How is time to death calculated in terminally I’ll patients?
In: 510
It depends.
When patients are first given the diagnosis of a lifethreatening disease, it is based on statistics collected from lots and lots of other cases.
So if someone gets diagnosed with ALS, the doctors can say that on average a person has 5 good years after diagnosos (as an example).
Or if a person gets diagnosed with metastatic liver cancer, the doctors can say “the majority of patients with similar cancers die within six months”.
Later in the process, it’s more an evaluation of that specific patient, their symptoms and the progression of their disease between scans/checks.
In the late terminal stage, doctors and nurses will have a general idea of whether we are talking hours, days or weeks before death based on the symptoms and how the patient looks/sounds/smells.
Speaking as a nurse, this is also how we try to frame it for the relatives. “It’s a matter of days, not weeks” for instance will give families are very concrete sense of how much time is left.
I tried to get a doctor to tell me when grandpa would die, was trying to get week, month, day out of him (not imfeont of grandpa so siblings, etc could come say goodbye) & he shook his head and said. Soon but the exact date is impossible to know. If they tell you hospice than usually a month or so & the hospice people know when it is going to happen. They said within “24hrs” and they were right.
We don’t give people “x months to live”. It generally ends up being inaccurate and creates a lot of confusion and stress.
I can tell you when you or your family will die in minutes or hours and am reasonably accurate for a few days. All of that is really just an eyeball guess based on objective data and the natural course of whatever disease the patient has.
For longer term diseases, there are various mortality calculators and data to guess when someone will die, but it’s not really a guess. All we know is that “‘x’% of people die in ‘y’ years”. We often don’t know how well or poorly someone will do because the data is not that individualized.
Sadly, we are typically terrible at prognostication for time horizons any longer than what I mentioned above. People die earlier and later than we think.
Source: I’m an IM doctor.
Edit: a few sentences changed for clarity
From what perspective are you asking?
There’s a a group of statistical methods broadly known as survival analysis or time-to-event that are used to, amongst other things, determine the survival (or time to death) of terminally ill patients. Obviously it depends on several factors and can only be expressed in terms of estimates for the population based on observed data (true of all research) so you can never be certain of time to death in an individual. As others have noted, you might be able to say something like, for a female with stage 3 breast cancer of your age, 50% of people live longer than X years. But you could never say to someone, you will die in 3 months. Obviously the closer a person gets to death from a terminal illness, the more obvious it becomes and you can be more certain.
From another perspective, as others have noted, doctors have experience with various illnesses and can make general statements about how long a person might expect to live. But they won’t (shouldn’t) give very precise estimates because it’s very easy to be wrong. You don’t have to look very hard to find examples of people being told they only had months to live and are still kicking around years later. Unfortunately the opposite can also be true (die much sooner than expected).
Definitely time travel!
They travel forward and check the death records!!
5 yo’s will believe anything¿¿?!
Seriously I’d say they just guess based.in experience of other patients and only give a range.
Statistics. If you’re thinking its like an ongoing estimate you’re mistaken. Its a one time thing when someone gets diagnosed, its just the average life span with that diagnosis if it is terminal, and the odds of beating it if it isn’t.
This is actually a very up and coming area of artificial intelligence and machine learning. Doctors use various metrics, along with personal experience, but algorithms are being developed that can look at hundreds of different variables, include genetics, family history, medical tests, rates of change, etc… to improve accuracy.
Being a doctor, I can confidently say that no doctor in the world will be able to tell when a patient is going to die accurately.
I’ll never give deadlines to patients. Patients whom i thought will die in a day have lived longer than a year and vice-versa. But I am able say if a patient is going to die in the final hours(~4-5hrs)….once they hit the point of no recovery I have never seen a patient coming back …
If you’re talking about the prognosis when someone has cancer and the doctor “gives them six months”, that’s a calculation called the median survival tjme. It’s based on how long other patients with this cancer have survived past diagnosis.
If the median survival time is six months for a particular kind of cancer, that means that half the patients survived more than six months after they were diagnosed, and the other half died before six months.
There’s a great essay by Stephen Jay Gould called “[The Median Isn’t the Message](https://journalofethics.ama-assn.org/article/median-isnt-message/2013-01)”, about how he could have been terrified by his cancer diagnosis with its short median survival time except that he knew how to interpret statistics and understood that some of those patients lived *far* beyond the eight months that was quoted. He did ultimately survive a couple of decades beyond his diagnosis, if I remember correctly.
We guess.
We see a lot of people die so we can form estimates based on how long other people held on when they were at a similar stage.
For example, if someone near death stops eating, they likely have less than 2 weeks to live. If their skin is starting to mottle and they’re agonal breathing, then hours to days.
However, I’ve personally been wrong so many times that I’ve learned to not give direct estimates. I tell the patient “hours-to-days”, “days to weeks”, or “weeks to months” based on the stage of their disease and other signs they’re showing.
Latest Answers