William Dean v. Wexford Health Sources, Inc.

• Court: Court of Appeals for the Seventh Circuit
• Decided: November 10, 2021
• Docket: 20-3139
• Status: Published

Opinion

In the

United States Court of Appeals For the Seventh Circuit ____________________ Nos. 20-3058 & 20-3139 WILLIAM DEAN, Plaintiff-Appellee, Cross-Appellant, v.

WEXFORD HEALTH SOURCES, INC., et al., Defendants-Appellants, Cross-Appellees. ____________________

Appeals from the United States District Court for the Central District of Illinois. No. 17-CV-3112 — Sue E. Myerscough, Judge. ____________________

ARGUED MAY 19, 2021 — DECIDED NOVEMBER 10, 2021 ____________________

Before WOOD, ST. EVE, and KIRSCH, Circuit Judges. ST. EVE, Circuit Judge. William Dean developed kidney cancer while incarcerated at Taylorville Correctional Center in central Illinois. Seven months after he first presented symp- toms, Dean had kidney-removal surgery. Unfortunately, the cancer had already spread to his liver, so Dean remains termi- nally ill. In this lawsuit Dean sues two of the doctors involved in his care: Dr. Abdur Nawoor and Dr. Rebecca Einwohner. He also sues their employer—Wexford Health Sources, Inc.— 2 Nos. 20-3058 & 20-3139

a private corporation that contracts with Illinois to provide healthcare to Illinois inmates. Dean’s lawsuit focuses on delays in the diagnosis and treatment of his kidney cancer. He blames the delays on his doctors’ failure to arrange timely offsite care and on Wex- ford’s “collegial review” policy, which requires Wexford’s corporate office to preapprove offsite care. Dean submits that the defendants’ actions were not merely negligent but delib- erately indifferent to his serious medical needs in violation of the Eighth Amendment. The case went to trial, and the jury sided with Dean, awarding him $1 million in compensatory damages and $10 million in punitive damages against Wex- ford. After trial, the district court reduced the punitive dam- ages award to $7 million. The defendants now appeal, chal- lenging the jury’s verdicts on the Eighth Amendment claims. We reverse. Dean has endured great suffering, but he did not produce enough evidence at trial to hold any of the de- fendants liable for violating his Eighth Amendment rights. Dean’s claim against Wexford hinged on the Lippert reports— two expert reports from another case that critique the medical care, and process for medical care, that Illinois provides, through Wexford, to its prisoners. The Lippert reports are hearsay, but the district court allowed Dean to use them for a non-hearsay purpose: to prove that Wexford had prior notice of the experts’ negative assessments of collegial review. The problem with the district court’s ruling is that the second Lip- pert report postdated all events relevant to this case and thus could not have given Wexford prior notice of anything. And even if the court did not abuse its discretion in admitting the first report—an issue we need not resolve—the first report alone was insufficient to hold Wexford liable under the Nos. 20-3058 & 20-3139 3

exacting requirements of Monell v. Department of Social Ser- vices, 436 U.S. 658 (1978), in this single-incident case. Dean fares no better at proving that the doctor-defendants were de- liberately indifferent, so we reverse and direct judgment as a matter of law across the board on the Eighth Amendment claims. We do not upset the jury’s findings that the defendants were negligent, but a new jury must reassess the issue of dam- ages. I. Background Dean has been incarcerated at the Taylorville Correctional Center in Taylorville, Illinois since 2012. Dean’s lawsuit cen- ters on the timing of the offsite care that he received at Taylor- ville in connection with the diagnosis and treatment of his kidney cancer. We describe the timeline in some detail before unpacking the evidence presented at trial. We recount all facts and evidence in the light most favorable to the jury’s verdict for Dean. See J.K.J. v. Polk Cnty., 960 F.3d 367, 371 (7th Cir. 2020) (en banc). A. Factual Background Even before developing kidney cancer, Dean had serious health issues. Among other ailments, Dean had heart disease, diabetes, and a history of kidney stones. Dean also had passed blood in his urine, though it was usually invisible. In August 2014 and July 2015, Dean had CT scans. A CT scan (short for “computerized tomography scan”) uses X-ray imaging and computer technology to create detailed cross-sectional images of internal parts of the body. A CT scan can be an invasive procedure, and it carries risks for patients with severely com- promised kidney function. The CT scans performed on Dean in 2014 and 2015 did not detect cancer. 4 Nos. 20-3058 & 20-3139

On December 19, 2015, Dean noticed visible blood in his urine. The medical term for blood in the urine is “hematuria.” Visible blood in the urine is “gross hematuria.” Dean’s gross hematuria was initially painless, but he later experienced sig- nificant pain from passing blood clots in his urine. On December 23, 2015, Dean went to the Taylorville med- ical clinic seeking help for his gross hematuria. Dr. Nawoor, Wexford’s medical director at Taylorville, physically exam- ined Dean and ordered testing of his blood and urine. The testing confirmed that there was blood in Dean’s urine. Dr. Nawoor told Dean to stay hydrated. At this early stage, Dr. Nawoor believed that either a recurrence of kidney stones or cancer was the cause of the gross hematuria. On January 7, 2016, Dean had a telemedicine visit with Dr. Einwohner. Dr. Einwohner is a nephrologist (kidney doctor) based in Wexford’s Pittsburgh corporate office. She provides telemedicine support for Wexford’s primary care doctors on kidney issues. After her visit with Dean, Dr. Einwohner emailed Dr. Stephen Ritz suggesting a collegial review for Dean. Dr. Ritz is the corporate medical director for utilization management at Wexford. Collegial review is Wexford’s procedure for discussing and approving offsite care for inmates. When a Wexford doc- tor decides that an inmate needs offsite care, the doctor sub- mits a referral request form to the medical director of the in- mate’s facility. If the medical director agrees with the request, the medical director presents the request at a collegial review. The medical director, too, can submit referral request forms. The collegial review itself is a call between the medical direc- tor and a Wexford doctor in Pittsburgh who either approves the offsite service or suggests an alternate plan of care. If Nos. 20-3058 & 20-3139 5

collegial review approves the offsite care, Wexford sends the prison an authorization number, at which point the prison calls the offsite specialist to schedule an appointment. Colle- gial reviews generally occur on a weekly basis. If collegial re- view approves offsite care, it usually takes an additional 24 to 48 hours for Wexford to send the prison an authorization number. Dr. Ritz is the Pittsburgh doctor who participates in collegial reviews for patients at Taylorville. In her email to Dr. Ritz, Dr. Einwohner discussed Dean’s present condition, history of kidney stones, and past CT scans. She concluded by suggesting a collegial review “with consideration of re-imaging and urology eval.” Dr. Einwoh- ner also followed up with a nurse at Taylorville to ensure that a collegial review would take place. Six days later, on January 13, 2016, Drs. Nawoor and Ritz had a collegial review to discuss the possibility of offsite care for Dean. This collegial review was apparently done at Dr. Nawoor’s request. At the collegial review, Drs. Nawoor and Ritz decided that Dean should undergo a kidney ultrasound. An ultrasound uses high-frequency sound waves to produce images of internal parts of the body. Ultrasounds generally are less invasive and provide less detailed information than CT scans. They are also cheaper than CT scans for Wexford because, unlike CT scans, they can be done onsite. Wexford gets reimbursed for onsite care whereas it pays out of pocket for offsite care. Dean had the ultrasound on February 2, 2016. A third- party imaging company came to Taylorville to perform it. A third-party radiologist reviewed the ultrasound and reported that the results were normal, with no evidence of masses on the kidneys. This reading by the third-party doctor was 6 Nos. 20-3058 & 20-3139

wrong; the ultrasound showed that Dean’s right kidney was abnormally large, potentially indicating a mass on the right kidney. 1 Dr. Einwohner saw Dean again on February 8, 2016. At this visit, Dr. Einwohner learned of the (incorrect) ultra- sound reading. After the visit, Dr. Einwohner followed up with a Taylorville nurse to ensure that Dean would receive another collegial review. Relying on the incorrect ultrasound reading, Drs. Nawoor and Ritz determined in a collegial review on February 10, 2016, that Dean might have bladder issues. As such, they de- cided to send him to an offsite urologist for a cystoscopy (a procedure that scopes the bladder). On March 10, 2016, Dean saw a urologist, Dr. William Severino. Dr. Severino recom- mended a CT scan and cystoscopy. Drs. Nawoor and Ritz con- sidered Dr. Severino’s recommendations during a March 22, 2016 collegial review. They approved the cystoscopy that day and approved the CT scan eight days later, on March 30, 2016, at a separate collegial review. Dean had the CT scan on April 12, 2016. The CT scan re- vealed that Dean had renal cell carcinoma (kidney cancer) on his right kidney. The CT scan also revealed that the cancer had potentially extended to Dean’s vena cava (a large vein that carries blood from the lower body to the heart). It did not show that the cancer had metastasized to any other organs. Dr. Severino told the prison that Dean would need kidney- removal surgery. Drs. Nawoor and Ritz approved the surgery at a collegial review nine days later, on April 21, 2016. The

1As best we can tell, no one noticed that the ultrasound reading was wrong until Dean filed this lawsuit. Nos. 20-3058 & 20-3139 7

next day, Dean had chest X-rays, which ruled out lung cancer. On May 6, 2016, Dean had a cardiology evaluation. Dr. Severino’s office originally scheduled the surgery for May 11, 2016, but Dr. Severino decided to push it back be- cause he needed more time to consult and recruit other doc- tors regarding the surgery—which would be very complex, given that the cancer appeared to extend to Dean’s vena cava. Dr. Severino also wanted Dean to have additional imaging and evaluations before going into surgery. Dr. Severino’s of- fice ultimately rescheduled the surgery for July 19, 2016. On June 1, 2016, an offsite radiologist (enlisted by Dr. Sev- erino) requested another CT scan. A collegial review ap- proved the CT scan the next day. The CT scan happened on June 8, 2016. This CT scan showed that the cancer had ex- tended up Dean’s vena cava past his diaphragm. In consulta- tion with a vascular surgeon, Dr. Severino decided to bring in a cardiothoracic surgeon. On June 13, 2016, the cardiothoracic surgeon asked to see Dean. A collegial review approved the visit the next day. Dean saw the cardiothoracic surgeon on June 20, 2016. On June 28, 2016, Dean had a second cardiology evaluation. In advising Dean about the upcoming surgery, Dr. Sev- erino told Dean that he would not leave prison alive without the surgery. (At the time of trial, Dean was slated to be re- leased in September 2020.) Dean decided to proceed with the surgery, despite the serious risks that it entailed. Dr. Severino told Dean that he had only a 50 percent chance of surviving the surgery. His chances of survival had decreased signifi- cantly since April, when Dr. Severino told him there was a 15 to 25 percent chance that he would “die on the table.” Even if 8 Nos. 20-3058 & 20-3139

Dean survived the surgery, Dr. Severino told him he had only a 35 percent chance of living five more years. Dean’s surgery took place as planned on July 19, 2016. All told, it lasted nine hours and required a ten-person surgical team with three surgeons. Dr. Severino described it as one of the most complex surgeries of his career. The surgery re- quired opening Dean’s chest, stopping his heart for half an hour, and bleeding him out. By all accounts, the surgery went as well as possible under the circumstances. During the sur- gery, however, Dr. Severino noticed spots on Dean’s liver that he suspected were metastatic cancer. The hospital discharged Dean on July 28, 2016. On August 3, 2016, a collegial review approved follow-up visits with Dr. Severino and the cardiothoracic surgeon. Dr. Severino evalu- ated Dean on August 11, 2016. He thought Dean was doing “unbelievably well.” He recommended that Dean see an on- cologist for cancer treatment moving forward. A collegial re- view approved the oncologist visit a week later. Dean saw the oncologist, Dr. Perry Guaglianone, on August 26, 2016. Dr. Guaglianone requested imaging and a follow-up visit. A col- legial review approved these measures five days later. On September 9, 2016, Dean had a CT scan of his abdomen and liver. On September 22, 2016, Dr. Guaglianone requested a liver biopsy. A collegial review approved the biopsy five days later. The biopsy occurred on October 3, 2016. It revealed that Dean had metastatic cancer on his liver. On October 19, 2016, Dr. Guaglianone prescribed a chem- otherapy drug called Votrient to treat Dean’s metastatic can- cer. Votrient was an expensive, “non-formulary” drug that re- quired a separate review process at Wexford. If approved, it would cost Wexford $15,000 per month. On October 24, 2016, Nos. 20-3058 & 20-3139 9

a Wexford pharmacist wrote to others at Wexford that Dean would “benefit more from palliative care” than from “aggres- sive treatment,” given his “background of multiple organ me- tastases and other chronic co-morbidities.” On November 8, 2016, Dr. Nawoor presented Dean with forms that would al- low Dean to refuse further treatments and opt instead for end- of-life care. Dean indicated, however, that he wanted all measures taken to prolong his life. Wexford ultimately ap- proved the Votrient prescription on November 14, 2016. Dean received his first dose on November 18, 2016. On March 2, 2017, Dr. Guaglianone determined that the Votrient was not controlling the growth of Dean’s metastatic cancer. He recommended a different chemotherapy drug, called Opdivo. On March 13, 2017, a corporate medical direc- tor at Wexford wrote in an internal email that there should be a collegial review to discuss Dean’s continued treatment plan. He added that “hospice care may be possibly more appropri- ate but we should be discussing the options and the status of the Patient.” Wexford ultimately approved the Opdivo pre- scription, and Dean began taking Opdivo on March 22, 2017. The Opdivo did not work either, so Dean began taking a third drug—Torisel—in May 2017. A few months later, Dr. Guagli- anone determined that Torisel was working to control (but not cure) Dean’s metastatic cancer. B. Procedural Background Dean filed this lawsuit in April 2017. He sued Wexford, Dr. Nawoor, and Dr. Einwohner under 42 U.S.C. § 1983 for delib- erate indifference in violation of the Eighth Amendment. He also sued Wexford for institutional negligence and the doctor- defendants for medical malpractice. 10 Nos. 20-3058 & 20-3139

1. Admission of Lippert Reports Before trial, the defendants filed a motion in limine to ex- clude the Lippert reports. Lippert v. Ghosh, 10-cv-4603 (N.D. Ill., filed July 23, 2010), was a class-action lawsuit against the Illi- nois Department of Corrections (IDOC) and Wexford alleging systemic deficiencies in medical care at IDOC facilities. The parties in Lippert stipulated to Wexford’s dismissal from that lawsuit on December 19, 2013. The district judge presiding in Lippert appointed two ex- pert witnesses agreed to by the parties—Dr. Ronald Shansky and Dr. Mike Puisis—to prepare reports analyzing the consti- tutional adequacy of the IDOC’s medical care for inmates. The experts reviewed eight IDOC facilities. Taylorville was not one of them. Both experts’ reports were lengthy and compre- hensively reviewed the IDOC’s healthcare system. We discuss only limited excerpts of the reports, addressing collegial re- view, that the district court admitted at trial. Dr. Shansky submitted his report in the Lippert case in De- cember 2014. Relevant here, Dr. Shansky found “breakdowns in almost every area” of collegial review, “starting with de- lays in identification of the need for the offsite services, delays in obtaining an authorization number, delays in being able to schedule an appointment timely, delays in obtaining offsite paperwork and delays or the absence of any follow-up visit with the patient.” He also found that, “although some of the facilities were tracking these steps fairly conscientiously, oth- ers were not, creating much less dependable outcomes.” Dr. Shansky offered a list of recommendations to improve the process. Among other things, he recommended tracking the entire collegial review process in a logbook and ensuring both that collegial reviews happen within one week and that Nos. 20-3058 & 20-3139 11

authorization numbers issue within one business day. He ad- vised, too, that “[s]cheduling should be based on urgency. Ur- gent appointments must be achieved within 10 days; if emer- gent, there should be no collegial review and there should be immediate send out. Routine appointments should occur within 30 days.” Dr. Shansky did not review the collegial re- view procedures at Taylorville as part of his report. Dr. Puisis submitted his report four years later, in October 2018. His charge was to determine whether any of the “sys- temic deficiencies” that Dr. Shansky had identified still ex- isted and to make additional findings and recommendations, as necessary. With respect to collegial review, Dr. Puisis con- cluded that “[t]here was no improvement since the First Court Expert’s Report.” He described collegial review as “a patient safety hazard [that] should be abandoned until such time that patient safety is ensured.” He outlined problems with record- keeping, delays, physician qualifications, and the substance of the collegial review calls. He described collegial review as “ineffective” and “a barrier to timely care” for many patients. The calls themselves involved “no clinical collegial discus- sion” and were “more of an approval process.” The defendants argued that the Lippert reports were inad- missible hearsay, irrelevant, and unfairly prejudicial. Dean countered that the reports were admissible for the non-hear- say purpose of showing that Wexford had notice of the “un- constitutional customs” that caused his injuries. The defend- ants replied that the 2018 report could not have given Wex- ford notice because it did not exist at the times relevant to this case (2015–17). The district court sided with Dean, reasoning that the reports were admissible to prove that Wexford was on notice of their contents. Thus, the court admitted the 12 Nos. 20-3058 & 20-3139

excerpts discussed above, while issuing a limiting instruction which we discuss below. 2. Trial The case proceeded to a seven-day jury trial in December 2019. The evidence at trial centered on Dean’s condition, col- legial review, and the doctor-defendants’ actions. We summa- rize the evidence relevant to this appeal. Dean testified at length about his physical suffering and emotional distress while waiting for treatment after his initial presentation of symptoms. He described urinating blood in- termittently and passing painful blood clots the size of “gummy worms.” Sometimes the blood clots were so large that they made the bathroom look like “a murder scene.” He testified to feeling fear, frustration, and hopelessness as he waited for treatment without knowing if or when it would come. Dean called the doctor-defendants as adverse witnesses. Dr. Nawoor testified that he and Dr. Ritz decided in their Jan- uary 13, 2016 collegial review that Dean should undergo an ultrasound, rather than a more invasive CT scan, because Dean had a history of kidney stones, decreased kidney func- tion, and he had already undergone two lithotripsies (proce- dures to destroy kidney stones using sound waves). He acknowledged that the standard of care for treating a patient with painless gross hematuria is to refer the patient to a urol- ogist for a CT scan and cystoscopy. Even so, he maintained that the standard of care in a particular case depends on the unique characteristics of the patient. And here, an ultrasound was an appropriate first step for Dean, given his medical his- tory. On this point, Dr. Nawoor stressed that Dean’s Nos. 20-3058 & 20-3139 13

ultrasound did in fact show an abnormality on his right kid- ney. Dr. Nawoor testified that he would have suggested a CT scan “right away” if the third-party radiologist had correctly read the ultrasound. At the same time, Dr. Nawoor admitted that an ultrasound cannot rule out cancer. And he knew even before ordering an ultrasound that Dean would need to see a urologist. Dr. Nawoor did not deny that there were problems in Dean’s care, but he generally blamed them on collegial re- view. He testified that Wexford “failed” by not including Dr. Einwohner in collegial review. He added that he could not make any diagnostic decisions (e.g., ultrasound, urology re- ferral) without first “calling Pittsburgh” for approval. He tes- tified that it took months rather than weeks to diagnose Dean’s kidney cancer because of Wexford’s practices: Q. So from the time of the approval of the CT scan until a clean, clear diagnosis of kidney cancer is two weeks? A. Mm-hmm. Q. And the reason it’s three months later instead of two weeks is because of delays caused by Wexford practices; right? A. Well, that’s what they do. I mean I have to com- ply with this. Dr. Einwohner acknowledged that she knew as early as January 2016 that Dean might have cancer. She also acknowl- edged that she thought Dean should see a urologist within a couple weeks. Yet, apart from her email suggesting reimaging and a urology visit, she did not reach out to Dr. Ritz or Dr. Nawoor to ensure that these steps were taken. In general, Dr. 14 Nos. 20-3058 & 20-3139

Einwohner portrayed her role in Dean’s care as limited. She provided support for Wexford’s primary care doctors, but she lacked the power to order diagnostic tests or make referrals on her own. The best she could do was suggest a collegial re- view, although she could not participate in such review. Dean also presented the testimony of other physicians in- volved in his care, including Dr. Ritz and Dr. Severino. Dr. Ritz testified about collegial review and the Lippert reports. He testified that collegial review is Wexford’s way of implement- ing its contractual requirement to “[a]ggressively manage all offsite services for appropriate utilization and cost-effective- ness.” He testified that “emergency” matters need not go through collegial review, and for “urgent” matters the medi- cal director can call Pittsburgh directly without waiting for the usual process to play out. Whether a situation qualifies as emergent or urgent is up to the medical director. Importantly, Dr. Ritz testified on direct examination that the exception for urgent or emergent cases did not exist until January 14, 2016, when an updated version of the collegial review policy went into effect. As Dr. Ritz explained it, the dual purposes of collegial re- view are to consider requests for offsite care and to serve as a forum for doctors to share perspectives and discuss cases. He denied that collegial review is designed to delay care or cut costs. He agreed, however, that 85 to 90 percent of collegial reviews approve the requested offsite care “at face value.” Dr. Ritz did not have any independent memory of his involve- ment in Dean’s care. But he testified that an ultrasound is, in his experience, generally the first step in treating a patient with a history of kidney stones. He added that Wexford’s Nos. 20-3058 & 20-3139 15

expectation is that “we provide all services on site that can be reasonably provided onsite.” To counter Dr. Ritz’s testimony about the purpose of col- legial review, Dean introduced an exhibit from Wexford’s website describing collegial review as “a proactive, physician- led process designed to reduce offsite care costs.” The excerpt explained that “collegial review results in fewer requested offsite referrals, a reduction in the percentage of referrals de- nied as inappropriate, and an overall decrease in specialty consults.” Dean’s counsel questioned Dr. Ritz at length about the Lip- pert reports. Dr. Ritz admitted that Wexford was aware of the reports and that they raised “serious” concerns that Wexford took seriously. He added, however, that Wexford disagreed with many of their findings. He conceded that the reports were prepared by “independent” court-appointed experts ra- ther than experts hired by the parties in Lippert. Dr. Severino testified that Dean’s surgery was “as complex as it gets,” due to the cancer invading the vena cava. When asked about the timing of the surgery, Dr. Severino testified that it was not an emergency. Rather, the timeline for the sur- gery depended on the availability of all parties involved: Q. Explain the time frame in which you’d schedule a surgery like this? A. As soon as we can get everybody on board. You know, get everybody seen and get it scheduled. You know, this is a ten-hour case or whatever. You’ve got to have three surgeons that are available for ten hours on the same day. There’s a little bit of complexity with 16 Nos. 20-3058 & 20-3139

that, because again, if it’s not done right, you’re guar- anteed to have a dead patient on the table. When asked specifically whether he thought the delay from April to July was reasonable, Dr. Severino said yes, explain- ing: “I would rather have [it] scheduled correctly than have a patient die on the table because you want to try and hurry up.” Dean called three experts to testify to the adequacy of the defendants’ actions in diagnosing and treating his cancer. Dr. Adam Metwalli, a urologic oncologist, testified that an ultra- sound was “not the ideal test in this case” and that a CT scan “should have been done earlier.” He explained that the risk of a CT scan for Dean was low, such that there was no reason to opt for an ultrasound. Still, he agreed on cross-examination that “whether or not to order an ultrasound in a specific situ- ation involves an exercise of clinical judgment.” In Dr. Metwalli’s view, the seven-month delay between Dean’s presentation of symptoms and his surgery violated the standard of care. He explained that “time is of the essence” with kidney cancer and that “the idea of waiting seven months before you get someone like that to the operation room is virtually unheard of in my experience.” In his words, “nobody in their right mind would wait that long.” He also testified that the delay harmed Dean. As a general matter, he explained that Dean “underwent a more complicated opera- tion that required a longer recovery and was higher risk than he needed to because of the delay.” More specifically, he tes- tified—after reviewing the April CT scan alongside the June CT scan—that, as of April 2016, Dean’s cancer had not yet ex- tended far enough up his vena cava to require opening his chest and stopping his heart. Between April and June, “the Nos. 20-3058 & 20-3139 17

tumor burden actually increased substantially,” which in turn required a more complex surgery. At the same time, Dr. Met- walli testified that there was no evidence that the delay from December to July caused the metastasis of cancer. Indeed, he opined that the cancer had likely already spread to Dean’s liver by December 2015, even though it was not visible on the April 2016 CT scan. Dr. Nivedita Dhar, a urologist, testified similarly. In her opinion, the delay preceding the urology referral and CT scan violated the standard of care. She testified that the standard of care for an elderly patient presenting painless gross hema- turia is to assume cancer until proven otherwise and that the way to rule out cancer is to refer the patient to a urologist for a CT scan within two weeks. She agreed that an ultrasound would be an appropriate first step for a patient with severely compromised kidney function. But she did not think that Dean fell within that exception. She also quantified the quali- tative differences between ultrasounds and CT scans: ultra- sounds can miss kidney masses up to 50 percent of the time whereas CT scans catch them 95 to 98 percent of the time. Like Dr. Metwalli, Dr. Dhar testified that she does not do surgeries as complex as Dean’s because “we make the diagnosis sooner.” Dean’s final expert was Dr. Bruce Barnett, a medical con- sultant specializing in correctional healthcare. Dr. Barnett was highly critical of collegial review at Wexford. He described it as “dangerous as applied” in Dean’s case. He stressed that Wexford’s policies allow a prison’s medical director to go out- side the normal process in urgent and emergent situations, but that never happened in Dean’s case. As a result, “instead of acting to protect the patient and make sure that he’s getting 18 Nos. 20-3058 & 20-3139

the proper kind of care, [collegial review] only acted to harm the patient and keep him from getting the care that he needed.” He described collegial review as a “barrier” to proper patient care. While collegial review is supposed to “do good” by helping identify appropriate treatment, “the colle- gial review process at Wexford is broken.” Dr. Barnett also testified about Dean’s harm. He explained that “not having the proper procedure, the proper testing, and then the proper treatment meant that [Dean] suffered for seven months before relief.” This suffering included the physical pain and “psychic agony” of passing blood clots, as well as “the growth of can- cer.” On cross-examination, Dr. Barnett acknowledged that some of the delays between April and July were attributable to Dr. Severino. Still, he faulted collegial review for adding delays to each step of the process. He also criticized collegial review for excluding specialists like Dr. Einwohner from the review process. Dr. Barnett was equally critical of the doctor-defendants’ care of Dean. Similar to Dean’s other experts, Dr. Barnett tes- tified that the standard of care is a urology referral and CT scan and that Dr. Nawoor breached the standard of care by opting instead for an ultrasound. Dr. Barnett also faulted Dr. Nawoor for failing to advocate for Dean. In his view, Dr. Na- woor did not adequately convey the facts to Dr. Ritz, and he “should have acted more aggressively than he did.” Like Dr. Metwalli, Dr. Barnett described the ultrasound as “less ideal” than a CT scan, though he conceded that “arguably the renal ultrasound is not a terrible thing to do.” The real problem with the ultrasound, in his view, was that it took six weeks to perform. On cross-examination, he agreed that an Nos. 20-3058 & 20-3139 19

“ultrasound can be an appropriate part of the workup for he- maturia.” As for Dr. Einwohner, Dr. Barnett described her email to Dr. Ritz suggesting a CT scan and urology evaluation as “ut- terly inadequate” given that she never followed up. He ex- plained: [T]his is an extraordinary, severe circumstance. This is a person in front of her who she knows is quite likely to have renal cell cancer, which means that untreated, he’ll die. That’s the most serious thing that you could be facing as a physician. Your job is to prevent the death. And to me, it’s like seeing your neighbor’s house on fire and leaving a message on his phone machine in- stead of calling the police department or the fire de- partment. The defendants countered Dean’s case with their own ex- pert witnesses. Dr. Richard Kosierwoski, an oncologist, testi- fied that Dr. Nawoor and Dr. Ritz complied with the standard of care under the circumstances. He testified that it was rea- sonable for Dr. Einwohner, as a telemedicine physician acting in a support role, to make recommendations to the doctors in charge of Dean’s care—i.e., Dr. Nawoor and Dr. Ritz. And he agreed with Dr. Nawoor’s assessment that the CT scan posed a significant risk to Dean, which made an ultrasound a better option. At the same time, he agreed that the CT scan “could have been done tighter” and that there was “plenty of blame” to go around. He conceded that he was unaware of any rea- son—other than collegial review—why Dr. Nawoor could not have ordered an ultrasound, CT scan, or urology referral on 20 Nos. 20-3058 & 20-3139

December 23, 2015, when he first learned of Dean’s symp- toms. The defendants’ second expert was Dr. Michael Racen- stein, a radiologist. Dr. Racenstein discussed Dean’s imaging and the progression of his cancer over time. He opined that the August 2014 CT scan did not show a mass, but the mass on Dean’s right kidney was already visible in the July 2015 CT scan. By February 2016, the mass on the right kidney had grown and was “very well displayed” on the ultrasound. The April 2016 CT scan showed that the cancer had spread to the vena cava. And the June 2016 CT scan showed that the cancer had spread to Dean’s liver. Like Dr. Metwalli, Dr. Racenstein testified that Dean’s cancer could have metastasized earlier, even though the April 2016 CT scan did not show it. More generally, Dr. Racenstein agreed that Dean’s cancer had spread between December 2015 and July 2016. As for the ad- equacy of the defendants’ care, Dr. Racenstein opined that it was reasonable to start with either an ultrasound or a CT scan, but he agreed that the timeframe for treating Dean’s cancer was “surprisingly too long.” In closing arguments, Dean’s counsel devoted much of his time to the deliberate indifference claim against Wexford and the evidence regarding collegial review. He argued that colle- gial review “built” delay into each step of Dean’s care, not for any medical reason but rather to control costs. He walked through each collegial review that occurred during the diag- nosis and treatment of Dean’s cancer and concluded that the review process added “substantial delay into Mr. Dean’s care.” Counsel used the Lippert reports to demonstrate that Wexford knew independent experts had reported harmful de- lays resulting from collegial review. He referenced the reports Nos. 20-3058 & 20-3139 21

again when asking the jury to impose punitive damages against Wexford. In its final instructions to the jury, the district court in- cluded a limiting instruction on the Lippert reports. Per the in- struction, the jury could consider the reports “only in decid- ing whether [Wexford] had notice and knowledge of the in- formation in the reports, not whether the information in the reports is true.” The court added that Wexford “disputes the truth of those reports and has not admitted liability in that case.” The jury found for Dean on both the negligence-based claims and the deliberate indifference claims. It awarded him $1 million in compensatory damages to cover his physical pain and suffering ($100,000), emotional pain and suffering ($500,000), disability and loss of normal life/diminished life expectancy ($100,000), and future medical expenses ($300,000). It also awarded him punitive damages in the amount of $10 million against Wexford, $12,500 against Dr. Einwohner, and $25,000 against Dr. Nawoor. 3. Post-Trial Rulings After trial, the defendants moved for judgment as a matter of law or a new trial on the Eighth Amendment claims. They argued that the court had improperly admitted the Lippert re- ports and that the evidence could not sustain the verdict against them on the Eighth Amendment claims. They did not challenge the sufficiency of the evidence on the negligence claims. The district court denied the defendants’ motion for a new trial. It defended its admission of the Lippert reports on the ground that both reports “were relevant to Wexford’s notice 22 Nos. 20-3058 & 20-3139

from independent court experts that its procedures, including collegial review, caused significant and unnecessary delays in the delivery of off-site care.” The court viewed the 2018 report as “a continuation” of the 2014 report, such that both reports were relevant. Admitting the reports was not unfairly preju- dicial, moreover, because the reports were “critical of Wex- ford but not inflammatory, and any potential unfair prejudice was mitigated by the Court’s limiting instruction and by the opportunity of Wexford representatives to testify that they disagreed with the reports’ conclusions.” The district court further concluded that, even if it had erred in admitting the reports, the error was harmless because “ample other evi- dence” supported a finding that Wexford was deliberately in- different. The district court denied the motion for judgment as a matter of law for similar reasons. It reasoned that the expert testimony on the proper treatment of painless gross hematu- ria, along with the doctor-defendants’ demeanors at trial, per- mitted a finding that they were deliberately indifferent. And with or without the Lippert reports, the evidence permitted a finding that Wexford was deliberately indifferent. Namely, the jury could find that collegial review on its face “would ob- viously and inevitably delay urgently needed care for some inmates, including Plaintiff” for no medical reason. The district court agreed, however, to reduce the punitive damages award against Wexford to $7 million, finding that the jury’s $10 million award violated due process. The defendants now appeal the district court’s denial of their post-trial motions. Their appeal focuses on the Eighth Amendment claims; they do not directly challenge the jury’s Nos. 20-3058 & 20-3139 23

verdict on the negligence claims. Dean cross-appeals the re- duction of punitive damages. II. Discussion The defendants maintain that Dean failed to prove their liability on the Eighth Amendment claims, such that the dis- trict court should have entered judgment as a matter of law in their favor. We review the district court’s denial of judgment as a matter of law de novo. Burton v. E.I. du Pont de Nemours & Co., Inc., 994 F.3d 791, 817 (7th Cir. 2021). Judgment as a matter of law is warranted only if “a reasonable jury would not have a legally sufficient evidentiary basis to find for the party on that issue.” Fed. R. Civ. P. 50(a)(1). A. Eighth Amendment Claim Against Wexford We start with the Eighth Amendment claim against Wex- ford. Before reaching the merits, however, we must address the defendants’ argument that the district court abused its discretion in allowing Dean to use the Lippert reports as evi- dence of Wexford’s deliberate indifference. 1. Lippert Reports The defendants contend that the Lippert reports are inad- missible hearsay, irrelevant, and unfairly prejudicial. As an initial matter, the defendants preserved these arguments by raising them in a written motion in limine before trial. Dean does not suggest otherwise. The dissent faults the defendants for objecting only to “relevance” when Dean offered the 2014 report at trial. But the court and the parties discussed the de- fendants’ written motion in limine at the start of trial, and the defendants told the court they would rely on the argument in their original motion rather than file a reply to Dean’s 24-page response. In its written ruling on the defendants’ motion in 24 Nos. 20-3058 & 20-3139

limine, the court addressed each of the defendants’ argu- ments, including unfair prejudice. In these circumstances, we see no reason to raise forfeiture sua sponte—especially given Dean’s 24-page response to the written motion. See Henry v. Hulett, 969 F.3d 769, 785 (7th Cir. 2020) (en banc) (noting that prejudice to the opposing party is the rationale for finding for- feiture). We review the district court’s admission of the re- ports for abuse of discretion. Burton, 994 F.3d at 812. Dean does not dispute that the Lippert reports would have been inadmissible hearsay if he had offered them to prove the truth of their contents. See Fed. R. Evid. 801. Indeed, we held as much in Wilson v. Wexford Health Sources, Inc., 932 F.3d 513, 522 (7th Cir. 2019). Dean points out, however, that he did not offer the reports for the truth of their contents. Rather, he of- fered them to prove that Wexford was on notice of their con- tents while treating him. The defendants do not dispute that a statement that is used only to show that the opposing party had notice of it is not hearsay. Marseilles Hydro Power, LLC v. Marseilles Land & Water Co., 518 F.3d 459, 468 (7th Cir. 2008); see also Daniel v. Cook Cnty., 833 F.3d 728, 735–36 (7th Cir. 2016). Rather, they maintain that the Lippert reports did not provide them with relevant notice of anything. With respect to the 2018 report, we agree. The 2018 report did not even exist during the period relevant to this lawsuit— late 2015 through early 2017. Dr. Puisis produced the report in late 2018, a year and a half after Dean filed suit. Echoing the district court, Dean contends that the 2018 report was a “con- tinuation” of the 2014 report. But we do not follow that rea- soning. For one thing, the 2018 report was not merely a con- tinuation of the 2014 report. It was a separate review by a sep- arate team of doctors designed to reassess the problems that Nos. 20-3058 & 20-3139 25

the 2014 report had identified, and to identify new problems if any existed. Some of Dr. Puisis’s findings tracked Dr. Shan- sky’s findings; others did not. In any event, even if the 2018 report were merely a continuation of the 2014 report, it would still be irrelevant in this case. Put simply, the findings of a 2018 report could not have put Wexford on notice regarding its actions prior to 2018 or affected Wexford’s decision to maintain collegial review in 2015, 2016, or 2017. In the dissent’s view, the 2018 report was relevant because it addressed Wexford’s polices as of 2014 and shed light on Wexford’s knowledge at that time. We respectfully disagree. As an initial matter, although the 2018 report was admitted for notice purposes only, this view sounds like the truth of the matter. Regardless, the admitted excerpts from the 2018 re- port did not discuss the 2014 policies. The admitted 2018 Lip- pert report consisted of six heavily redacted pages. It con- cluded that “the health program is not significantly improved since the First Court Expert’s report.” The admitted excerpts then went on to conclude that the “collegial review process of accessing specialty care is a patient safety hazard and should be abandoned until patient safety is ensured.” That opinion pertained to 2018, not 2014. Similarly, the remaining admitted excerpts discussed the “current findings” and conditions at the facilities, not the 2014 findings. Even if the 2018 report described the 2014 policies or prob- lems with those polices, it did not say or even suggest that Wexford had knowledge of such issues at the time. More im- portantly, the admitted excerpts of the 2018 report did not re- count any events occurring in 2014. As noted, the jury saw a heavily redacted version of the report that contained only Dr. Puisis’s findings and conclusions—most notably, his “patient 26 Nos. 20-3058 & 20-3139

safety hazard” conclusion. Dr. Puisis published those conclu- sions in late 2018, so Wexford could not have known about them before. To borrow the dissent’s analogy, a person acting in 1972 could not have known what conclusions a historian would reach in 2002, even if the historian were reviewing events that occurred in 1972. Because the 2018 report is irrele- vant to Dean’s claims, the district court abused its discretion in admitting it. See Fed. R. Evid. 402 (“Irrelevant evidence is not admissible.”). Hoping to salvage the admission of the 2018 report, Dean contends that it was also admissible under Federal Rule of Ev- idence 807. Rule 807 creates a residual exception to the rule against hearsay. It provides that a sufficiently trustworthy hearsay statement is admissible if “it is more probative on the point for which it is offered than any other evidence that the proponent can obtain through reasonable efforts.” Fed. R. Evid. 807(a). We construe the Rule 807 requirements nar- rowly. Burton v. Kohn L. Firm, S.C., 934 F.3d 572, 583 (7th Cir. 2019). Dean maintains that the 2018 report is more probative as to notice than any other evidence that he could obtain through reasonable efforts, given the defendants’ supposed failure to produce relevant discovery. But, as we have already explained, the 2018 report was irrelevant to notice so it could not have been “more probative” than anything. Moreover, Rule 807 applies to hearsay statements, and by Dean’s own account he offered the 2018 report for a non-hearsay purpose. Rule 807 therefore does not apply. More generally, Rule 807 is not a proper vehicle to remedy a discovery violation. If the defendants were late with discovery, Dean could have filed a motion to compel before the district court. Nos. 20-3058 & 20-3139 27

The analysis for the 2014 report is more complicated. Dr. Shansky produced the 2014 report only a year before Dean first presented symptoms, and the report highlighted a prob- lem—systemic delays in medical care resulting from collegial review—that, at least at first glance, seems closely linked to the problem at the heart of Dean’s lawsuit against Wexford. On the other hand, Taylorville was not one of the facilities that Dr. Shansky examined in connection with his report, and Dr. Shansky acknowledged that his finding of systemic delays was somewhat facility specific. Beyond that, the collegial re- view policy in effect when Dr. Shansky conducted his review did not have an exception for urgent or emergent cases. In that respect, it differs from the policy in effect for nearly all events relevant to this lawsuit. So, there is room to debate the rele- vance of the 2014 report. See Fed. R. Evid. 401 (providing that evidence is relevant if it has “any tendency” to make a mate- rial fact “more or less probable than it would be without the evidence”). Federal Rule of Evidence 403 adds another layer of com- plexity to the analysis for the 2014 report. That rule allows a court to “exclude relevant evidence if its probative value is substantially outweighed by a danger of one or more of the following: unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.” Fed. R. Evid. 403. The defendants con- tend, and we cannot deny, that the 2014 report poses signifi- cant dangers of “confusing the issues” and “misleading the jury” in this case. And that is true with or without a limiting instruction. While “we assume that limiting instructions are effective in reducing or eliminating unfair prejudice,” United States v. Vargas, 552 F.3d 550, 557 (7th Cir. 2008), they are not always a perfect solution, see, e.g., United States v. Gomez, 763 28 Nos. 20-3058 & 20-3139

F.3d 845, 860 (7th Cir. 2014) (en banc). The 2014 report reflects the opinion of an independent court-appointed expert that collegial review causes systemic delays in medical care for Il- linois inmates. In a case alleging systemic delays in medical care resulting from collegial review, telling jurors to ignore the truth of the report is somewhat “like telling jurors to ig- nore the pink rhinoceros that just sauntered into the court- room.” United States v. Jones, 455 F.3d 800, 811 (7th Cir. 2006) (Easterbrook, J., concurring). Even so, we recognize that it is usually necessary in Monell cases to introduce evidence of a prior pattern of similar con- stitutional violations. See Bd. of Cnty. Comm’rs v. Brown, 520 U.S. 397, 407–08 (1997). By its nature, evidence of a defend- ant’s past violations creates a risk that the jury will infer cul- pability in the present case from culpability in past cases. See 22A Charles Alan Wright & Arthur R. Miller, Federal Practice & Procedure § 5216.1 & n.10.50 (2d ed. 1987 & Supp. 2021). A court should hesitate to hold that evidence is inadmissible un- der Rule 403 when the governing law makes it necessary to introduce evidence that might confuse the issues or mislead the jury. See Thompson v. City of Chi., 722 F.3d 963, 976 (7th Cir. 2013) (danger of introducing evidence of other officers’ “out- rageous conduct” was “heavily discounted” because “that risk is always present in a conspiracy claim”). In addition, while the defendants protest that they had no way to chal- lenge Dr. Shansky’s findings because he was prohibited from testifying, the defendants were free to challenge the report in other ways. In the end, we need not decide whether the district court abused its discretion in admitting the 2014 report because, as we explain below, Dean failed to prove that Wexford was Nos. 20-3058 & 20-3139 29

deliberately indifferent with or without the 2014 report. For now, we emphasize that district courts have considerable dis- cretion in deciding whether to exclude relevant evidence un- der Rule 403, United States v. Burt, 495 F.3d 733, 740 (7th Cir. 2007), but in a case like this the Lippert reports pose serious dangers. Before admitting them, a court should carefully bal- ance their probative value against these dangers. See United States v. Eads, 729 F.3d 769, 777 (7th Cir. 2013). The governing law and other evidence in the case will also affect this analy- sis. See Thompson, 722 F.3d at 971, 976; Old Chief v. United States, 519 U.S. 172, 184–85 (1997). Limiting instructions are presumptively effective, Vargas, 552 F.3d at 557, but they are not foolproof. Gomez, 763 F.3d at 860; Wright & Miller §§ 5224 & 5224.1. 2. Monell Liability We now address whether the evidence at trial permitted the jury to find Wexford liable for violating Dean’s Eighth Amendment rights. For purposes of this analysis, we assume without deciding that the district court permissibly admitted the 2014 Lippert report. The Eighth Amendment’s ban on “cruel and unusual pun- ishments” obligates prison officials to provide medical care to prisoners in their custody. Estelle v. Gamble, 429 U.S. 97, 103 (1976). To prevail on an Eighth Amendment claim for inade- quate medical care, a prisoner must show that a prison official acted with deliberate indifference to the prisoner’s objectively serious medical need. Farmer v. Brennan, 511 U.S. 825, 834 (1994). Section 1983 creates a private right of action against any “person” who violates the plaintiff’s federal rights while 30 Nos. 20-3058 & 20-3139

acting under color of state law. 42 U.S.C. § 1983. In Monell, 436 U.S. at 690, the Supreme Court held that municipalities are “person[s]” who may be sued under § 1983. But the Court added an important caveat: Municipalities are not vicariously liable for the constitutional torts of their employees or agents. Id. at 691–94. “Instead, it is when execution of a government’s policy or custom, whether made by its lawmakers or by those whose edicts or acts may fairly be said to represent official policy, inflicts the injury that the government as an entity is responsible under § 1983.” Id. at 694. In other words, a munic- ipality is liable under § 1983 only “for its own violations of the federal Constitution and laws.” First Midwest Bank ex rel. LaPorta v. City of Chi., 988 F.3d 978, 986 (7th Cir. 2021) (empha- sis added). Monell governs Wexford’s liability in this case be- cause we, like our sister circuits, treat private corporations acting under color of state law as municipalities. Iskander v. Vill. of Forest Park, 690 F.2d 126, 128 (7th Cir. 1982); see also Shields v. Ill. Dep’t of Corr., 746 F.3d 782, 789–96 (7th Cir. 2014) (tracing the development of the doctrine and questioning its foundations). 2 We recently reiterated the elements of a Monell claim in LaPorta. To begin, a § 1983 plaintiff must always show “that he was deprived of a federal right.” LaPorta, 988 F.3d at 987. Beyond that, the plaintiff must trace the deprivation to some municipal action (i.e., a “policy or custom”), such that the

2 In a short footnote, Dean asks us to overrule Iskander and its progeny and

hold that private corporations, unlike their municipal counterparts, are subject to respondeat superior liability. To persuade us to overturn our prec- edent, however, Dean must offer more than a few conclusory sentences in a footnote. Dean does not even engage with Shields, where we ourselves canvassed the arguments for overturning Iskander. Nos. 20-3058 & 20-3139 31

challenged conduct is “properly attributable to the municipal- ity itself.” Id. at 986. There are at least three types of municipal action that may give rise to municipal liability under § 1983: “(1) an express policy that causes a constitutional deprivation when enforced; (2) a widespread practice that is so permanent and well-settled that it constitutes a custom or practice; or (3) an allegation that the constitutional injury was caused by a person with final policymaking authority.” Id. (quoting Spie- gel v. McClintic, 916 F.3d 611, 617 (7th Cir. 2019)). Inaction, too, can give rise to liability in some instances if it reflects “a con- scious decision not to take action.” Glisson v. Ind. Dep’t of Corr., 849 F.3d 372, 381 (7th Cir. 2017) (en banc); accord J.K.J., 960 F.3d at 378. Next, the plaintiff must show that “the policy or custom demonstrates municipal fault,” i.e., deliberate indif- ference. LaPorta, 988 F.3d at 986. “This is a high bar.” Id. at 987. If a municipality’s action is not facially unconstitutional, the plaintiff “must prove that it was obvious that the municipal- ity’s action would lead to constitutional violations and that the municipality consciously disregarded those conse- quences.” Id. Finally, the plaintiff must show that the munici- pal action was “the ‘moving force’ behind the federal-rights violation.” Id. (quoting Brown, 520 U.S. at 404). This “rigorous causation standard” requires “a ‘direct causal link’ between the challenged municipal action and the violation of [the plaintiff’s] constitutional rights.” Id. (quoting Brown, 520 U.S. at 404). In short, a Monell plaintiff must show that some municipal action directly caused him to suffer a deprivation of a federal right, and that the municipality took the action with conscious disregard for the known or obvious risk of the deprivation. 32 Nos. 20-3058 & 20-3139

Dean relies on an express policy (collegial review), and we assume for present purposes that he has shown a constitu- tional deprivation. Even so, Dean has not shown municipal fault or moving-force causation—two indispensable prereq- uisites to Monell liability. Dean concedes that collegial review is not unconstitu- tional on its face. We held as much in Howell v. Wexford Health Sources, Inc., 987 F.3d 647, 659 (7th Cir. 2021). His theory in- stead is that collegial review caused unconstitutional delays as applied to him. This type of claim presents “difficult prob- lems of proof.” Brown, 520 U.S. at 406. As early as 1985, a plu- rality of the Supreme Court made clear that a plaintiff seeking to hold a municipality liable for a facially lawful policy gen- erally must prove a prior pattern of similar constitutional vi- olations resulting from the policy. As the Court put it: “[W]here the policy relied upon is not itself unconstitutional, considerably more proof than the single incident will be necessary in every case to establish both the requisite fault on the part of the municipality, and the causal connection between the ‘pol- icy’ and the constitutional deprivation.” City of Okla. City v. Tuttle, 471 U.S. 808, 824 (1985) (plurality) (emphasis added); see also Calderone v. City of Chi., 979 F.3d 1156, 1164 (7th Cir. 2020) (holding based on Tuttle that “[o]ne single incident can- not suffice; rather, Calderone must show ‘a series of constitu- tional violations’”) (quoting Est. of Novack ex rel. Turbin v. Cnty. of Wood, 226 F.3d 525, 531 (7th Cir. 2000)). The Court discussed the rationale behind this requirement in Brown. First, a prior pattern of similar violations puts the municipality on notice of the unconstitutional consequences of its policy, such that its “continued adherence” to the policy might “establish the conscious disregard for the consequences Nos. 20-3058 & 20-3139 33

of [its] action—the ‘deliberate indifference’—necessary to trigger municipal liability.” Brown, 520 U.S. at 407; accord Jack- son v. Marion Cnty., 66 F.3d 151, 152 (7th Cir. 1995) (explaining that “a series of bad acts” allows the inference that “the poli- cymaking level of government was bound to have noticed what was going on and by failing to do anything must have encouraged or at least condoned, thus in either event adopt- ing, the misconduct of subordinate officers”). Similarly, a pat- tern of violations may show that the policy itself, “rather than a one-time negligent administration of the program or factors peculiar to the officer involved in a particular incident, is the ‘moving force’ behind the plaintiff’s injury.” Brown, 520 U.S. at 407–08 (quoting City of Canton v. Harris, 489 U.S. 378, 390 (1989)). To be sure, there are limited exceptions to this rule. In some “rare” cases, the risk of unconstitutional consequences from a municipal policy “could be so patently obvious that a [municipality] could be liable under § 1983 without proof of a pre-existing pattern of violations.” Connick v. Thompson, 563 U.S. 51, 64 (2011); accord J.K.J., 960 F.3d at 380–81; Glisson, 849 F.3d at 382. In Glisson, moreover, we hypothesized that “[a] single memo or decision showing that the choice not to act is deliberate could also be enough.” Glisson, 849 F.3d at 381; see also King v. Kramer, 680 F.3d 1013, 1021 (7th Cir. 2012) (seven news articles and policymaker’s admitted knowledge of re- ported problems permitted inference of deliberate indiffer- ence). Again, these cases are the exception. Regardless of the exact form of proof, the question is always whether the mu- nicipal policy reflects a conscious disregard for a known or obvious risk of the constitutional deprivation. See J.K.J., 960 F.3d at 381. “A showing of simple or even heightened negli- gence will not suffice.” Brown, 520 U.S. at 407. 34 Nos. 20-3058 & 20-3139

The dissent contends that we have collapsed the critical distinction between the existence of a policy and the effects of that policy. According to the dissent, pattern or practice evi- dence is only necessary when the presence of an official pol- icy, custom, or practice is in question. Not so. The Supreme Court certainly did not hold that in Tuttle. See 471 U.S. at 823– 24 (“Proof of a single incident of unconstitutional activity is not sufficient to impose liability under Monell, unless proof of the incident includes proof that it was caused by an existing, unconstitutional municipal policy … .”). Moreover, we recently recognized the distinction between a policy that is unconstitutional on its face and one that is not in Calderone. See 979 F.3d at 1164–65. There, the plaintiff brought an as-applied constitutional challenge to the City’s personnel rules under Monell. The existence of the policy–the written personnel rules–was not at issue. Id. at 1164. Nonethe- less, we rejected Calderone’s Monell claim because she failed to demonstrate causation and culpability based on her single incident of an alleged constitutional violation. Since she could not establish that the personnel rules were unconstitutional on their face, she had to show a “series of bad acts[,] creating an inference that municipal officials were aware of and con- doned the misconduct of their employees.” Id. We empha- sized that “[o]ne single incident cannot suffice; rather, Calde- rone must show ‘a series of constitutional violations.’” Id. Dean did not introduce any substantive evidence of a pat- tern or practice of similar violations. He did not offer substan- tive evidence that collegial review had caused unconstitu- tional delays for other prisoners. He only offered substantive evidence of collegial review causing unconstitutional delays in his own healthcare. Nor does he contend on appeal that his Nos. 20-3058 & 20-3139 35

is one of those “rare” cases where the risk of unconstitutional delays is “patently obvious” even without proof of other vio- lations. Connick, 563 U.S. at 64. The district court relied on the “obviousness” theory, reasoning that the jury could find that collegial review on its face “would obviously and inevitably delay urgently needed care for some inmates, including Plain- tiff” for no medical reason. But as we discuss below, Wexford allows its medical directors to go outside the normal collegial review process in urgent or emergent situations, so it could not have been obvious from the face of the policy that collegial review would delay urgently needed care. If offsite care was urgent, the policy provided an exception to prevent harmful delays. Rather than offer substantive evidence of deliberate indif- ference, Dean relies solely on the Lippert reports. His theory is that the Lippert reports put Wexford on notice that independ- ent experts in another case had found that collegial review was causing systemic delays in medical care and that Wexford consciously disregarded that risk in adhering to collegial re- view at the Taylorville facility. We have held that the 2018 re- port was inadmissible, so the question is whether the 2014 re- port gave the jury a sufficient basis for finding that Wexford acted with deliberate indifference. 3

3 Even if the 2018 report was admissible as the dissent contends, it makes no difference to our analysis. Dean seeks to impose Monell liability based on a facially lawful policy and therefore must prove a prior pattern of sim- ilar constitutional violations resulting from the policy. This requires “con- siderably more proof than the single incident.” Tuttle, 471 U.S. at 824. The 2018 Lippert report was deficient to establish deliberate indifference for the same reason that the 2014 Lippert report does not provide this crucial evidence. The dissent asserts there are only twelve pages between allow- ing Dean to keep the damages in this case. To the contrary, Dean did not 36 Nos. 20-3058 & 20-3139

It did not. Although we have not directly confronted this issue before, our prior cases suggest that evidence admitted only for notice cannot establish that a municipality acted with deliberate indifference unless the plaintiff also has substan- tive proof that the “noticed” problems actually existed. In Daniel, we held that a Department of Justice report finding that medical care at the Cook County Jail fell below constitu- tional standards was probative of deliberate indifference— but we stressed that the plaintiff there had also produced sub- stantive evidence showing that the problems in the report ex- isted. 4 Daniel, 833 F.3d at 735–36; see also J.K.J., 960 F.3d at 372, 382 (deeming letter admitted for notice probative of deliberate indifference where there was also substantive evidence). Here, outside of Dean’s own injury, his only evidence of Wex- ford’s deliberate indifference is a hearsay report admitted for notice. Even assuming, moreover, that notice-only evidence can prove deliberate indifference for Monell liability, the 2014 re- port nonetheless falls short. The 2014 report provides notice— but notice of what? Taylorville was not one of the facilities re- viewed in the 2014 report, so the report could not have given Wexford notice of any specific problems occurring there. Cf. Wilson, 932 F.3d at 522 (upholding exclusion of Lippert reports

introduce any substantive evidence of similar constitutional violations in this case. 4The dissent claims that Daniel is unhelpful because it did not involve a written policy but instead a custom or practice. The discussion upon which we rely, however, pertains to Daniel’s ability to prove the munici- pality’s notice, not the existence of the policy. Nos. 20-3058 & 20-3139 37

because they revealed general problems at Wilson’s prison “without linking those problems to Wilson’s personal experi- ence”). Relatedly, the report portrayed the problem of delays as facility specific: some facilities were “conscientiously” ad- ministering collegial review, while others were not. Because the 2014 report did not link any problems to Taylorville spe- cifically, it gave Wexford no particular insight into whether Taylorville was one of the facilities where collegial review was causing problems, nor did Dean introduce any evidence that it was. See Walker v. Wexford Health Sources, Inc., 940 F.3d 954, 967 (7th Cir. 2019) (“Wexford’s knowledge that some referrals slipped through the cracks is not the same as Wexford’s knowledge that constitutionally necessary referrals were not happening with such frequency that it ignored an obvious risk of serious harm.”). Furthermore, the admitted excerpts of the report said nothing about the harm (if any) resulting from the reported delays, making it difficult to infer solely from the report that Wexford knew of any unconstitutional conse- quences from the delays and consciously disregarded the risk of those consequences while caring for Dean. Most critically, the 2014 report reviewed a materially dif- ferent version of Wexford’s collegial review policy. Undis- puted evidence at trial—which Dean introduced—showed that the policy in effect in 2014 and reviewed by the 2014 re- port did not contain an exception for urgent or emergent cases. In fact, the 2014 Lippert report recommended creating such an exception, and Wexford did just that when it adopted a new version of the policy that went into effect on January 14, 2016 (the day after Dean’s first collegial review). The up- dated policy provides that medical directors can fast-track ur- gent or emergent cases. Emergencies do not require collegial review; for urgent cases, collegial review can occur the same 38 Nos. 20-3058 & 20-3139

day. This exception for urgent or emergent cases (whether motivated by the 2014 report or not) is directly responsive to the possibility that collegial review might cause harmful de- lays in these cases. As such, even if the 2014 report gave Wex- ford notice that its prior policy would cause constitutional vi- olations, it could not have given Wexford notice that its up- dated policy suffered from the same deficiencies. The up- dated policy contained a critical exception that the earlier pol- icy did not. Granted, the exception for urgent or emergent cases does not automatically insulate Wexford from liability. If, for ex- ample, Dean could show that the exception was not prevent- ing harmful delays—and that Wexford was deliberately indif- ferent to this known risk—then Dean might be able to hold Wexford liable despite the exception. But that is not what the evidence shows. Dean argued that the 2014 report, standing alone, put Wexford on notice that collegial review would vio- late his constitutional rights. The problem with his argument is that collegial review had changed by the time of his treat- ment. Dean cannot show that Wexford “consciously disre- garded” a known risk of constitutional violations while treat- ing him when, only weeks after Dean first presented symp- toms, Wexford adopted a policy change that was directly re- sponsive to that risk. The dissent calls the 2016 policy change “cosmetic,” point- ing to similarities in the language of the 2014 and 2016 poli- cies. But the portions of the policies that the dissent cites cover “Emergency/Hospital Notification” procedures. Different sections of the policies deal with collegial review. And it is those sections that differ in a material way. The 2016 policy provides an exception for “more urgent cases” that the 2014 Nos. 20-3058 & 20-3139 39

policy does not, as Dr. Ritz testified. Policy changes aside, the dissent suggests there is no evidence that Wexford’s practices changed after it implemented the 2016 policy change. But Dean’s Monell claim focuses on Wexford’s express policy of collegial review—not on its informal customs or practices. For these reasons, the 2014 report alone did not permit the inference that Wexford consciously disregarded a known or obvious risk that collegial review would violate Dean’s con- stitutional rights. For related reasons, Dean fails to show moving-force cau- sation. Dean must show that collegial review itself—not simply the actions of the employees administering it—di- rectly caused his constitutional deprivation. Yet his correc- tional healthcare expert, Dr. Barnett, described collegial re- view as “dangerous as applied” to Dean, and specifically tes- tified that the exception for urgent or emergent cases should have prevented the delays in Dean’s care. He did not testify (nor did anyone else) that collegial review violated the consti- tutional rights of other inmates. Dr. Barnett’s testimony strongly suggests that the delays in Dean’s care resulted from the negligent actions of Wexford’s agents, and not from colle- gial review. Monell requires more; Dean must show that Wex- ford itself directly caused the constitutional violation. Dean also points to Dr. Nawoor’s testimony that Wexford’s “prac- tices” were to blame for the delays in Dean’s care. But Dr. Na- woor never testified (nor did anyone else) that collegial re- view had caused similar problems for other inmates, so his testimony falls short of establishing that collegial review itself was the moving force behind Dean’s constitutional injury. See Tuttle, 471 U.S. at 824; Brown, 520 U.S. at 407–08. 40 Nos. 20-3058 & 20-3139

Consistent with the Supreme Court’s guidance, we have repeatedly rejected Monell claims that rest on the plaintiff’s individualized experience without evidence of other constitu- tional violations. See, e.g., Quinn v. Wexford Health Sources, Inc., 8 F.4th 557, 567-68 (7th Cir. 2021); Howell, 987 F.3d at 659; Cal- derone, 979 F.3d at 1165; Ruiz-Cortez v. City of Chi., 931 F.3d 592, 599 (7th Cir. 2019); Chatham v. Davis, 839 F.3d 679, 685 (7th Cir. 2016); Hahn v. Walsh, 762 F.3d 617, 638 (7th Cir. 2014); Cal- houn v. Ramsey, 408 F.3d 375, 381 (7th Cir. 2005); Turbin, 226 F.3d at 531; Robles v. City of Fort Wayne, 113 F.3d 732, 737 (7th Cir. 1997). We do so again here. While we are sympathetic to Dean’s experience, his only substantive proof relates to the delays in care that he himself experienced. He has not proven a pattern of similar constitutional violations or a patently obvious risk of such violations. See Tuttle, 471 U.S. at 824. We acknowledge, as we did in Glisson, that there may be other pathways to Mo- nell liability based on a facially lawful policy. But this case does not require us to further explore that possibility. Dean’s only other evidence is the 2014 report, which shows, at most, that Wexford knew an expert in another case had concluded that a materially different version of its collegial review policy had caused delays at some other IDOC facilities. This notice evidence alone cannot establish that Wexford knew of, and consciously disregarded, the risk that collegial review would likely violate Dean’s constitutional rights. Nor has Dean shown that collegial review itself rather than “a one-time neg- ligent administration of the program” was the moving force behind his constitutional injury. Brown, 520 U.S. at 407–08. Anticipating our holding that the Lippert reports do not es- tablish Wexford’s deliberate indifference, Dean contends in Nos. 20-3058 & 20-3139 41

the alternative that other evidence at trial supported a finding of deliberate indifference. Specifically, Dean points to the tes- timony of Nurse Lisa Mincey. But Nurse Mincey testified that she complained about Dr. Nawoor’s conduct. Indeed, Dean points to her testimony again when defending the jury’s find- ing that Dr. Nawoor was deliberately indifferent. Nowhere in the cited testimony does Nurse Mincey fault Wexford. So, we disagree with Dean that other evidence at trial proved Wex- ford’s deliberate indifference. For its part, the dissent contends that Wexford’s knowledge can be inferred from the fact that there were ten collegial reviews in the 207 days between Dean’s initial presentation of symptoms and his surgery. But why? The question is not whether Wexford knew that Dean’s offsite care requests had to go through collegial review. The question is whether Wexford knew that collegial review would likely vi- olate Dean’s constitutional rights. We cannot infer such knowledge from the mere fact that Wexford applied collegial review in Dean’s case, even if it did so repeatedly. The dissent also relies on Dr. Barnett’s testimony, but as we have explained, Dr. Barnett testified about problems with Dean’s care. He did not testify that collegial review generally causes similar problems. Nor did he testify that Wexford knew of such problems. As evidence of Wexford’s knowledge, the dissent points to Dr. Barnett’s comment about his “under- standing that [Wexford is] aware their system is flawed, and that people are bringing it to their attention but they’re not- withstanding [sic] that attention.” But, as the dissent seems to recognize, this testimony was predicated on a counterfactual hypothetical. Dr. Barnett was testifying to what his opinion of collegial review would have been if Nurse Mincey had 42 Nos. 20-3058 & 20-3139

complained earlier about problems with Dean’s care. Dean cannot prove that Wexford was deliberately indifferent by changing the relevant facts and eliciting a hypothetical opin- ion based on those nonexistent facts. More broadly, the dissent suggests that we are invading the jury’s province and improperly reweighing the evidence. Not so. In this sufficiency of the evidence challenge, our role is to police the evidentiary boundary for Monell liability. As explained above, Dean introduced no evidence permitting a jury to conclude that Wexford knew in advance that collegial review would violate Dean’s constitutional rights in this sin- gle-incident case. Even though the jury instructions are un- challenged, we must ensure that the jury had a legally suffi- cient evidentiary basis for holding Wexford liable. Finally, the dissent claims that “twelve pages” of the Lip- pert reports are all that stands in the way of affirming the jury’s verdict on the Eighth Amendment claim against Wex- ford. But that is not true. We hold that the 2018 report was inadmissible and that the 2014 report—assuming it was ad- missible—was not enough to prove Wexford’s knowledge. The fundamental problem with Dean’s claim, however, is that he has no evidence of Wexford’s knowledge. Dean’s Monell claim fails because he lacks critical proof, not because he in- troduced the Lippert reports. C. Eighth Amendment Claims Against Doctor-Defendants The individual defendants, Drs. Nawoor and Einwohner, also seek judgment as a matter of law on the Eighth Amend- ment claims against them. Because the doctors are individuals rather than entities, the standards for holding them liable are more straightforward. To prevail against them, Dean had to Nos. 20-3058 & 20-3139 43

prove that (1) he had an objectively serious medical need (2) to which they were deliberately indifferent. Farmer, 511 U.S. at 834. We have established that Dean had an objectively seri- ous medical need, so we focus on the doctors’ mental states. Dean must show that the doctors acted with deliberate in- difference—that is, consciously disregarded a serious risk to his health. Petties v. Carter, 836 F.3d 722, 728 (7th Cir. 2016). Deliberate indifference requires more than negligence or even objective recklessness. Id. Dean “must provide evidence that an official actually knew of and disregarded a substantial risk of harm.” Id. Of course, medical professionals rarely admit that they deliberately opted against the best course of treat- ment. Id. So in many cases, deliberate indifference must be in- ferred from the propriety of their actions. Id. We have held that a jury can infer deliberate indifference when a treatment decision is “so far afield of accepted professional standards as to raise the inference that it was not actually based on a med- ical judgment.” Norfleet v. Webster, 439 F.3d 392, 396 (7th Cir. 2006). But where the evidence shows that a decision was based on medical judgment, a jury may not find deliberate in- difference, even if other professionals would have handled the situation differently. See id. (“[A] difference of opinion among physicians on how an inmate should be treated cannot support a finding of deliberate indifference.”); Petties, 836 F.3d at 729 (“[E]vidence that some medical professionals would have chosen a different course of treatment is insufficient to make out a constitutional claim.”). Even if Dean can show deliberate indifference, he must also demonstrate that the doctors personally caused a viola- tion of his constitutional rights. Walker, 940 F.3d at 964. In other words, Dean must show that “the defendant’s actions 44 Nos. 20-3058 & 20-3139

or inaction caused the delay in his treatment,” id., and that “the delay exacerbated the injury or unnecessarily prolonged pain,” Petties, 836 F.3d at 730–31. 1. Dr. Nawoor Dean’s claim against Dr. Nawoor centers on Dr. Nawoor’s decision to order an ultrasound. Dean contends that Dr. Na- woor knew the standard of care and consciously breached it by ordering an ultrasound rather than a CT scan. Deliberate indifference claims require proof of subjective knowledge, so it is worth recalling what Dr. Nawoor knew when he decided on the ultrasound. There is no dispute that Dr. Nawoor knew from the beginning that Dean might have kidney cancer. On December 23, 2015, when Dean first pre- sented with painless gross hematuria, Dr. Nawoor thought that either a recurrence of kidney stones or kidney cancer was to blame. Dr. Nawoor also knew the undisputed standard of care for treating painless gross hematuria: refer the patient to a urologist for a CT scan and cystoscopy. He knew, in addi- tion, that an ultrasound could not rule out the possibility of kidney cancer and that Dean would need to see a urologist eventually. Keeping this knowledge in mind, we must ask whether Dr. Nawoor’s decision to order an ultrasound reflected an ex- ercise of medical judgment or a complete abandonment thereof. If Dean can show that Dr. Nawoor chose an ultra- sound solely because it was cheaper or easier, then Dean has met his burden. See Petties, 836 F.3d at 730 (“If a prison doctor chooses an ‘easier and less efficacious treatment’ without ex- ercising professional judgment, such a decision can also con- stitute deliberate indifference.” (quoting Estelle, 429 U.S. at Nos. 20-3058 & 20-3139 45

104 n.10)). If, however, Dr. Nawoor’s decision to order an ul- trasound reflected his medical judgment, then he was not de- liberately indifferent as a matter of law. See id. The evidence at trial compels the conclusion that Dr. Na- woor’s decision to order an ultrasound was an exercise of medical judgment. Dr. Nawoor testified that he and Dr. Ritz decided on an ultrasound because Dean had a history of kid- ney stones (which an ultrasound could detect) and decreased kidney function (which increased the risk of a CT scan). He acknowledged the standard of care but explained that an ul- trasound made more sense for Dean under the circumstances. See Roe v. Elyea, 631 F.3d 843, 859 (7th Cir. 2011) (“[D]eliberate indifference claims based on medical treatment require refer- ence to the particularized circumstances of individual in- mates.”). After all, Dean had a recent history of kidney stones, and recent CT scans had not detected kidney cancer. Dean’s experts agreed that Dr. Nawoor’s decision to order an ultrasound was an exercise of medical judgment. On cross- examination, Dr. Metwalli conceded that “whether or not to order an ultrasound in a specific situation involves an exercise of clinical judgment.” This concession comports with the Su- preme Court’s acknowledgement that a choice among differ- ent types of diagnostic tests is a “classic example of a matter for medical judgment.” Estelle, 429 U.S. at 107. “A medical de- cision not to order an X-ray, or like measures, does not repre- sent cruel and unusual punishment. At most it is medical mal- practice.” Id.; accord Pyles v. Fahim, 771 F.3d 403, 411 (7th Cir. 2014). Of course, the evidence could in theory show that Dr. Na- woor’s decision to order an ultrasound was so out of bounds that it could not have reflected medical judgment on these 46 Nos. 20-3058 & 20-3139

facts. See Norfleet, 439 F.3d at 396. But Dean made no such showing. His experts offered only mild criticism of Dr. Na- woor’s decision to order an ultrasound. Dr. Metwalli testified that an ultrasound was “not the ideal test in this case” and a CT scan “should have been done earlier.” Dr. Dhar testified that Dr. Nawoor breached the standard of care by ordering an ultrasound. She agreed that an ultrasound would have been an appropriate first step if Dean had severely compromised kidney function, but in her view, he did not. Similarly, Dr. Barnett testified that “the standard of care was the CT scan” and the ultrasound “was less ideal.” Even then, he conceded that “arguably the renal ultrasound is not a terrible thing to do.” And he agreed that an “ultrasound can be an appropriate part of the workup for hematuria.” Viewing this testimony in the light most favorable to Dean, it reflects a difference of opinion among medical pro- fessionals, which cannot support a deliberate indifference claim. Norfleet, 439 F.3d at 396; Petties, 836 F.3d at 729. In ret- rospect, an immediate urology referral may have been more prudent than an ultrasound. “But this is just to reiterate the standard for medical malpractice, which falls short of deliber- ate indifference.” Duckworth v. Ahmad, 532 F.3d 675, 681 (7th Cir. 2008). Apart from the decision to order an ultrasound, Dean of- fers a list of other examples of Dr. Nawoor’s supposed delib- erate indifference. But all his citations are to Dr. Nawoor’s own testimony, where he admitted to not taking certain ac- tions while caring for Dean. Nowhere in the cited testimony did Dr. Nawoor testify that he consciously opted against what he knew was the best treatment. Dean cannot establish delib- erate indifference simply by citing to things that Dr. Nawoor Nos. 20-3058 & 20-3139 47

did not do. Without any evidence of his mental state, Dr. Na- woor’s inaction demonstrates, at most, negligence. Dr. Na- woor is therefore entitled to judgment as a matter of law on the Eighth Amendment claim against him. 2. Dr. Einwohner Dean’s case against Dr. Einwohner fails for different rea- sons. Even assuming that she was deliberately indifferent, Dean’s claim against her collapses on causation. Dean cannot show that Dr. Einwohner was responsible for any of the chal- lenged delays. Dr. Einwohner testified, without contradiction, that she lacked the power to change the course of his treat- ment. She worked in Pittsburgh and provided telemedicine support for the primary care doctors at Taylorville and other facilities. As a matter of Wexford policy, she was not invited to collegial reviews. Nor could she order a urology referral or CT scan on her own. The best she could do was make recom- mendations to the doctors who participated in collegial re- views—which she did. Immediately after seeing Dean on Jan- uary 7, 2016, she emailed Dr. Ritz suggesting additional im- aging and a urology evaluation. Dean concedes, of course, that her recommendation was on target. But he faults her for not following up about it, and more generally for failing to communicate with Drs. Nawoor and Ritz about his care. In other words, Dean’s claim against Dr. Einwohner rests on the assumption that, if she had followed up more, Dean would have seen a urologist sooner. But that is pure specula- tion. No evidence supports the inference that Dr. Einwohner could have changed the course of Dean’s treatment if she had been more persistent. Dean himself elicited testimony from her that she lacked the power to force Wexford to send him to a urologist in January 2016. Thus, the jury did not have a 48 Nos. 20-3058 & 20-3139

sufficient evidentiary basis for holding Dr. Einwohner liable for the challenged delays. See Walker, 940 F.3d at 966. The district court reasoned after trial that the doctor-de- fendants’ demeanors at trial supported an inference of delib- erate indifference. But a witness’s demeanor at trial is not sub- stantive evidence. A jury may consider a witness’s demeanor in deciding whether the credit the witness’s testimony. But the doctor-defendants’ demeanors while testifying shed no light on the substantive question of whether they acted with deliberate indifference while caring for Dean years earlier. So, we reject the district court’s reasoning. *** For these reasons, the defendants are entitled to judgment as a matter of law on the Eighth Amendment claims. In pass- ing, the defendants ask us to go further and hold that the im- proper admission of the 2018 report requires at least a new trial on the negligence-based claims. On one hand, we agree that the Lippert reports were significant to the institutional negligence claim against Wexford. But the defendants do not explain how they had any effect on the medical malpractice verdicts. And the unchallenged jury instructions required the jury to find Wexford negligent if it found that either of the doctor-defendants (i.e., Wexford’s agents) had engaged in medical malpractice. As such, we do not upset the jury’s find- ings of liability on the negligence-based claims. We need not reach the parties’ arguments about punitive damages. With the Eighth Amendment claims off the table, punitive damages are no longer available. See 735 ILCS § 5/2- 1115; Woodward v. Corr. Med. Servs. of Ill., Inc., 368 F.3d 917, 930 (7th Cir. 2004). Although we leave the negligence verdicts Nos. 20-3058 & 20-3139 49

intact, a second jury must decide compensatory damages anew. See Beard v. Wexford Health Sources, Inc., 900 F.3d 951, 954–55 (7th Cir. 2018). III. Conclusion Nothing we have said should be taken to cast doubt on the gravity of Dean’s pain and suffering. Dean has shown re- markable resilience in battling terminal cancer from prison. The jury found that the defendants were negligent in caring for Dean, and we do not disturb those findings. Deliberate in- difference is a high hurdle, however, and the evidence at trial did not surmount it. Accordingly, we reverse the judgment on the Eighth Amendment claims and remand for the district court to enter judgment as a matter of law for the defendants on those claims. We vacate the judgment on the negligence- based claims and remand for a new trial on those claims lim- ited to the issue of damages. 50 Nos. 20-3058 & 20-3139

WOOD, Circuit Judge, dissenting. Twelve pages. Twelve pages admitted into evidence subject to a careful limiting in- struction. That is the difference, according to the majority in this case, between allowing plaintiff William Dean to keep the $1 million in compensatory damages and $7 million in puni- tive damages that the jury awarded him (after a $3 million re- duction by the court), and overriding the jury’s judgment to take it away. The majority takes the position that the district court’s admission of those twelve pages from the so-called Lippert Reports requires this override. But it is not our role to reassess the evidence and second-guess the jury’s conclu- sions. Even if the court erred in admitting those twelve pages (and in my view it did not), they were not so prejudicial either by themselves or alongside the rest of the evidence to require this radical step. I therefore respectfully dissent. I While he was an inmate at Illinois’s Taylorville Correc- tional Center, Dean developed an aggressive, metastatic kid- ney cancer. This case involves the care he received—or more to the point, did not receive—at the hands of Wexford Health Sources, which provided the health services at Taylorville. The majority opinion contains an overview of the pertinent facts, but when our review of a judgment is governed by the standard of review that applies to a jury’s verdict, details mat- ter. As we noted years ago, “[a]ttacking a jury verdict is a hard row to hoe. … Our inquiry is limited to whether the evidence presented, combined with all reasonable inferences permissi- bly drawn therefrom, is sufficient to support the verdict when viewed in the light most favorable to the party against whom the motion is directed … .” Sheehan v. Donlen Corp., 173 F.3d 1039, 1043–44 (7th Cir. 1999) (cleaned up). Nos. 20-3058 & 20-3139 51

With this in mind, I first consider what Dean must prove in light of the fact that he brought an as-applied challenge un- der Monell v. New York City Department of Social Services, 436 U.S. 658 (1978). I then briefly review the critical evidence, be- fore turning to the two points on which the majority rests: the proposition that the Lippert Reports played a pivotal role in the trial; and the proposition that there was little evidence other than the Lippert Reports on which the jury might have relied in reaching its verdict in favor of Dean. Neither of these points holds up when one looks at the evidence that was in the record through the proper lens—that is to say, in the light most favorable to the verdict. Such a review shows that there was ample evidence on which the jury properly relied when it returned a verdict for Dean and carefully allocated the com- pensatory and punitive damages among the culpable defend- ants. Dean was not a healthy person. Even before his kidney cancer manifested itself, he suffered from heart disease, dia- betes, and kidney stones. In August 2014 and July 2015, he had CT scans performed, but at that time no one reading the re- sults detected cancer. In December 2015, however, Dean spot- ted blood in his urine. This prompted him to go to Taylor- ville’s medical clinic, where the medical director for the prison, Dr. Abdur Nawoor, gave him a physical examination and ordered testing of his urine and blood. The tests con- firmed the presence of blood in the urine (formally, hematu- ria). Expert testimony at trial indicated that painless hematu- ria—which is what Dean was experiencing—is associated more with cancer than with kidney stones. Although Dr. Na- woor himself did not admit that fact, he did immediately con- clude that one of two things lay behind Dean’s condition: 52 Nos. 20-3058 & 20-3139

either Dean had a recurrence of his kidney stones, or he had cancer. Approximately two weeks later, Dean had a telemedicine appointment with Dr. Rebecca Einwohner, a nephrologist employed by Wexford in its Pittsburgh corporate office. She heard enough to cause her to email Wexford’s corporate di- rector for utilization management, Dr. Stephen Ritz, and sug- gest that Wexford conduct a “collegial review” about Dean’s case to see if another image of his kidneys should be taken. Wexford responded with a half-measure: Drs. Ritz and Nawoor met on January 13, 2016, and agreed that they would send Dean out for a renal ultrasound, which (according to ev- idence the jury was entitled to credit) was cheaper than a CT scan, but not as likely to detect cancer (one of Dr. Nawoor’s two tentative diagnoses). The ultrasound was performed on February 2 by a group called Precise Specialties, but the tech- nician erroneously saw no evidence of cancer. (All remaining dates from this point were in the year 2016.) On February 10 Drs. Ritz and Nawoor recapitulated their “collegial review,” after which Dr. Nawoor told Dean that he needed a CT scan. The wheels turned slowly, however. At no time did Wexford signal that Dean’s case demanded emergency measures, or even urgent care. Around the first of March, however, a nurse called the urology specialist, Dr. William Severino, to see if he might see Dean sooner than March 21. That call apparently prompted some action: Dean saw Dr. Severino on March 10, and the doctor ordered a CT scan and cystoscopy (i.e. an ex- amination of the inside of his bladder and urethra). Given Wexford’s system, however, there was no direct re- sponse to Dr. Severino’s order, because as I noted, no one ap- pears to have regarded Dean’s condition as requiring urgent Nos. 20-3058 & 20-3139 53

attention. Instead, Drs. Ritz and Nawoor held yet another “collegial review” session on March 22. Eight days later they approved Dr. Severino’s proposal, and another two weeks later, on April 12, Dean finally got the CT scan he had been waiting for since February 10. It revealed that he had a serious case of kidney cancer. Dr. Severino recommended that Dean have the affected kidney surgically removed (i.e. a nephrec- tomy). But again, his recommendation was not enough in it- self. More delays ensued as Wexford again subjected the spe- cialist’s advice to “collegial review.” Ultimately (on April 21) Wexford approved the nephrec- tomy. Over the next three months, from April 22 to July 19, Dean underwent a number of pre-surgical tests. These tests revealed that his kidney cancer had metastasized to his liver and beyond, up the vena cava (the largest vein in the body) and past his diaphragm. By the time he had his surgery on July 19–207 days after his first appointment with Dr. Nawoor and cancer was first suspected–it was an extraordinarily lengthy and risky procedure. A few months after the opera- tion, oncologist Dr. Guaglianone confirmed that Dean still had metastatic cancer of the liver, for which he received chem- otherapy drugs. We were informed at oral argument that Dean is still alive. II Before considering the evidence, it is crucial to be clear about what Dean, having brought an as-applied claim under Monell v. New York City Department of Social Services, 436 U.S. 658 (1978), must prove. It was not Dean’s burden to litigate on behalf of others; this was not a class action, and he did not have to prove that Wexford’s policy always led to catastrophic results. He had to show only that Wexford’s unwavering 54 Nos. 20-3058 & 20-3139

policy requiring collegial review amounted to deliberate in- difference to his condition. My colleagues have effectively collapsed the critical dis- tinction between the existence of a policy and the effects of that policy by insisting, at every turn and for each of Monell’s elements, that Dean demonstrate a prior pattern of constitu- tional harm wrought by the collegial-review process. Monell requires no such thing. “Pattern or practice” evidence of a problem or failure is necessary in as-applied challenges only when the presence of an official policy, custom, or practice— Monell’s threshold question—is in doubt. In those cases, pat- tern evidence substitutes an inference from a long-standing practice for the certainty of a written policy. See Glisson v. In- diana Department of Corrections, 849 F.3d 372, 382 (7th Cir. 2017) (“The critical question under Monell remains this: is the action about which the plaintiff is complaining one of the in- stitution itself, or is it merely one undertaken by a subordinate actor.”). In Dean’s case, collegial review is an explicit, official policy followed by Wexford; everyone, including my col- leagues, readily recognizes this fact. No one denies that Dean was attacking Wexford’s own policy and actions; he was not making a subterranean vicarious liability argument that would not be cognizable under Monell, nor was he claiming that Wexford dealt with his case pursuant to an implied policy or custom separate from collegial review. A quick overview of some of the key Monell precedents demonstrates the properly circumscribed role of pattern evi- dence. In Board of County Commissioners of Bryan County v. Brown, 520 U.S. 397 (1997), a case upon which the majority heavily relies, the Supreme Court considered whether a sher- iff’s “inadequate screening” of an officer with a prior criminal Nos. 20-3058 & 20-3139 55

record rendered the county liable for the officer’s use of exces- sive force. For obvious reasons, the county did not have an explicit policy commanding officers to use excessive force, nor did it have an explicit policy requiring the sheriff to eschew a close review of future officers’ criminal records. The question was therefore whether the county had an implied policy of not reviewing criminal history that reflected the county’s con- scious choice. The Court determined that the isolated hiring decision, in the absence of a more extensive pattern of similar past incidents, did not support such an inference. The “failure to train” claim in City of Canton v. Harris, 489 U.S. 378 (1989), also involved a case where no official policy existed. There, the plaintiff alleged that the city systematically failed to train officers in medical-response techniques, and the Court asked “[u]nder what circumstances can inadequate training be found to be a ‘policy’ that is actionable under § 1983?” Id. at 383. The Court remanded the case for a new trial without detailing the type of proof needed to support a find- ing of an implied policy, though Justice O’Connor separately opined that the lack of pattern evidence in the record would make it difficult to show “that the city had an unwritten cus- tom of denying medical care to emotionally ill detainees.” Id. at 398 (O’Connor, J., concurring in part and dissenting in part). Justice O’Connor drew upon Tuttle, which like Brown asked whether an instance of excessive force was executed pursuant to an unwritten custom or policy. See id. at 400-01 (“To infer the existence of a city policy from the isolated mis- conduct of a single, low-level officer, and then to hold the city liable on the basis of that policy, would amount to permitting precisely the theory of strict respondeat superior liability re- jected in Monell.” (quoting Oklahoma City v. Tuttle, 471 U.S. 56 Nos. 20-3058 & 20-3139

808, 831 (1985) (Brennan, J., concurring in part and concurring in judgment))). This is not to say that pattern evidence has no bearing on Monell’s additional elements requiring notice and “moving- force” causation. In some contexts, particularly when unwrit- ten customs or practices are being challenged, the question whether the municipality has recognized an official policy of- ten logically overlaps with these elements, such that pattern evidence provides a clear route to proving each. But this does not mean that pattern evidence is required when, as here, the official policy is not in doubt and the notice and causation re- quirements are analytically separable. Because Dean’s case in- volves an official policy—collegial review—there is nothing more that needs to be said on that point. I thus move on to the notice and causation elements. In situations such as those presented in Brown, Canton, or Tuttle, the existence of the policy and the municipality’s knowledge of the implied policy’s risks are two sides of the same coin. The pattern of deficiency shows both the existence of an implicit policy and the municipality’s awareness of that policy. Cf. Jackson v. Marion County, 66 F.3d 151, 152 (7th Cir. 1995) (“The usual way in which an unconstitutional policy is inferred, in the absence of direct evidence, is by showing a series of bad acts and inviting the court to infer from them that the policymaking level of government was bound to have noticed what was going on and by failing to do anything must have encouraged and at least condoned, thus in either event adopting, the misconduct of subordinate officers.”). But in Dean’s situation, where the official policy of collegial review is firmly established, the question of notice—was Wexford aware of collegial review’s risks—is all that is left. I do not Nos. 20-3058 & 20-3139 57

doubt that one way of showing notice would be through pat- tern evidence—with each additional delay caused by collegial review, Wexford would have been more likely to realize the policy’s risks. But awareness can also be proven more di- rectly. For instance, a public report from a respected author- ity, such as the court-appointed experts who prepared the Lip- pert Reports or the Department of Justice, would without doubt grab a municipal entity’s attention too. Monell’s “moving-force” causation inquiry asks whether the policy itself, as opposed to a negligent act of an officer out- side the policy or some other intervening cause, precipitated the constitutional injury. Using pattern evidence to prove the cause of a single instance is something that can be done only with care, and only with close attention to the facts of the in- cident in question. Direct evidence of the cause of the single incident will always suffice; pattern evidence is not essential. When the Supreme Court referred to causation in Brown, it noted only that “the existence of a pattern of tortious conduct by inadequately trained employees may tend to show that the lack of proper training … is the ‘moving force’ behind the plaintiff’s injury.” Brown, 520 U.S. at 407–08 (emphasis added). That is what makes evidence of past events relevant, though not necessary, in Dean’s case. Because collegial review is an established policy of Wex- ford, Dean must show only that Wexford was aware of colle- gial review’s risks of harmful delays and that it was collegial review, not individual-officer negligence or some other inter- vening cause, that lay behind the deliberate indifference to the urgency of Dean’s medical need. The majority takes the posi- tion that Dean failed on both those scores; I do not agree. 58 Nos. 20-3058 & 20-3139

Indeed, in my view, as I now explain, Dean has satisfied both elements with or without the Lippert Reports. III Before turning to the content and impact of the two Lippert Reports, we must look at the evidence wholly apart from them. This independent evidence was enough, by itself, to support a finding that Wexford’s policy of collegial review was, as applied to Dean, intentionally applied, structurally defective, and the cause of the harm he suffered (both in the form of much more serious illness and in the form of greatly increased risk associated with treatment). The strength of this evidence is critical to our evaluation of the use and impact of the excerpts from the Lippert Reports permitted by the trial judge. Contrary to Wexford’s insistence in its submissions to this court, this limited use of the Lippert Reports did not taint all the rest of the evidence. They were far from the only sup- port for the jury’s decision. Whether it was error to admit them, and whether, if there was error, it was harmless, can be assessed only by taking the full context into account. Many witnesses testified that the delays in Dean’s treat- ment, as outlined above, were caused by Wexford’s collegial- review process. And they did not mince words. Some of this testimony focuses on the application of the policy to Dean: in his case, they said, collegial review caused the harm, even though in theory it might foster patient care in other settings. Other testimony was not so cabined—some witnesses said that collegial review suffers from general defects by its very nature. Dr. Barnett, who served as an expert witness for Dean, provided the most compelling evidence that in Dean’s case, collegial review failed as a system. Here are some highlights: Nos. 20-3058 & 20-3139 59

• [I]f I could put it simply, [the collegial review pro- cess] is dangerous as applied. [The process] … has … allowances for urgent and emergency situa- tions where the medical director at the prison … would be able to act immediately; … [b]ut as it’s been applied in this case, it never happened. • [T]he collegial review process … only acted to harm the patient and keep him from getting the care that he needed. • [Responding to the question whether anything ex- plained the delay between Dean’s first presentation with hematuria and the ultrasound]: Just the awk- wardness of the collegial review process. And oth- erwise, there was nothing intervening. • [Responding to a question about his opinion of the operation of the process in Dean’s case]: Danger- ous. • The process, as I observed it, did not promote proper health care. … [I]t caused [Dean] to continue to suffer with the conditions that he had been treated for. These excerpts demonstrate a close causal relationship be- tween the scheme and Dean’s injury (or so the jury was enti- tled to infer), and there is much more to the same effect. In addition, and contrary to the majority’s characterization that Dr. Barnett confined his analysis to Dean’s case, his testi- mony swept more broadly. First registering his own past ex- periences—“I understand the process. I’ve been involved in it intimately.”—Dr. Barnett then likened Wexford’s process to a system where an ambulance with a gunshot victim woodenly 60 Nos. 20-3058 & 20-3139

stops at every traffic light, instead of activating its sirens and flashing lights and rushing to the hospital. He goes on to link Dean’s particular injury to general problems with the system: • [T]he opportunity for collegial review to do good was missed over and over again. So I think the rea- son why is because this collegial review process at Wexford is broken. It’s—to my way of interpreting the data that I’ve been looking at. Traffic lights are fine, normally, but there has to be a way of overriding them. Dr. Kosierowski, who was a medical oncologist serving as an expert witness for the defense, admitted that there was no reason “other than the collegial process” that prevented Dean from getting a referral to a urologist the very same day when he first saw Dr. Nawoor (December 23, 2015)—the day when, as noted earlier, Dr. Nawoor first recognized that Dean’s symptoms may have been the result of either cancer or kidney stones. Collegial review was also the only reason Dr. Kosierowski could think of for choosing the less effective ul- trasound, in the face of those competing possibilities, rather than the more effective (and more expensive) CT scan. Overall, the timeline established by independent trial evi- dence showed that more than half of the delay Dean experi- enced prior to his operation was directly attributable to colle- gial review. Even Dr. Nawoor, who claims to have acted at all times pursuant to the requirements of Wexford’s collegial-re- view program, indicated that it was collegial review, not his own actions, that was to blame for Dean’s enhanced illness and less-than-successful surgical outcome, which left in place the metastatic cancer in his liver. Nos. 20-3058 & 20-3139 61

And there was still more. One way to think of Dean’s case was as a cascading series of negative incidents, as a result of which earlier harms put later events in context and supplied notice of the urgency of his problem. There were ten collegial reviews over the 207 days that elapsed between Dean’s first contact with Dr. Nawoor and his surgery. The jury was enti- tled to find (as it did) that at some point along the way, Wex- ford became aware that its repeated use of that procedure amounted to deliberate indifference to Dean’s medical condi- tion. My colleagues suggest that the jury could not infer from the repeated use of collegial review alone that Wexford ever knew collegial review was causing delays, and that these de- lays added up to deliberate indifference to Dean’s medical needs, in violation of the Eighth Amendment. But upon each of its reviews, Wexford learned about the absence of mean- ingful medical intervention for a case of metastatic cancer, which rapidly becomes worse over time. This is critical, in light of the fact that the constitutional deprivation that Dean alleged is the deliberate delay in diagnosis and treatment, full stop. The jury was not compelled to accept the majority’s po- sition that Wexford had no idea that accumulated delays in care were endangering Dean’s health every day. And there was additional evidence outside the Lippert Re- ports that put Wexford on notice of the harm that collegial re- view was imposing on Dean. For example, Nurse Lisa Mincey attempted to intervene to move the process along, but to no avail. Dean’s counsel asked Dr. Barnett what his assessment of the Wexford system would be if Nurse Mincey had been pushing the Wexford doctors to do more during the relevant time (December to July). He replied that if that were the case (and other evidence indicated that she actually was trying to do so), “it reinforces my understanding that they [i.e. 62 Nos. 20-3058 & 20-3139

Wexford] are aware their system is flawed, and that people are bringing it to their attention but they’re [sic] notwithstand- ing that attention.” Nurse Mincey confirmed that there were problems with ordering and scheduling, and that Wexford was the entity in charge of those functions. She reported those problems to Dr. Matticks (Dr. Nawoor’s superior), and she alerted the Wexford officials to the problems in Dean’s case. My colleagues write off this understanding as resting exclu- sively on Dr. Barnett’s counterfactual reasoning, but they miss its relation to Nurse Mincey’s testimony. Together, this evi- dence supports a reasonable inference that Wexford was in fact notified of delays in Dean’s care. The jury was entitled to credit this and the other testimony I noted above, and this is enough to satisfy Monell’s notice and causation requirements. It is of no moment at this stage in the proceedings that there was also testimony to the contrary. Jury trials are not elections, and so the victory does not neces- sarily go to the side that has presented the greater number of witnesses to testify on its behalf. See, e.g., Weiler v. United States, 323 U.S. 606, 608 (1945) (“Our system of justice rests on the general assumption that the truth is not to be determined merely by the number of witnesses on each side of a contro- versy.”); Pennsylvania R.R. Co. v. Chamberlain, 288 U.S. 333, 338 (1933) (“It, of course, is true, generally, that where there is a direct conflict of testimony upon a matter of fact, the question must be left to the jury to determine, without regard to the number of witnesses upon either side.”). See also Federal Civil Jury Instructions of the Seventh Circuit, Instruction 1.17 (2017 rev.) (“Number of Witnesses: You may find the testi- mony of one witness or a few witnesses more persuasive than the testimony of a larger number. You need not accept the tes- timony of the larger number of witnesses.”). The only Nos. 20-3058 & 20-3139 63

remaining question is whether the two Lippert Reports, dating from 2014 and 2018, had such a poisonous effect on the trial that the only remedy is to set aside the jury’s verdict entirely on Dean’s Eighth Amendment argument—not even giving him a new trial on that point!—and remit him to his negli- gence claims. As I will now show, neither Lippert Report had such an effect. IV A Before turning to the substance of the reports, it is worth a quick look at the question whether Wexford properly pre- served the objections it is now making to the use of those ma- terials at trial. In a motion in limine before trial, Wexford’s law- yer referred once to Federal Rule of Evidence 403, the rule that applies when concededly relevant evidence may be so preju- dicial that it ought not to be admitted. This is what counsel said: “The reports’ probative value is substantially out- weighed by the danger of unfair prejudice to Defendants, and this evidence should be barred under Federal Rule of Evi- dence 403. Based upon the foregoing legal precedent, it is ev- idence that the Lippert Reports are inadmissible, and would unfairly prejudice the Defendants.” Seventh Motion in Limine (Entry 137) at 5. (Counsel briefly repeated this point, without a citation, in his Reply to Dean’s Response to the Seventh Mo- tion in Limine.) But counsel did not reiterate his Rule 403 ob- jection at trial—at least, not in so many words. Instead, the following exchange took place: Q (Dean’s Counsel): With respect to the collegial review process, you are familiar with the 64 Nos. 20-3058 & 20-3139

Shansky report in the Lippert versus Godinez litigation; correct? A (Wexford’s Counsel): I’ll object to the rele- vance. A (Mincy’s counsel): I’ll join. THE COURT: The objection is overruled. Transcript at 872–73. There are a few other scattered spots where Wexford’s counsel either refers the district court to the motion in limine or objects to admissibility on other grounds, including rele- vance and the propriety of questions directed to Dr. Ritz. My colleagues think that these vague references were enough to preserve an objection under Rule 403, but I am dubious. To state the obvious, a rule that begins with the words “[t]he court may exclude relevant evidence if its probative value is substantially outweighed” by various problems, does not ad- dress the admission of irrelevant evidence. Evidence Rules 401 and 402 perform the latter function: Rule 401 describes what is relevant, and Rule 402 says that “[r]elevant evidence is ad- missible” unless an authoritative source provides otherwise. Wexford has been represented by experienced counsel throughout these proceedings—counsel who know the differ- ence between a relevance objection and an objection that re- quires the balancing described in Rule 403. As Evidence Rule 102(b) permits and is commonly the case, the court’s pretrial rulings on the motion in limine were subject to reconsidera- tion. I would be prepared to find that counsel’s brief mentions of Rule 403 did not suffice to preserve this point. On that basis, the most straightforward path for affirming the judgment is Nos. 20-3058 & 20-3139 65

to find that Wexford failed to preserve its Rule 403 objection with the necessary specificity. B If Wexford did not forfeit its Rule 403 objection, however, we would turn to the merits. There, too, the record reveals no reversible error. Applying the proper deferential standard of review to the court’s decisions about the admissibility of evi- dence, I begin with the 2018 Report before moving to the 2014 Report. In admitting the 2018 Report, the district court did not make the foolish mistake of thinking that something written in 2018 was capable, on its own, of providing notice to some- one acting in 2016 who knew about the 2014 Report. In put- ting this spin on the district court’s actions, my colleagues set up a straw man. That is not the purpose for which the six pages from the 2018 report were admitted. Instead, they came in because they addressed Wexford’s policies as of 2014, and in that sense, were backward-looking. It is as if an historian wrote a book in 1972 about the effect of the Vietnam War on the U.S. economy, and another wrote a book on the same topic in 2002. We would not say that the 2002 book has nothing to say about 1972, because it was written long after those effects had dissipated. Both books deal with the same period and the same topic, and so the later book legitimately might discuss what people knew at the time. The question is a factual one: what was known in 2014? And, contrary to the suggestion made by the majority, the later document (here, the 2018 Re- port) does not depend on an after-the-fact assessment of the earlier facts. No one is saying that Wexford should have real- ized in 2014 what its system would look like to someone in 2018. 66 Nos. 20-3058 & 20-3139

The experienced district judge instead admitted the ex- cerpts from the 2018 Report not for the truth of its contents (i.e. whether systemic deficiencies necessarily existed), but be- cause it shed light on Wexford’s contemporaneous knowledge of how its system may, or may not have been, working. Here is the actual instruction: You may consider these reports only in deciding whether Wexford Health Sources, Inc. had no- tice and knowledge of the information in the re- ports, not whether the information in the re- ports is true. In other words, if the jury learned from the 2018 Report that systemic delays were repeatedly alleged between 2014 and 2018, it was entitled to consider that information as it evalu- ated Wexford’s knowledge in 2014. That is precisely how the 2018 Report was used, and I find no error in the judge’s deci- sion to permit this. The events evaluated in the 2018 Report took place before Wexford encountered Dean’s case, and thus the jury could conclude that the 2018 Report supported a find- ing that Wexford was fully aware of the relevant problems by the time Dean’s problem came along. Another way of understanding the role played by the 2018 Report is that it helped contextualize the 2014 Report, which remained the touchstone of analysis for purposes of notice. If, for instance, Wexford had put forth evidence from 2019 that the 2014 Report was methodologically defective and its con- clusions no longer trusted, or that it reflected a fringe view of Wexford’s system at the time, this would obviously have un- dermined Dean’s use of the 2014 Report—and no one would doubt that evidence’s relevance. In making clear that the 2014 Report has since been regarded as a recognized authority, the Nos. 20-3058 & 20-3139 67

2018 Report supported the salience of the 2014 Report’s alle- gations between 2014 and 2016. My colleagues additionally suggest that, even if the 2018 Report alleged earlier problems with collegial review, it can- not itself demonstrate that Wexford had knowledge of these problems. I take this point also to apply to the 2014 Report. But there is ample transcript evidence that Wexford was aware of the issues chronicled within the Reports, which each were ordered by courts to assist with litigation involving Wexford. Consider the following points in Dr. Ritz’s testi- mony: • I’m familiar with the existence of [the 2014 Report]; yes. • Yes, I’m familiar with [the 2018 Report’s] existence. • [In response to the question: “and you and other Wexford executives are aware of those reports; cor- rect?”] Yes. • I was aware that there were several conclusions, opinions made by the reviewing physicians. I did not read the reports in fine detail. We were made aware by the DOC, Department of Corrections, that there were concerns that were raised in these re- ports; yes. • We would take [the concerns raised in the Reports] seriously; yes. Dr. Ritz then represented that he did not “completely” agree with some of the Reports’ “characterization[s]” of collegial re- view’s “purpose and intention”: 68 Nos. 20-3058 & 20-3139

I don’t agree with all of the contents of these re- ports, no. I think that the reports don’t com- pletely characterize and mischaracterize, not necessarily willingly, the purpose and intention of the collegial review process. So no, I don’t agree with all of the components of them. But we are certainly aware of them; yes. In other words, Dr. Ritz believed Wexford’s underlying “in- tentions” went underappreciated. But he never disavowed, or even cast doubt upon, the Reports’ concerns regarding the ac- tual functioning of collegial review. As my colleagues ob- serve, Wexford was free to contest the Reports or its knowledge of the associated allegations at trial, but it failed to do so. In turn, the jury was entitled to conclude from the Re- ports and this contextualizing testimony that Wexford was fully aware of—and indeed believed to be serious—the risk of harmful delays in care posed by collegial review. Wexford’s knowledge that serious health risks attended collegial review is more than sufficient for satisfying the Mo- nell notice requirement. We have never suggested that a gov- ernment entity must know with something close to certainty that the application of a policy will cause a constitutional vio- lation—such an extreme view would foreclose Monell liability for facially lawful policies which, by definition, can in theory be applied lawfully. In fact, all that must be shown is that Wexford knew of collegial review’s potential patient-safety hazards at the time it applied the policy to Dean, and here the evidence is overwhelming. My colleagues seek to impose a new condition on Monell plaintiffs such as Dean: in addition to demonstrating that Wexford knew of possible problems with collegial review, Nos. 20-3058 & 20-3139 69

they contend that Dean needed to provide substantive proof that those problems in fact existed. Framing this as a question of first impression, the majority acknowledges that we have never previously recognized such a condition. But even if this requirement should be created, Dean satisfied it through Dr. Barnett’s testimony, which drew upon his past experiences with and observations of collegial review to conclude that the system suffered from general defects. If problems are built into the fabric of collegial review, it follows that they neces- sarily existed in some form, however inchoate, prior to Dean’s case, and the jury could reasonably have reached this conclu- sion. Likewise, Dr. Ritz’s own testimony in response to the Reports—“I don’t agree with all of the components of [the Re- ports]. But we are certainly aware of them; yes.”—could be construed as having validated the existence of longstanding problems with the system. It is therefore not the case that Dean put forth no substantive evidence for the noticed prob- lems, as my colleagues insist. Substantive proof that problems materialized is not an in- dependent requirement for notice. As I previously noted, it is true that evidence of earlier problems makes it more likely that a municipal entity has learned that its policy is defective. In other words, this kind of evidence is a rough proxy for knowledge, and it may be helpful at the margins if a defend- ant disputes whether it was aware of a policy’s potential risks. But Wexford never contested the Reports’ proof of notice, and so this additional evidence is unnecessary. Moreover, even if Wexford had contested the Reports, the jury would have been within its rights to credit them, and thus make a finding of notice, without additional substantiating evidence. Unless the majority has some advanced problem in epistemology in mind—namely, that Wexford was falsely under the belief that 70 Nos. 20-3058 & 20-3139

its system had risks—I see no logical reason for this require- ment. The majority draws this requirement from Daniel v. Cook County, 833 F.3d 728 (7th Cir. 2016), where the plaintiff had produced both a Department of Justice report (which, like the Lippert Reports, was admitted by the district court only for no- tice) as well as additional evidence substantiating the prob- lems documented in the report. But my colleagues fundamen- tally misread Daniel. Daniel concerned a Monell challenge to Cook County’s informal customs and practices, not to any ex- plicit policy it had. See id. at 733 (“Plaintiff Daniel claims that Cook County and its Sheriff violated their duties under the Due Process Clause by acting with deliberate indifference to- ward his serious health needs as the result of inadequate cus- toms and practices.”) (emphasis added); id. at 736 (concluding that “it would be reasonable … to infer an official custom, pol- icy, or practice”). As I have stressed, the need for pattern evi- dence to prove a policy’s existence arises only when there is no written policy. In Daniel, separate proof was necessary not for notice or for “moving-force” causation (which was dealt with later in the opinion), but to infer an official policy. See id. at 734–35 (“To prove an official policy, custom, or practice within the meaning of Monell, Daniel must show more than the deficiencies specific to his own experience, of course. … If Daniel meets this mark, he must then show that a policymaker or official knew about these deficiencies and failed to correct them.” (citations omitted) (emphasis added)). Because Dean is challenging an explicit policy, either Lippert Report stand- ing alone would be sufficient to demonstrate that Wexford was on notice of collegial review’s risks. Nos. 20-3058 & 20-3139 71

Having dealt with the 2018 Report and the knowledge concerns animating both Reports, I now turn to the 2014 Re- port. My colleagues contend that the collegial-review process itself changed in 2016, when (they say) Wexford articulated an explicit exception for emergencies. It follows, they argue, that the 2014 Report was incapable of alerting Wexford to the pervasive flaws in the collegial-review system that applied to Dean in 2016. But they are mistaken on the underlying facts. Understanding why this is so requires some additional back- ground into collegial review’s structure and a closer assess- ment of what Wexford’s own witness, Dr. Ritz, testified to at trial. In 2016, Wexford modified its collegial-review manual. (It made this modification in the course of its internal annual re- view, as it notes on page 2 of the manual; we do not know how often it undertook such reviews.) Two changes were dis- cussed at trial: first, the fact that the 2016 manual includes an express exception for “emergent” and “urgent” cases, and second, that the name of the manual was changed from “Uti- lization Management Policies and Procedures” to “Utilization Management Guidelines.” As Dr. Ritz described in his testi- mony, the collegial process is circumvented in cases of emer- gency because “[e]mergent or emergency would be some- thing that you have to call 9-1-1 about” and “[o]bviously, we’re not going to make somebody call to get approval for that.” Transcript at 778–79. An “urgent” case “wouldn’t be 9- 1-1, but it really can’t wait for the phone call, the collegial call next week.” Id. Instead, as Dr. Ritz noted earlier in the same exchange, urgent cases “can be reviewed usually the same day.” Id. at 708–09. These are exceptions to the regular collegial review process that continued to apply to the vast majority of cases. 72 Nos. 20-3058 & 20-3139

It is important in this connection to keep a few points in mind. First, the question whether there was a material conti- nuity in policy is a factual one embedded within the jury’s notice finding. It is not a separate threshold inquiry subject to a lower standard of review, or some all-or-nothing determi- nation that denies all evidence of known risks arising before a certain date. Municipal entities modify internal and external policy language all the time in ways large and small, and it cannot be that any such change inevitably brings a “new” pol- icy into existence. The question is, instead, whether a policy has been materially altered with respect to the particular point at issue. The jury would have been within its right to find that any purported policy change in 2016 was partially, though not completely, responsive to the many risks and recommen- dations regarding delays in care reflected in the 2014 Report, and it could accordingly have reduced without entirely viti- ating the 2014 Report’s proof of notice. Second, I question the majority’s suggestion that we should look away from Wex- ford’s actual practices for purposes of this determination. How else are we to measure whether modifications to a cor- porate manual amount to a material break? The majority ap- pears to draw this view from the fact that plaintiffs can bring Monell claims against either an express policy or, in the ab- sence of an express policy, an informal practice. But it does not follow that evidence about the application of an express policy can no longer be considered in evaluating the policy’s contents or continuity. Whether the 2016 exception for emergent or urgent cases was new in substance as compared with the 2014 version of the manual was a question as to which the jury could have made reasonable inferences in either direction. All Dr. Ritz agreed to is that the 2014 document lacks an “emergency or Nos. 20-3058 & 20-3139 73

urgent provision for collegial review.” Transcript at 715 (em- phasis added). But neither his nor anyone else’s testimony connects the presence of this single provision to a new Wex- ford practice, or otherwise makes any comparison between Wexford’s practices in the pre- and post-2016 periods. For in- stance, nobody suggests that urgent or emergent cases were dealt with more efficiently, or otherwise any differently, after 2016. Neither does anyone indicate that collegial review was unavoidable in extreme cases before 2016. This is unsurpris- ing, as it is difficult to believe that a prisoner who crashed to the floor from a heart attack or stroke in 2015 would have been forced to wait through weeks of collegial review before going to the hospital. And indeed, a closer inspection of the 2014 manual shows that it, too, provided for emergent and urgent situations. As with the 2016 manual, Section UM-002 of the 2014 manual is titled “Emergency/Hospital Notification,” and sets forth a workaround to collegial review “to ensure timely notifica- tions of all emergent off-site care” including “ER, hospitaliza- tion, urgent office, urgent procedures, and radiology.” (Empha- sis added.) This is not to say the two manual versions track one another word-for-word, but the substance on this critical point appears to be roughly equivalent. The 2016 version of Section UM-004, which governs the standard collegial review process, is more detailed than Section UM-004 in the 2014 ver- sion. In one new provision, which I take to be the collegial review “more urgent cases”/“emergency” exception referred to by Dr. Ritz at trial and now by my colleagues, the 2016 man- ual at page 10 stipulates that “in more urgent cases an imme- diate Collegial Review should be undertaken by the Site Med- ical Director.” That provision then links to another Section (UM-005), which at page 12 directs the Site Medical Director 74 Nos. 20-3058 & 20-3139

to submit either a “Referral Request form marked urgent or an Emergency/Hospitalization Notification form.” But the 2014 Request form, included at page 24 of the 2014 manual, is nearly identical to the Request form included at page 37 of the 2016 manual. Both include an “Urgent Yes/No” option for the Site Medical Director to select. And the Emer- gency/Hospitalization Notification forms, included at page 19 of the 2014 manual and page 26 of the 2016 manual, are carbon copies. In other words, and contrary to the majority’s sugges- tion that I am looking at the wrong portions of the manual, the collegial review “urgent cases” provision cross-refer- ences, and operates through, other sections of the manual that have gone unchanged since 2014. A doctor seeking urgent re- view after 2016 would have been directed by the manual to use the same mechanisms—either a Request form marked “urgent” or an Emergency/Hospitalization Notification form—that a doctor in 2014 would have been directed to use. Therefore, even if the jury were somehow obligated to assume a perfect congruence between a corporate manual and a cor- poration’s practice (which, of course, it was not), it rationally could have concluded that the differences between the 2014 and the 2016 versions were immaterial. And that is exactly the assessment Dr. Ritz reached. Ultimately, the notion that only starting in 2016 did Wexford recognize the possibility of emer- gent and urgent cases requiring immediate or same-day care is unsupported. Even if, contrary to fact, we were to assume that the 2016 version represents a substantive change in practice for urgent or emergent situations, that does nothing to undermine Dean’s case. What matters here—or so the jury was entitled to find—is that the default collegial review process, to which Nos. 20-3058 & 20-3139 75

Dean was subjected, went unchanged. On this point, Dr. Ritz repeatedly suggested that the shift in label from “policy” to “guidelines” was nominal rather than a real alteration to the process. The jury was entitled to understand his testimony this way notwithstanding the fact that Dean’s counsel, per- haps in an effort to show an acknowledgment of the flaws in the earlier system, suggested at trial that the change might have meant something. In a series of exchanges, which I re- produce in full in the Appendix to this dissent, Dean’s counsel asks for the reason for the language change. Dr. Ritz’s re- sponses include the following: • The guideline verbiage, I think, reflects more the way things function in the real world. (Appendix A) • [The name of the document changed] I think to bet- ter reflect a real-world process. (Appendix B) • The term guideline just better reflected what the purpose of the document was. (Appendix B) Dr. Robert Matticks, the Lead Illinois Regional Medical Direc- tor for Wexford, was also asked about the purported change, to which he responded: • I really can’t speak to that. I was not involved in the genesis and review of this particular document. … [W]hat I’m pointing out is that I was not involved in the genesis of these documents or if they were changed throughout, you know, the years, except for perhaps for reasons of clarification of what the guidelines should be. (Appendix C) None of the expert testimony, Dr. Ritz’s included, suggests that collegial review under the pre-2016 “policy” was 76 Nos. 20-3058 & 20-3139

bureaucratically stymied in a way that was later cured in the post-2016 “guidelines.” Both versions had some degree of flex- ibility. But, as the 2014 Report revealed to Wexford, this resid- ual flexibility was not enough to keep the default collegial- review process from causing harmful delays in care, as it did when the process was applied to Dean. An example may help to illustrate the key point. Suppose a health care company has known for some time that Medi- cine A is ineffective. It creates Medicine B to be used in ex- treme circumstances. But it leaves Medicine A in place, un- changed (aside from an amendment to the label), as the com- mon treatment. The creation of Medicine B does not alter Medicine A nor does it negate the company’s knowledge of Medicine A’s defects. The same is true in our situation. First, as I noted earlier, the policy in effect in 2014 also provided for urgent and emergent situations. Second, the clarification of that exception in 2016 did not transform collegial review’s core practices. The default practice remained exactly the same, and so the 2014 Report continued to put Wexford on notice of the unsatisfactory way in which that default process operated. Perhaps if there were some evidence that Dean’s case had been regarded at the time as emergent or urgent under the 2016 standards and was thus handled outside the standard process, the majority’s argument against the notice value of the 2014 Report might have some purchase. But there is not a hint that at any time between December 2015 and July 2016 Wexford regarded Dean’s case as something requiring imme- diate action. Nothing suggests that Dean’s diagnostic tests and eventual surgery caused Wexford to think that he pre- sented a “9-1-1” (i.e. emergency) or same-day (i.e. urgent) Nos. 20-3058 & 20-3139 77

situation. Indeed, my colleagues make this point when invok- ing Dr. Severino’s testimony that the surgery was not an emer- gency. Instead, Dean was caught up in Wexford’s longstand- ing and primary procedural arrangement. All of this is to say that the jury was entitled to conclude that Wexford made no material changes to the policy in the wake of the 2014 report. Dr. Ritz resisted the conclusion that the 2016 report represented a departure from its predecessor. When Dean’s counsel directed him to the line in the manual regarding “more urgent cases,” he acknowledged that the text was new. But he then insisted repeatedly that the language was changed only to approximate collegial review’s actual operation better. In turn, the change in labels from “policy” to “guidelines” was inconsequential for Dean’s purposes, and the additional manual language regarding an emergency/ur- gency exception was just a cosmetic change (or so the jury per- missibly could have inferred from the testimony). If, to the contrary, there was a genuine change, it was limited to the emergency/urgency exception, which was of no relevance to Dean’s case. Given those facts, the excerpts from the 2014 Re- port were admissible for the narrow purpose identified by the district court: notice. For all these reasons, I would find that the district court did not abuse its discretion by admitting carefully selected ex- cerpts from both the 2014 and the 2018 Lippert Reports. The record of the trial as a whole provides no basis for thinking that the jury was bowled over by these modest submissions, either when they were presented at trial or when counsel al- luded to them during closing arguments. * * * 78 Nos. 20-3058 & 20-3139

Suppose, however, that I am wrong about this, and it was such a serious mistake to admit these 12 pages into evidence that it amounted to an abuse of discretion. That is only the beginning of the analysis, not the end. Federal Rule of Evi- dence 103 directs that “[a] party may claim error in a ruling to admit or exclude evidence only if the error affects a substan- tial right of the party and [a proper objection or offer of proof was made].” FED. R. EVID. 103(a). And if that were not enough, Federal Rule of Civil Procedure 61 also stipulates that “[u]nless justice requires otherwise, no error in admitting or excluding evidence … is ground for granting a new trial, for set- ting aside a verdict, or for vacating, modifying, or otherwise disturbing a judgment or order.” (Emphasis added.) As I will now show, on this record the only rational conclusion is that any error with respect to the Lippert Reports was harmless. In considering the impact of the two reports, it is critical to bear in mind, as I have stressed throughout, that the judge admitted only brief excerpts into evidence. Dean accom- plished this through Dr. Ritz’s testimony. Only six non- blank/non-cover pages of the 2014 report (PTX-194) were ad- mitted into evidence; the full report spans 46 pages. Similarly, only six non-blank/non-cover pages of the 2018 report (PTX-193) were admitted into evidence; that report is 150 pages long. Neither the content nor the volume of the admit- ted materials amounted to the kind of bombshell that was ca- pable of overwhelming the jury. As Dean said in his Response in Opposition to Defendants’ Motion to Supplement the Rec- ord on Appeal: Plaintiff only ever moved 24 heavily redacted pages into evidence. Of these 24 pages, 12 were either report cover pages or altogether blank Nos. 20-3058 & 20-3139 79

and included only for the sake of consecutive pagination. Many of the remaining pages had well over half of their contents blanked out en- tirely. Nor is it the case that these 12 pages, combined, played an outsized role in the closing arguments. True, Dean’s counsel referred to the admitted excerpts briefly during his closing ar- gument. The transcript of his closing argument runs for 53 pages, but references to the Lippert Reports appear in only two of those pages, and those references were not inflammatory. Counsel put PTX-194 (pages from the 2014 Report) back up on the screen for the jury. He reminded the jury that the report was prepared for a different case, and that Wexford contested the conclusion in that report. He went on as follows: And Shansky [the author]—by the way this re- port spans hundreds of pages, but we’re just re- ducing a couple of snippets for you. They find breakdowns in almost every area starting with delays in the identification and need for offsite services. Delays in obtaining authorization numbers, delays in being able to schedule an ap- pointment, delays in obtaining offsite paper- work, and delays in the absence of any follow- up visits with the patient. Sounds pretty famil- iar, right? Wexford knows what its collegial pro- cess does. With respect to the 2018 Report, counsel reminded the jury, Dr. Puisis came to the same conclusions about the way that collegial review was operating during the period covered by the 2014 report. His final reference was to the 2014 Report; he concluded that Wexford had “been on notice of this kind of 80 Nos. 20-3058 & 20-3139

conduct for a very long time, since at least 2014, and they ha- ven’t changed. They need to be punished.” Before we could find that this modest use of these limited materials was prejudicial to Wexford, we would have to be prepared to say both that this information was not cumulative and that there was something in it likely to sway a properly instructed jury. Neither conclusion is supported by the rec- ord. As I indicated at the outset, the conclusion drawn in these pages had already been articulated by other expert witnesses for the jury. The Lippert evidence was thus cumulative, and there was nothing special about the fact that an expert pre- pared it. There is also no reason to think that the jury was im- properly influenced in its verdict. To the contrary, the district court carefully instructed the jury about the limited use to which it was allowed to put the reports: You have heard evidence about reports filed in a different case regarding the delivery of health care to inmates in the Illinois Department of Corrections. Defendant Wexford Health Sources, Inc. disputes the truth of those reports and has not admitted liability in that case. You may consider these reports only in deciding whether Wexford Health Sources, Inc. had no- tice and knowledge of the information in the re- ports, not whether the information in the re- ports is true. Remember, the issue is whether defendants violated plaintiff’s rights as I de- scribe those rights to you in these instructions. Nothing in this record persuades me that this is the rare case in which we set aside our normal rule under which we as- sume the jury follows the court’s instructions. Nos. 20-3058 & 20-3139 81

V The jury had many paths to finding in favor of Dean. Set- ting aside the Lippert Reports, it could have found causation for Monell purposes from Dr. Barnett’s and Dr. Kosierowski’s testimony, which described how there was no explanation other than collegial review for the harmful delays in care. The notice element could have been satisfied by viewing Dean’s saga as a series of individual failures that increasingly re- vealed the shortcomings of Wexford’s rigid system; over ten collegial reviews, Wexford directly learned of the lack of sig- nificant medical intervention and the arc of Dean’s cancer’s progression, yet still did not act efficiently or effectively. Al- ternatively, the notice element could have been satisfied by Nurse Mincey’s testimony in combination with Dr. Barnett’s. Though unnecessary to the jury’s ultimate determination, the Lippert Reports—whether the two are taken together or the 2014 Report is taken alone—shore up this conclusion. The 2016 update to the manual does not reflect a relevant, material break. The jury easily could have seen the additional manual language on which my colleagues rely as elaborative, as Dr. Ritz said. Even if this were not the case, Dean was treated only pursuant to the default collegial review process; his case was never expedited as an urgent or emergency matter, and so any purported material change in policy was simply irrelevant to him. I therefore conclude where I began: this was a vigorously fought jury trial. The jury certainly would have been within its rights to find in favor of Wexford, but it found the evidence on Dean’s side to be more persuasive. I would affirm its ver- dict across the board. I therefore respectfully dissent from the 82 Nos. 20-3058 & 20-3139

majority’s decision setting aside the jury’s verdict on Dean’s Eighth Amendment count. Nos. 20-3058 & 20-3139 83

Appendix A Exchange Between Dean’s Counsel and Dr. Ritz: Q. So in the—under this—this one, by the way, for some reason, the 2014 document is entitled utilization management policies and proce- dures, and then in 2016 it changes to utilization management guidelines. But this—so under this policy and procedure, the referral request form should have any and all supporting documen- tation attached to it; correct? A. That's what this states, yes. Q. Right. And the reason—I assume the reason for that is if you're going to do a collegial review or going to do any review, you want to be able to look at, you know, a referral request form and then look at what the supporting documents are, whatever is necessary to look at; right? A. The expectation is that as much information is submitted with the referral as possible to help support the referral and to help guide the deci- sion and the conversation that would result from the referral request. Sometimes the infor- mation, things that are listed here, sometimes it may not be available. For example, we may not have all the diagnostic reports or consultation reports that we might want. And I would say that’s one of the reasons why, as it came up 84 Nos. 20-3058 & 20-3139

before, why this was changed from a policy and procedures which tends to imply that this must—every single component of it must be done every single time as versus a guideline. The guideline verbiage, I think, reflects more the way things function in the real world. Transcript at 716–17.

B Exchange Between Dean’s Counsel and Dr. Ritz: Q. So then with respect to a little bit of cleanup here. With regard to the policy and guideline that we were talking about earlier. Do you have an understanding—you talked a little bit—let me say it like this: You talked a little bit about how it went from a policy to a guideline because the policy—policy sounds like it’s a must and the guideline sounds like it’s a should, if you will. Is that your understanding as to why the policy changed? A. In general, that’s what has changed the name of the document, yes. And I think to better re- flect a real-world process. Q. Were there any other—did it change as a re- sult of a lawsuit? A. Not that I’m aware of. Q. You actually didn’t participate in changing the policy; correct? A. No. Nos. 20-3058 & 20-3139 85

Q. That was Dr. Lehman, who’s your boss? A. That’s correct. Q. Okay. So you don’t actually know the reason for the change from policy to guideline? A. I don’t know specifically, but in general, that’s what the discussion we talked about whether it was medical policies, UM policies, rather than have it being a concrete policy doc- ument, it states that is the guideline. The term guideline just better reflected what the purpose of the document was. Transcript at 739–40.

C Exchange Between Dean’s Counsel and Dr. Matticks: Q. Now, when we got into this little line of ques- tioning, I had asked you do you have any un- derstanding as to why the policy changed. And as you correctly pointed out, it went from policy to guideline. So I’ll ask this to you a different, more accurate way, Doctor. Do you have any understanding as to why the policy that was re- flected in the—in PTX103 changed as reflected in PTX102 on January 14th, 2016? A. No, I really can’t speak to that. I was not in- volved in the genesis and review of this partic- ular document. That was done at the corporate level. 86 Nos. 20-3058 & 20-3139

Q. So those—in preparing for all of the testi- mony that you've done in 24 depositions or more or your trial testimony in other cases, you have no understanding as to why these policies changed? A. No. I have—what I'm pointing out is that I was not involved in the genesis of these docu- ments or if they were changed throughout, you know, the years, except for perhaps for reasons of clarification of what the guidelines should be.” Transcript at 657–58.