“How people die remains in the memories of those who live on.”1 This statement explains a basic tenet of palliative medicine: the family is the unit of care. Unfortunately, the medical literature seems to grow exponentially with evidence of inadequate medical care at the end of life. Too often, medical prognosis and goals of care are not discussed, symptoms such as pain are underassessed and undertreated, and psychosocial or spiritual issues are ignored. As well as resulting in patient suffering, the memories of those who live on may be of intense suffering by their loved ones. In response, measuring quality of care has become a priority in end-of-life research. Identifying the best method, however—that is, who, when, where, and how—remains a tremendous and unresolved challenge.2 Uncontrolled pain is associated with depression, anxiety, functional loss, economic hardship, suffering, and stress for the caregiver. Unsurprisingly, research into palliative care universally includes pain control as one of many possible items to measure quality of care at the end of life. Through retrospective assessment by family members, Tolle and associates assess pain control in hospitalized patients at the end of life. The advantage of this method is to reduce research time and cost. Also, it ensures a completed pain assessment because some patients are cognitively impaired during the last week of life. Unfortunately, studies by Higginson and colleagues and Hinton suggest that retrospective pain assessments by proxies may not be valid.3,4 In fact, proxies consistently report more severe pain than the patient does. Because only 26% of eligible family members agreed to participate in this study, the results may not be representative of the larger population. This low participation rate is puzzling, given the reported acceptance and emphasis of end-of-life care in Oregon. In a North Carolina study with similar methods, interviews were completed with 66% of eligible family members.5 Results of the Tolle study include a 54% rate of moderate to severe pain in hospitalized patients during the last week of life. These results are similar to those of the SUPPORT trial in which about 50% of conscious patients had moderate to severe pain at least 50% of the time before death.6 Interestingly, 90% of families in the Tolle study also reported that attention to comfort needs was good or excellent. These results are similar to those of the study of Hanson et al in which almost 70% of family comments on hospital end-of-life care were positive.5 The SUPPORT trial did not appear to have an equivalent question. Thus, we are left with an obvious discrepancy in the present study. Although pain control appeared inadequate, family satisfaction with attention to comfort was high. Is this adequate quality of care? Given the inherent limitations of the methods, this study may better assess family perception than actual quality of care as experienced by the patient. Although an understanding of its limitations is important, this research must be acknowledged for its value. First, family perceptions of quality of care are important independent of patient experience. Second, hypotheses are generated that can be subsequently evaluated with more methodologically stringent research. This may include direct assessment of patient pain and other quality-of-care indicators obtained prospectively in the last week of life. Studies have shown that as many as 50% of patients are awake in the last week of life and may, therefore, be able to give a history.7 Such research would require a large amount of time, effort, and most important, funding. With a common destiny, can we afford less?
Read full abstract