Physicians' perceptions of suffering in people with dementia at the end of life

Jenny T. Van Der Steen, Luc Deliens, Raymond T.C.M. Koopmans, Bregje D. Onwuteaka-Philipsen

Research output: Contribution to journalArticleAcademicpeer-review

14 Citations (Scopus)


Objective: Our aim was to describe physicians' perceptions of the suffering of their patients who are dying with dementia, many of whom are incompetent with regard to decision making and have difficulty with communicating about the source of their distress and with identifying related factors. Method: We analyzed data from the nationally representative observational Dutch End-of-Life in Dementia (DEOLD) cohort study (2007-2011), which involved 34 long-term care facilities. A total of 103 physicians completed questionnaires about 330 patients with dementia who had died in a participating facility. Suffering during the last six hours of life was defined as a patient being disturbed by or aware of symptoms, suffering until the end or death was a struggle - all related to objective indicators of lack of comfort. We employed generalized estimating equation models to assess associations of suffering with the characteristics of physicians and patients, the patient's death, and the decision-making process. Results: In 13.8% of cases, the physician felt that the patient had suffered. An unexpected death and death with pneumonia were strongly (an odds ratios close to 6) associated with suffering, and suffering was also independently associated with the physician's perception of worse quality of end-of-life care, death with cardiovascular disease, a less experienced physician, no palliative sedation, and a younger patient. Significance of Results: Most patients with dementia did not suffer during their final hours of life, according to their physicians. There are a number of factors associated with suffering, among them death with pneumonia and unexpected death. We may not be able to have much influence on death from pneumonia, but quality of care and an unexpected death are reasonable targets for intervention. Earlier identification of the beginning of the dying process would allow time to better prepare for approaching death, which would provide a source of comfort.

Original languageEnglish
Pages (from-to)587-599
Number of pages13
JournalPalliative and Supportive Care
Issue number5
Publication statusPublished - 1 Oct 2017


  • Dementia
  • End of life
  • Nursing homes
  • Palliative care
  • Quality of life

Cite this