From 'live fast, die young' to 'live longer, die slowly'
After being in danger of "imminently dying", a cancer patient defied her prognosis and continued to attend yoga classes for many years after.
The unpredictability of death is why MP Madeleine Moon wants benefits changes.
Currently, to qualify for special payments, a patient must be expected to die within six months.
But Mrs Moon, who lost her husband to motor neurone disease, broke down in tears saying patients should not spend their last months worrying about money.
So, how easy is it to predict how long somebody will live for?
Specialist Dr Mark Taubert - a senior lecturer and clinical director at Velindre Hospital, which treats cancer patients in Cardiff - admits at times he has misjudged a prognosis "quite spectacularly".
Here he gives his personal thoughts.
Many years ago, I worked in a hospice and admitted a severely ill lady. She had cancer. We all felt that she may be imminently dying.
We gave her antibiotics and fluids at the hospice, and helped her with her nausea and severe breathlessness.
Many thought she would not survive the night.
She recovered and improved, albeit slowly. She required morphine infusions to help with breathlessness and pain, and for a while required strong sedatives and high flow oxygen.
- '25% of entitled patients miss out' on palliative care
- MP Madeleine Moon in tears in terminal illness speech
We eventually were able to wean her off most of the supporting medication and discharged her from the hospice.
We stayed in touch, and she came to the hospice day centre very regularly, where she enjoyed joining in the yoga sessions and exercises.
I bumped into her five months ago, at a meeting and we had a coffee together.
We talked about what had happened and she could not believe she was still around, and I had to admit to her that I, too, had felt she would die that night many years back.
In fact, I had told her husband that same night that I wanted him to prepare for the worst.
"I know," she said, "and he remains grateful that you were so frank."
She went on to tell me that she was aware the cancer was still spreading, and that she had learned to make peace with the uncertainty of not knowing how long she might have left to live.
And I wasn't going to rush to make any predictions.
This is a more unusual example, of course. I can, however, recall a number of situations when I have misjudged a prognosis quite spectacularly, and have either been overly optimistic, or overly pessimistic.
Medicine has become more sophisticated, and our knowledge from years ago is not keeping up to date with many recent developments.
So the averages and statistics from studies done even five years ago, are not really that useful, in light of this constant flux and change.
We all die, but serious advances in medicine and technology in recent years have enabled us to live longer.
And with longer lives, we also have a growing population and people are living with more long-term serious conditions.
The coming years will see a sharp rise in the number of people dying, presenting a challenge about how we care for those at the end of life.
Numbers of people dying each year in England and Wales are set to go up by 20% over the next 20 years.
Dementia has become the most common cause of death. The old expression "live fast, die young" can be replaced with "live longer, die slowly".
Sometimes people live with two or three conditions at the same time.
Many of these are life threatening, life-shortening, but predicting when a sudden or slow onset of dying may start, is hard for even the most experienced healthcare practitioners.
People are admitted to hospital repeatedly, receive medical treatments and of course one of these many admissions during the final years of life will be the final one, but predicting when that might be is guesswork.
When I meet people in palliative care settings, I listen to many stories.
Not uncommonly, a patient, let's call him Des, will recall a confident senior clinician, who, with a kind look and professional head tilt, broke the bad news (it's pancreatic cancer) and gives a prognosis of not more than 12 months.
"That was two years ago", Des winks at me.
He is still going strong, but it doesn't prevent me from talking about what his wishes are for a time when he becomes less well in the future.
"Have you ever thought about what happens at the end?"
"Of course I have, doctor". Des knows, but he hasn't told anyone else. Because it's hard.
I have been asked by many relatives and patients how long I think they might have left. "You must know, doctor".
As a junior doctor, I was perhaps more confident reeling off statistics, but soon started appreciating the fallibility of giving mathematical averages.
If the statistics state that people with Condition A can live from between two months and 48 months (and in rare cases even longer, but some actually die after just a few days), then that information is very hard to make much sense of.
We know that in breaking bad news scenarios, a large part of the information transmitted from clinician to patient is lost.
If I gave the above statistic, the patient's mind, unbeknownst to me, may be going into hyper drive after I've said "between two months and...", and the last thing they truly heard and understood was two months.
"It's going to be over in two months," he or she thinks and walks away. No one I have met has ever found these numbers particularly helpful, when they reflected on it later.
Where there is a serious illness, there is the prospect of death.
It may sound obvious, but it is perhaps a sign of our sometimes death-denying times that even those with advanced cancer sometimes haven't allowed themselves to contemplate the possibility of dying.
Or a well-meaning relative may interject and say "don't talk like that, just be brave, and battle on. Keep fighting".
But this fighting talk sometimes detracts from the valuable conversations about where you'd like to be when you die and with whom.
Not just the when, but the where and how. Perhaps even what music you most definitely don't want played when you are lying on your deathbed.
I personally do not need to bow out with Justin Bieber playing in the background. You may laugh, but it is quite important to me.
When you allow for such conversations to happen, naturally the realisation of the possibility of a death that may happen soon hits many people hard.
In my specialty of palliative care, we work, sometimes from diagnosis onward, to try and address a lot of the symptoms (eg pain, nausea, breathlessness) that people can get with their illness, or as a result of the treatments given.
But we don't just focus on physical symptoms; the individuals concerned are often frightened and have a lot of questions.
So when someone asks: "How long have I got left to live?", healthcare professionals have to address this very common and reasonable question sensitively.
Unpicking what someone is really asking requires skill and time.
Why is it so difficult to give a prognosis?
A breast cancer behaves differently to a lung cancer. A lung cancer that has spread to the bones, is very different to a bladder cancer that has spread to the lungs.
In dementia, it may well be an overwhelming infection that takes the person in the end, but of course we are swift to treat this each time with powerful antibiotic drugs.
Will they work? Not always. And then who knows precisely what influence the patient's long-standing diabetes, kidney problems and heart disease have on the overall picture?
The truth is that no one really knows the right answer, no one is perfect at predicting the outcomes of a life-limiting illness other than to say that it is likely to lead to death within years, months or perhaps within weeks, but even this can be on shaky ground.
Last 48 hours
Perhaps there is already, somewhere, a supercomputer that can give us the precise answer on when we will all die. Would you use it? I personally would avoid the temptation.
But I realise that it is natural to want to know and sometimes the uncertainty is overwhelming.
I find that it is not uncommon to discover that there are other fundamental underlying questions to the one above.
"Will you be frank and honest about what may lie ahead, when I'm ready to ask?"
"Would you tell my family what may lie ahead? I can't."
"What are the last 48 hours of life usually like? Will I be aware, or will I be asleep?"
Once we have talked this through and I have admitted to my own uncertainty in giving any firm predictions on how long a piece of string is, we usually relax, and sometimes we talk about the life that there is left to live.
There always is, even if it's short.