Markers and Timelines for End of Life Reconsidered: an attempt to bring order to chaos

mike stone 14/05/15 Dignity Champions forum

I have recently exchanged e-mails with the Association of Ambulance Chief Executives (the people who run England's 999 ambulance services), and it seems they are thinking about some of the issues which have concerned me for several years - they wrote in their letter to me:

'Firstly, thank you for raising your concerns. These are particularly valid and relate directly to our work.'

I sent e-mails to England's Ambulance Services after noticing the AACE's admission to a group of MPs that sometimes 999 paramedics will attempt CPR when, with hindsight, it would appear unethical for CPR to be attempted in the circumstances. My questions were was pushed 'upwards' from the individual AS Chief Executives to the AACE, which gave me a 'joint' response. The reasons for, and solutions to, that problem of 'the attempting of CPR looks unethical with hindsight' are complex - 'complex' is true for many of the issues I write about.

In their reply to me, the AACE wrote in answer to my question 'What do you mean by end of life ?' [when they used the phrase in their evidence to the MPs]:

'We don't have a specific or robust definition, and believe that it is actually difficult to define. We currently rely on information provided to us by the patient, relatives and carers, patient notes and other documentation made by professionals involved in that person's care. This process is due to be fully considered and evaluated by the national group I mentioned above.'

The AACE is correct: end-of-life is a term which potentially means a number of different things, and is currently being used to mean different things depending on who is using the term. NHS/Expert guidance/advice about the final year(s) of a person's life, is crammed full of terms which mean different things to different authors/readers.

Biology and chemistry struggled to advance until the classification of plants and the periodic table of the elements had been developed: thermodynamics did not develop until concepts such as 'heat' and 'energy' had become reasonably well-understood and were correctly defined.

'End-of-Life' needs a 'terminology set' which is precise, not unnecessarily complex or large, and which is being used by everyone - until that happens, progress will be very difficult, and discussions will be impeded. And because it is necessary if balanced behaviour is to be achieved, for the perspectives of both professionals and also patients and family carers/relatives to be taken into account and the needs of everyone to be properly weighted, the terminology used by the professionals should be reflective of 'everyday language' unless this is impossible - otherwise, the professionals and laymen cannot easily talk to each other.

Many clinicians write about end-of-life mainly in terms of the clinical situation, and most EoL timelines written by clinicians focus of 'how healthy or unwell the patient is at this time'. I do not believe that is the most useful way to describe the 'EoL Fundamentals' - I will now describe a limited set of 'markers' which I think would be more useful, are simple, could be applied (with the necessary 'buy-in' from GPs, etc) throughout England, and are (crucially !) as precise as it is possible to be, and do not 'imply more understanding of the situation, than is logically possible'.


The first marker for 'end of life', should the 'The time of a terminal diagnosis'.

The assumption is that the patient is told his diagnosis is terminal, when the doctor decides the prognosis is terminal.

My definition of 'is terminal' would be that 'the doctor believes there is at least a 50% likelihood that the illness described as terminal, will appear as the cause of death on the patient's death certificate'.

The point about this marker, is that surely 'the way the patient thinks about his life and future' will be altered, once he is told of a terminal diagnosis.


The final marker for 'end of life', should be the 'end of the relative's 'deep grieving'' - so some months or years, after the patient has died. And probably so vague, as to be more of a 'guiding concept' than a point in time. Which doesn't really matter all that much, compared to the other markers on my timeline.


The next of my markers, is very obvious - 'the patient's death' is a very significant marker (and it is the only unequivocal 'clinical' marker, which will be present on the EoL Timeline of every patient).


My next two markers, are necessary to make post-mortem behaviour for EoL Home Death more balanced and rational - they have the added benefit, of removing the need for a separate 'definition of' 'expected death', and also of separating 'somewhat early but during known EoL deaths' from 'genuinely unexpected/sudden death'.

These are two notes which the GP (a consultant could do the same for hospital patients, but the problem these resolve is mainly when death in the community occurs) has to write in the medical notes as they become true - and CRUCIALLY all of the patient's carers, both professional and family carers, AND ALSO anybody else sharing a home with the patient, needs to know about these two notes. The two notes are:

'I (the GP) would no longer be surprised by the natural death of this patient, but I would need to attend post-mortem before deciding whether to certify the death'

and, at a later stage of clinical deterioration

'I (the GP) will now certify any death which is not apparently unnatural, even if I am unable to attend post-mortem'.

The second of those, is actually 'a statement of expected death' [as the concept is used in post-mortem behaviour] but if you use the statement, there is NO NEED to separately define 'expected death'.

There is a problem with the combination of 'expected death' and Deprivation of Liberty (DoL) - this would require the second statement to be written like this, for patients with a DoL(S) in place:

'I (the GP) would now certify any death which is not apparently unnatural, even if I were unable to attend post-mortem, except that the presence of a DoL prevents me from certifying the death and the law requires coronial investigation for any death with a DoL in place'.


The final one of my markers, is a division of the post-mortem period into two sections. The first section is:

'The period between the death and attempting to go to sleep.'

The rest of the post-mortem period is:

'The longer-term 'grieving process' in the days, months or years after the death.'

The reason for that division, is this: I believe that the day of the death itself 'is about the creation of long-term memories', whereas the days and months after that, are about 'dealing with those memories'.

One of my major claims, is that at the moment, the way that police behave when involved in 'early EoL home death' is biased towards 'investigation of possible wrong-doing' because it neglects the fact that if entirely innocent live-with relatives are treated 'like suspects' in the hours immediately after a death, that will inevitably strengthen their [in essence, traumatic] memories of the death itself - and the reason 'people seem a bit weird just after a death' is surely because the relative's mind is trying to weaken memories of the death itself.

Post a reply

mike stone 19/05/15

I will explain the problem with contemporary Community End-of-Life Death policies, and also why it would help if everyone - all of England's GPs and all of England's regions - adopted my 'marker' system, and also recorded and analysed deaths in terms of the presence or absence of these markers.

At the moment, the behaviour for EoL death at home, varies from region to region, and depends on what the local NHS has agreed with the local coroner. The only policies in existence are invariably 'Expected Death' policies, but the term doesn't mean, as a family member might expect it to mean' 'my dad was expected to die'. It actually means 'the GP has indicated to the coroner, that even if the GP cannot attend the deceased patient, the GP will certify the death (i.e. provide a note to the coroner, stating 'what the patient died from') unless it appeared to be obviously unnatural. .

In other words, the marker on my timeline of:

'I (the GP) will now certify any death which is not apparently unnatural, even if I am unable to attend post-mortem'

indicates an 'expected death'.

The point of 'expected death' is to keep the police away from 'imminent natural deaths'. So any well-organised and competent region, will have in place arrangements which allow people such as trained nurse to confirm that the patient is indeed dead (Verify the death) and to arrange for the body to be removed from the house, and for the relatives to be given the necessary paperwork, etc, even if the GP cannot turn out.

All other EoL deaths, are currently described as 'unexpected' or 'sudden', and the police tend to be involved in these deaths, unless the GP turns up before the police, and the GP decides to Certify the death. I have described the problem with this, in quite strong terms, on the BMJ website:

Now, in England this promise that the death will be certified even if the GP cannot attend post-mortem, requires that the GP has visited the patient within 14 days before the death: the GP can decide to certify the death even if the most recent visit was more than 14 days previousy, but ONLY IF the GP attends the death and examines the deceased.

If the death wasn't 'expected', and the GP doesn't visit post-mortem and decide to certify the death because for some reason the GP cannot visit, the GP might still end up certifying the death, after a discussion with the coroner - this made an attempt to 'count the perecentage of unexpected deaths' very complicated, in a paper you can find online.

Within the paper, the authors explain the complication of deciding what is or isn't 'an unexpected death', and they comment that 'it could be assumed, however, that referrals to the coronial service where no inquest and no post mortem were required were in fact not unexpected deaths because a death certificate could be issued following a discussion between the clinician and coroner only,' - and if that assumption is made they came up with a figure of 22% of deaths as being 'unexpected', whereas if you include 'there was a
conversation with the coroner' in the description of an unexpected death, the figure for 'unexpected' deaths was 46%.

Coroners seem to be very reluctant to clearly explain 'expected death' - perhaps because in effect it amounts to 'a very light 'investigation' of deaths which the GP knows are going to happen soon, in order to keep the police away, because the police when involved, tend to 'harass the just-bereaved relatives''. Doubtless coroners are aware that some people, and some media, would scream 'You might be letting people get away with murder !' - which is a possibility, but the balance on the other side, is that 'aggressive police investigation' into the vast majority of home EoL deaths, which are overwhelmingly natural, leaves entirely innocent relatives with 'long-term trauma memories'.

But, one consequence of the reluctance of coroners to explain this properly, is that many nurses believe that when they (nurses) verify an expected death, they are 'supposed to be linking the death to the known terminal illness'. Those nurses should only be confirming that the patient is dead, and noticing any obvious signs of an unnatural death - 'linking the death to the known illness' is CERTIFICATION, and ONLY a doctor who could certify the death, should be doing that. People such as nurses and paramedics, should definitely NOT 'be offering an opinion about why the patient died' [except for drawing attention to signs of an unnatural death].

The fundamental problem with 'expected death', is twofold: nobody is 100% sure 'how imminent a death should be, to qualify as 'expected'', and also doctors are not at all good, at predicting exactly when a patient will die.

If the idea is to keep the police away from 'imminent deaths', the death would need to become 'expected' at least a few days before the patient died - it would be pointless the GP 'classifying a death as expected' when the GP expected the patient to die within 24 hours.

As the GP must visit post-mortem if he has visited during the previous 14 days, 'expected to die within 14 days' looks like the upper limit.

So, let us assume that 'when the GP expects the patient to die within 14 days', the death becomes expected, and the GP writes in the patient's notes:

'I (the GP) will now certify any death which is not apparently unnatural, even if I am unable to attend post-mortem'

In reality, no GP can be sure of when, a patient will die: if a GP says 'I think this patient will die in 10 days time', the GP could not honestly be surprised if the patient dies in 7 days time, or in 15 days time.

It is easier 'for the thinking' of the next part, to switch the prediction to a different type. Imagine that the patient has got a fixed chance of dying, on every day, until he dies. So imagine that someone is very ill but is not 'getting worse' - so if the patient has a 1-in-5 chance of dying today, and if he does not die today, he has the same 1-in-5 chance of dying tomorrow (and so on - a 1-in-5 chance of dying, on every day until he actually dies).

What, thinking about this patient, does the chance need to be, for the patient to be 'expected death' ? It turns out, that a 1-in-10 chance of dying on every day, is approximately 'death within 14 days' (about 4 in every 5 patients with a 1-in-10 chance of dying on every day, will be dead by the 14th day).

The way EoL death at home is treated at the moment, the patient who the GP considered to have a 1-in-10 chance of dying on every day, would be 'expected death'. A patient with a 1-in-11 chance of dying would be 'unexpected death'. Even if the GP could be certain of the chance of dying [which the GP cannot be] - HOW DOES THAT MAKE SENSE ?

If the GP believes his patient has a 1-in-10 chance of dying, then the patient's death is treated 'sensitively', the police are kept away from the death, and the live-with relatives are treated with care and consideration. If the GP believes the patient has a 1-in-11 chance of dying, then it is quite likely that the live-with relatives will be confronted by police officers, who far from being 'sympathetic and caring' tend to treat them 'as if they were potential suspects in a murder case'. This is CRAZY and it treats the 'early but known EoL death of a patient' as if it is very similar to 'the unexpected death of a thought-to-be-healthy 25 yr old'.

The obvious way around this - to me - is for GPs to record that 'I would no longer be surprised statement':

'I (the GP) would no longer be surprised by the natural death of this patient, but I would need to attend post-mortem before deciding whether to certify the death'

AND FOR THE RELATIVES TO KNOW THAT. Then, if police 'start to interrogate the relatives', the relatives can say 'why do you seem to be surprised by my husband's death - the GP wouldn't have been surprised that he has died, so why do you seem so surprised ?

I have no objection to police attending these 'earlier deaths', nor for the body to be retained by the coroner until the coroner has been satisfied about the cause of death, nor about the police asking relatives ONCE 'what happened'. But I object HUGELY to the police behaving as if these just-bereaved relatives are 'primarily suspects', repeatedly and insensitively questioning them at length when there is NO ACTUAL INDICATION of any wrong-doing, and seriously interfering with the relatives in the hours after the death (when, in my opinion, long-term memories are being formed, and when 'messing just-bereaved people about' leaves them with 'strengthened, which means worse, memories of the death of their loved one'. Not only, in my opinion, does that type of overly-intrusive police questioning leave bad memories of the questioning, but the lengthy and intrusive questioning stops the bereaved person's mind from 'distracting itself from the death', and you are consequently left with stronger memories of the death itself - and NOBODY 'wants to remember the death itself').

If the presence and times of these markers, and the time and cause of the death, were collected for all deaths, then many currently 'unanswerable' questions about EoL, could in theory be answered. A few examples are:

It is accepted that the difficulty of accurately predicting when a patient is going to die, differs according to the illness. If you filter deaths according to the illness, you can then look at the two GP statements for each type of death, and that would tell you which types of illness had [in relative terms] harder-to-predict deaths: you might , for example, discover that it was much more difficult for GPs to predict when a patient suffering from heart failure would die, than for a patient with lung cancer.

If a GP Practice had 4 GPs, then by comparing the times between the markers and the deaths for each GP's patients, you could see how varied or similar different GPs were in terms of how they predicted the deaths of their patients.

By doing the same type of analysis but across different regions of the country, you could see 'if behaviour for community EoL death was the same in different parts of the country'.

Etc - it is very easy to see how that type of analysis could be performed, if everyone used the 'marker system' and the results were available for analysis.

Although I most fervently support this idea, because I strongly object to the way that 'somewhat early but known end-of-life deaths', seem to often be treated, especially by police officers, as if they were 'genuinely suspicious' despite there being no real evidence that the death is unnatural.

mike stone 21/11/15

I am currently in discussion, with the authors of the 'EPaCCS core guidance' - EPaCCS is an electronic online database, containing information about patients whose GPs have identified as 'likely to die within a year'.

I have several 'issues with' the guidance as it currently stands, one being the 'single main carer' issue - see:

I am also 'somewhat worried' by 'field no 34', which is 'Likely prognosis', and where the options are:

A clinical judgement indicating the anticipated period of time until death.
Last days of life (definition: death is likely in <1 week)
Last weeks of life (definition: death is likely in <1 month)
Last months of life (definition: death is likely in <1 year)

The EPaCCS guidance does correctly point at uncertainty in such predictions:

'Used to denote those who are nearing their last months/weeks/days of life but with recognition of the uncertainty related to such a judgement.'

But, I am very unclear, about how useful that information is, and I am worried that it might be incorrectly conflated with:

a) Certification/Verification of death (see the earlier sections of this piece, where I have analysed EoL Home Death with regard to certification of the death);

b) Resuscitation behaviour (there is no logically NECESSARY connection between 'how soon the patient is expected to die' and 'whether CPR should be attempted').

The current guidance around both certification of home death (and immediate post-mortem behaviour) and also CPR/DNACPR is far from perfect - I will be 'very annoyed' if section/field 34 of EPaCCS, makes the situation even worse !

mike stone 28/04/16

Within the last few minutes, I have uploaded to the Dignity in Action Facebook page, and also to my own recently set-up Facebook page about end-of-life, a one-page (in theory, 'landscape A3' sized) PDF which shows the 'markers' I've discussed in this piece as a timeline. It also has a comment from a group of GPs I've discussed this with in it - basically, they agree that I might have a valid point about the 'lack of proportionality in the behaviour'.

I would have uploaded the file here, but that isn't an option any more, since the website was redesigned.

The PDF file is titled:

Markers and Timelines for End of Life Reconsidered

Keith Comley 29/04/16

Great discussion

I remember someone once saying "Life is terminal and death is the likely cause of end of life in 100% of cases."

In my PERSONAL opinion End of Life Care in a person cantered care environment should be exactly what the individual wants to happen, within reason even if death can possible predicted to happen whilst in someone's care.

Having a DNAR is the individuals choice as long as they have capacity. If no DNAR is in place everyone must assume they would like CPR attempted even if it will only result in one or two more breaths of life or to feel the warmth of the sun through a window one last time otherwise there would be a valid DNAR

My opinion is it is only "unethical for CPR to be attempted" if there is a valid DNAR in place. I still teach this.

If someone expresses a desire for CPR before they loose capacity would it be morally right to change their wishes and put a DNAR in place if you are the next of kin or a clinician? If a request for CPR is recorded in their notes you could argue this it precludes any future DNAR.

I will give a hypothetical situation. A older person in nursing care is admitted to hospital who has become seriously ill but has a CPR request on their notes although this is 6 months old. The clinician says the likelihood of survival is low and gets the family to sign a DNAR and endorses it. The patient survives retains capacity and returns to their nursing home.

What does everyone think the reaction would be from the individual if they found out about the DNAR?
At what stage should the DNAR be removed?
What may happen if the DNAR goes back with the notes to the Nursing Home?
Considering the original request for CPR in the notes from the Nursing Home did the clinician or family have the moral right to override it regardless of the expected outcome of the illness that led to admission?

I remember having to discuss this exact hypothetical situation a few years ago. Being new to the site would be interested in every ones thoughts?

mike stone 30/04/16

Hi Keith,

I think I agree with you.

There is a problem with 'DNACPR forms' which I pointed at in one of my BMJ rapid responses:

There is also a problem, for EoL at home, with verbal refusals of CPR (sorry - but they ARE 'legally binding' !) which I've covered in my post at 11/04/13 - 13:32 to my piece at:

You might - I suspect you would - find the file you can download from my piece at:


Would you like to write something for my Facebook group - only just set up, and intended to promote some much-needed discussion of various EoL issues ?

The url is:

But I am not sure (I'll check) if that lets you get to my group, or not (I'm very new to - and frequently confused by - Facebook).

Thanks for your comment, Mike.

mike stone 01/05/16

Cardiopulmonary resuscitation (CPR) - and this is 'off-topic' for the original theme of this piece, but it was introduced by Keith Comley - is almost a 'stand alone' issue for various reasons.

I wrote about the way that some clinicians seem to be intruding into 'quality of life' [which isn't their decision to make] when they write about CPR, in my piece at:

In that piece I commented:

'But there are many issues: I am only happy with 'clinical DNACPRs' if the term means 'attempted CPR would definitely not restart the heart'. But many clinicians wish to add 'or would be futile', with 'would be futile' meaning 'or would only restart the heart for a few hours or days'. How, objectively, is that 'few' to be decided: an hour, a day, 3 days or what ?

And it isn't for clinicians to decide whether restored life but of a 'clinically poor quality' would be acceptable to the patient: that decision falls to either the patient if capacitous, or to a proper Mental Capacity Act 'best interests' decision (which cannot be made by clinicians in isolation from 'those who know the patient as a person': in fact, the ideal 'best interests decision' would be a decision that everyone involved supported and agreed with - in which case, it would not be 'made' by any particular individual).'

More recently, Professor David Oliver has written in BMJ about what he predicts will be unfortunate operational consequences of two recent court rulings (Tracey and Winspear) which strongly push clinicians towards discussing and revealing 'clinical' DNACPR (I use the term 'clinical DNACPR' to describe the 'the clinicians do not intednt o attempt CPR because they consider it would not succeed in restarting the heart') decisions with/to patients and/or family/friends. Professor Oliver also added a rapid response (the BMJ's version of comments to articles) to the series about his original article - his claim, is that the two court rulings which were intended to push clinicians into properly discussing CPR with the laymen, will actually lead to a worsening of the present situation - the thrust of David's position is contained in his words:

'... doctors will leave the person "For CPR" until they can discuss it properly. As I argued, this will lead to numerous futile CPR attempts unlikely to be in the patient's best interests. ... The result [of the court rulings]? Rushed, perfunctory assessments of capacity and discussions of DNACPR very close to the person's initial admission, often without them fully understanding it in order to "tick the box" on the form that CPR has been discussed. And often in a very busy, short staffed emergency floor.'

I will give the link to the rapid responses to the article here - it is easy to get from those to the article, you simply click on 'article':

mike stone 20/06/16

Just noticed that my proof-reading is a hopeless as ever in this piece - in my second post I wrote:

As the GP must visit post-mortem if he has visited during the previous 14 days, 'expected to die within 14 days' looks like the upper limit.

when of course it should have been

As the GP must visit post-mortem if he has NOT visited during the previous 14 days, 'expected to die within 14 days' looks like the upper limit.

I recently attacked my own proof-reading on BMJ:

nicola ambler 22/06/16

I am the manager from two care homes and we feel it is the opinion of the resident to choose about DNAR, but at times we do have to decide in the best interest for the resident, and their quality of life especially if they are coming to the end of life. We always discuss this with everyone who is involved in the care, and if relatives and doctors make this decision about the resident if they lack capacity we record that we have had this discussion

Ken Nicholas 22/06/16

Hi Mike,
This is indeed, a complicated subject and I speak as a member of front-line ambulance services, being an ambulance technician of many years. I was involved in bringing this issue to the attention of my service management some years ago and liaised with local doctors, a specialist palliative care doctor and trust representatives to ensure cross-county recognition of DNACPR certificates and advanced decision notices.
Without having sight of either of these documents, I am obliged to begin resuscitation attempts because it is not possible to immediately determine whether the chronic life-limiting condition has brought about cardio-pulmonary arrest or an idiopathic cause is the origin.
Part of the solution is to ensure that those responsible for the person's welfare in primary care explain the implications of the medical condition to the patient, family and carers to ensure the patients needs and wishes are met. My understanding is that those patients who were placed on the now discredited (wrongly, in my opinion) were informed that they would automatically be regarded as DNACPR, making it clear to clinicians that resuscitation need not commence.
In my role, I often have contact with those suffering a life-limiting condition and I discuss their illness with them and ask about their wishes, should they suffer an arrest. This often leads to me speaking to their own GP in order to facilitate their wishes, and I advise that the relevant paperwork should always be available.
In no way, would I ever presume that some-one should not be resuscitated simply because they had a long term illness and there is the human factor involved in family relationships i.e. that many relatives take great comfort in the fact that they believe that 'everything possible was done to save the patient'.
Hope this viewpoint is helpful.

Ken Nicholas 22/06/16

Sorry, Discredited process is Liverpool Care Pathway !!!

mike stone 24/06/16

Hi Keith,

Thanks for your contribution, which I only just spotted.

You conflated two different things - not just you, a lot of clinicians do this - with your:

'Without having sight of either of these documents, I am obliged to begin resuscitation attempts because it is not possible to immediately determine whether the chronic life-limiting condition has brought about cardio-pulmonary arrest or an idiopathic cause is the origin.'

The long-term condition - or the terminal illness - isn't the reason why a DNACPR decision would be made: such decisions are based on the spectrum of possible post-CPR 'successful' outcomes, compared to the known outcome (death) of no attempted CPR. And if a patient is at home, all you know is that the best you would ever be clinically post-CPR, is 'the same as you were before the CPA' - usually, you are likely to be worse, and often much worse.

I think the guidance for 999 paramedics is flawed - there was a series of BMJ rapid responses discussing this type of issue, and in one of mine at:

I pointed out that it seemed likely that a 999 paramedic would be the first clinician 'in the firing line' for a prosecution for 'assaulting the patient by attempting CPR'. This seems rather unfair, on 999 paramedics - although I would support the prosecution, as it looks to be the only way to get clinicians to properly follow the principles of the MCA.

I also highlighted a few issues around CPR, in my piece 'It would be much simpler if techniques for CPR did not exist' at:

mike stone 25/06/16

By the way, just for clarity: the thing I'm discussing in this piece, is post-mortem behaviour in connection with Verification and Certification of Home EoL Death.

It is nothing to do - AT ALL ! - with 'should CPR be attempted ?, which is an entirely different issue.

The separation between those two different issues, is covered in my piece 'DNACPR is not correctly 'a proxy marker for' 'expected death' at:

There is 'an implicit belief' in some clinical writing, that 'if the death could not be certified, CPR should be attempted': this is not consistent with Patient Autonomy (patients can refuse CPR even if the subsequent death could not be certified) nor is it consistent with the MCA's 'best-interests 'rules'.

mike stone 14/07/17

There is a one-page (more A3 than A4) PDF covering this, which I wrote last year, and which I have just attached here.

Associated files and links: