Guest Post from Timaty – You have X days left to live
October 6, 2017
My gran passed away last year. I didn’t realise it at the time, but looking back, she kept defying her doctors each time they told her that she had “X days left to live”. Since then I have spoken to a few friends and they have recounted similar stories, and it got me wondering what might be causing doctors to systemically underestimate remaining life expectancy…?
Maybe there is an element of survivorship bias? People tend to remember or place more emphasis on events where loved ones have defied the odds, and so there is a natural bias to think the doctors have underestimated more often than not. I think this might explain an element of this, but not the whole picture.
Some people that I have argued that Doctors might do this on purpose out of compassion. They want to make sure that both the person dying and their loved ones are prepared for when the moment comes. This sounds plausible and makes for a moving argument. But how must it feel to a dying person and their close friends and family to count down the time for their death and for it not to come? Possibly even get some false hope that things are going to get better when they defy the physician’s life expectancy statements time after time? I appreciate that underestimating may be preferable to overestimating, but again, I don’t think this account for the whole picture.
Doctors are trained to be doctors and not statisticians
In my opinion the piece of the puzzle complementing the above two reasons is that doctors have trained long and hard to be good at what they do. Sure there is an element of compulsory statistics when training to be a doctor, but being a good statistician is not a core part of the job requirement, and a lack of subsequent use of this statistical training will make it difficult for doctors to use the data around them to interpret the life expectancy for a particular patient.
Most doctors will establish remaining life expectancy for their patient by consider their past patients that have been in similar positions. The trouble is that information about past patients is from the past. By not explicitly considering improvements in medical procedures and medical care doctors are likely underestimating life expectancy, albeit unintentionally. This is compounded for the prognoses of longer term terminal illnesses.
Well, probably nothing. Nothing is likely to change and the status quo is okay – whatever the actual underlying reason may be, a slight bias to underestimating is preferable to overestimating remaining life expectancy. This was just another pondering on life.