One previous systematic review, the Toolkit (Teno 2005
), described a large number of patient-centered measures potentially applicable to the end of life and we identified 64 additional measures with published psychometric properties. Data on reliability or validity are available for only about one-third of published measures. Reliability and validity testing was often limited in scope and performed in small populations that were often not representative of the dying population as a whole.
Researchers faced with the challenge of selecting instruments for study will find a number of robust instruments (in the Toolkit and our review) that address pain and other symptoms, QOL, and quality of or satisfaction with terminal care described by after-death report. We identified several measures beyond those previously put forward by the Toolkit that were particularly noteworthy on the basis of conceptual grounding, psychometric evaluation, and/or acceptance in the field of palliative care. These measures include: the QUAL-E instrument, the QODD instrument, and the POS.
The current evidence base for palliative and end-of-life care is anchored in cancer care, and progress requires reliable and valid measures of the patient and caregiver experience for other conditions. At present, most measures and evaluations are limited to a single setting, usually the hospital or hospice. As patients at the end of life often use multiple sites of care, measures should be useful longitudinally over the time and settings. Culture may influence the patient and caregiver end-of-life experience, but measure development addressing population differences such as race or ethnicity are uncommon.
We also found that measures mostly addressed the domains of QOL, quality of care and satisfaction, and pain and physical symptoms. Gaps in measuring important domains of end-of-life care include continuity of care, advance care planning, spirituality, and caregiver well-being. In more developed domains, such as QOL or satisfaction, different projects almost always used different measures. The large number of measures of uncertain quality makes it difficult to compare findings or to synthesize insights across research or quality improvement studies; we recommend uniformity and use of the highest quality measures as important goals among others noted by the 2004 National Institutes of Health State-of-the-Science Conference Statement on Improving End-of-Life Care (available at http://consensus.nih.gov/
In the published intervention studies we reviewed, <9 percent of measures were used in more than two studies. Prior recommendations for measures in the field of palliative care (e.g., the Toolkit), as well as the existing interventional literature, have capitalized on the existing measures that were often developed for other uses. Future interventional research should increasingly emphasize the use of measures that are more specific and appropriate for palliative care. Support for collaborative research may be one helpful approach to facilitating robust development of the strongest measures and comparisons across studies.
Our systematic review has several limitations. For some purposes, our broad definition of “end of life” would be overinclusive; and for other purposes, our exclusions might have left out some important elements. Some measures developed for other populations might be useful for measuring certain aspects of the end-of-life patient and caregiver experience, although our review focused specifically on the extent of progress in measurement in the field of end-of-life care. Neither the Toolkit nor our systematic review identified all of the measures used to assess the outcomes of intervention studies, a fact that suggests that many of the measures of interest may not be indexed as “end of life” or previously developed or used in the end-of-life care population. The current review was unable to address literature on technical interventions, nor was it feasible to address children, and those limitations highlight the need for additional focused review of those areas. Our recommendations for measures are limited as the field lacks an agreed upon method for grading instruments and our implicit process criteria may not consider all the appropriate factors; our recommendations should be considered on the basis of these limitations.
Priorities for future research include developing measures of continuity of care, advanced care planning, spirituality, and grief and bereavement. Studies should compare the highest quality measures across diseases, settings, and among important populations. Measures are also needed that are appropriate for a variety of study purposes—clinical research, health services, and quality assessment and improvement. Research funders should emphasize the use of high-quality, comparable metrics.
Advancement of the quality of end-of-life measurement will aid reforms to improve the quality of end-of-life care. Better availability of well-developed, widely evaluated end-of-life-relevant measures is a critical step in improving the knowledge base and the quality of end-of-life care. This review provides a reference for researchers seeking guidance in choosing domains and the highest quality measures for evaluating the outcomes of interventions relevant to the end of life.