consistent with patient preferences is, When? Guidelines that recognize different needs at different points along the disease trajectory are necessary, especially those that are of particular importance when a person accepts that he or she is dying, such as spirituality and transcendence.
Patient preferences and satisfaction are important at every stage of treatment, but they take on added significance at the end of life. The measurement tools now available are based on review of medical records or administrative data. New measures are needed that incorporate the extent to which a patient’s care is based on informed preferences, that measure whether patients receive psychological support if needed and wanted, and that they assess whether care is both coordinated and compassionate. The perceptions of the dying patient and family provide an important perspective on each of these aspects of medical care. These surveys should be developed according to a conceptual model that is based on guidelines and the concerns voiced by dying persons and their families.
Some work has been started toward surveys of bereaved family members. One effort (Teno et al., 2001a; Teno, et al., 2001b) uses current guidelines and results of focus groups from around the country to develop questions on unmet needs and on the family’s perspective of the quality of care delivered to the dying person and to themselves. A second survey (Patrick and Curtis, 2001) focuses on the quality of dying. As these and other tools are developed, some questions will be applicable to all dying persons, but there will also be a need for disease-specific questions (e.g., management of toxicity from chemotherapeutic agents is a very important concern, and the specifics of management are different for cancer patients than for those dying from other causes).
The initial work has focused on retrospective surveys of surviving family members largely because the denominator is easily defined (based on cancer registry or death certificate data) and family members are often the only ones able to be interviewed in the last month of a patient’s life. Surveys that directly capture the patients’ perspective are needed, as well, however. The design of such surveys could be linked to sentinel events or triggers (e.g., admission to palliative care or hospice program, reaching a certain disease stage), with consideration give to which domains are included and the point (or points) along the patient’s disease trajectory at which questions should be asked.
An important tension that the developers of surveys will face is between respondent’s burden and the desire to be comprehensive. The eventual goal is to minimize the respondent burden, but initially a larger number of items will be tested and a winnowing process used to arrived at a parsimonious set of questions.
The mode of survey administration is another important research question: can valid information be gathered through a self-administered ques-