Kerr’s study: 59 dying patients and the dreams and visions they described every day

Step 1

From 1926 to 2011 · The question

For almost a century, what the dying saw was told by other people

People have long said that some of the dying see their dead before they die. In 1926, the physicist William Barrett collected accounts of this kind in a book. Later, in the 1960s and 1970s, the researchers Karlis Osis and Erlendur Haraldsson asked hundreds of doctors and nurses what they had seen in their patients.

Almost everything that was known came from relatives, doctors, and nurses. They told what the patient had said to them, sometimes long afterward. Hardly anyone had asked the patients themselves while they could still answer.

A hospice doctor decided to ask them.

Step 2

Before the study · The author

A doctor at a hospice in Buffalo wants to convince his colleagues

Christopher Kerr was born in Toronto, Canada. He is a doctor with a PhD in neurobiology, the study of the nervous system. In 2020, when he published a book about these experiences, he was the chief executive officer and the chief medical officer of the hospice in Buffalo.

In his book, he says that a veteran nurse at his hospice used to tell him about the patients’ dreams. Many doctors took them for hallucinations caused by drugs, fever, or confusion. Kerr also says he wanted to gather data to convince his colleagues that these experiences mattered to patients.

To do that, he prepared a few simple questions.

Step 3

January 2011 to July 2012 · How it was done · As the study tells it

Every day, a researcher asks the inpatients seven questions

The team invited the patients staying in the hospice’s unit. They had to be adults, able to give their consent, and spend almost all day in bed or in a chair. Patients with a serious mental illness such as schizophrenia were left out, and so were those who couldn’t communicate. In all, 66 patients agreed to take part.

A researcher visited each patient every day until the patient died or went home. The researcher asked whether the patient had dreams or visions, whether they happened asleep or awake, who the patient saw, and how many times a day. The researcher also asked whether they seemed real, whether they were comforting or distressing, and whether the people the patient saw were going somewhere. The researcher marked the answers on a form. Then the researcher asked the patient to describe it in their own words.

On 90 occasions, the patient didn’t respond and couldn’t be interviewed. Patients also weren’t asked when talking made them suffer or when confusion kept them from answering. Almost half of the patients had been admitted precisely because they were confused. The study didn’t use a test to measure whether each patient was confused on the day they answered.

The team carried out 453 interviews.

Step 4

What the patients described · As the study tells it

Of 59 patients, 52 describe dreams or visions, mostly of relatives and friends who had died

Of the 66 patients, the team studied 59. Six were still alive when the study ended. Another patient had answered only once, many months before dying. The average age was 75, and two in three were women. Six in ten had cancer, and 28 of the 59 had been admitted because they were confused. The team spoke with each patient an average of almost eight times.

Of the 59, 52 described at least one dream or vision. Almost half of these experiences happened during sleep. About 16 in 100 happened while the patient was awake, and about 39 in 100 happened sometimes asleep and sometimes awake. Almost all seemed real to the patients. They said so for 267 of the 269 dreams or visions they were asked about.

What they saw most often were relatives or friends who had died. They appeared in 46 in 100 accounts. The living appeared in 17 in 100, and strangers in 10. In almost 4 in 10 accounts, the people the patients saw were going somewhere or getting ready to leave.

The article gives examples with changed names. A man aged 88 dreamed that he was driving toward a place he didn’t know. He was comforted to hear his mother tell him that everything was fine and that she loved him. A woman aged 64 dreamed of the dog she had as a child. She stopped dreaming of it when her health improved. Two patients said that someone in a dream told them when they would die, and they died six to seven days later. The article doesn’t say how many announcements like this did not come true.

The team also measured how the patients felt about these experiences.

Step 5

What the patients described · As the study tells it

Most of the dreams and visions comfort, and those with the dead comfort the most

The patients rated each dream or vision from 1 to 5. A 1 was very distressing and a 5 was very comforting. Out of every 100, 60 comforted the patients, 19 distressed them, and 21 made them feel neither comforted nor distressed.

Those that comforted the most were the ones with the dead, with an average of 4.1. Those with the living came to 3.2, and those with other people and things, to 2.9.

Some dreams were distressing. One woman dreamed of a fire. A man aged 82 dreamed that he was trying to fix the post of a fence that was crumbling, and he couldn’t. A man aged 88 who had fought in the Normandy landings saw dead soldiers around him.

The team looked at whether they changed as death approached. Looked at day by day, no change showed up. Grouped by weeks, one did. In the last weeks, there were more comforting dreams of the dead.

With these data, the authors drew their conclusions.

Step 6

2014 · The conclusion · As the study tells it

The authors conclude that these experiences are common and ask that they not be treated as confusion

According to the authors, these dreams and visions are common as death approaches, are felt as real, and carry a great deal of meaning for the patient. They write that they can be a great source of comfort. They warn that treating them with drugs as if they were confusion may take away something that helps the patient.

They also write that many patients had these dreams and spells of confusion at the same time. For them, the two are different things. Confusion muddles thinking and is often frightening, while these dreams are remembered well and bring peace. They take this difference from other studies. This study did not measure it.

The authors acknowledge the limits. Many patients already had these dreams before they entered the study, so it couldn’t be seen when they began. The team also couldn’t invite everyone who met the conditions. The authors ask that future studies control confusion better.

There is one problem the article doesn’t explain. Some figures in the final section don’t match those in the results. For example, the final section says 72 in 100 for meetings with loved ones who had died, and the results say 46 in 100. It may be that some figures are counted per patient and others per dream, but the article doesn’t make this clear. On this page, we use the figures from the results.

The same team kept studying these experiences.

Step 7

From 2015 to 2026 · What came next

Kerr’s team measures confusion in its next study and takes its work into a book and interviews

In 2015, the team published what the patients in the same study had said in their own words. It sorted the accounts into six types, such as getting ready to leave or loved ones who are waiting.

From 2013 to 2015, it carried out another study, with 83 patients cared for at home. This time, it gave a confusion test at each visit. It is a short test used in hospitals to find out whether a patient is confused. In 2020, the team published what those patients described. Relatives who had died kept appearing, often at a different age or in better health than when they died.

In 2020, Kerr described these experiences in a book, Death Is But a Dream. His work was also featured in a documentary. The book has no new data. It repeats the 2014 figures and tells the cases of some of his patients. Kerr writes that, in most of the patients who got better, the dreams stopped. In interviews, Kerr has given higher figures for comfort than the article does, such as 80 or 88%. The article gives 60%.

In 2025, Kerr and two colleagues reviewed the studies on these experiences. They write that these experiences suggest the mind and the brain might separate at death. They ask for studies with comparison groups and brain tests to find out.

Other teams carried out similar studies outside Buffalo.

Step 8

From 2012 to 2026 · Outside Buffalo

In India, Brazil, and England, the figures are very different

In 2016, the doctor Abhijit Kanti Dam published a study of 60 patients cared for at home in India. He asked them questions similar to Kerr’s, once and then again two days later. Of the 60, 38 described dreams or visions. Of these, 30 saw relatives, friends, or acquaintances who had died. For 32 of the 38, they were distressing. Dam also found that the dreams appeared when the patient was feeling worse physically. When the team eased the pain or nausea, the dreams became less frequent or went away.

In 2026, a team in Brazil used the same questions as Kerr with 85 patients at three hospices. Its first author, Taís Oliveira da Silva, also works at Kerr’s hospice. This time, a confusion test was given. Of the 85, 29 described these experiences. In half, relatives or friends who had died appeared. Out of every 100 experiences, 36 were comforting and 27 were distressing. According to the test, 78 in 100 of the experiences were described by patients who were not confused. The authors think the share came out lower than in Buffalo because they could talk with each patient far fewer times, fewer than two on average. In Buffalo, Kerr’s team spoke with each patient almost eight times.

In 2012, the English palliative care doctor Averil Fountain and the sociologist Allan Kellehear compared the figures from several studies. Visions of this kind appeared in fewer than 10 in 100 patients in two studies done at hospices in England. In India and in Moldova, families reported many more. They propose two causes for the difference. One is morphine and similar drugs, widely used in England and little used in the other two countries. The other is the religious culture of each place.

That leaves the question we started with: how many patients have these dreams and visions, what do they see, and how do they feel about it? Let’s look at what is known and what isn’t.