In Canada hospital beds have been reduced in number, and there is increased fiscal pressure for patients with advanced terminal illness to be cared for in their own homes until death. In this issue (see pages 361 to 367) Drs. Ian R. McWhinney and Martin J. Bass and Ms. Vanessa Orr report that people who die at home rather than in hospital are more likely to be cared for by family members other than a spouse and to have the services of a private duty nurse. The literature has shown that health problems of elderly spouses, occupational and other responsibilities of family members, and the physical, psychologic and financial strain of providing home care can make it difficult to honour a terminally ill person's wish to die at home. The findings of McWhinney and colleagues point to the existence of a two-tiered health care system in which those who have access to private duty nursing are able to stay at home to die. Their study also raises three key questions that must be addressed in the assessment of patient preferences as to place of death: Should family members be included in the assessment? How should preferences be measured? and What is an appropriate time frame for such an assessment? Although McWhinney and colleagues identify characteristics of care associated with place of death and underline the need for careful assessment of patient preferences regarding place of death, further research is needed to build on these findings. In the current context of health care reform, we need to examine more closely the type and intensity of services needed to support patients and their families in the final stages of a terminal illness.