its own penological experience -- it would be

able to address that future risk appropriately through

psychotherapy and antidepressants.

We begin by discussing the district court's conclusions

regarding the objective prong. We then examine de novo the

question whether the treatment offered was constitutionally

adequate.

1. The district court's medical prudence determination

The district court ruled that SRS was a medically

necessary treatment, and that Dr. Schmidt's alternative belief was

outside the bounds of medical prudence.9 However, the court's

finding that Dr. Schmidt's views were medically imprudent was based

on several erroneous determinations.

First, the court ruled that, unlike prudent medical

professionals, Dr. Schmidt did not "follow" the Standards of Care

in his treatment of GID. This finding ignored critical nuance in

Dr. Schmidt's testimony and based its conclusion on a severely

strained reading of Dr. Levine's expert testimony.

9
For the sake of clarity, we reiterate that medical imprudence --
without more -- is insufficient to establish an Eighth Amendment
violation. See Estelle, 429 U.S. at 105-06; Watson, 984 F.2d at
540. Instead, a prisoner must satisfy both prongs of the Eighth
Amendment inquiry, proving that the level of care provided is
"sufficiently harmful to evidence deliberate indifference to
serious medical needs." Estelle, 429 U.S. at 106.

-48-
 As an initial matter, the Standards of Care themselves

admit of significant flexibility in their interpretation and

application. They state, for example, that "[t]he Standards of

Care [a]re Clinical Guidelines" and are "intended to provide

flexible directions" to medical professionals in crafting treatment

plans. Standards of Care at 1 (emphases added). The Standards of

Care also specifically warn that "[a]ll readers should be aware of

the limitations of knowledge in this area." Standards of Care at

1. "Individual professionals and organized programs," the

Standards of Care continue on, "may modify [the standards]" as

appropriate. Id. at 2. Dr. Levine's testimony acknowledged this

flexibility:

[DR. LEVINE]: [T]he "Standards of Care" was a
consensus document from people from seven
different countries or something, you know,
who come from different systems, and it was a
political process that forged together a set
of standards . . . . So "prudent" is a
wonderful word, but it's not like it has one
simple definition.

. . .

THE COURT: But is this an area in which you
think prudent professionals can reasonably
differ as to what is at least minimally
adequate treatment for this condition?

[DR. LEVINE]: Yes, and do.

Moreover, the district court put great weight on the fact

that the Standards of Care require that patients receive two

letters of recommendation prior to SRS. The court concluded,

-49-
therefore, that "prudent professionals who treat individuals

suffering from severe gender identity disorders write such letters

of recommendation," and it faulted Dr. Schmidt as imprudent for his

failure to engage in this practice. In so doing, the court relied

on Dr. Levine's testimony, which it believed stated that a prudent

professional would not "[refuse] to write letters of

recommendation."

Dr. Schmidt's testimony, however, makes clear that

although he does not advocate or recommend surgery to his patients,

if a patient chooses to seek SRS, he releases all of their medical

files to a surgeon and writes that surgeon a letter confirming that

the patient is eligible for surgery. Insofar as Dr. Schmidt had

not advocated for the surgery, this neutrality aligns with what Dr.

Levine describes as the accepted practice for doctors in the

treatment of GID: "[i]f the patient meets eligibility requirements

. . . we then write a letter of support . . . I understand how

others may perceive this as a recommendation . . . [but] we tell

ourselves we are opening a gate to their decision." Therefore,

whatever the semantic force of the district court's distinction, we

see no material difference between the letters written by Dr.

Schmidt confirming a patient's readiness for surgery and what the

Standards of Care refers to as a letter of recommendation.

The district court next concluded that Dr. Schmidt was

imprudent because antidepressants and psychotherapy alone are

-50-
inadequate to treat GID. Again, the court claimed that it relied

on the testimony of Dr. Levine, but misconstrued his testimony in

support of its conclusion. Dr. Levine did in fact state that

"gender dysphoria is not significantly ameliorated . . . by

treating [patients] with a prozac-like drug alone." He continued

on, however, to explain that he did not believe this was the

treatment plan advocated by Dr. Schmidt or the DOC. To the

contrary, he understood that Kosilek would continue to receive

ameliorative treatment for her GID and, if she entered a depressive

or suicidal state based on her inability to receive SRS,

antidepressants and psychotherapy would be used to help stabilize

her mental state so as to alleviate the risk of suicide while

working with her to craft new perspectives and life goals beyond

surgery. He felt that the treatment might well be successful in

this capacity, when combined with the direct alleviative treatments

currently provided.

Finally, the district court found Dr. Schmidt imprudent

because he did not believe that a real-life experience could occur

in prison, given that it was an isolated, single-sex environment.

The district court disagreed, stating that it had concluded a real-

life experience could occur in prison, as Kosilek would remain

incarcerated for her entire life. In reaching this determination,

the court made a significantly flawed inferential leap: it relied

on its own -- non-medical -- judgment about what constitutes a

-51-
real-life experience to conclude that Dr. Schmidt's differing

viewpoint was illegitimate or imprudent. Prudent medical

professionals, however, do reasonably differ in their opinions

regarding the requirements of a real-life experience -- and this

reasonable difference in medical opinions is sufficient to defeat

Kosilek's argument. Cf. Bismark v. Fisher, 213 F. App'x 892, 897

(11th Cir. 2007) ("Nothing in our case law would derive a

constitutional deprivation from a prison physician's failure to

subordinate his own professional judgment to that of another doctor

. . . ."); Estate of Cole by Pardue v. Fromm, 94 F.3d 254, 261 (7th

Cir. 1996); Bowring v. Godwin, 551 F.2d 44, 48 (4th Cir. 1977).

In fact, Dr. Levine noted that an incarcerative

environment might well be insufficient to expose Kosilek to the

variety of societal, familial, and vocational pressures foreseen by

a real-life experience. This viewpoint aligned with that of Dr.

Schmidt and Osborne. And, although Dr. Forstein's written report

appears to presume Kosilek had completed a real-life experience, it

echoed this same point: "being in prison has helped [Kosilek]

consolidate her desire . . . simplifying the issues, without the

stressors and choices that she would have had to make out in the

outside real world." We find no support for the district court's

conclusion that no reasonable medical expert could opine that

Kosilek lacked real-life experience, particularly in light of the

contrary testimony from medical experts concerning the range of

-52-
social, environmental, and professional considerations that are

necessary to constitute a real-life experience under the Standards

of Care. The district court thus erred by substituting its own

beliefs for those of multiple medical experts.10

The district court's finding of medical imprudence relied

heavily on inferences we do not believe can rightly be drawn from

Dr. Levine's testimony; this finding also ignored significant

contrary evidence regarding the breadth and variety of acceptable

treatments for GID within the medical community.11 Its conclusion

that the Fenway Center's recommendation constituted the sole

acceptable treatment plan is, thus, contradicted by the record.

10
There are obvious reasons for the range of judgments in this
area. Although the medical experts disagreed over whether
experience in a prison setting could qualify as real-life
experience, none of the experts who opined that it could do so
appear to have considered the fact that after SRS, Kosilek would
most likely be housed in the drastically different setting of a
female facility. This distinction was reflected in Dr. Forstein's
report, which stated that "[Kosilek's] 'real life experience' leads
her to the conclusion that so long as she is in a male prison . . .
she cannot perceive herself as a true woman." This statement
acknowledges that any real-life experience available to Kosilek was
shaped by her current, all-male prison environment. Kosilek
introduced no evidence to show that her experience there would
satisfy the requirement that she have real-life experience in her
post-operative housing environment.
11
The district court ignored or minimized significant portions of
Dr. Levine's testimony on the theory that the doctor had based his
evaluation of medical prudence on the "erroneous assumption[]" that
Kosilek may not have had a real-life experience in prison and faced
no other extrinsic obstacles to surgery. As explained above, in
doing so the court improperly supplanted a question of medical
opinion -- on which experts may differ -- with its own decision
based on a layman's view, and terming all contrary views imprudent.

-53-
 2. Adequacy of the DOC's treatment plan

Regarding the medical adequacy of Kosilek's treatment,

the district court held that psychotherapy and antidepressants

alone would not adequately treat Kosilek's GID. This finding

mischaracterizes the issues on appeal and unduly minimizes the

nature of the DOC's preferred treatment plan. The DOC does not

claim that treating Kosilek's GID merely with therapy and

antidepressants alone would constitute adequate care. Cf. Fields

v. Smith, 653 F.3d 550, 556 (7th Cir. 2011) (accepting, in the

absence of contrary evidence, expert testimony that "psychotherapy

as well as antipsychotics and antidepressants . . . do nothing to

treat the underlying disorder [of GID]"). In fact, since Kosilek

I the DOC has acknowledged the need to directly treat Kosilek's

GID. Beginning in 2003, it has provided hormones, electrolysis,

feminine clothing and accessories, and mental health services aimed

at alleviating her distress. The parties agree that this care has

led to a real and marked improvement in Kosilek's mental state.

There is also no dispute that this care would continue, whether or

not SRS is provided.

The question before our court, therefore, is not whether

antidepressants and psychotherapy alone are sufficient to treat

GID, or whether GID constitutes a serious medical need. Rather,

the question is whether the decision not to provide SRS -- in light

of the continued provision of all ameliorative measures currently

-54-
afforded Kosilek and in addition to antidepressants and

psychotherapy -- is sufficiently harmful to Kosilek so as to

violate the Eighth Amendment. It is not. See Smith v. Carpenter,

316 F.3d 178, 186 (2d Cir. 2003) ("[I]t's the particular risk of

harm faced by a prisoner due to the challenged deprivation of care,

rather than the severity of the prisoner's underlying medical

condition, considered in the abstract, that is relevant for Eighth

Amendment purposes."); see also Estelle, 429 U.S. at 106 (requiring

proof of "acts or omissions sufficiently harmful" as to illustrate

deliberate indifference to a serious medical need); Estate of

Bearden ex rel. Bearden v. Anglin, 543 F. App'x 918, 921 (11th Cir.

2013); Leavitt, 645 F.3d at 497.

Kosilek admits that the DOC's current treatment regimen

has led to a significant stabilization in her mental state.

Kosilek's doctors testified to the same, highlighting her "joy

around being feminized." This claim is also borne out by the

passage of significant time since she exhibited symptoms of

suicidal ideation or attempted to self-castrate. In addition to

alleviating her depressive state, this treatment has also resulted

in significant physical changes and an increasingly feminine

appearance.

The significance of a future risk of suicidality is not

one that this court takes lightly, and Kosilek is right to note

that a clear risk of future harm may suffice to sustain an Eighth

-55-
Amendment claim. See Helling, 509 U.S. at 35 (determining that an

"unreasonable risk" of future harm may amount to an Eighth

Amendment violation); Baze, 553 U.S. at 49; Roe, 631 F.3d at 858.

Nonetheless, the risk of suicidal ideation is born from Kosilek's

GID-related mental distress. Therefore an assessment of the

gravity of that risk, and its appropriate treatment, must encompass

the entirety of the DOC's treatment plan, not merely the potential

addition of psychotherapy and antidepressants.

Kosilek is provided hormones, facial hair removal,

feminine clothing and accessories, and access to regular mental

health treatment. The DOC also stands ready to protect Kosilek

from the potential for self-harm by employing its standard and

accepted methods of treating any prisoner exhibiting suicidal

ideation. Trial testimony established that this plan offers real

and direct treatment for Kosilek's GID. It employs methods proven

to alleviate Kosilek's mental distress while crafting a plan to

minimize the risk of future harm. See Carpenter, 316 F.3d at 186.

It does not wantonly disregard Kosilek's needs, but accounts for

them. See Torraco, 923 F.2d at 235.

The law is clear that where two alternative courses of

medical treatment exist, and both alleviate negative effects within

the boundaries of modern medicine, it is not the place of our court

to "second guess medical judgments" or to require that the DOC

adopt the more compassionate of two adequate options. Layne v.

-56-
Vinzant, 657 F.2d 468, 474 (1st Cir. 1981) (quoting Westlake v.

Lucas, 537 F.2d 857, 860 n.5 (6th Cir. 1976)); Bismark, 213 F.

App'x at 897; Medrano v. Smith, 161 F. App'x 596, 599 (7th Cir.

2006); Sanchez v. Vild, 891 F.2d 240, 242 (9th Cir. 1989); Bowring,

551 F.2d at 48.

That the DOC has chosen one of two alternatives -- both

of which are reasonably commensurate with the medical standards of

prudent professionals, and both of which provide Kosilek with a

significant measure of relief -- is a decision that does not

violate the Eighth Amendment.12

Kosilek warns, however, that upholding the adequacy of

the DOC's course of treatment in this case -- despite her medical

12
This holding in no way suggests that correctional administrators
wishing to avoid treatment need simply to find a single
practitioner willing to attest that some well-accepted treatment is
not necessary. We do not establish here a per se rule allowing a
dissenting medical opinion to carry the day. Rather, our
determination is limited to the particular record on appeal, which
involves a medical condition that admits of a number of valid
treatment options. This fact was testified to by Dr. Levine,
recognized by the UMass doctors in their correspondence with the
DOC, and corroborated by Dr. Forstein in his written report.

The DOC did not engage in a frenzy of serial consultations aimed
at finding the one doctor out of a hundred willing to testify that
SRS was not medically necessary. Rather, it made a considered
decision to seek out a second opinion from an expert previously
considered in its initial selection process. Our opinion rests on
the facts presented in this record, and we find merely that the
regimen of care provided by the DOC -- which includes hormonal
treatments as well as feminine products, clothing, and hair
removal, and which has successfully alleviated the severity of a
prisoner's distress -- is not sufficiently harmful to Kosilek to
constitute an Eighth Amendment violation.

-57-
history and record of good behavior -- will create a de facto ban

against SRS as a medical treatment for any incarcerated individual.

We do not agree. For one, the DOC has specifically disclaimed any

attempt to create a blanket policy regarding SRS. We are confident

that the DOC will abide by this assurance, as any such policy would

conflict with the requirement that medical care be individualized

based on a particular prisoner's serious medical needs. See, e.g.,

Roe, 631 F.3d at 862-63 (holding that the failure to conduct an

individualized assessment of a prisoner's needs may violate the

Eighth Amendment).

For another, this case presents unique circumstances; we

are simply unconvinced that our decision on the record before us

today will foreclose all litigants from successfully seeking SRS in

the future. Certain facts in this particular record -- including

the medical providers' non-uniform opinions regarding the necessity

of SRS, Kosilek's criminal history, and the feasibility of post-

operative housing -- were important factors impacting the decision.

D. The Subjective Prong: Deliberate Indifference

1. The DOC's reliance on medical experts

The subjective element of an Eighth Amendment claim for

injunctive relief requires not only that Kosilek show that the

treatment she received was constitutionally inadequate, but also

that the DOC was -- and continues to be -- deliberately indifferent

-58-
to her serious risk of harm. See Farmer, 511 U.S. at 844-45.13 On

the record presented, this is a burden Kosilek cannot meet. Even

if the district court had been correct in its erroneous

determination that SRS was the only medically adequate treatment

for Kosilek's GID, the next relevant inquiry would be whether the

DOC also knew or should have known this fact, but nonetheless

failed to respond in an appropriate manner. See Wilson v. Seiter,

501 U.S. 294, 298 (1991). In answering this question, it is not

the district court's own belief about medical necessity that

controls, but what was known and understood by prison officials in

crafting their policy. Id. at 300 (requiring a showing of

purposefulness or intent on the part of prison administrators).

In this case, the DOC solicited the opinion of multiple

medical professionals and was ultimately presented with two

alternative treatment plans, which were each developed by different

medical experts to mitigate the severity of Kosilek's mental

distress. The choice of a medical option that, although disfavored

by some in the field, is presented by competent professionals does

13
Although the DOC has not specifically argued that the
conflicting medical opinions preclude a finding of subjective
deliberate indifference, we do not find this argument waived. As
we have explained above, the subjective and objective analyses
overlap. See supra note 7; see also Leavitt, 645 F.3d at 498. The
DOC's contention that the district court erred in deeming SRS
medically necessary and in rejecting Dr. Schmidt's approach as
imprudent necessarily entails the DOC's subjective belief that SRS
was unnecessary. The contrary position -- i.e., that SRS is not
objectively necessary but that the DOC did not disagree as to the
need for SRS -- would be wholly illogical.

-59-
not exhibit a level of inattention or callousness to a prisoner's

needs rising to a constitutional violation.14 Cf. Torraco, 923 F.2d

at 234 ("[T]his court has hesitated to find deliberate indifference

to a serious need '[w]here the dispute concerns not the absence of

help, but the choice of a certain course of treatment,' [but]

deliberate indifference may be found where the attention received

is 'so clearly inadequate as to amount to a refusal to provide

essential care.'" (internal citations omitted)). Moreover, a later

court decision -- ruling that the prison administrators were wrong

in their estimation of the treatment's reasonableness -- does not

14
If the prison itself should have been aware that some of the
medical advice it was receiving was imprudent -- that is, if any
layperson could have realized that the advice was imprudent -- then
the decision to still follow that advice may qualify as deliberate
indifference. See Farmer, 511 U.S. at 846 n.9 ("If, for example,
the evidence before a district court establishes that an inmate
faces an objectively intolerable risk of serious injury, the
defendants could not plausibly persist in claiming lack of
awareness . . . ."); Hadix v. Johnson, 367 F.3d 513, 526 (6th Cir.
2004) ("If [the challenged prison conditions] are found to be
objectively unconstitutional, then that finding would also satisfy
the subjective prong because the same information that would lead
to the court's conclusion was available to the prison officials.").
The facts of this case, however, are highly distinct from such a
scenario.

Nor did the district court's conclusion render the DOC's
continued refusal to provide SRS deliberately indifferent. On the
contrary, the evidence was conflicting as to the medical need for
SRS. The choice between reasonable medical views was not for the
district court to make, and the DOC remained entitled to reasonably
rely on Schmidt's and Osborne's expert opinions. Moreover, even
assuming arguendo that the DOC was on notice that its treatment was
insufficient, the DOC's continued refusal also rested on valid
security concerns, discussed below, such that its actions did not
amount to deliberate indifference in any event.

-60-
somehow convert that choice into one exhibiting the sort of

obstinacy and disregard required to find deliberate indifference.

Cf. Nadeau v. Helgemoe, 561 F.2d 411, 417 (1st Cir. 1977) (refusing

to "substitute the values and judgment of a court for the values

and judgment of the . . . prison administration").

2. The DOC's security concerns

The subjective prong also recognizes that, in issues of

security, "[p]rison administrators . . . should be accorded wide-

ranging deference in the adoption and execution of policies and

practices that in their judgment are needed to preserve internal

order and discipline and to maintain institutional security."

Bell, 441 U.S. at 547. Although we cannot "abdicate our

responsibility to ensure that the limits imposed by the

Constitution are not ignored," Blackburn v. Snow, 771 F.2d 556, 562

(1st Cir. 1985), we do not sit to substitute our own judgment for

that of prison administrators, see Nadeau, 561 F.2d at 417. As

long as prison administrators make judgments balancing security and

health concerns that are "within the realm of reason and made in

good faith," their decisions do not amount to a violation of the

Eighth Amendment. Battista, 645 F.3d at 454.

The DOC officials explained that they believed SRS would

create new security issues, the most significant being the

provision of safe housing options for Kosilek after her surgery.

They further explained the importance of keeping other inmates from

-61-
believing that they could use threats of suicide to extract

concessions from the prison administration. Nonetheless, rather

than deferring to the expertise of prison administrators, the

district court ignored the DOC's stated security concerns,

reasoning both that Kosilek could be housed safely and that the DOC

had not acted out of a legitimate concern for Kosilek's safety and

the security of the DOC's facilities. As explained below, this was

in error.

a. The DOC's concerns about safety and security
were reasonable

Recognizing that reasonable concerns would arise

regarding a post-operative, male-to-female transsexual being housed

with male prisoners takes no great stretch of the imagination. See

Farmer, 511 U.S. at 848-49 (summarizing evidence that a prison's

refusal to provide segregated housing to a pre-operative

male-to-female transsexual could pose significant security

concerns). At the same time, as particularly relevant in Kosilek's

case, the DOC's security report reflected that significant concerns

would also arise from housing a formerly male inmate -- with a

criminal history of extreme violence against a female domestic

partner -- within a female prison population containing high

numbers of domestic violence survivors. Nonetheless, in dismissing

the DOC's concerns, the district court relied heavily on the fact

that security issues have not yet arisen within MCI-Norfolk's

general population. Rejecting the testimony of multiple

-62-
individuals with decades of penological experience -- all of whom

acknowledged the risk of housing a female prisoner at MCI-Norfolk

-- the district court reasoned that Kosilek's past safety was

indicative of a likelihood that she could reside safely at the

prison after her operation.

This reasoning wrongly circumvents the deference owed to

prison administrators: the appropriate inquiry was not whether the

court believed that Kosilek could be housed safely, but whether the

DOC has a reasoned basis for its stated concerns. Indeed, that

Kosilek had so far been safe within MCI-Norfolk's prison population

does not negate the DOC's well-reasoned belief that safety concerns

would arise in the future after SRS. Cf. Jones v. N.C. Prisoners'

Labor Union, 433 U.S. 119, 132-33 & n.9 (1977) (holding, in the

First Amendment context, that the rights of prisoners may be

abridged based on a reasonable belief that future harm or

disruption may occur); cf. Hudson v. Palmer, 468 U.S. 517, 526-27

(1984) (requiring prison administrators to implement prophylactic

solutions to foreseeable security issues reasonably within the

scope of their expertise). Moreover, the fact that, pre-

operatively, Kosilek has not been subject to assault or threats

does not vitiate the concern that she would be victimized after

receiving SRS.15

15
These concerns were obvious to more than just those individuals
within the DOC with significant penological experience. The
likelihood that issues surrounding secure housing would arise after

-63-
 The district court also reasoned that "the DOC [could]

reasonably assure the safety of Kosilek and others after sex

reassignment surgery by housing Kosilek in a segregated protective

custody unit." It then noted, however, that there existed a strong

argument that such isolation would amount to "a form of

extrajudicial punishment that is prohibited by the Eighth

Amendment." This warning echoes the very concerns highlighted by

the DOC, which expressed disagreement with the use of long-term

isolation as a housing solution for Kosilek, based on its potential

negative effects on her mental health. See also Battista, 645 F.3d

at 454 (explaining that creating a segregated treatment center to

house a GID prisoner would "pose administrative difficulties and be

isolating"). The deference awarded to prison administrators cannot

be defeated by such circular reasoning, which dismisses the DOC's

concern in one breath only to recognize its validity in the next.

The prison administrators in this case have decades of

combined experience in the management of penological institutions,

and it is they, not the court, who are best situated to determine

what security concerns will arise. See Bell, 441 U.S. at 548

("[J]udicial deference is accorded [in part] because the

administrator ordinarily will . . . have a better grasp of his

domain than the reviewing judge . . . ."). The DOC's judgment

SRS was also acknowledged by Kosilek's treating psychologist, Mark
Burrows, and by the Fenway Center doctors in their initial report.

-64-
regarding post-operative housing is without doubt "within the realm

of reason," Battista, 645 F.3d at 454, and the district court's

alternative belief as to the possibility of safely housing Kosilek

does not suffice to undermine this reasonableness.

The DOC officials also expressed concern that providing

Kosilek SRS would incentivize the use of suicide threats by

prisoners as a means of receiving desired benefits. Although the

district court determined that, in this case, Kosilek's risk for

suicidal ideation was very real, this finding does not invalidate

the DOC's reasonable belief that providing SRS might lead to

proliferation of false threats among other prisoners.

The DOC's concern -- regarding the unacceptable precedent

that would be established in dealing with future threats of suicide

by inmates to force the prison authorities to comply with the

prisoners' particular demands -- cannot be discounted as a minor or

invalid claim. Such threats are not uncommon in prison settings

and require firm rejection by the authorities, who must be given

ample discretion in dealing with such situations. Given the

circumstances presented here, we cannot say that the DOC lacks

reasonable security concerns.

b. Deference to the DOC's reasonable concerns
about safety and security

The district court ultimately dismissed the DOC's

concerns as pretextual, reasoning that DOC was in fact acting in

response to "public and political criticism." The primary evidence

-65-
on record tending to support this theory includes a press interview

by Commissioner Dennehy, Dennehy's relationships with a state

senator and the lieutenant governor, and the acknowledgment that

the DOC was aware of negative news coverage regarding Kosilek's

request for surgery.

In her testimony, Dennehy denied being influenced by such

media and political pressures, and stated that the decision not to

provide SRS was founded in bona fide security concerns alone. The

district court, however, found this testimony non-credible, and

this credibility finding is the sort of determination to which our

court gives deference. See Fed. R. Civ. P. 52(a)(6). Even

accepting that Dennehy's motivations were colored by political and

media pressure, however, does not take Kosilek's claim as far as it

needs to go.

As an initial matter, the fact that Dennehy was motivated

in part by concerns unrelated to prison security does not mean that

the security concerns articulated by the DOC were irrelevant,

wholly pretextual, or -- most importantly -- invalid on the merits.

In Battista, our court held that deference to the decisions of

prison administrators could be overcome where those administrators

admittedly relied on inflated data, identified a security concern

only several years after refusing to provide treatment for an

acknowledged medical need, and engaged in a pattern of changing

positions and arguments before the court. Battista, 645 F.3d at

-66-
455. Such gross delays and misstatements were not present here.16

Rather, the DOC testified consistently that it believed the post-

operative security concerns surrounding Kosilek's treatment were

significant and problematic.17 Even if not entitled to deference,

see id., those concerns still matter insofar as they are reasonable

and valid, and Kosilek did not put on any evidence showing that

they wholly lacked merit.18

16
Great weight was placed on the fact that Dennehy told a reporter
that there were significant security concerns about post-operative
housing three days before she met with Superintendents Spencer and
Bissonnette. The record reveals, however, that discussions about
housing had previously occurred at Executive Staff Meetings, and
Dennehy testified that she had conducted phone calls with both
Superintendents prior to meeting to formalize their security
report. This timeline, therefore, is far from sufficient to
establish that the DOC's security assessments were unprincipled or
invalid.
17
That the DOC may have, in the district court's assessment,
engaged in a pattern of prevarication regarding whether they
understood that SRS was being recommended by UMass as medically
necessary, does not undercut the consistency with which they
identified safety and security concerns -- concerns which are
within their expert province -- that would arise from the surgery.
18
Kosilek did cross-examine Commissioner Clarke to show that a
transgendered prisoner had safely been housed in a Washington State
prison under his supervision. Left unexplored, however, were the
numerous ways in which MCI-Norfolk's environment, facilities, or
population might be distinct from this prison in Washington.
Neither was there a comparison between that prisoner's criminal
history and the criminal history of Kosilek. That an individual
was housed safely by Commissioner Clarke while employed in another
state does not rebut Superintendent Bissonnette's testimony that
moving her to MCI-Framingham would cause climate problems in that
particular prison. See Feeley v. Sampson, 570 F.2d 364, 371 (1st
Cir. 1978) (rejecting uniform housing conditions for detainees,
without regard to their disparate criminal history, because
"Constitutional rights cannot be defined in terms of literal
comparisons of this nature").

-67-
 Second, when determining the appropriateness of

injunctive relief, our focus must include "current attitudes and

conduct." Farmer, 511 U.S. at 845 ("'[D]eliberate indifference[]

should be determined in light of the prison authorities' current

attitudes and conduct': their attitudes and conduct at the time

suit is brought and persisting thereafter." (quoting Helling, 509

U.S. at 36)). Dennehy has not served as DOC Commissioner since

2007. Given the age of this litigation and the changes in DOC

leadership that have occurred since the suit was filed, the

district court's assumption that Dennehy's attitudes necessarily

carried over to her successors and governed their actions is

unsupported by the record. Although consideration of Dennehy's

motivation is surely relevant, it is insufficient to show that the

DOC continued to be motivated by public pressure even after her

departure, or that this is what motivates the DOC presently.

Indeed, it was Commissioner Clarke -- and not Dennehy --

who made the decision here. And the only evidence tending to show

that Commissioner Clarke may have considered public and political

criticism were two letters received by Clarke -- who did not

respond -- from Massachusetts legislators. These letters, however,

relate almost in their entirety to concerns about the cost of SRS,

and the district court soundly rejected any argument that the DOC,

or Clarke specifically, had adopted its safety and security

measures as a pretextual means of addressing the cost concerns

-68-
raised by state legislators. Moreover, Clarke was never found by

the court to be noncredible.19

The district court improperly imputed its belief that

Commissioner Dennehy had acted out of concern for public and

political pressure to its assessment of the motivations of future

DOC Commissioners. This error ignores the requirement, in cases of

injunctive relief, that a court consider the attitudes and beliefs

of prison administrators at the time of its decision. Id. at 845-

46. The effect of this error is particularly clear given that

Clarke has now been replaced by Commissioner Spencer, so that

Dennehy is now several administrations and more than seven years

removed from the decisionmaking process. Without proof that the

DOC remains motivated by pretextual or improper concerns with

public pressure, even if it was assumed that Dennehy was improperly

motivated, the district court's finding that injunctive relief was

required is unsupportable.

III. Conclusion

We are not tasked today with deciding whether the refusal

to provide SRS is uncompassionate or less than ideal. Neither

finding would support Kosilek's claims of a constitutional

violation. The Eighth Amendment, after all, proscribes only

medical care so unconscionable as to fall below society's minimum

19
We further note that the DOC has not defended this case based
on cost considerations relating to the provision of SRS.

-69-
standards of decency. See Estelle, 429 U.S. at 102-05. In this

case, the DOC has chosen to provide a form of care that offers

direct treatment for Kosilek's GID. Cf. Leavitt, 645 F.3d at 498

(acknowledging that the effects of treatment decisions may be

relevant to consideration of the subjective component of the Eighth

Amendment). Moreover, it has done so in light of the fact that

provision of SRS would create new and additional security concerns

-- concerns that do not presently arise from its current treatment

regimen.

Given the positive effects of Kosilek's current regimen

of care, and the DOC's plan to treat suicidal ideation should it

arise, the DOC's decision not to provide SRS does not illustrate

severe obstinacy or disregard of Kosilek's medical needs.

DesRosiers, 949 F.2d at 19 ("[T]he complainant must prove that the

defendants had a culpable state of mind and intended wantonly to

inflict pain."). Rather, it is a measured response to the valid

security concerns identified by the DOC. Battista, 645 F.3d at 454

("Medical 'need' in real life is an elastic term: security

considerations also matter at prisons. . . ."); Cameron, 990 F.2d

at 20 ("Nothing in the Constitution mechanically gives controlling

weight to one set of professional judgments."). Having reviewed

the record before us, we conclude that Kosilek has failed, on these

facts, to demonstrate an Eighth Amendment violation. Accordingly,

we reverse the district court's order of injunctive relief and

-70-
remand this case to the district court with instructions to dismiss

the case.

Reversed and Remanded.

-Dissenting Opinions Follow-

-71-
 THOMPSON, Circuit Judge, dissenting. The majority turns

a blind eye to binding precedent, opting instead to cobble together

law from other circuits and non-Eighth Amendment jurisprudence to

formulate a standard of review that, though articulated as one of

variable exactitude, amounts to sweeping de novo review. Armed

with the ability to take a fresh look at findings that clearly

warranted deference, the majority easily steps into the trial

judge's shoes -- the inarguable superiority of the judge's ability

to marshal facts, assess motive, and gauge credibility all but

forgotten. The parameters set by the majority foretold the result.

It concludes that the Massachusetts Department of Correction did

not violate Michelle Kosilek's constitutional rights. That

conclusion is erroneous, the majority's analytical path to it is

misguided, and the fact that this case is even subject to en banc

scrutiny in the first place is wrong. And so I dissent.

I. En Banc Grant

The criteria for en banc relief are clear: it is not a

favored form of relief, and ordinarily should not be ordered unless

"(1) en banc consideration is necessary to secure or maintain

uniformity of the court's decisions; or (2) the proceeding involves

a question of exceptional importance." Fed. R. App. P. 35(a). My

colleagues' reasons for granting en banc review are not

articulated, but it seems clear that the maintenance of uniformity

piece is not in play. Therefore I can only assume they perceive an

-72-
issue of exceptional importance. This justification is

problematic.

As my colleague has explained in a series of thoughtful

dissents, in this circuit there has been what some might see "as

the recurring unprincipled denial and granting of petitions for

rehearing en banc, without any attempt to define and apply a set of

objective criteria to determine when a case is of exceptional

importance." Kolbe v. BAC Home Loans Servicing, LP, 738 F.3d 432,

474 (1st Cir. 2013) (Torruella, J., dissenting); see also Igartúa

v. United States, 654 F.3d 99, 105 (1st Cir. 2011) (Torruella, J.,

filing opinion concerning denial of en banc consideration); United

States v. Vega-Santiago, 519 F.3d 1, 7 (1st Cir. 2008) (Torruella,

J., dissenting). I am at a loss to see what objective criteria

warranted review in this case.

While the relief ordered by the district court, and

affirmed by a majority of the original panel, was unprecedented,

Kosilek's case is not a legally complicated one. Rather it is a

fact-intensive dispute, which required the original panel to

determine whether the district court's take on the significant

amount of evidence, and its ultimate holding as to the existence of

an Eighth Amendment violation, was erroneous. I fail to see what

in this framework made this case worthy of en banc review.

I am not implying this case is unimportant. This

litigation is significant to Kosilek, the DOC, and many others, and

-73-
the rights afforded under the Eighth Amendment are crucial. But if

those things alone were enough, nearly every case would attract the

full court's attention. And a good deal more cases would be heard

en banc if disagreeing with the result reached by the original

panel, or simply desiring to weigh in, were valid grounds for

awarding en banc review. They are not, but unfortunately I suspect

they were the grounds that carried the day here. See, e.g., Kolbe,

738 F.3d at 474 (Torruella, J., dissenting) ("En banc consideration

is not for the purpose of correcting panel decisions.") (citing

Calderón v. Thompson, 523 U.S. 538, 569 (1998) (Souter, J.,

dissenting)).

This case does not satisfy the well-settled requirements

for a grant of en banc. Lamentably, a majority of this court

decided otherwise. Similarly, a majority has decided that the

district court got it wrong. That conclusion is fundamentally

flawed, starting with the level of scrutiny paid to the lower

court's decision.

II. Standard of Review

The issue of what standard of review should be employed

is a significant point of divergence for me, and indeed one that

permeates the entirety of my discord with the majority. The

majority, undoubtedly aware that it could more handily toss aside

the district court's findings if it utilized a non-deferential

standard of review, formulates its standard by borrowing liberally

-74-
from other circuits and non-Eighth Amendment jurisprudence while

disregarding on-point case law from this circuit. The end result

is a standard that, in theory, afforded minimal deference to the

lower court's finding, and in the majority's actual application,

afforded essentially none.

Let me start with our common ground. I agree with the

majority that different standards of review are in play. When

deciding a post-bench-trial appeal, this court takes up questions

of law de novo, but reviews findings of fact for clear error only.

Wojciechowicz v. United States, 582 F.3d 57, 66 (1st Cir. 2009).

On the latter point, this means we accept the court's factual

findings, and the inferences drawn from those facts, unless the

evidence compels us to conclude a mistake was made. Janeiro v.

Urological Surgery Prof'l Ass'n, 457 F.3d 130, 138 (1st Cir. 2006).

With inquiries that are more of a mixed bag, there is a continuum.

Johnson v. Watts Regulator Co., 63 F.3d 1129, 1132 (1st Cir. 1995).

The more fact-intensive the question, the more deferential our

review. Id. Conversely, the more law-dominated the query, the

more likely our review is de novo. Id.

That is where the congruity ends. The majority,

undoubtedly with a certain end result in mind, maneuvers the

standard of review into its most favorable form. While it

correctly acknowledges that factual and legal issues are

implicated, the majority utterly favors the de novo end of the

-75-
spectrum.20 This approach does not accord with our case law

(although to read the majority you would think we had very little

on-point jurisprudence in this circuit).

For one, the majority posits that the issue of deliberate

indifference is a legal one to be reviewed de novo. It relies on

Fourth Amendment jurisprudence, citing criminal cases that, in the

context of deciding the validity of searches and seizures, hold

that reasonable suspicion and probable cause determinations should

receive de novo appellate review. See Ornelas v. United States,

517 U.S. 690, 699 (1996); United States v. Camacho, 661 F.3d 718,

724 (1st Cir. 2011). I do not see how these cases are analogous to

Kosilek's challenge, nor why we should look to Fourth Amendment

cases rather than our Eighth Amendment jurisprudence.

In the context of the Eighth Amendment, we have explained

that the existence of deliberate indifference is a "state-of-mind

issue" that usually presents a jury question, Torraco v. Maloney,

923 F.2d 231, 234 (1st Cir. 1991), or in other words, an issue for

20
The majority's decision to give little deference to the district
court is undoubtedly a boon to the DOC, and given the DOC's
garbled treatment of the standard of review issue on appeal, it is
a downright windfall. In violation of our rules, the DOC did not
include a standard of review in its opening brief. See Fed. R.
App. P. 28(a)(8)(B). In its reply brief, the DOC gave us a bit
more, arguing that the appropriateness of medical care called for
de novo review but neglecting to indicate what scrutiny a
deliberate indifference finding necessitated. In its petition for
en banc review, the DOC's position continued to evolve. It
contended that a heightened standard of review should be applied
because this case involves intertwined issues of law and fact.

-76-
the finder of fact. This makes sense. Often intertwined in state-

of-mind issues are determinations about credibility and motivation;

those are classic examples of the judgment calls to which we give

deference. See Fed. R. Civ. P. 52(a)(6) ("[T]he reviewing court

must give due regard to the trial court's opportunity to judge the

witnesses' credibility."); Monahan v. Romney, 625 F.3d 42, 46 (1st

Cir. 2010). See also Janeiro, 457 F.3d at 138-39 (explaining that,

following a bench trial, "if the trial court's reading of the

record [with respect to an actor's motivation] is plausible,

appellate review is at an end") (alteration in original).

The majority recognizes Torraco, citing it for the narrow

proposition that "issues of culpability in a deliberate

indifference inquiry are usually questions for a jury," in

connection with its discussion about what standard of review

findings of fact garner. But this is a mischaracterization of what

Torraco held. Rather, the case states that "the existence of

deliberate indifference," is a state-of-mind issue, which makes it

a typical juror question. Torraco, 923 F.2d at 234 (emphasis

added). The majority's slight spin on this holding allows it to

ignore Torraco, and lean on Fourth Amendment jurisprudence instead

to support the notion that deliberate indifference gets a fresh

look from this court.

Similarly erroneous is the majority's position that we

review de novo the district court's ultimate determination as to

-77-
whether an Eighth Amendment violation occurred. For support it

cites to a series of Eighth Amendment cases from other circuit

courts. See, e.g., Thomas v. Bryant, 614 F.3d 1288, 1307 (11th

Cir. 2010); Hallett v. Morgan, 296 F.3d 732, 744 (9th Cir. 2002).

At first blush, there is some surface appeal to this position. If

nothing else, the existence of a constitutional violation sounds

like something that would fall closer to the question-of-law end of

the spectrum. The problem though is that the ultimate

constitutional question is inextricably tied up with the factual

details that emerged at trial, the credibility of the witnesses,

and the questions of motivation. This counsels against pure de

novo review and our own case law supports this notion.

As explained above, a state-of-mind issue such as the

existence of deliberate indifference is typically left to the

finder of fact. Torraco, 923 F.2d at 234. And when reviewing a

trial judge's determination on the adequacy of medical treatment

following a bench trial, this court has applied the deferential

clearly erroneous standard. DesRosiers v. Moran, 949 F.2d 15, 19-

20 (1st Cir. 1991). On top of this, it is well established that

"elusive issues of motive and intent" (relevant here in connection

with the Eighth Amendment's subjective prong) are typically fact-

bound ones subject to the clearly erroneous rule. Fed. Refinance

Co. v. Klock, 352 F.3d 16, 27-28 (1st Cir. 2003); see also McIntyre

ex rel. Estate of McIntyre v. United States, 545 F.3d 27, 40 (1st

-78-
Cir. 2008). Thus the major pieces of the puzzle in an Eighth

Amendment inquiry -- adequacy of medical care, the existence of

deliberate indifference, and the parties' motive and intent -- are

subject to the clearly erroneous standard, making unqualified de

novo review a bad fit.

Policy concerns do not counsel otherwise, making the

majority's reliance on Ornelas, 517 U.S. at 690, a Fourth Amendment

case, not particularly persuasive.21 Ornelas, which characterized

the ultimate reasonable suspicion and probable cause determination

as a mixed question of law and fact, decided that de novo review

was the best fit for its resolution. Id. at 696-97. The Supreme

Court, as the majority points out, emphasized that "[i]ndependent

review" by appellate courts can help "to maintain control of, and

to clarify, the legal principles" in reasonable suspicion and

probable cause cases. Id. at 697. While I do not disagree that as

an appellate court we are often required to clarify legal

21
The majority also relies on United States v. Bajakajian, 524
U.S. 321 (1998), an Eighth Amendment excessive punishment and fines
case, for the same proposition it cites Ornelas for. Specifically,
the majority states that in Bajakajian, the Supreme Court reasoned
that the "'application of a constitutional standard to the facts of
a particular case' . . . may appropriately require de novo
appellate review to ensure consistency in the law's development."
Bajakajian does not say this. The Court there did not address the
concept of consistency of the law; it simply cited Ornelas for the
narrower proposition that de novo review attaches to the issue of
whether a fine is constitutionally excessive. See id. 336, n.10.
For that reason I focus on Ornelas.

-79-
principles and ensure continuity of the law's development, this is

not a persuasive justification for employing de novo review here.

As noted by the dissent in Ornelas, "[l]aw clarification

requires generalization, and some issues lend themselves to

generalization much more than others." Id. at 703 (Scalia, J.,

dissenting). The issues here do not. Cases dealing with the

constitutional adequacy of medical care under the Eighth Amendment

are incredibly fact-specific, resulting in distinctive issues. The

trial judge must, among other things, have a handle on the

prisoner's medical condition, the treatment sought, the treatment

provided (if any), what treatment medical providers recommended,

what the defendant knew and when, and what motivated its decisions.

This court cannot hope to match the district judge's expertise in

these areas, nor can I fathom why we would want to try. The

"extremely fact-bound nature" of these cases means that "de novo

review [will] have relatively little benefit," id. at 700 (Scalia,

J., dissenting), leaving us unmoved by the uniformity-of-the-law

considerations raised by the majority.

So where does all this leave us with regard to the

standard that attaches to the determination of whether the Eighth

Amendment has been violated? It is clear (and the majority agrees)

that with questions of varying exactitude, the "standard of review

applied depends, in the last analysis, on the extent to which a

particular question is fact-dominated or law-dominated." Turner v.

-80-
United States, 699 F.3d 578, 584 (1st Cir. 2012) (internal

quotation marks omitted); see also In re IDC Clambakes, Inc., 727

F.3d 58, 64 (1st Cir. 2013); Dugas v. Coplan, 506 F.3d 1, 8 (1st

Cir. 2007). Drawing the distinction between law-heavy versus fact-

heavy questions is sometimes a tricky thing to do, and given that

establishing an Eighth Amendment claim involves a mixed question of

law and fact, it is a thicket into which we must enter. Luckily,

I do not think it is a particularly thorny one in this case.

Here, before reaching its ultimate constitutional

conclusion, the trial court heard testimony from no fewer than

nineteen witnesses (e.g., medical providers, medical experts,

prison officials, and Kosilek) over the course of a trial that

ultimately extended two years. The court scrutinized events that

had transpired over a twenty-year period, including those relating

to what treatment Kosilek had requested, what treatment had been

recommended, and what care was ultimately provided. The court

considered evidence about the DOC's security review, how it was

conducted, and the concerns it raised. It assessed the credibility

of Kosilek, DOC officials, and the medical experts. The court

reviewed a copious amount of exhibits, such as Kosilek's medical

records, Kosilek's prison records, DOC policies, DOC contracts, DOC

manuals, reports from Kosilek's medical providers, reports penned

by each side's experts, DOC staff meeting notes, security reports,

medical literature, correspondence, and deposition testimony. The

-81-
end result was pages upon pages of factual findings made by the

trial judge.22

In other words, the district court "engaged in a careful

and close analysis of the trial evidence," Turner, 699 F.3d at

584, to make its ultimate determination that the DOC, without any

valid penological purpose, refused to provide medically necessary

treatment for Kosilek's life-threatening condition. Given the

clearly fact-intensive nature of the court's review, our own

examination into whether the court was correct that the DOC

violated the Eighth Amendment should be deferential, as opposed to

the fresh look the majority proposes.23 See id.; Fed. Refinance

Co., 352 F.3d at 27 (explaining that the more fact-intensive the

question, the more deferential our review). As ably said by the

Supreme Court, "deferential review of mixed questions of law and

fact is warranted when it appears that the district court is better

positioned than the appellate court to decide the issue in

22
Indeed the majority dedicates over thirty pages of its opinion
to the factual and procedural background in this case. This is not
surprising; those facts are integral to the resolution of the
constitutional question. What is surprising is the majority's
failure to see the significance of the factually concentrated
nature of this case.
23
Plus, even assuming that the conclusion that the DOC's refusal
to provide care constituted an Eighth Amendment violation lands
closer to the law side of the mixed-question spectrum, a measure of
deference is still appropriate. See Battista v. Clarke, 645 F.3d
449, 454 (1st Cir. 2011) ("The legal labels applied to facts are
reviewed on appeal more closely than a district court fact-finding,
but often with some deference to the district judge.").

-82-
question, or that probing appellate scrutiny will not contribute to

the clarity of legal doctrine." Salve Regina Coll. v. Russell, 499

U.S. 225, 233 (1991) (internal quotation marks omitted).

The majority's articulation of a standard of review that

runs afoul of our case law is not the only problem. There is also

its application. While the majority's skewed standard allows

minimal aspects of the lower court's decision to garner clear error

review, namely factual findings and credibility determinations, in

actual application essentially no deference was paid. The only

conclusion of the district court that the majority concedes

warranted deference was the judge's determination that Commissioner

Kathleen Dennehy's testimony was not credible.24 Given the

voluminous record in this case, and the breadth of the lower

court's findings, it is simply unfathomable that the majority did

not consider a single other fact-drawn inference, credibility

finding, or motive determination, all of which warrant deference.

For instance, the district court drew inferences from the

various medical providers' testimony to decide what constituted a

prudent approach. It also considered what Commissioner Harold

Clarke's motivations were for denying sex reassignment surgery. It

drew inferences from the DOC's conduct (e.g., the timing of

security reviews and the DOC's communications with Kosilek's

24
Of course the majority then goes on to explain why the court's
adverse credibility determination does not matter, a point I will
get into later.

-83-
medical providers) to determine that the DOC had engaged in

prevarication and delay. The majority; however, does not appear to

adjust its consideration of these issues to reflect any deference

to the trial judge. Rather it decides anew what inferences should

be drawn from the facts attested to at trial. Even under the

majority's standard, this is not proper.

Without doubt, the level of scrutiny applied by a court

permeates its analysis and guides the outcome. The impact here is

clear. The Eighth Amendment is violated when prison officials fail

to provide an inmate with adequate medical care, such that "their

'acts or omissions [are] sufficiently harmful to evidence

deliberate indifference to serious medical needs.'" Leavitt v.

Corr. Med. Servs., Inc., 645 F.3d 484, 497 (1st Cir. 2011) (citing

Estelle v. Gamble, 429 U.S. 97, 106 (1976)). The district court

concluded that the evidence established the DOC had committed such

a violation. The majority says otherwise but its analysis is

plagued with flaws, starting with its determination as to the

objective prong.

III. Eighth Amendment: Objective Prong

Whether the so-called objective component of the Eighth

Amendment inquiry is satisfied turns on whether the alleged

deprivation is "objectively, sufficiently serious." Farmer v.

Brennan, 511 U.S. 825, 834 (1994) (internal quotation marks

omitted); Leavitt, 645 F.3d at 497. In this context, a prisoner

-84-
with a "serious medical need," Mahan v. Plymouth Cnty. House of

Corr., 64 F.3d 14, 17-18 (1st Cir. 1995), is entitled to adequate

medical care, i.e, "services at a level reasonably commensurate

with modern medical science and of a quality acceptable within

prudent professional standards." United States v. DeCologero, 821

F.2d 39, 43 (1st Cir. 1987).

That gender identity disorder is a serious medical need

which warrants treatment, is not, as the majority notes, disputed.

The disagreement -- both between the parties and amongst this en

banc court -- centers around whether the district court correctly

found that the DOC's proffered regimen of care was inadequate, and

that sex reassignment surgery is the only appropriate treatment for

Kosilek. Based on the record, and when one employs the proper

standard of review, that conclusion was generously supported by the

evidence.

A. Dr. Schmidt's Prudence

To start, despite the majority's qualms, the district

court's conclusion that the DOC's expert, Dr. Chester Schmidt, was

not a prudent professional was not clearly erroneous. In his

testimony, Dr. Schmidt expressed a good deal of disagreement with

the Harry Benjamin Standards of Care, which were widely relied upon

by the other medical providers who testified below and which have

been generally accepted by courts. See, e.g., De'Lonta v. Johnson,

708 F.3d 520, 522-23 (4th Cir. 2013) (describing the Standards of

-85-
Care as "the generally accepted protocols for the treatment of

GID"); Soneeya v. Spencer, 851 F. Supp. 2d 228, 231 (D. Mass. 2012)

(noting that the "course of treatment for Gender Identity Disorder

generally followed in the community is governed by the 'Standards

of Care'"); O'Donnabhain v. Comm'r of Internal Revenue, 134 T.C.

34, 65 (U.S. Tax Ct. 2010) (indicating that the Standards are

"widely accepted in the psychiatric profession, as evidenced by the

recognition of the standards' triadic therapy sequence as the

appropriate treatment for GID and transsexualism in numerous

psychiatric and medical reference texts").

While, as the majority notes, the Standards of Care have

a built-in flexibility, that pliancy appears to stem from the

uniqueness of patient needs and the evolution of the gender

identity disorder field.25 Dr. Schmidt's departure from the

Standards appeared more fundamental. For instance, the Standards

of Care explained that sex reassignment surgery is not

"experimental, investigational, elective, cosmetic, or optional in

any meaningful sense." Standards of Care, Version 6, at 18. Dr.

Schmidt disagreed. In his expert report, he wrote that sex

reassignment surgery was a "voluntary, elective choice[] and

procedure[]," calling the steps towards reassignment "equivalent to

25
The Standards state: "Clinical departures from these guidelines
may come about because of a patient's unique anatomic, social, or
psychological situation, an experienced professional's evolving
method of handling a common situation, or a research protocol."
Standards of Care, Version 6, at 2.

-86-
a variety of elective cosmetic non-surgical procedures and elective

cosmetic surgical procedures." Another example: the Standards of

Care provide that, for persons with severe gender identity

disorder, sex reassignment surgery is effective, and when paired

with hormone therapy and a real-life experience, "medically

indicated and medically necessary." Standards of Care, Version 6,

at 18. Dr. Schmidt again was not on board. He testified that

generally he does not believe that sex reassignment surgery is

medically necessary and his practice manifests this philosophy. In

the approximately 300 patients he had evaluated, Dr. Schmidt never

recalled seeing even one case of gender identity disorder serious

enough to warrant surgery.26

For Dr. Schmidt, there was an additional wrinkle. In Dr.

Schmidt's opinion, a real-life experience living as the opposite

gender could not be effectively replicated in prison, and this

counseled against surgery for Kosilek. The district court found

that this viewpoint was not prudent. The majority claims that in

doing so the court "relied on its own -- non-medical -- judgment

26
The majority makes much of the district judge faulting Dr.
Schmidt for not writing letters of recommendation for patients
seeking sex reassignment surgery, suggesting that the judge did not
appreciate the nuance between opening the door for surgery and
advocating for it. I suspect the judge was more broadly concerned
with the fact that Dr. Schmidt did not think sex reassignment was
ever medically necessary, nor had he ever seen a case where it was
warranted. And despite having this strident perspective, Dr.
Schmidt nonetheless opened the door for patients to undergo this
major medical procedure.

-87-
about what constitutes a real-life experience." This is not

accurate.

The court based its determination, back in Kosilek I, on

the testimony of Dr. Marshall Forstein and Dr. George Brown, who

"convincingly testified [that] Kosilek's 'real life' is prison."

Kosilek v. Maloney, 221 F. Supp. 2d 156, 167 (D. Mass. 2002). Then

in Kosilek II, the court found the "credible evidence in the

instant case confirmed the conclusion in Kosilek I that a person

can have a 'real life experience' in prison." Kosilek v. Spencer,

889 F. Supp. 2d 190, 232 (D. Mass. 2012). Evidence before the

court in Kosilek II included an expert report from Dr. Forstein,

and testimony from Dr. Randi Kaufman, both of whom indicated that

Kosilek had undergone a real-life experience in prison. There was

also the February 24, 2005 report from the Fenway doctors, Dr.

Kevin Kapila and Dr. Kaufman, which explained that Kosilek had

moved successfully through the steps outlined by the Standards of

Care. Then, in their October 7, 2005 report, the Fenway doctors

explained at length why Cynthia Osborne's review subtly distorted

the concept of the real-life experience, and why Kosilek had

completed the real-life test -- a test made even more stringent by

the fact that she was living as a female in an all-male prison.

Dr. Brown echoed a similar sentiment. He testified that Kosilek

had not only met the minimum real-life experience but had exceeded

it. Dr. Brown focused on the significant amount of information

-88-
that existed regarding Kosilek's time in prison, a record that his

patients in the outside world would never have.27

The record is clear. The district court's determination

that Dr. Schmidt's viewpoint about the feasibility of a real-life

experience in prison was not based on the judge's own lay opinion.

It was, as the district court alluded to, grounded in a significant

amount of evidence offered by competent medical professionals, all

of whom disagreed with Dr. Schmidt.28

The same can be said about what course of treatment was

appropriate for Kosilek. Dr. Schmidt testified that Kosilek had

"made an excellent adaptation" on her current treatment regimen and

that surgery would not "confer any additional functional

capability." Surgery was not, according to Dr. Schmidt, medically

necessary for Kosilek. To minimize the risk of future harm to

27
The majority mentions that none of the experts who opined that
Kosilek completed a real-life experience considered that she might
be housed in a female facility post-surgery. This is hardly
surprising as this is a theory of my colleagues' own making. The
DOC never made any argument that a potential post-surgery housing
change rendered Kosilek unable to complete the real-life
experience, nor did any provider opine that it was even a
consideration.
28
The Seventh Version of the Standards of Care came out in 2011.
Notably it contains a new section devoted to scenarios where
persons with gender identity disorder are living in institutional
environments such as prisons or long-term care facilities.
Standards of Care, Version 7, at 67. It provides that those
individuals' health care "should mirror that which would be
available to them if they were living in a non-institutional
setting" and that "[a]ll elements of assessment and treatment as
described in the [Standards of Care] can be provided to people
living in institutions." Id.

-89-
Kosilek, Dr. Schmidt thought employing psychotherapy and medication

to reduce her dysphoria and, if needed, placing Kosilek in a

medical facility would be effective. A majority of the testifying

medical providers said otherwise though. When asked what they

thought about Dr. Schmidt's suggested regimen, Drs. Kapila,

Kaufman, Appelbaum, and Forstein all thought it unreasonable.29 The

common thinking was that Dr. Schmidt's approach was not likely to

effectively reduce Kosilek's risk of self harm, given that the

source of her dysphoria was her male genitalia.

In the Eighth Amendment context, the adequacy of medical

care is "measured against 'prudent professional standards.'" Nunes

v. Mass. Dept. of Corr., 766 F.3d 136, 142 (1st Cir. 2014) (quoting

DeCologero, 821 F.2d at 43). The district court here concluded

that Dr. Schmidt was not a prudent professional. Given the above,

I am not convinced that this determination was clearly erroneous.

Dr. Schmidt's significant disagreement with widely accepted

guidelines and the sharp contrast between his and the other well-

credentialed providers' opinions, offer strong support for the

court's finding.

29
Court-appointed expert, Dr. Stephen Levine, ultimately testified
that from a purely medical perspective (absent considerations
relative to the prison environment), a prudent professional would
not deny Kosilek sex reassignment surgery. However, Dr. Levine
initially opined that Dr. Schmidt's view was reasonable (if not
popular), a discrepancy that apparently arose from Dr. Levine
disregarding the district court's order to treat Kosilek as a
patient in free society. Considering this incongruity, I do not
list Dr. Levine as one of Dr. Schmidt's critics.

-90-
 B. Adequacy of the DOC's Treatment

In light of the court's determination as to Dr. Schmidt's

prudence, the question remains whether the evidence supported its

conclusion that the DOC's treatment was not medically adequate.

The majority's consideration of this issue begins with a faulty

premise. It states that the "district court held that

psychotherapy and antidepressants alone would not adequately treat

Koslilek's GID," a finding the majority calls an incorrect

characterization of the issues, and a minimization of the DOC's

proffered treatment plan. It is the majority who is wrong.

The district court was of course well aware that the DOC

was suggesting a more comprehensive treatment plan beyond therapy

and medication. Nonetheless, as it repeatedly explained, it found

that all treatment other than sex reassignment surgery was

inadequate for Kosilek. See, e.g., Kosilek, 889 F. Supp. 2d at

202, 233, 236, 238, 240. This included the DOC's past treatment,

as well as its intended treatment going forward. In other words,

the court did not minimize the DOC's regimen. Based on the

testimony and evidence presented, it simply found the regimen did

not, and would not going forward, adequately treat Kosilek's gender

identity disorder. This finding was well within the court's

purview to make. The fact that the DOC fashioned some treatment,

in the form of hormone therapy, electrolysis, and access to

feminine items does not insulate it from liability. In De'Lonta v.

-91-
Johnson, the Fourth Circuit Court of Appeals found that an inmate,

who sought sex reassignment surgery after her gender identity

disorder failed to resolve despite receiving hormones, stated a

plausible deliberate indifference claim. 708 F.3d at 522, 525.

The court concluded that, though the Virginia Department of

Corrections had provided the inmate with hormone therapy and

psychological counseling consistent with the Standards of Care, "it

does not follow that they have necessarily provided her with

constitutionally adequate treatment." Id. at 522, 526 (emphasis in

original).

The majority nonetheless would have us believe the care

provided by the DOC can withstand constitutional scrutiny. It

endeavors to convince by giving little weight to the attested to

shortcomings in Kosilek's treatment plan, and instead focusing

heavily on the improvement Kosilek has made since being provided

hormones, electrolysis, feminine garb and gear, and mental health

treatment. This is not in dispute; Kosilek has indeed progressed.

However, despite the short shrift the majority pays it, there was

ample evidence supporting the district court's conclusion that this

improvement was not sufficient to ease Kosilek's suffering to a

point where she was no longer facing a life-threatening risk of

harm.

Though the DOC has been treating Kosilek for many years,

the district court found that she "continues to suffer intense

-92-
mental anguish." Kosilek, 889 F. Supp. 2d at 202. The court

chronicled the evidence: Kosilek's own testimony about her

continued distress,30 the Fenway Center report indicating Kosilek's

ongoing angst over her male genitalia and the high likelihood of

another suicide attempt, and the along-the-same-lines testimony of

Kosilek's treating psychologist, Mark Burrowes. See id. at 226.

There was also Dr. Kaufman's testimony that, even with the

treatment the DOC provided, Kosilek still suffered from clinically

significant distress and severe dysphoria, a fact she found "quite

notable." Dr. Brown testified similarly, explaining that Kosilek's

treatment to date, including the hormones, had not obviated her

need for surgery. Further, there was evidence that Kosilek's

improvement was tangled up in her continuing hope that sex

reassignment surgery would be provided. Dr. Brown testified: "And

without that hope, the [DOC's] treatments are -- I wouldn't say for

naught, but they are not going to continue her level of improvement

where she is now."

Thus, even with Kosilek's documented improvement, Drs.

Brown, Kaufman, Forstein, Kapila, and Appelbaum all testified

unequivocally that sex reassignment surgery was medically necessary

and the only appropriate treatment for Kosilek. They further

30
The court found Kosilek testified credibly that although hormone
treatments had helped, she was distressed by her male genitalia and
believed that she needed surgery. Antidepressants and
psychotherapy, according to Kosilek, would not alter the fact that
she did not want to continue living with her male genitalia.

-93-
agreed that there was a serious risk of harm, most likely suicide,

should Kosilek not receive the surgery, which was a concern the

Fenway doctors voiced as early as 2005. As the majority says, this

potentiality matters because the Eighth Amendment's protections

extend beyond present suffering to future harm. See Helling v.

McKinney, 509 U.S. 25, 33-34 (1993); Leavitt, 645 F.3d at 501.

The DOC's assertion that this future risk could be curbed

with medication and psychotherapy cannot carry the day. As the

district court found, treating the underlying disorder and its

symptoms are two very different things, a distinction also drawn by

the Seventh Circuit. See Fields v. Smith, 653 F.3d 550 (7th Cir.

2011). In Fields, the court found a Wisconsin statute that

prohibited the state's correctional department from providing

transgender inmates with hormones and sex reassignment surgery

unconstitutional. Id. at 552-53, 559. The court, discussing how

some patients require hormone therapy, found the department of

corrections had not effectively rebutted the evidence that an

offering of medication and psychotherapy would "do nothing to treat

the underlying disorder." Id. at 556. In the instant matter, Drs.

Appelbaum and Kapila testified that the preferred approach is to

treat the underlying problem -- Kosilek's gender identity disorder

-- as opposed to the symptoms it might produce. As chronicled

above, the consensus was that the only way to adequately treat that

problem was with sex reassignment surgery.

-94-
 Lest we forget, the procedural posture of this case bears

another mention. The DOC is challenging the district court's grant

of injunctive relief following a bench trial, meaning that due

regard is paid to the judge's factual findings and credibility

determinations. See Monahan, 625 F.3d at 46. When the evidence

yields competing inferences or two permissible views, we cannot

second guess, "even if, had we been sitting as triers of the facts,

we might have arrived at a different set of judgments." N. Ins.

Co. of N.Y. v. Point Judith Marina, LLC, 579 F.3d 61, 67 (1st Cir.

2009). Here the judge concluded that the DOC's present treatment

regimen, with the added medication and therapy to cushion the post-

surgery-denial fallout, would not reduce Kosilek's suffering to the

point where she did not have a major medical need. Rather, sex

reassignment surgery was the only adequate treatment for Kosilek's

life-threatening disorder. As detailed above, these findings were

supported by the un-objected to testimony of multiple eminently

qualified doctors, by widely accepted, published standards, and by

the testimony of Kosilek herself. The factfinder found this

evidence convincing; he found the DOC's evidentiary offering less

so. It is not for us to re-weigh the evidence and second-guess

this determination, but that is exactly what the majority does.

What's more, by upholding the adequacy of the DOC's

course of treatment, the majority in essence creates a de facto ban

on sex reassignment surgery for inmates in this circuit. Its

-95-
attempt to repudiate this notion is not compelling. For instance,

the fact that the DOC has "disclaimed any attempt to create a

blanket policy regarding SRS" is a non-starter. The issue is not

whether correctional departments will voluntarily provide the

surgery, it is whether the precedent set by this court today will

preclude inmates from ever being able to mount a successful Eighth

Amendment claim for sex reassignment surgery in the courts.

Equally unconvincing is the majority's assertion that the "unique

circumstances" presented by Kosilek's case will prevent any de

facto ban. The first so-called anomaly cited by the majority --

the divergence of opinion as to Kosilek's need for surgery -- only

resulted from the DOC disregarding the advice of Kosilek's treating

doctors and bringing in a predictable opponent to sex reassignment

surgery. It is no stretch to imagine another department of

corrections stealing a page from this play book, i.e., just bring

in someone akin to Osborne. It is hardly a matchless scenario.

The same goes for Kosilek's criminal history and post-surgical

housing options, which the majority also points to. Rare will be

the prisoner who does not pose some type of security concern, or

harbor some potential for causing climate unrest. So the question

remains, if Kosilek -- who was time and again diagnosed as

suffering from severe gender identity disorder, and who was

uniformly thought by qualified medical professionals to require

-96-
surgery -- is not an appropriate candidate for surgery, what inmate

is?

In sum, the majority's conclusion that the district court

wrongly found that Kosilek satisfied the objective component of the

Eighth Amendment inquiry is, in my opinion, flatly incorrect. I am

no more convinced by the majority's examination of the subjective

component.

IV. Eighth Amendment: Subjective Prong

A satisfied subjective prong means that prison officials

had "a sufficiently culpable state of mind" in that they showed

deliberate indifference to an inmate's health and safety. Farmer,

511 U.S. at 834; Leavitt, 645 F.3d at 497. The officials were both

"aware of facts from which the inference could be drawn that a

substantial risk of serious harm exists" and they drew that

inference. Ruiz-Rosa v. Rullán, 485 F.3d 150, 156 (1st Cir. 2007).

The majority posits that the DOC, because it was faced with

conflicting medical opinions about what treatment was appropriate

for Kosilek, and because it proffered reasonable security concerns,

was not deliberately indifferent to Kosilek's risk for serious

harm. Both theories fail to convince.

-97-
 A. Conflicting Medical Opinions

The idea that incompatible medical opinions serve to

insulate the DOC from a deliberate indifference finding is a

concept not advanced by the DOC, which rests on several faulty

propositions and has very problematic implications.

The majority concedes that the DOC never made this

particular argument, but charitably claims it is not waived because

"[t]he DOC's contention that the district court erred in deeming

SRS medically necessary and in rejecting Dr. Schmidt's approach as

imprudent necessarily entails the DOC's subjective belief that SRS

was unnecessary."31 This is a stretch. An argument advanced on

appeal years after surgery was denied is not the equivalent of the

DOC's subjective belief that sex reassignment surgery was

unnecessary when it denied it. Moreover, the mere existence of

contradictory medical opinions does not necessarily mean that the

DOC did not deny Kosilek surgery for purely pretextual reasons. It

is certainly conceivable that a correctional department could seize

on an opinion from a medical provider, whether or not it found it

compelling, as a means to justify denying treatment.

31
When the DOC first informed the district court that it would not
be providing Kosilek with surgery (back in June 2005 under
Dennehy's watch), Dr. Schmidt had not even evaluated Kosilek let
alone communicated his findings. At the time, the DOC was only
armed with the report of Cynthia Osborne who had not met with
Kosilek but rather had simply peer reviewed the Fenway Report.

-98-
 Another even more serious flaw in the majority's theory

is that it is contradicted by the evidence. Commissioner Dennehy

testified multiple times, and submitted a report to the same

effect, that it was security concerns that motivated her decision

to deny Kosilek surgery. During Dennehy's first round of

testimony, when she was still claiming ignorance about whether

UMass (the DOC's contracted health-services provider) was

recommending surgery, she testified that based "strictly [on]

safety and security concerns" she would still veto the surgery even

if UMass told her that it was medically necessary. Then, once

UMass's position that surgery was medically necessary became

pellucid to Dennehy, she submitted a report to the court indicating

that she was standing firm in her decision to deny surgery based on

"alarming and substantial" safety and security concerns. Her final

time on the stand, Dennehy testified that the only thing, in her

view, preventing surgery for Kosilek was safety and security

concerns; absent such concerns, Dennehy would have no reason to

interfere with any medical order for treatment.

The evidence with regard to Commissioner Clarke's stance

on the issue was similar. In his report to the court, Clarke

disclaimed any ability to render an opinion on the validity of the

medical opinions expressed at trial, and went on to explain his

view that "the safety and security concerns presented by the

prospect of undertaking sex reassignment surgery for Michelle

-99-
Kosilek are insurmountable." Clarke then hammered home his

security concerns on the stand. Therefore, even though there was

contradictory opinions on whether surgery was medically necessary

for Kosilek, both Dennehy's and Clarke's decision to deny the

procedure was, as they put it, based solely on security concerns.

The majority's presumption that the existence of varying

medical opinions should insulate the DOC is not only an

unpreserved, unsupported argument but it has very troubling

implications. It gives correctional departments serious leeway

with the Eighth Amendment. If they do not want to provide a

prisoner with care recommended by one or more than one medical

provider, they need only find a doctor with a differing mind set

(typically not a difficult task). It is no stretch to think that

might be what happened here. The DOC had the treatment

recommendation of Drs. Kaufman and Kapila, a local psychiatrist and

psychologist who had evaluated Kosilek. The doctors themselves

were recommended by the DOC's own medical provider, UMass. Yet the

DOC took the unusual step of having the Fenway doctors'

recommendation peer reviewed by Cynthia Osborne, an out-of-state

social worker with a known opinion about sex reassignment surgery.

It seems highly unlikely that the DOC was simply looking for a more

complete picture of Kosilek's treatment options, and that Osborne's

predictable opposition to Kosilek being provided with surgery was

a non-factor. The DOC knew that Osborne was working with the

-100-
Virginia and Wisconsin departments of corrections to help defend

lawsuits filed by transgender prisoners, and internal DOC meeting

minutes noted that Osborne "may do more objective evaluations" and

was "[m]ore sympathetic to DOC position." Predictably, Osborne was

one-hundred percent sympathetic.

B. Security Concerns

There is no dispute that "security considerations . . .

matter at prisons," leaving "ample room for professional judgment."

Battista v. Clarke, 645 F.3d 449, 453, 454 (1st Cir. 2011). "Any

professional judgment that decides an issue involving conditions of

confinement must embrace security and administration and not merely

medical judgments." Id. at 455 (quoting Cameron v. Tomes, 900 F.2d

14, 20 (1st Cir. 1993) (emphasis in original)). But it is also

true that at some point a defendant forfeits the advantage of

deference, for instance following a "pattern of delays, new

objections substituted for old ones, misinformation and other

negatives."32 Id. The district court determined that the DOC had

done just this, causing undue delay in Kosilek's treatment regimen,

32
The pattern in Battista -- a case in which a transgender inmate
sued the Massachusetts DOC for failing to provide doctor-
recommended hormones -- included an initial failure to take the
inmate's diagnosis and hormone request seriously, the years it took
for a solid security justification to be made, and the DOC's claim
that withholding hormones or placing the inmate in severely
constraining protective custody were the only two options. In
other words, there are some marked similarities between that case
and this one. That is, apart from their outcomes. In Battista,
this court affirmed the district court's deliberate indifference
determination.

-101-
manufacturing security concerns, and orchestrating a half-hearted

security review. The record amply supported these conclusions, yet

the majority too easily discounts them, especially given the

deferential look this issue warrants. See, e.g., Torraco, 923 F.2d

at 234 (explaining that deliberate indifference is usually a jury

question); Monahan, 625 F.3d at 46 (providing that due regard is

given to credibility calls); Fed. Refinance Co., 352 F.3d at 27-28

(noting that a clear error look makes sense when there are

questions of motive and intent).

Of course, it has been many years since medical providers

began considering the propriety of surgery for Kosilek. Back

during Kosilek I, Dr. Forstein recommended that Kosilek be allowed

to consult with a surgeon who specialized in sex reassignment

surgery. Then in 2003, Dr. Seil said Kosilek should be allowed to

meet with a specialist after a year on hormones. But right when

she started as commissioner, Dennehy made a curious move. She

reassessed the care being provided to all inmates suffering from

gender identity disorder, despite the DOC's contract with UMass

placing that medical care squarely in UMass's purview. Then once

the Fenway doctors opined in 2005 that Kosilek should be allowed to

have surgery, the DOC frittered away time claiming not to

understand that UMass recommended surgery for Kosilek. The

majority does not quibble with the court's finding that the DOC

prevaricated in this respect because it "does not undercut the

-102-
consistency with which they identified safety and security

concerns." This misses the point. To establish a subjective

intent, "it is enough for the prisoner to show a wanton disregard

sufficiently evidenced 'by denial, delay, or interference with

prescribed health care.'" Battista, 645 F.3d at 453 (quoting

DesRosiers, 949 F.2d at 19); see also Johnson v. Wright, 412 F.3d

398, 404 (2d Cir. 2005) (A "deliberate indifference claim can lie

where prison officials deliberately ignore the medical

recommendations of a prisoner's treating physicians."). That is

precisely what the district court found happened here, and the

evidentiary support for this determination is in the record.

The same goes for the court's conclusion that the DOC's

security reviews were rushed and results-driven. Dennehy told a

news outlet that the DOC would deny Kosilek's request for surgery

despite only having "generalized discussions" and phone calls with

the relevant players; she had not yet received written reports or

convened a formal security meeting. When the DOC did meet, there

was just a week left before its court-ordered security report was

due -- a report that was then penned predominantly by trial counsel

and reviewed by Dennehy only a day or two before its filing. Once

trial was underway, the hurriedness continued. A mere nine days

before expert disclosures were due, Dennehy contacted the director

of the Federal Bureau of Prisons looking for a security expert.

And the experts the DOC ultimately did present at trial seemed ill

-103-
prepared, failing to take into account important details about

Kosilek's medical and disciplinary history.

For the district court, another reason not to esteem the

DOC's proffered security concerns was the fact that they were

"largely false" and "greatly exaggerated." This finding is not

clearly erroneous. Yet the majority easily dismisses it, in part

by limiting its focus to what it presumably perceives as the DOC's

more valid security concerns -- where to house Kosilek post-

operatively and the deterrence of false suicide threats by inmates.

The majority is conveniently forgetting the throw-it-up-and-see-

what-sticks approach taken by the DOC below. It was this approach,

in part, that led the court to question whether the DOC could be

trusted to give an accurate picture of security concerns consequent

to surgery.

For instance, the DOC repeatedly claimed that

transporting Kosilek to surgery out of state would pose an

insurmountable security risk. It is hardly surprising the district

court thought this was an embellished concern. Kosilek had been

transported to multiple doctor's appointments without issue, and it

is illogical to think Kosilek would attempt to flee en route to the

surgery she has dedicated decades of her life to obtaining. Also

eminently unlikely is that during the transport home from highly

invasive surgery, a sixty-five-year-old, recovering Kosilek would

be able to escape the grasp of DOC personnel. Even Clarke thought

-104-
it near certain that Kosilek could safely be transported to and

from surgery.

With regard to housing Kosilek in a female prison, the

DOC painted Kosilek as a highly-polarizing escape risk who could

not possibly safely reside in MCI-Framingham's general population.

It pointed to the comparatively weaker perimeter of MCI-Framingham,

alleging that Kosilek's superior male strength and life sentence

made her a flight risk. One can easily see why the district court

was not buying this. Kosilek was advanced in age, physically

slight, had taken female hormones for years, and had an excellent

disciplinary record. And MCI-Framingham successfully housed

approximately forty life offenders. The court also had reason to

be skeptical of the DOC's adamant contention that Kosilek would

cause inmate climate issues at MCI-Framingham due to the fact that

she murdered her wife. Undoubtedly inmates find other inmates

offensive for a plethora of reasons, such as, race, religion, gang

affiliation, sexual orientation, or the crime committed. Prisons

deal with these situations on a routine basis and the evidence

established that MCI-Framingham had procedures in place to do just

that.

The DOC even admitted at oral argument that had a post-

operative, transgender person out in free society committed murder,

the DOC would have to figure out where to house that person. The

DOC; however, did not think this a particularly important point,

-105-
protesting that Kosilek presents unique concerns that separate her

from this hypothetical inmate. I am unmoved. The fact that

Kosilek's crime was one of violence against a woman could equally

apply to another potential inmate. And the fact that Kosilek

gained notoriety by litigating against the DOC all these years --

in other words, successfully pursuing her constitutional right to

adequate medical care -- hardly seems a compelling consideration.

For the district court, also blunting the DOC's fervent

cries of overwhelming security concerns were the alternatives to

placing Kosilek in the general population of a Massachusetts

prison. There was the option of transferring Kosilek to an out-of-

state prison (though this scenario appears to have been left

largely unexplored by the DOC). In fact, the evidence established

that Clarke's former employer, the Washington Department of

Corrections, housed a post-operative female transgender inmate,

also serving a life sentence for murdering a female relative,

without security or climate issues. The inmate's housing was so

unremarkable that Clarke was not even aware of it during his tenure

in Washington. Further, there was evidence that Kosilek's safety

could be ensured by placing her in a segregated housing unit.

The DOC's past conduct was also relevant to the district

court's credibility assessment. In connection with Kosilek I, then

Commissioner Michael Maloney hammered the serious security concerns

surrounding Kosilek remaining at MCI-Norfolk while receiving

-106-
hormones, theorizing that an inmate living as a female (with female

attributes) among sex offenders would create a risk of violence.

However, once the DOC actually stopped to conduct a security

review, it determined there were no current security concerns with

Kosilek being provided estrogen therapy. Indeed no security issues

ever arose. Kosilek has been safely housed at MCI-Norfolk for many

years presenting herself as female. The DOC's reversal on this

issue calls into question its stance before this court about the

non-feasibility of housing a post-surgical Kosilek at MCI-Norfolk.

The DOC also expressed concern that providing Kosilek

with surgery would encourage inmates to utilize suicide threats to

receive a desired benefit, and the majority deems this concern

reasonable. I am not convinced, and neither was the district

court. Not only is there absolutely no evidence that Kosilek is

trying to game the system, but the DOC, as it emphasized at trial,

employs mental health professionals and has policies in place to

deal with suicidality. Presumably, these tools can be used by the

DOC to assess whether an inmate's particular suicide threat is

manufactured or real, and it can be dealt with accordingly. That

the DOC does not want to be inundated with a hypothetical influx of

false suicide threats hardly seems a valid reason to deny a

prisoner care deemed medically necessary.

For the district court, the public and political

disapproval of Kosilek's surgical pursuit was another factor. It

-107-
did not believe Dennehy's and Clarke's claims that the avoidance of

controversy played no role in the DOC's decision to deny surgery.

The majority concedes that it must give deference to the court's

finding that Dennehy's motivations were colored by public pressure

and so, instead, the majority hypes up the role of Commissioner

Clarke by characterizing him as the ultimate decision maker. I see

a few flaws with the majority's reasoning.

For one, the majority says the district court improperly

imputed Dennehy's motivations to Clarke, thus ignoring the

injunctive-relief requirement that it take into account the DOC's

then present-day stance.33 See Farmer, 511 U.S. at 845 (quoting

Helling, 509 U.S. at 36) (The court considers deliberate

indifference "'in light of the prison authorities' current

attitudes and conduct,' . . . their attitudes and conduct at the

time suit is brought and persisting thereafter."). The majority

has it wrong. The court took testimony from Clarke, reviewed his

written report, and spoke extensively in its decision about why it

was not convinced that Clarke denied Kosilek surgery based on

33
While a defendant's attitudes and conduct at the time a decision
is rendered are relevant, what motivates the DOC today is not.
This fact may be less than clear given the majority's reference to
the DOC's present stance ("proof that the DOC remains motivated by
pretextual or improper concerns") and the fact that Dennehy is now
seven years removed from the decision-making process. To be clear,
we are reviewing the district court's decision that the DOC,
through Dennehy and Clarke, denied Kosilek surgery based on
pretextual reasons. Indeed it would be an amazing feat of
prescience for the district court to anticipate what the DOC's
viewpoint would be two years after penning its decision.

-108-
legitimate penological concerns. Of note, it was Kosilek who

sought to have Clarke inform the court of his position, and the

DOC, which stipulated at trial that Dennehy was the operative

decision maker, actually objected to Clarke even testifying as he

was simply "maintain[ing] the position set forth by the DOC through

former Commissioner Dennehy."

Furthermore, though the majority defers to the court's

take on Dennehy, it refuses to do so for Clarke, claiming that

"Clarke was never found by the court to be non-credible." This is

not entirely accurate. Clearly the import of the court's

conclusion that Clarke's articulated security concerns were either

false or exaggerated as a pretext to deny surgery means that the

court did not think Clarke a completely credible witness. See

Kosilek, 889 F. Supp. 2d at 241 ("[T]he purported security

considerations that Dennehy and Clarke claim motivated their

decisions to deny Kosilek sex reassignment surgery are largely

false and any possible genuine concerns have been greatly

exaggerated to provide a pretext for denying the prescribed

treatment.") In fact, the court specifically found certain claims

made by Clarke not to be credible. See id. at 244 (finding that

"neither Dennehy nor Clarke has provided a credible explanation for

their purported belief that if Kosilek's genitalia are altered the

risk to him and others at MCI Norfolk will be materially magnified"

and "[t]he claims of Dennehy and Clarke that they have denied sex

-109-
reassignment surgery for Kosilek in part because MCI Framingham is

not sufficiently secure to prevent an escape by Kosilek, who has

never attempted to flee, are not credible.") Therefore, as it did

with Dennehy, the majority should be giving due regard to the

court's conclusion that Clarke was not believable.

The majority also misses the mark with its contention

that the "only evidence" tending to show Clarke may have considered

public and political criticism were the two letters from the

unhappy Massachusetts legislators. This is not the whole picture.

In addition to the letters, what convinced the court that Clarke

was improperly motivated was his advancing inflated security

concerns following a hasty review, suggesting that he did not

operate with an open mind. Having already detailed the evidence

supporting the court's distrust of the DOC's proffered security

concerns, I will not rehash.

As for the thoroughness of Clarke's review, the court

criticized Clarke for not consulting with Luis Spencer, who was

Superintendent of MCI-Norfolk at the time, and for not reviewing

the DOC's security-expert trial testimony, prior to deciding

whether to deny surgery. The DOC counters that Clarke, pursuant to

the court's order, was not required to do either of those things.

It is both conceivable that Clarke's review was too cursory, or

that he felt constrained by the court's order, though the fact that

Clarke did not know significant details such as Kosilek's age and

-110-
excellent disciplinary record favors the former possibility.

Either way, both views are permissible, which means that the

district court's choice of one of them cannot be clearly erroneous.

See Monahan, 625 F.3d at 46. Nor is it appropriate for us to

second-guess the court's tenable perception of Clarke's

motivations, as deference extends to "inferences drawn from the

underlying facts, and if the trial court's reading of the record

[with respect to an actor's motivation] is plausible, appellate

review is at an end." Janeiro, 457 F.3d at 138-39 (internal

quotation marks omitted) (alteration in original).

Ultimately, there was adequate evidentiary support for

the court's determination that the DOC was deliberately

indifferent. The court did not obviously err in concluding the DOC

delayed implementing medical treatment recommended by its own

providers, sought out a more favorable medical opinion, engaged in

a hasty, result-driven security review, offered a host of poorly

thought out security concerns, and then denied surgery based not on

any legitimate penological concerns but on a fear of controversy.

Whether I, or any of my colleagues, would have drawn these same

conclusions had we been presiding over this trial is irrelevant;

our review is circumscribed. It is enough that the district court

had a reasonable basis for its judgment. The district court's

determination that the Eighth Amendment's subjective prong was

satisfied should stand.

-111-
 V. Conclusion

I am confident that I would not need to pen this dissent,

over twenty years after Kosilek's quest for constitutionally

adequate medical care began, were she not seeking a treatment that

many see as strange or immoral. Prejudice and fear of the

unfamiliar have undoubtedly played a role in this matter's

protraction. Whether today's decision brings this case to a close,

I cannot say. But I am confident that this decision will not stand

the test of time, ultimately being shelved with the likes of Plessy

v. Ferguson, 163 U.S. 537 (1896), deeming constitutional state laws

requiring racial segregation, and Korematsu v. United States, 323

U.S. 214 (1944), finding constitutional the internment of Japanese-

Americans in camps during World War II. I only hope that day is

not far in the future, for the precedent the majority creates is

damaging. It paves the way for unprincipled grants of en banc

relief, decimates the deference paid to a trial judge following a

bench trial, aggrieves an already marginalized community, and

enables correctional systems to further postpone their adjustment

to the crumbling gender binary.

I respectfully dissent.

-112-
 KAYATTA, Circuit Judge, dissenting. Reading the

majority's lengthy and oft-revisited discussion of the applicable

standard of review, one would think that this case posed issues of

law or application of law to fact. This is plainly not so.

There is not a comma, much less a word, of the applicable

law that the district court did not expressly and correctly explain

and apply. All the parties and all the judges in this case,

including the trial judge, agree on the controlling principles of

law, long ago established by the Supreme Court. See Estelle v.

Gamble, 429 U.S. 97, 104 (1976); Farmer v. Brennan, 511 U.S. 825,

837 (1994); see also United States v. DeCologero, 821 F.2d 39, 43

(1st Cir. 1987). Under that law, a prison must supply medical care

to its prisoners "at a level reasonably commensurate with modern

medical science and of a quality acceptable within prudent

professional standards." DeCologero, 821 F.2d at 43. The failure

to provide such care, moreover, does not constitute an Eighth

Amendment violation unless it rises to the level of "deliberate

indifference" to a "serious medical need." "Deliberate

indifference" means that the prison official "knows of and

disregards an excessive risk to inmate health or safety." Farmer,

511 U.S. at 837. A "serious medical need" is defined as, among

other things, "one that has been diagnosed by a physician as

mandating treatment." Gaudreault v. Municipality of Salem, Mass.,

923 F.2d 203, 208 (1st Cir. 1990).

-113-
 Our decision in this case therefore necessarily turns on

the facts themselves. And we begin our review knowing that Kosilek

does indeed have a serious medical need, and the prison's own

doctors, as well as the specialists retained by those doctors,

informed DOC that treatment of Kosilek's medical condition in

accordance with prudent professional standards requires sex

reassignment surgery (SRS).34 That leaves only two factual

questions: (1) Are the DOC's doctors correct that SRS is the only

treatment for Kosilek's condition that is commensurate with modern

medical science as practiced by prudent professionals;35 and, if so,

(2) Did prison officials nevertheless deny that treatment not

because they disbelieved their own doctors, and not because of

prison security considerations, but rather simply because they

feared public ridicule. If the answer to each of these two

questions is "yes," Kosilek should win. Otherwise, she loses.

Were I the trial judge charged with answering these two

factual questions based solely on the written record, I would

likely find against Kosilek on the first question for the reasons

stated by the trial court's appointed independent expert, Dr.

Levine. In a nutshell, Dr. Levine, who participated in drafting

34
None of these witnesses face challenge on the grounds that their
opinions are outside the bounds of accepted science. See Daubert
v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589-90 (1993).
35
In the majority's words, "whether SRS is a medically necessary
component of Kosilek's care, such that any course of treatment not
including surgery is constitutionally inadequate." Slip Op. at 46.

-114-
the Standards of Care, provided carefully nuanced and persuasive

testimony that medical science has not reached a wide,

scientifically driven consensus mandating SRS as the only

acceptable treatment for an incarcerated individual with gender

dysphoria. But I am not the trial judge in this case. Nor are my

colleagues. And that is the rub.

The experienced jurist who was the trial judge in this

case, and who actually sat and listened to the live testimony,

found as a matter of fact that:

(1) Commensurate with modern medical science, no prudent

professional would recommend any treatment for Kosilek other than

SRS; and

(2) Prison officials nevertheless denied the treatment

not because they rejected the accuracy of the medical advice

tendered by their own doctors, and not because of security issues,

but rather because they feared public ridicule. Their reasons for

denying the necessary treatment were thus in bad faith.

The majority never explains why these two findings are

not pure findings of fact, and are not therefore subject solely to

review for clear error. Nor can it. After all, we are talking

about, first, what the medical--not legal--standard of care is for

a particular affliction, and second, whether Dennehy and Clarke

were truthful in describing their security objections, such as

their claim that they feared that Kosilek, after trying to get this

-115-
surgery for twenty years, would escape on the way to the operating

room or, fresh from the surgeon's knife, overpower her guards and

run away. Let me be plain on this point: Until today, there was

absolutely no precedent (and the majority cites none) for reviewing

such quintessentially factual findings under anything other than

the clear error test.

As Judge Thompson carefully explains, there is a

considerable amount of evidence that directly supports the trial

court's findings on these two points, depending on which witnesses

one believes. I write separately only to stress that even if one

agrees with the majority that the district court got the fact-

finding wrong, we should defer unless the result is clearly

erroneous. Of course, deferring to the trial judge's fact-finding

happens to produce a result in this case that some of us find

surprising, and much of the public likely finds shocking.

Scientific knowledge advances quickly and without regard to settled

norms and arrangements. It sometimes draws in its wake a reluctant

community, unnerved by notions that challenge our views of who we

are and how we fit in the universe. The notion that hard-wired

aspects of gender may not unerringly and inexorably correspond to

physical anatomy is especially unnerving for many.

The solution, I think, is to trust our trial judges to

resolve these factual issues when the evidence supports a finding

either way. Some will get it wrong; most will get it right. The

-116-
arc of decision-making, over time, will bend towards the latter.

For each instance of error in fact-finding, such as possibly this

case itself, $25,000 or so may be lost. But doctors and lawyers

will refine their presentations and other trial judges will make

their own findings, not bound in any way by the fact-finding in

this case.

Instead, by deciding the facts in this case as an

appellate court essentially finding law, the majority ends any

search for the truth through continued examination of the medical

evidence by the trial courts. It locks in an answer that binds all

trial courts in the circuit: no prison may be required to provide

SRS to a prisoner who suffers from gender dysphoria as long as a

prison official calls up Ms. Osborne or Dr. Schmidt.36

Acknowledging that the majority may well be correct on the facts,

I nevertheless decline the invitation to join the majority in

embracing the authority to decide the facts. I suspect that our

court will devote some effort in the coming years to distinguishing

this case, and eventually reducing it to a one-off reserved only

for transgender prisoners.

36
No prisoner is likely to have a more favorable record than
Kosilek.

-117-