Doctors and nurses who regularly work around life-sustaining treatment may think about their own end-of-life care differently from people who spend less time in intensive care.
New research suggests that the difference is not simply that ICU professionals are more willing to refuse treatment.
Their choices appear to depend more strongly on the long-term consequences of survival, and less on mortality risk alone.
A study published in Scientific Reports on 24 September 2026 surveyed 743 French hospital professionals using nine hypothetical end-of-life scenarios.
The sample included 195 ICU caregivers, 349 non-ICU caregivers and 199 non-clinical hospital professionals.
For each scenario, participants indicated whether they would want life-sustaining treatment continued or withdrawn if they personally were the patient. They also indicated what they believed a collegial medical team would decide for a patient in the same situation.
In seven of the nine scenarios, ICU caregivers were more likely to choose withdrawal of life-sustaining treatment for themselves than they believed would typically be chosen for patients.
The finding raises an uncomfortable but important question: does regularly witnessing the burdens and consequences of intensive care change what health professionals consider an acceptable outcome for themselves?
The study separated the chance of dying from the consequences of surviving
End-of-life decisions are often framed around mortality.
A treatment may be described in terms of the probability that a patient will survive or die.
But survival can lead to very different outcomes.
One person may recover to independent life, while another may survive with severe disability, prolonged hospitalisation, dependence on others or substantial loss of autonomy.
The nine scenarios used in the study were designed to vary both mortality risk and long-term consequences.
This allowed the researchers to examine whether participants were responding mainly to the probability of death or to the quality and burden of the life that might follow survival.
Long-term consequences mattered more than mortality risk
Across the professional groups, long-term consequences influenced treatment preferences more strongly than mortality risk alone.
This pattern was particularly pronounced among ICU caregivers.
In other words, the possibility of surviving with serious functional limitations often mattered more to their decision than a high statistical risk of dying.
This does not mean ICU professionals were indifferent to mortality.
It means they appeared to evaluate treatment in a more multidimensional way, taking into account what survival might actually look like.
That distinction is central to many debates about end-of-life care because a treatment can successfully prolong biological life while leaving the patient with an outcome they personally would not consider acceptable.
ICU professionals were not uniformly more willing to stop treatment
The findings are more nuanced than the idea that ICU staff simply prefer less treatment.
After statistical adjustment, ICU caregivers were more likely than other hospital professionals to choose withdrawal in a scenario involving prolonged hospitalisation and burden on others.
But they were less likely to choose withdrawal in a scenario dominated by high mortality risk.
The researchers interpreted this as evidence of more context-sensitive decision-making.
Rather than applying a simple rule such as “high mortality means stop treatment”, ICU professionals appeared more responsive to the particular combination of survival probability and long-term consequences.
Experience may change how treatment burdens are understood
ICU professionals routinely see treatments that most people encounter only during major illness.
Mechanical ventilation, resuscitation, prolonged sedation, invasive monitoring and organ support can save lives.
They can also involve substantial physical and psychological burdens.
ICU staff are therefore exposed not only to successful recoveries but also to prolonged dying, severe functional decline and situations in which survival comes with major dependence.
The study cannot prove that this professional exposure caused the differences in preferences.
People who choose to work in intensive care may already differ from other professionals in their values or attitudes toward risk and treatment.
But the pattern is consistent with the possibility that repeated exposure to treatment outcomes changes how clinicians weigh the benefits and burdens of continued care.
Personal preferences differed from perceived clinical practice
One of the most striking parts of the study was the comparison between personal preference and perceived professional decision-making.
Participants were not asked only what they would want for themselves.
They were also asked what they believed a collegial medical team would decide for a patient in the same scenario.
Among ICU caregivers, the personal preference leaned toward withdrawal more often than the perceived patient decision in seven of nine scenarios.
This does not show that clinicians are actually overtreating patients.
The study measured perceptions of collegial practice rather than reviewing real treatment decisions in patient records.
Still, the discrepancy raises questions about whether patients’ own preferences are always elicited early enough and clearly enough to guide care when they can no longer speak for themselves.
Only 13.3% of ICU caregivers had written advance directives
The study also exposed a gap between knowledge and preparation.
All ICU caregivers were aware of advance directives, yet only 13.3% reported having written their own.
Advance directives allow people to record preferences about future medical treatment in case they later lose the ability to communicate.
For professionals who regularly participate in end-of-life decisions, the low completion rate is notable.
Knowledge of the importance of advance planning did not automatically translate into completing it personally.
This mirrors a broader difficulty in end-of-life planning: people may recognise that preferences matter while still postponing the act of documenting them.
Quality of life and survival are not interchangeable outcomes
The study touches on a longstanding ethical tension in medicine.
Saving a life is usually treated as an obvious benefit.
But medicine increasingly recognises that patients may value independence, cognition, communication and freedom from severe suffering as highly as survival itself.
There is no universal threshold at which continued treatment becomes unacceptable.
One person may accept severe disability for a small chance of additional life. Another may prioritise independence and reject treatment that is likely to leave them permanently dependent.
Neither preference can be derived from mortality statistics alone.
That is why patient values are central to ethically defensible end-of-life decisions.
Professional experience can create knowledge that patients do not have
ICU professionals may understand the likely consequences of intensive treatment differently because they see those outcomes repeatedly.
A member of the public may imagine resuscitation primarily as the restoration of a heartbeat.
A clinician may simultaneously think about neurological injury, prolonged ventilation, rehabilitation, frailty and the probability of returning to independent life.
This difference in knowledge creates an ethical challenge.
Clinicians need to communicate realistic outcomes without replacing the patient’s values with their own.
Professional experience can improve understanding of likely outcomes, but it does not give health professionals authority to decide what quality of life another person should consider acceptable.
The study does not show what patients themselves would choose
This is an important limitation.
The participants were hospital professionals, not patients facing the clinical situations described.
The study compared ICU caregivers, non-ICU caregivers and non-clinical hospital professionals.
It therefore cannot establish how closely any group’s preferences resemble those of the general public or of seriously ill patients.
Patients may place different weight on family relationships, religious beliefs, disability, pain, independence or even a small chance of recovery.
The findings are best interpreted as evidence that professional context is associated with different end-of-life preferences, not as evidence that one group’s preferences are more correct.
Hypothetical choices may differ from real decisions
The nine scenarios were hypothetical.
Participants had time to consider situations described in a questionnaire without being critically ill themselves.
Real end-of-life decisions can occur under emotional pressure, uncertainty and rapidly changing medical circumstances.
People may also change their preferences after experiencing serious illness or disability.
The study therefore measures stated preferences rather than behaviour under real clinical conditions.
That does not make the responses meaningless, but it limits how directly they can be translated into actual treatment decisions.
The study was conducted within a French hospital context
The participants were French hospital professionals, with the research team based at Saint-Denis University Hospital on Réunion Island.
Legal frameworks, professional norms and cultural attitudes toward withdrawal of life-sustaining treatment differ between countries.
The results should therefore not be assumed to represent health professionals in South Africa or elsewhere.
The comparison is still useful because it illustrates how end-of-life preferences can vary even among people working within the same broader healthcare system.
The result supports earlier conversations about patient values
The practical message is not that ICU clinicians should make treatment decisions based on what they would personally choose.
It is the opposite.
If professionals’ own preferences differ from those of colleagues, patients or families, then assumptions become dangerous.
Explicit conversations about values become more important.
Questions about acceptable disability, dependence, prolonged hospitalisation and the burdens of treatment may need to be discussed alongside the probability of survival.
These conversations are easier when they occur before a crisis rather than during one.
Advance directives are useful only when they reflect a person’s actual priorities
The low rate of completed advance directives among ICU caregivers also highlights the difference between a document and a meaningful conversation.
An advance directive is most useful when it reflects what matters to the person, not merely a list of procedures they do or do not want.
Someone may not know in advance whether they would want mechanical ventilation in every possible situation.
They may, however, know that permanent inability to communicate or complete dependence would be unacceptable to them.
Recording those values can help families and clinicians interpret treatment preferences when the exact scenario was never anticipated.
The bigger lesson is that survival statistics cannot answer a values question
Medicine can estimate whether a treatment is likely to prolong life.
It can sometimes estimate the probability of disability, dependence or long-term complications.
What it cannot determine scientifically is whether a particular outcome is worth living through for a particular person.
The new study shows that even health professionals exposed to the same healthcare system weigh these trade-offs differently depending on their work environment and experience.
That makes end-of-life decision-making partly a medical problem, but also an ethical and personal one.
The closer a decision moves toward questions of dependence, dignity and acceptable quality of life, the less safely anyone can assume that another person would make the same choice they would make for themselves.
Source Information
Study Title: Life-sustaining treatment preferences at end of life: a comparison between ICU and non-ICU hospital professionals
Authors: Eva Bembekoff, Babacar Tounkara, Marie Baron, Nicolas Allou, Cyril Ferdynus and Jérôme Allyn
Journal: Scientific Reports
Published: 24 September 2026
Sample: 743 French hospital professionals, including 195 ICU caregivers, 349 non-ICU caregivers and 199 non-clinical hospital professionals.
Method: Cross-sectional anonymous survey using nine validated hypothetical clinical scenarios that varied mortality risk and long-term consequences. Participants chose whether they would continue or withdraw life-sustaining treatment for themselves and what they believed a collegial medical team would decide for a patient. The researchers used multivariate logistic regression and decision-pattern analyses.
Main finding: Long-term consequences influenced end-of-life preferences more strongly than mortality risk alone, particularly among ICU caregivers. In seven of nine scenarios, ICU caregivers chose withdrawal of life-sustaining treatment for themselves more often than they believed a collegial team would choose it for patients. Only 13.3% of ICU caregivers had written advance directives despite universal awareness of them.
DOI: 10.1038/s41598-026-70416-6
