How Much Time Have I Got, Doc? The Problems With Predicting Survival At End Of Life
Doctors are often asked to give prognoses, but they can’t be sure of how long patients will survive. 
Christine Gleason/Flickr, CC BY-SA

Predicting how long a patient will survive is critically important for them and their families to guide future planning, yet notoriously difficult for doctors to predict accurately. While many patients request this information, others do not wish to know, or are incapable of knowing due to disease progression.

Fuelling this complexity are families who prefer the patient not to be told for fear of torpedoing hope and reducing the quality of time remaining. Conversely, patients may want to know themselves, but do not want to distress their loved ones with this knowledge.

We can’t ever be sure

Central to these scenarios is whether accurate prognostication at end of life is actually possible. Providing a meaningful survival time to patients is often genuinely challenging for doctors. Accuracy declines further the longer the patient is expected to live.

A number of studies indicate clinicians tend to be over-optimistic in predicting survival times. Research from 2011 indicated surgeons’ prognosis for survival time for patients with abdominal malignancies was accurate in 27% of cases, too optimistic in 42% and too pessimistic in 31% of cases.


innerself subscribe graphic


This is one of the reasons some doctors are reluctant to attempt to predict survival time at end of life. This has traditionally been seen as part of the doctor’s special domain of knowledge to be communicated at the doctor’s discretion (if and when it is the right time to tell the patient, so it is not going to cause harm).

This archaic and paternalistic view melds conveniently with popular positive thinking, replete with militaristic vocabulary such as “fighting to the end”, which some doctors share.

This mindset views meaningful discussions of prognosis as harmful, as it may cause the patient to lose hope and give up the fight. It underpins those not uncommon cases when a family requests the clinician not to disclose a prognosis or a diagnosis to their dying relative. Unfortunately, it may also shut down meaningful end-of-life discussion and planning and result in harm, including to the grieving who remain.

When patients do not want to know their prognosis, this should be respected. For others who ask and the outlook is poor, a relationship built on trust is important.

Author Bill, an emergency and palliative care physician, is asked to give a prognosis every day. A discussion about prognosis includes the caveat that the accuracy the patient usually seeks is elusive, if not impossible to nail down.

Much can be said, however, including an explanation of why there is no firm prognosis. If it is possible to estimate survival time (derived from a mixture of medical details about the clinical history, prior response to treatment, imaging results, pathology results, functional status of the patient and experience), this is best communicated in terms of a short number of months, (long months is very difficult), long weeks or short weeks, a week or a few days or a few hours.

The accuracy of survival time can become more obvious as time progresses, just as the doctor-patient relationship develops, enabling more explicit discussions about survival time. In emergency medicine, when there is no time for these relationships to develop and time is short, patients frequently seek honesty and are extremely good at telling if the doctor is hiding something. This may then lead to them imagining something worse than the reality.

If the doctor does get the prognosis wrong, there is surprisingly little Australian authority as to whether a doctor will be liable. Considering the general principles of medical negligence is useful here. It suggests that if a doctor provides a prognosis that is widely accepted as competent professional practice, shared by other respected clinical peers, then that prognosis is not negligent.

Even if the doctor provided a prognosis that was not widely accepted as competent professional practice, provided the incorrect prognosis did not cause additional damage to the patient, then no liability will follow.

How long have I got, doc?

Most of us are going to have to ask this one day – presuming we have not confronted it personally or through close relationships already. Despite the understandable imperative for those who want to know, the answer is rarely as crisp or accurate as the original diagnosis.

Breaking bad news to a patient is much more a process than an event, unfolding as symptoms develop and viable treatments recede. Best medical practice aims consistently for open, honest communication that is delivered sensitively.

Most doctors try to provide accurate information if able, despite clinical uncertainty. The aim is to maximise the good and minimise harm. When a prognosis appears wildly inaccurate, is not supported by a group of peer doctors and causes significant harm, Australians may be able to pursue the matter through legal action.

Author: Sarah Winch, Health Care Ethicist and Sociologist, The University of Queensland

This articled originally appeared on The Conversation

Books by this Author

at InnerSelf Market and Amazon