(LOL Oscar from Lauren.)

While I expressed skepticism the other day regarding the media reports that a cat named Oscar could predict which patients at the nursing home in which he resides were within hours of death, some of you believed it, some even going so far as to speculate that not only could Oscar detect impending death but that he hangs out by the dying because he wants a snack.

But none have gone so far as Mighty Ponygirl in speculating about Oscar’s true motivation.

Personally, I like my explanation that it’s just confirmation bias better. It’s less–shall we say?–disturbing. I also tend to agree with Clark about the potentially pernicious effects that the belief among the nursing home staff in Oscar’s predictive ability could have on patient care.

Another pernicious effect is that the staff now call the families of patients to whom Oscar pays attention to tell them that they should come in because their loved one’s death is imminent. In other words, the staff is starting to base health care decisions (whether to call family or not) on the cat’s behavior. If that were my relative in the nursing, my first question would be: Why do you think death is imminent? If they told me it was because the cat was curling up with him or her, personally, I’d start looking for another nursing home.


  1. #1 sailor
    July 29, 2007

    Orac you kno of this one:
    Some dogs can apparently also detect siezures before they happen.
    Confirmation bias has to be the initial hypothesis with the doom kitty, but it would be interesting to test it and see if anything else is happening.

  2. #2 trrll
    July 29, 2007

    If they told me it was because the cat was curling up with him or her, personally, I’d start looking for another nursing home

    I think most people would do the same thing…unless, of course, they found that they no longer had a need for a nursing home.

    Which is why I’m skeptical of confirmation bias as an explanation. People tend to be very annoyed by false alarms of this nature, which makes it hard to ignore any “failures.”

  3. #3 Clark
    July 29, 2007

    Thanks for the mention Orac.


    Maybe they do have false alarms. Why would that be advertised by Dr. Dosa? Why would families feel the need to come out and tell the media? Why would the media care? They had their story and were clearly happy with it. Nobody wants to read or hear about a cat that can’t predict death.

  4. #4 nosugrefneb
    July 29, 2007

    Of course, an alternative hypothesis would be that he’s causing patients’ deaths…do any of these patients have severe cat allergies? 🙂

  5. #5 daedalus2u
    July 29, 2007

    When people die, they don’t die instantly, especially people dying of terminal degenerative diseases (not like trauma which can be quick). It is a slow death. People may be in a coma, the circulation starts to fail, some organs go hypoxic, necrotic even. A little at first, and then more as the heart and other organs go downhill. Extremeties may die and become necrotic while the core stays alive.

    As a carnivore, cats are going to be sensitive to the odors that dying things give off. Most likely they would be attracted to them.

    Maybe cats in general have this ability. Maybe this is why cats have historically been characterized as familiars of witches. When cats had access to dying people, they would sense death, they would show up to see what was going on, the person would finally die and the cat would get blamed.

    Lots of organisms can sense death. Quite a few different types of flies that lay eggs on carrion can, buzzards can, it wouldn’t be a surprise if cats could too. Maybe that is why they clean themselves off so much, to ensure they don’t have any little bits of dead flesh on them which would overwhelm the more subtle scent of something dead far off.

  6. #6 marion
    July 29, 2007

    nosugrefneb, given that they’re all in end-stage terminal dementia, their immune systems could only hope to be strong enough to react to something like cat dander.

    Not sure why the idea that an animal could identify the smell of a dying human and react to it is causing such powerful disbelief among so many people. As for looking for another nursing home, Orac…if the nursing home were wrong, then you’d be right to do so, but if the nursing home were right, would you complain that they had called you? And how would you feel if they called, you refused to believe that a cat could predict death, and someone you loved died without you being there? (If your answer is, “Well, if the someone had terminal dementia and had long stopped knowing who I and everyone else were, not bad,” then I certainly understand…but do keep in mind that many people feel otherwise.)

    I’m not a mystical person, but if I had a loved one in terminal dementia at a nursing home and that home told me that a certain animal had started reacting differently to him or her without some new diagnosis or medicine, I’d start thinking that SOMETHING was up. I also think that what most family members fear most acutely when they put someone they love in a nursing home, aside from any financial issues, is that that person will die without his or her family there. The people I know who have put aged loved ones in nursing homes who haven’t managed to make it to their deathbeds in time tend to feel very guilty. If you’re immune from this, I certainly understand, but when a staff is dealing with an entire population of people who are inches away from death, relatively speaking, it can be very difficult to know if someone is going to die TODAY.

    As for this affecting the quality of care…given that the deaths appear to happen within a few hours of Oscar’s “sign,” given that there are typically family members around for most of this time, and given that these patients are all in end-stage dementia when they come to this facility, exactly what do you think the medical professionals involved could be failing to do during these few hours that would keep these patients alive significantly longer? We are talking literally hours. Assuming that they don’t stop IVs/medications/treatments that the patients have been receiving and don’t actively seek to hasten death, I fail to see how their actions in a matter of *hours* could bring about the deaths of these people – who are all fairly close to death anyway – when said deaths wouldn’t have happened for a significant amount of time otherwise. And actually, if the staff does believe in Oscar’s “ability,” wouldn’t Oscar failing to “notice” a patient indicate to them that that patient isn’t “scheduled” to die yet, and thus needs more attentive care?

    If someone wants to claim that Oscar has some sort of mystical powers, I’m perfectly happy to argue with him or her. So far, the only claim I see is that Oscar is sensing and reacting to something the humans around him are not. I have cats, and I can tell you that they sense and react to things all of the time that I barely notice. I think the nursing home staff may be anthropomorphizing the cat a bit much, but I doubt they’d be making this much of a fuss over him if he weren’t really making their lives easier in some way.

  7. #7 Alan Kellogg
    July 29, 2007

    I know what Oscar is doing, he’s sitting Kadish. 🙂

  8. #8 Dianne
    July 30, 2007

    I have some doubts about the “Oscar just wants a snack and is tired of kibbles” hypothesis. Cats are predators, not scavengers. They don’t politely wait for their prey to expire before digging in, they start eating as soon as the prey can no longer fight back. If Oscar really wanted to eat Granny, he’d probably start as soon as she was comotose and no longer able to fight back. I’d go back to confirmation bias as the first hypothesis with acknowledgement that it is only a hypothesis and other hypotheses (change in care when the cat comes in, subtle smell associated with dying person, etc) may be correct.

  9. #9 Luis
    July 30, 2007

    Some followup, from an online discussion with Dr. Dosa

    I feel compelled to mention that bashing on the staff for putatively letting the cat dictate care (!) strikes me as a bit much. Not only is it a sad, soggy little ad hominem in the service of what seems like an overvalued idea, but it also assumes facts not in evidence regarding the professionalism and critical thinking skills of the staff.

    Salt Lake City, Utah: The article said that Oscar has presided over the deaths of more than 25 people. Is that a small or large percentage of the people who have died during the time period that Oscar has been there?[…]

    Dr. David M. Dosa: Oscar has been at each and every death that occurs in the facility in one capacity or another—there are a few occasions where family have not wanted him in the room. In those cases, he is never far away.
    Montgomery Village, Md.: Does Oscar ever spend any significant amount of time curled up with somebody who isn’t about to die?

    Dr. David M. Dosa: Not generally, no–He keeps to himself and avoids staff/ more robust residents on the unit[.]

  10. #10 Warren
    July 30, 2007

    Montgomery Village, Md.: Does Oscar ever spend any significant amount of time curled up with somebody who isn’t about to die?
    Dr. David M. Dosa: Not generally, no–He keeps to himself and avoids staff/ more robust residents on the unit[.]
    Posted by: Luis

    And that’s the money quote. Cats almost invariably prefer to curl up and sleep next to someone who isn’t moving around a lot, who won’t stir and disturb them, or shove them off, or possibly roll atop them. Anyone who’s ever owned a cat knows this.

    This is confirmation bias, clearly; and I share Orac’s concern that the staff of the nursing home are now skewing their patient care when they notice the cat overing around someone for a while.

    You’d think they’d know better. In an elderly-care facility like this one, the odds of the cat snuggling up to someone who is due to die in a few hours must be considerably higher than they are anywhere else; over time, in fact, his “accuracy” will approach 100%. At some point or another EVERYONE he sleeps near is going to die.

  11. #11 Luis
    July 30, 2007


    In an elderly-care facility like this one, the odds of the cat snuggling up to someone who is due to die in a few hours must be considerably higher than they are anywhere else;

    According to the story, it’s not as simple as that. The cat apparently AVOIDS people who are not about to die as well as snuggling up (“purring and nuzzling”) to those who are about to die. There appears to be a sharp discontinuity, not some sort of behavioral-shaping thing.

    This “confirmation bias” thing is starting to look like a bit of an overvalued idea. Interestingly, I can’t think of a way to disconfirm the confirmation-bias hypothesis that wouldn’t be argued around by someone or other here. Which is telling in itself.

    Empiricism doesn’t mean running away screaming from anything that looks like woo, or merely assuming that it IS woo. It means being willing to check out whether or not it is, and keeping an open mind as to one’s assumptions. Sadly, that doesn’t seem to be going on here.

    I don’t have any problem with people suggesting that confirmation bias may explain the Goodbye Kitty phenomenon. I do start to get itchy when that hypothesis is forwarded and thereafter assumed to be true, and more, the ONLY true explanation. I also get very itchy when something similar, but more pernicious, happens with respect to people’s notions about what the staff may or may not be doing. No one’s checked on those things, so, ideally, people who give a damn about empiricism ought to be critically minded about their own assumptions.

    But who am I fooling? This is the Internet.

  12. #12 Orac
    July 30, 2007

    My statement about what the medical staff is doing based on Oscar’s behavior is based on the reports about Oscar:

    Such is Oscar’s apparent accuracy – 25 consecutive cases so far – that nurses at the US home now warn family members to rush to a patient’s beside as soon as the cat takes up residence there.

    And, from the NEJM article itself:

    Making his way back up the hallway, Oscar arrives at Room 313. The door is open, and he proceeds inside. Mrs. K. is resting peacefully in her bed, her breathing steady but shallow. She is surrounded by photographs of her grandchildren and one from her wedding day. Despite these keepsakes, she is alone. Oscar jumps onto her bed and again sniffs the air. He pauses to consider the situation, and then turns around twice before curling up beside Mrs. K.

    One hour passes. Oscar waits. A nurse walks into the room to check on her patient. She pauses to note Oscar’s presence. Concerned, she hurriedly leaves the room and returns to her desk. She grabs Mrs. K.’s chart off the medical-records rack and begins to make phone calls.

    You’re the one who apparently hasn’t read up on this story, Luis. I stated nothing that wasn’t in the NEJM essay or the news story about Oscar. The staff is making medical decisions whether to call a patient’s family or not based on the behavior of the cat. It’s just that simple.

    As for your implication that I’m not willing to consider that the cat may be able to detect something, that’s just a load of B.S. I have said on more than one occasion, including the original post about Oscar that this could be a real phenomenon but that the anecdotes just aren’t enough to demonstrate whether it is:

    As I said, this looks like nothing more than confirmation bias. Of course, it is possible that the cat somehow senses something common to people who are within hours of death, but to determine whether this is indeed the case, at the very minimum, as Phil Plait points out, this is what we would need to know:

    Nicholas Dodman, who directs an animal behavioral clinic at the Tufts University Cummings School of Veterinary Medicine and has read Dosa’s article, said the only way to know is to carefully document how Oscar divides his time between the living and dying.

    And that is a question that has not been looked at systematically. It probably never will be, given that no one seems particularly interested in knowing whether Oscar is really detecting impending death or something related to impending death (such as more attention to the patient or simply, as Dodman put it in his comments, more warm blankets placed on dying patients), or whether this is just a classic case of confirmation bias.

    The bottom line: Without a lot more objective information, this looks a lot more like confirmation bias than any real phenomenon. Could it be a real phenomenon? Sure. I’ve said as much. However, it is up to the persons arguing that it is a real phenomenon to provide more than memories and anecdotes to back up their belief. Moreover, the NEJM should be ashamed for publishing such credulous tripe and thus putting its imprimatur on it. I doubt the press would have lapped it up the way it did if this had been published in a lesser medical journal.

  13. #13 Clark
    July 30, 2007

    Yes, like Prevention Magazine, People, Parade or anything that has ever been shoved under my windshield wiper.

  14. #14 Sam
    July 31, 2007

    No, Orac, it is not up to the persons arguing that it’s real to provide you with the evidence just because your skepticism requires you to doubt. Why do they owe this duty to you? They’re not asking you to buy anything! They have made an observation, they haven’t speculated wildly, and of course it could be similar to Hans the Talking Horse and even if valid there could be a confirmation bias effect.

    You say it’s confirmation bias. Now it’s you duty to go out and test your conjecture. Or explain how is it that it’s always Someone Else’s Problem to provide “convincing” evidence? That you are allowed to speculate but not other people? You poo-poo the story and offer nothing in return. Shades of the treatment of plate tectonics!

  15. #15 Orac
    July 31, 2007

    Nice try at shifting the burden of proof, but you’ve got it wrong.The person making an extraordinary claim (for example, that a cat has a flawless ability to tell when a person is within hours of death) has the duty to demonstrate his hypothesis, not the person pointing out one mundane possible explanation for the observed “phenomenon.”

  16. #16 vera
    July 31, 2007

    Hi Friends,

    I Find Absolutely FREE PlayBoy & Penthouse:


    If I find something else I’ll inform you.

    Best Regards,

  17. #17 Bronze Dog
    July 31, 2007

    The difference between the hypothesis of confirmation bias versus a death-detecting cat is equivalent to the difference between a hidden switch in your refrigerator door and a little gnome that turns out the light.

    If you’re going to claim that the fridge light goes out because of the gnome, you have to demonstrate the gnome. We already know that hidden switches exist in similar devices so it’s not far fetched at all for there to be one in fridges.

  18. #18 Oliver
    July 31, 2007

    Except we have no gnome here but an animal of which we know it has some senses surpassing those of a human by far. Contrary to Orac’s claim, we do not have an extraordinary claim here, especially since the sample size is relatively low. We also have no case of a “fridge gnome” here. This is not the case of “hidden switch” vs. “fridge gnome” since we know “similar devices” for both points -which was already pointed out but happily ignored by Bronze Dog. The hypothesis of confirmation bias is one that has the same requirements as any other hypothesis. It’s not a default. There are no “default hypotheses”. To act in such a fashion is patently unscientific. The proper answer to “there is no sufficient evidence that the cat can actually detect dying people” is NOT “so it HAS to be confirmation bias” but “we can’t tell at this point”.

    Now please stop acting as if cats are humans. Just because you can’t see a direct causative connectíon doesn’t mean there is any, and to act as if that’s the case, sorry, is a pretty good disqualification for scientific reasoning. I have yet to see a proton, but boy, am I conviced they exist.

  19. #19 Orac
    July 31, 2007

    Oh, please, give me a break with the straw men.

    I never denied that there is the possibility that the Death Kitty’s talent may be real. My first post on the subject said so explicitly. What I have said all along is that, absent a careful and unbiased study of how the cat spends his time, there isn’t even any credible evidence that there is a real phenomenon or ability here. Instead, all we have is anecdotes and memories, all of which are prone to confirmation bias, among other sources of confounding.

    Absent better (i.e., more objective) evidence, the most likely explanation for the cat’s apparent ability is confirmation bias. Since it is the kitty’s supporters who are making the claim that the cat can unerringly detect when a person is within hours of death, it is up to them to show at the very least that there is a real phenomenon here to study that is not due to other more likely explanations.

  20. #20 Oliver
    July 31, 2007

    Orac, repeating something that is false doesn’t make it right. Could it be that you are the victim of confirmation bias here?

    Science doesn’t have the “horror vacui” that you promote here. Science can say “As of now, I don’t have a f***ing idea what’s going on here”.

    Absent better (i.e., more objective) evidence, the most likely explanation for the cat’s apparent ability is confirmation bias

    False. Absent better evidence, there is no statement possible at all. The explanation of confirmation bias is a distinct hypothesis that needs to be tested just as much as any other explanation for the observation.

    Since it is the kitty’s supporters who are making the claim that the cat can unerringly detect when a person is within hours of death, it …

    Where’s the evidence for claims of “unerring”? So far, we have a very small sample size. We could find plenty of errors with more samples. But precisely because we have a small sample size, we cannot say anything significant either way.

  21. #21 Orac
    July 31, 2007

    Orac, repeating something that is false doesn’t make it right.

    Exactly my thought when I read your comment to which I responded, except that I was too nice to say it.

  22. #22 Buffybot
    August 1, 2007

    Could it be a combination of confirmation bias and learned behaviour on the part of the cat? If the moggy is being fussed over and otherwise rewarded for hanging with very sick and/or immobile people, then the behaviour will be reinforced.

  23. #23 vlad
    August 1, 2007

    I think there should be a test done to confirm if the cat is actually smelling something or not. I’s say but an RF collar on the cat and record his movements for a month. Unfortunately easier said than done as these collars are made of plastic which may alter the cats sense of smell. I’m all for performing the study as long as the fur ball remains un-harmed. I believe the dementia center would likely have the collars (ankle bracelets) and the telemetry to perform the tests. I think some dementia centers (the best funded ones) have individual telemetry systems so they can track specific patients.

  24. #24 Arlene
    August 4, 2007

    I am pleased there is such controversy and doubt over what this cat is doing and why. If the scientific community were to fully accept the notion the cat is accurately predicting death, tens of thousands of cats would end up in laboratories with their brains exposed so humans might intelligently discover the source of their ability. That is just how diligent we are in our search for truth and factual evidence.
    Remember this when you feel your intelligence is superior to animals…Animals are not putting the planet at risk of annihilation. Without humans, they would have had a chance.

New comments have been temporarily disabled. Please check back soon.