Intended for healthcare professionals

Views And Reviews Acute Perspective

David Oliver: What to say when patients are “sick enough to die”

BMJ 2019; 367 doi: (Published 16 October 2019) Cite this as: BMJ 2019;367:l5917
  1. David Oliver, consultant in geriatrics and acute general medicine
  1. Berkshire
  1. davidoliver372{at}
    Follow David on Twitter: @mancunianmedic

One weekend, as a consultant for acute admissions, I spoke to the families of three patients and explained that their relative was critically ill and would probably die soon. Two of those patients rallied sufficiently to return home days later.

The next week on my ward I had similar chats with three other patients and their families. Death followed within hours, their symptoms well palliated. We also got two more patients home, whose admitting doctors had said to expect the worst and who had moved away from active intervention to supportive care.

I’ve practised at the acute end of geriatrics and internal medicine for 30 years. Like many colleagues I find such conversations and scenarios commonplace yet critical. Very sick people present acutely or deteriorate quickly, necessitating time critical assessments. But those few days made me reflect on my practice, even as a veteran. Had I become too eager to give people the direct, unvarnished truth? Was I too hasty in deciding that death was imminent? Was my judgment failing me?

I reflected further and spoke to colleagues. I re-read the NICE guidelines on Care of Dying Adults in the Last Days of Life,1 which state that recognising imminent death is a crucial first step in optimal care. I was reminded of some truths from my own experience and from the literature.

Discussion and involvement

First, patients’ families are far more likely to be distressed by clinical teams not explaining how sick their loved ones are, and not saying that they’re dying, than by doctors openly discussing the issue and involving them in decisions.2 Some people won’t welcome unexpected conversations about dying, however sensitively handled—but they’ll be in a minority. Most people value discussion and involvement. The National Survey of Bereaved People and the National Audit of End of Life Care in Hospital3 make that clear, as does the Royal College of Physicians’ Talking About Dying.45 Communication failings and insufficient information are a major cause of bad experiences.6

Second, although we should communicate sensitively, vague euphemisms don’t work. Unless you say very clearly that someone is dying, or at very high risk, patients and families are left in limbo. If doctors are resolved to have these difficult conversations we can’t equivocate or eschew plain language.

Third, it’s not enough to have the conversation once. To be clear and direct you need to go back and have it as many times as needed for people to come to terms with what can be shocking news.7

Finally, while our language should be direct and unequivocal, that doesn’t mean that we should be more confident in predicting imminent death than the evidence allows. Early warning scores for physiologically deteriorating patients, or scores such as APACHE for those with sepsis or intensive care needs, have strong correlations with risk of death and good predictive validity, but peer reviewed analyses of big datasets show that they’re not infallible.89 Besides, they depend on individual context: how much intervention we give and for how long.

Major variations

A 2016 systematic review by White and colleagues found that, even when experienced doctors were accurate in identifying someone as dying (especially in non-cancer diagnoses), there were major variations in their ability to predict accurately how long the patient had—a question families often ask us.10 We can be both under- and overoptimistic, but we tend towards the former. And more senior doctors aren’t necessarily more accurate.11

So, patients and their families do want honest, timely, clear conversations, and we need to ensure that they understand and accept what we’ve explained. But we also need to be honest about uncertainty. In Kathryn Mannix’s wonderful book With the End in Mind,12 based on her experiences in palliative medicine, she recommends using the phrase “sick enough to die” and then discussing both end-of-life care and survival scenarios.

It’s good advice, and it’s what I now do.