In April 2018, my wife of thirty-eight years died from complications of ovarian cancer. During her three-year ordeal, it seemed that every twist and turn of the disease could be predicted by a lab value or framed in terms of a statistical probability. Right up until the end. In her last days, it was not blood work or vital signs that foretold her death. It was a dream. If that sounds strange, it is because our society mostly shuns death and consequently knows little about dying.

At the time, I certainly knew little about the process of dying. Of course, I knew that the treatments Lisa was receiving were not working, but her doctors never spoke of how she would die or how close to the end she was. Only when she entered hospice were we told that she probably had weeks to live, not months. Even then, however, there was no discussion of how she would likely die or how we would know that death was very near. Thus, when she was admitted overnight to an in-patient facility to get control of symptoms we were not able to address at home, I was unprepared for the conversation I had with the hospice physician the next morning.

 “Medically,” Dr. D said, “Lisa is much better. Her vital signs are strong, and she is not experiencing any nausea. This is the good news. The bad news,” he continued, “is that your wife called the nurses in the middle of the night to say that she saw her parents on a boat outside the window beckoning her to come. I know this may not make sense,” he went on, “but we see this repeatedly in our patients. When patients report a vision like this, they almost always die within a day or two. I’m so sorry.” My wife died a little more than twenty-four hours later.

I have spent a lot of time since my wife’s death trying to make sense of this paradox. In the high-tech, evidence-driven world of contemporary medicine, it was a dream that led a physician to conclude that my wife was dying. How was that possible?

To try to answer that question I researched end-of-life experiences among hospice patients, but initially found little scholarly literature. I talked to friends and family about Lisa’s experiences and heard plenty of anecdotes about near-death experiences but not much about end-of-life dreams. Then I discovered the work of Christopher Kerr, MD, PhD. Although the experience of pre-death dreams and visions predicting death is well known in hospice circles, Kerr is one of the only physicians in the country who has studied this phenomenon in depth. He undertook the study of end-of-life dreams and visions (ELDVs) because he repeatedly encountered patients whose dreams and visions were important to them and were frequently predictive of death. Yet, without research findings that validated the importance of such dreams, Kerr could not get physicians to take them seriously.

Ask hospice nurses about this phenomenon and they will likely tell you about the many, many patients for whom they have cared who were comforted by end-of-life experiences. Yet the phenomenon is not well known outside hospice circles, and even in hospices many physicians treat such experiences as forms of delirium and medicate patients when they report them.

Kerr is not surprised when physicians do not take ELDVs seriously, for he was once such a physician. He tells the story of caring for a patient early in his career. He believed the patient was not about to die and had instructed a nurse to start a course of IV fluids and antibiotics. The nurse responded skeptically about the need for fluids and antibiotics because, she said, the patient was dying. When Kerr asked her how she knew, she responded that the patient was seeing his deceased mother.

Kerr eventually decided to study the end-of-life experiences of his patients to see if he could empirically document the role of dreams and visions in the dying. In a series of research studies conducted over the past ten years, Kerr has generated a substantial body of evidence that ELDVs are common, therapeutically important, and frequently predictive of imminent death. Yet, despite the rigor of the research, it has been an uphill battle to get clinicians to take end-of-life experiences seriously. I asked Kerry Egan, a former hospice chaplain who has also written on her experiences with dying patients, about the reception Kerr’s work has received among physicians, and she was very direct: most of them remain skeptical if not dismissive. “It breaks my heart,” she said, “but it is not surprising.” The account in her own book, On Living, of the role of the hospice chaplain may explain why physicians resist Kerr’s findings. The role of the chaplain is not to preach or teach; it is instead to “create a space—a sacred time and place—in which people can look at the lives they’ve led and try to figure out what it all means to them.” End-of-life dreams, she told me, help patients do that; physicians typically do not.

Given how generally irreverent Kerr is, he would probably not use the language of sacred time and space, but Kerr knows from both his experience and his research that end-of-life dreams help the dying find meaning. He came to believe that even if physicians resisted this critical knowledge, patients would welcome it, and so he began to see his primary audience as patients and their families and not just his fellow physicians. For that reason, to get his message out, he turned from academic writing to storytelling. His first effort in this direction was a TEDx talk on the topic of whether death can be illuminating. The talk was wildly successful, and the media coverage of the talk led to a book contract for Death Is But a Dream. Because he is a clinician and not a writer, the publisher wanted him to work with a ghostwriter to produce a manuscript. By all accounts, the collaboration did not go well, and Kerr turned to a friend and literary scholar Carine M. Mardorossian for help. Mardorossian is a professor of English and global gender and sexuality studies at the University of Buffalo. Though she had not written creative nonfiction before collaborating with Kerr, she knew from her academic work the power of listening to the marginalized.

It was a commitment to telling Kerr’s patients’ stories that made the collaboration work. The previous ghostwriter who worked with Kerr wanted to highlight themes in the research and not the stories of the patients. Mardorossian understood that, like medicine, the humanities are often death-denying. Scholars write with ease about illness and its treatment, but they are not comfortable writing about death itself. She knew that in listening carefully to his patients and their experiences at the end of life, Kerr was rejecting the strict binary of life versus death. Dying, he seemed to say, was a way of living. To listen to the dying was to attend to a form of living that is frequently ignored.

When I asked Mardorossian what it was like to collaborate with Kerr, her admiration and respect for him were palpable. “Chris constantly changed anything I wrote for him in the first person from ‘I’ to ‘we.’ I had to fight for the first person, because Chris always wanted to give credit to others.” This makes it seem like Mardorossian wrote and Kerr revised, but Mardorossian is clear that this was not the case. She learned as much about writing from Kerr as he did from her. “It was like he had a (narrative) chessboard in his head,” Mardorossian said. “I would see the next move in the section we were working on, and Chris would see the next five moves almost immediately.”

Mardorossian is particularly fond of a chapter about love in Death Is But a Dream. It tells the stories of some of Kerr’s elderly patients whose love confounds our culture’s focus on romantic love as only for the young. One of Kerr’s patients, Benny, suffered from what is commonly referred to as broken-heart syndrome, or in technical terms, stress-induced takotsubo cardiomyopathy. Eighty-seven years old and in good health when his wife, Gloria, died suddenly from an infection, Benny was inconsolable and visited his wife’s grave every day, sometimes several times a day. On Valentine’s Day, two months after Gloria died, Benny’s daughter found her father at the cemetery in subzero weather tracing an outline of a heart in the snow around Gloria’s tombstone. Benny’s health declined dramatically after his wife’s death, and this ill-advised trip to the cemetery accelerated the decline. Yet when he was admitted to hospice, Benny found a kind of peace, because almost immediately he began dreaming of Gloria. He could no longer visit his wife at the cemetery, but it was as if he visited her in his dreams, or so he insisted. Benny made it clear that when he dreamed, he was in his wife’s presence.

Kerr’s comment on this case helps explain why he is passionate about listening to the dreams and stories of his patients: there is wisdom at the end of life that is missed if one thinks dying is only about death and not also about life. “Old couples,” Kerr writes, “have much to teach us about true love. Their bond requires no big declarations, loyalty tests, or dramatic endings.… They continue to feel and believe in it even when the person through whom that love originated leaves them. For elderly patients especially, their love for their other half is who they are. Jobs, ambitions, hobbies, mortgages, and plans have come and gone. What is left and what matters is the relationships they have maintained, cherished, and tended to through a lifetime of small gestures and greetings, loving glances and humorous words, shared stories and forgiven faults.”

If Kerr’s academic research has not gotten much traction among clinicians, the more popular presentation of his work has struck a chord. The TEDx talk has had nearly five million views. The book that followed has sold extremely well and has been translated and distributed in ten other countries. A documentary based on the book has aired nationally on more than thirty PBS-affiliated television stations.

The more I read about Kerr’s work and watched videotapes of his interaction with patients, the more I wanted to meet the person himself. So I wrote to ask if we could meet via Zoom to discuss his work. During that meeting, I asked if I might come to Buffalo to interview him and some of his colleagues. Without hesitation, he agreed, and the next day sent me a list of seven colleagues who had also agreed to meet with me.

Although the stereotype of a hospice physician is that of a kindly grandfather, Kerr is sixty years old and radiates a kind of barely contained kinetic energy. In fact, knowing that he was originally from Canada, I couldn’t help but wonder if he was a former hockey player. When I asked, his characteristically self-deprecating response was: “No, but I guess my face does look like it has seen some pucks.” Except when he is at home on his working horse farm, you are likely to see him in an open-collar dress shirt and sport coat, with a stethoscope around his neck and his glasses pushed up on his head. But you get the sense that he wishes he could wear jeans, a t-shirt, and a baseball cap to work. His colleagues describe him as gruff, rebellious, someone who paints outside the lines, passionate, a fierce advocate for his patients, and a person with no tolerance for injustice.

Kerr’s mentor Robert Milch—a legendary figure in the history of Hospice and Palliative Care of Buffalo—used to tell the story that his father believed that a good surgeon needed three things: a sense of humility, a sense of humor, and an incision. The story might be adapted to say that a hospice physician needs three things as well: humility, humor, and a significant encounter with death. Kerr has the first two in abundance, and he traces his career trajectory to losing his father at the age of twelve.

Kerr is affable and engaging, and if you talk to one of his colleagues or friends, it won’t be long before you hear a favorite Chris Kerr story. One particular story I heard seemed to provide a measure of the man. It has to do with a patient named Ann Gadanyi, a devout Catholic who was dying after a twelve-year battle with breast cancer. Gadanyi was fifty-six years old when she entered in-patient hospice care around Thanksgiving. Her daughters, Juliana (seventeen) and Emily (twenty-five), pretty much took up residence with their mother in her room at Hospice Buffalo. As Christmas approached, Ann told Kerr that she wanted to spend Christmas Eve at home, go to Christmas Eve Mass to hear her daughter sing in the choir, and wake on Christmas morning at home.

Kerr explained that this plan would be difficult because Ann was receiving medication through IV lines to which she was tethered and without the IVs, pain and bleeding could be serious problems. However, he told her, if she had her heart set on this plan, they could probably make it happen. Kerr, who is neither Catholic nor particularly religious, concocted a plan that involved his drawing up syringes full of medications to take to church and home, accompanying Ann and her daughters to Mass, returning to their home afterward to give Ann the injections, and then returning later that day to get his patient back to hospice. Ann was so grateful for Kerr’s help that she called a reporter at the Buffalo News to ask if the reporter would do a story on Kerr. When the reporter asked Kerr why he had gone to such lengths to help, Kerr replied simply: “Sometimes the best medicine is not medicine.”

I like this way of putting the point, but it may be more accurate to say that sometimes the best medicine is not medicine as it has come to be practiced outside hospice. John Tangeman, the administrative medical director of Hospice Buffalo, believes that being a hospice physician is distinctive. “Being a hospice doctor is doctoring in its purest form. It is old-style doctoring, which involves listening both to the patient and to his or her body.” Hospice physicians rarely need to rely on technology; they can focus on the person before them. Listening and touching, Tangeman insists, are more important than technology.

Perhaps this is why Hospice and Palliative Care of Buffalo has such an extensive array of services that might initially strike some as non-medical. I spoke to Abby Unger, the director of expressive therapies at Hospice Buffalo, about her team’s work, which includes music, art, massage, and dance therapy. Prior to the pandemic, there was a twelve-member team that delivered a variety of techniques for engaging the bodies of patients. Unger told me that when these interventions are structured so that family members are present, everyone seems to benefit. As Unger puts it, when patients get a massage, family members seem to relax as if they, too, were getting a massage.

Unger makes another point that echoes what Tangeman and Kerr say about caring for hospice patients. “When people come to us,” she says, “their whole lives come.” This may be why home care is so important; home is often where patients’ families are, and caring for patients in their homes may, as Kerr puts it, be the best medicine. It is striking that Hospice Buffalo has twelve physicians who see patients in their homes.

I had come to Hospice Buffalo to better understand the phenomenon of pre-death dreams and visions and how they might be predictive of imminent death. What I discovered is that pre-death dreams and visions are so much more than predictors of death. I’m certain Kerr’s research group was glad to document that ELDVs foretell death, but that is not why they studied ELDVs. Instead, they were interested in these experiences for the same reason they were interested in the role of new medications for treating pain, fatigue, and depression. It is why they examined cost-saving measures with at-home hospice care or how music could help patients tell life stories or how different delivery systems for end-of-life medications might be more efficacious. In every case, the research was designed to demonstrate how a novel approach to dying patients could be shown to have better outcomes for patients and families. It is certainly what they found in studying pre-death dreams and visions.

Consider just some of the findings from the small mountain of data on end-of-life experiences Kerr’s research team has accumulated. In his studies, Kerr found that close to 90 percent of patients report having at least one dream or vision that could be classified as an end-of-life experience. These dreams are distinguished from regular dreams by being especially vivid. When asked to rate the degree of realism of such dreams, most rate them ten out of ten—the highest degree of realism. Patients often report that they are “more real than real.” They occur both during periods of sleep and periods of wakefulness, and they are easily distinguished from hallucinations or bouts of delirium.

Kerr has also tracked the content of the dreams, their frequency, their relationship to time of death, and the subjective significance that patients and their families placed in them. Here, too, the results were striking. Because Kerr began studying end-of-life dreams when patients first entered hospice and were not all imminently dying, he could monitor the dreams over many months. He found a predictable pattern. As patients approached death, their dreams increased in frequency and their content changed. Earlier in their time in hospice, patients reported dreams about living friends and relatives; as the patients approached death, the dreams were mostly filled with deceased family and friends. It was also clear that the dreams involving family members who had already died provided the most comfort. On a scale of 1 to 5, with 5 being the highest level of comfort, dreams with dead relatives and friends scored 4.08. Dreams with living friends and family provided much less comfort, scoring only 2.86.

Although Kerr’s research has focused primarily on the effects end-of-life dreams have on his patients’ acceptance of death, his research team has also examined whether these experiences affected the process of bereavement for loved ones. Here, too, he found a positive correlation between pre-death dreams and the ability to accept death and loss. Not surprisingly, the more at peace someone is with his or her own dying, the easier bereavement is for those left behind. Because end-of-life experiences often bring peace to those who are facing death, encouraging dying patients to talk about these experiences with their families is a way to care for those families and help facilitate the grieving process.

Kerr’s team is working on a “toolkit” on dreams and visions to give to patients and family members when a dying person enters hospice. The toolkit provides about as clear a statement of why such work is important as I have seen. “End-of-life Dreams and Visions (ELDVs),” the toolkit reads, “show that the experiences at the end of life can be full of dignity, strength, and grace. While many envision dying as a bleak or hopeless time, ELDVs reveal the heart of the human experience and re-contextualize death by emphasizing the connections between end-of-life experiences and the living.”

The evidence that Kerr and the research team at Hospice Buffalo has documented about the role ELDVs can play in providing meaning and comfort for individuals at the end of life is important and should be better known. But it is also important to note that not everyone has an end-of-life dream and not all such dreams are in fact comforting. Kerr’s team also emphasizes that their research does not seek to explain pre-death visions. They make it clear that their research focuses only on the experiences of patients and families in the face of the vivid dreams and visions they have documented. And most patients are not particularly concerned about explanations; it is enough that the experiences provide meaning and comfort.

In this regard, there is one finding of Kerr’s research that is somewhat surprising. Almost none of his patients had a dream or vision that was explicitly religious. I asked Kerry Egan whether that was true in her experience working with dying patients. She emphatically agreed with Kerr on this. “In all my years of listening to the dreams of my patients,” she said, “I never once heard a patient talk of seeing Jesus in a dream.” Yet, Egan believes that the dreams are nevertheless often profoundly spiritual. Although religious figures rarely appear in end-of-life dreams, deceased family members, particularly parents, are pervasive. And as Egan makes clear in On Living, when hospice chaplains spend time talking to patients about their families, they are having religious conversations. Talking about families, she writes, “is how we talk about God. This is how we talk about the meaning of our lives. That is how we talk about the big spiritual questions of human existence.”

I think this is why Egan says that the role of a hospice chaplain is to create a kind of sacred space in which people can examine their lives and try to make sense of them at the end. It seems to me that this is precisely what Kerr and his team are attempting to do when they ask patients to talk about their end-of-life dreams and visions and to share them with their families. To be sure, creating such a space requires treating the medical problems that may prevent patients from trying to make sense of their lives as they die, but it requires much more than that.

For this reason, I came to see the stethoscope that was always around Kerr’s neck less as a diagnostic tool than as a symbol of how he sees his role as a physician caring for the dying. He is at the bedside to listen to his patients—and not just to their heart and lungs. He is there to listen to the story of their lives. This is essentially what Egan said of hospice chaplains. They are with patients, she said, not as storytellers but as “story holders.” That’s not a bad way to think of the work of Christopher Kerr and his colleagues. They are a remarkable group of “story holders” who have discovered the power of end-of-life dreams to help the dying find meaning.

Learning about the therapeutic power of pre-death dreams has not answered my original question about how such dreams can predict death, but knowing that such dreams provide comfort to the dying is itself comforting. I like to think that my wife found joy in the vision she had of her parents shortly before she died. I also can’t help wondering whether, when the time comes, I will find myself in Lisa’s presence again. If the research of Kerr and his colleagues is any guide, the answer is likely yes.